EPA/600
September 1989
DRAFT
EXPOSURE ASSESSMENT METHODS HANDBOOK
Exposure Assessment Group
Office of Health and Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC 20460

-------
DISCLAIMER
This document has been reviewed in accordance with U.S. Environmental
Protection Agency policy and approved for publication. Mention of trade
names or commercial products does not constitute endorsement or
recommendation for use.

-------
TABLE OF CONTENTS
Page No.
Tables 		vii
Foreward 		xi
Preface 		xii
Abstract 		xi i i
Authors, Contributors, and Reviewers 		xiv
PART I
1.	INTRODUCTION 		2-1
1.1	Background 		1-1
1.2	Purpose and Scope 		1-2
1.3	Organization of the Handbook 		1-3
2.	OVERVIEW OF THE EXPOSURE ASSESSMENT PROCESS 		2-1
2.1	Methodological Framework 		2-1
2.2	Elements of Exposure Assessment 		2-1
2.2.1	General Information 		2-1
2.2.2	Estimation of Release 		2-3
2.2.3	Environmental Fate Analysis/Determination of
Environmental Concentrations 		2-4
2.2.4	Development of Exposure Scenarios 		2-5
2.2.5	Estimation of Exposure/Dose 		2-5
2.2.6	Estimation of Integrated Exposure 		2-9
3.	ASSESSMENT OF EXPOSURE VIA SURFACE WATER 		3-1
3.1	Release Analysis - Surface Water 		3-1
3.1.1	Screening of Release Potential 		3-1
3.1.2	Quantitative Release Estimation 		3-2
3.2	Fate Analysis - Surface Water 		3-15
3.2.1	Screening of Surface Water Fate Processes 		3-15
3.2.2	Quantitative Fate Estimation 		3-25
3.3	Estimation of Concentrations - Surface Water 		3-38
3.4	Exposure Estimation - Surface Water 		3-40
iii

-------
TABLE OF CONTENTS (continued)
Page No.
4.	ASSESSMENT OF EXPOSURE VIA GROUND WATER 		4-1
4.1	Release Analysis - Ground Water 		4-1
4.1.1	Screening of Release Potential 		4-1
4.1.2	Quantitative Release Estimation 		4-2
4.2	Fate Analysis - Ground Water 		4-10
4.2.1	Screening of Fate Processes 		4-10
4.2.2	Quantitative Fate Estimation 		4-19
4.3	Estimation of Concentration - Ground Water 		4-40
4-4 Exposure Estimation - Ground Water 		4-47
5.	ASSESSMENT OF EXPOSURE VIA FISH/SHELLFISH 		5-1
5.1	Release Analysis - Fish/Shellfish 		5-1
5.1.1	Screening of Release Potential 		5-1
5.1.2	Quantitative Release Estimation 		5-1
5.2	Fate Analysis - Fish/Shellfish 		5-1
5.2.1	Screening of Fate Processes 		5-1
5.2.2	Quantitative Fate Estimation 		5-2
5.3	Estimation of Concentrations - Fish/Shellfish 		5-10
5.4	Exposure Estimation - Fish/Shellfish 		5-10
6.	ASSESSMENT OF EXPOSURE VIA TERRESTRIAL PLANT AND ANIMAL FOOD
SOURCES 		6-1
6.1	Release Analysis - Terrestrial Food Sources 		6-1
6.1.1	Screening of Release Potential 		6-1
6.1.2	Quantitative Release Estimation 		6-1
6.2	Fate Analysis - Terrestrial Food Sources 		6-1
6.2.1	Screening of Fate Processes 		6-1
6.2.2	Quantitative Fate Estimation 		6-2
iv

-------
TABLE OF CONTENTS (continued)
Page No.
6.3	Estimation of Concentrations - Terrestrial Food
Sources 		6-8
6.4	Exposure Estimation - Terrestrial Food Sources 		6-12
7.	ASSESSMENT OF EXPOSURE VIA AMBIENT AIR 		7-1
7.1	Release Analysis - Ambient Air 		7-1
7.1.1	Screening of Release Potential 		7-1
7.1.2	Quantitative Release Estimation 		7-2
7.2	Fate Analysis - Ambient Air 		7-22
7.2.1 Atmospheric Degradation 		7-23
7.3	Estimation of Concentrations - Ambient Air 		7-32
7.4	Exposure Estimation - Ambient Air 		7-43
8.	ASSESSMENT OF EXPOSURE VIA DIRECT CONTACT WITH CONTAMINATED
SOILS 		8-1
8.1	Release Analysis - Soil 		8-1
8.1.1	Screening of Release Potential 		8-2
8.1.2	Quantitative Release Estimation 		8-5
8.2	Fate Analysis - Soil 		8-7
8.2.1	Screening of Soil Fate Processes 		8-7
8.2.2	Quantitative Fate Estimation 		8-11
8.3	Estimation of Concentrations - Soil 		8-11
8.3.1	Estimation of Airborne Concentrations Resulting
from Contaminated Soil 		8-11
8.3.2	Estimation of Concentrations in Surface Water
Resulting from Contaminated Soil 		8-16
8.3.3	Estimation of Concentrations in Ground Water
Resulting from Contaminated Soil 		8-16
8.4	Exposure Estimation - Soil 		8-18
v

-------
TABLE OF CONTENTS (continued)
Page No.
PART II
1.	EXPOSED POPULATIONS 		1-1
2.	CHARACTERIZATION AND ANALYSIS OF UNCERTAINTIES IN THE
SAMPLING, MONITORING, AND MODELING OF ENVIRONMENTAL
CONCENTRATIONS OF CONTAMINANTS 					2-1
2.1	Causes of Uncertainty 		2-1
2.2	Qualitative Analysis of Uncertainties 		2-3
2.3	Quantitative Analysis 		2-4
2.4	Presentation of Uncertainty Analysis Results 		2-16
3.	REFERENCES 		3-1
APPENDIX A - Glossary of Terms 		A-l
APPENDIX B - Exposure Assessment Worksheets 		B-l
APPENDIX C - Completed Exposure Assessment Worksheets for An
Example Chemical 		C-l
APPENDIX D - Transfer Coefficients for Assessing Exposures to
Contaminants in Animal Tissues and Animal Products .	D-l
vi

-------
LIST OF TABLES
Page No.
PART I
Table 3-1. Runoff Curve Numbers 	:	 3-4
Table 3-2. Magnitude of Soil Erodibility Factor K 	 3-6
Table 3-3. "C" Values for Permanent Pasture, Rangeland, and
Idle Land 	 3-10
Table 3-4. "C" Values for Woodland 	 3-11
Table 3-5. Values of P for Agricultural Land 	 3-12
Table 3-6. Values of P for Construction Sites 	 3-13
Table 3-7. Relative Importance of Processes Influencing Aquatic
Fate of Organic Priority Pollutants 	 3-17
Table 3-8. Estimated Adsorption of Chemicals to Suspended
Solids in Surface Waters 	 3-21
Table 3-9. Classes of Organic Functional Groups Potentially
Susceptible to Hydrolysis 	 3-23
Table 3-10. Classes of Organic Functional Groups Potentially
Resistant to Hydrolysis 	 3-24
Table 4-1. Parameter Values for Permeation Equation (at 25°C) . 4-6
Table 4-2. Polymer Categorization for Permeation of Water 	 4-7
Table 4-3. Permachor Values of Some Organic Liquids in
Polyethylene and PVC 	 4-8
Table 4-4. Water Permachor Value for Dry Polymers 	 4-9
Table 4-5. The General Relationship Between the Soil/Solution
Partition Coefficient Kj and Soil Mobility 	 4-15
Table 4-6. Representative Values for Saturated Moisture
Contents and Field Capacities of Various Soil
Types 	 4-25
vi i

-------
LIST OF TABLES (continued)
Page No.
Table 4-7. Representative Values of Hydraulic Parameters
(Standard Deviation in Parentheses) 	 4-26
Table 4-8. Representative Values of Saturated Hydraulic
Conductivity 	 4-27
Table 4-9. Saturated Hydraulic Conductivity Ranges for Selected
Rock and Soil Types 	 4-28
Table 4-10. Suggested Value for £et Relating Evaporation from a
U.S. Class A Pan to Evapotranspiration from 8- to
15-cm Tall, Well-Watered Grass Turf 	 4-31
Table 4-11. Crop Coefficients for Estimating
Evapotranspiration 	 4-32
Table 4-12. Features of Unsaturated Zone and Ground-Water Models 4-43
Table 4-13. Data Requirements for Unsaturated Zone and Ground-
Water Models 	 4-46
Table 5-1. Reasonable Worst-Case Bioconcentration Factors
(BFCs) for Metals in Fish/Shellfish 	 5-9
Table 6-1. Relative Accumulation of Heavy Metals into Edible
Plant Parts of Vegetable Crops 	 6-9
Table 6-2. Parameter Values for Estimation of Concentrations in
Animal Tissues/Products Using Transfer
Coefficients 	 6-13
Table 7-1. Environmental Variables and Model Parameters for the
Wind Erosion Equation 	 7-4
Table 7-2. Diffusion Coefficients of Selected Organic Compounds 7-13
Table 7-3. Approximate Absorption Regions for Organic Molecules 7-25
Table 7-4. Oxidant Concentrations in Water and Air 	 7-28
Table 7-5. Rate Constants for Oxidation by HO Radical in the
Atmosphere at 25°C 	 7-29
vi i i

-------
LIST OF TABLES (continued)
Page No.
Table 7-6. Rates of Oxidation by Ozone in the Atmosphere at
25°C 	 7-31
Table 7-7. Capabilities of Reviewed Models 	 7-33
Table 7-8. Preferred Models for Selected Applications in Simple
Terrain 	 7-36
Table 7-9. Resource Requirements and Information Sources:
Atmospheric Fate Models 	'	 7-37
Table 7-10. Features of Atmospheric Fate Models 	 7-41
Table 7-11. Data Requirements for Atmospheric Models 	 7-42
Table 8-1. Factors Necessary to Determine Releases from
Contaminated Soils 	 8-3
Table 8-2. Importance of Volatility from Dry Soils 	 8-10
Table 8-3. Default Assumptions for Calculation of Short-Term
Maximum Concentrations in Air 	 8-15
PART II
Table 2-1. Relative Errors and Sensitivity Coefficients for the
Gaussian Plume Concentration Model of Examples 1
& 2: C = (8.65xl07Q/1000ua)exp_1/2(H2/a2j 	 2_8
Table 2-2. Uncertainty Analysis Report for Example 3
Concentration Estimation 	 2-18
ix

-------
LIST OF FIGURES
Page No.
PART I
Figure 2-1. Framework and Components of an Exposure Assessment . 2-2
Figure 2-2. Representations of Microenvironment Contributions to
Total Inhalation Exposure to Nitrogen Dioxide 	 2-12
Figure 3-1. Slope Effect Chart Applicable to Areas A-l in
Washington, Oregon, and Idaho, and All of A-3 	 3-7
Figure 3-2. Soil Moisture-Soil Temperature Regimes of the
Western United States 	 3-8
Figure 3-3. Slope Effect Chart for Areas Where Figure 3-2 is Not
Applicable 	 3-9
Figure 4-1. Examples of the Range of Hydrolysis Half-Lives for
Various Types of Organic Compounds in Water at pH 7
and 25°C 	 4-17
Figure 7-1. Mean number of Days per Year With >0.01 Inches of
Precipitation (i.e., "Wet Days") 	 7-7
Figure 7-2. Decision Tree for Selection of Air Transport
Models 	 7-35
Figure 8-1. Horizontal Dispersion Coefficient as a Function of
Downwind Distance from the Source 	 8-13
Figure 8-2. Vertical Dispersion Coefficient as a Function of
Downwind Distance from the Source 	 8-14
x

-------
FOREWORD
The Exposure Assessment Group (EAG) of EPA's Office of Research and
Development has three main functions: (1) to conduct exposure assess-
ments, (2) to review assessments and related documents, and (3) to develop
guidelines for exposure assessments. The activities under each of these
functions are supported by and respond to the needs of the various program
offices. In relation to the third function, EAG sponsors projects aimed
at developing or refining techniques used in exposure assessments.
The purpose of this document is to provide guidance to exposure
assessors regarding accepted methodologies for estimating or modeling
concentrations of contaminants in various environmental media. This is
presented as a companion document to the Exposure Factors Handbook (EPA
Publication No. 600/8-89/043) which provides algorithms and exposure
parameter values for estimating exposures based on contaminant concentra-
tions, which are the output of the methods described in this document.
The methodologies presented in this document are generic and are based
upon our evaluation of the available literature and accepted methods
employed by most exposure assessors. Some modification of these methods
may be necessary for scenarios that differ significantly from those repre-
sented here. The document is published in the three-ring binder format so
that it can be easily updated as new methodologies and information become
available.
Michael A. Callahan
Director
Exposure Assessment Group
xi

-------
PREFACE
The Exposure assessment Group (EAG) of the Office of Health and Envi-
ronmental Assessment (OHEA) has prepared this handbook as a companion doc-
ument to the Exposure Factors Handbook (EPA Publication No. 600/8-89/043).
It was prepared in response to requests from many EPA program and regional
offices for guidance on how estimated concentrations of chemicals in
various environmental media should be calculated.
The purpose of this Handbook is to provide a summary of the available
and accepted methods for estimating and modeling concentrations of chemi-
cals in ambient air, surface water, ground water, fish and shellfish,
terrestrial food sources, and soil. In addition, the Handbook provides,
where appropriate, default values for parameters used for making these
calculations. While originally intended for assessing environmental
concentrations resulting from releases from uncontrolled hazardous waste
sites, the Handbook will also provide a common resource that exposure
assessors in all Agency programs can use. Thus, it should help improve
the consistency with which exposure assessments are conducted across the
Agency, but still allow different approaches that may be appropriate based
on considerations of policy, precedent, or other factors.
xi i

-------
ABSTRACT
This document provides a summary of the available algorithms and
parameter values that can be used to determine estimated or modeled con-
centrations of chemicals in various environmental media, including ambient
air, surface water, ground water, fish and shellfish, terrestrial food
sources, and soil. These methods are organized and presented by environ-
mental media for the convenience of the user. The chapters on ambient
air, surface water, and ground water include quantitative treatment of
losses in concentrations of chemicals due to environmental fate processes.
This Handbook also provides an overview of the exposure assessment process
and how these methods are integrated within that process, a discussion of
methods for characterizing and enumerating exposed populations, and a dis-
cussion of methods for quantitating the statistical uncertainties related
to estimation of environmental concentrations. In addition, the Handbook
provides a series of worksheets that contain the step-by-step procedure
for employing the methods described.
xi i i

-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS
The Exposure Assessment Group (EAG) within EPA's office of Health and
Environmental Assessment was responsible for the preparation of this
handbook. The document was prepared by Versar Inc. under EPA Contract
No. 68-02-4254, Work Assignment No. 249. Seong Hwang of EAG served as
the EPA Task Manager, providing overall direction and guidance.
AUTHORS
Gary Whitmyre	RiskFocus Division
Versar Inc.
Springfield, Virginia
James Konz	Exposure Assessment Division
Clay Carpenter	Versar Inc.
Dan Arrenholz	Springfield, Virginia
Arthur Clarke
Norman Gabel
Karen Li si
Willi am Myers
Sanjeev Aggarwal
Among the authors, Mr. Gary Whitmyre was the Task Manager responsible
for coordinating the overall effort; determining the scope and technical
content of the manual; providing assistance in development of the chapters
related to estimation of concentrations in terrestrial food sources and
in fish/shellfish and to analysis of uncertainties; and creation of the
exposure assessment worksheets.
Mr. James Konz prepared chapters related to terrestrial food sources
and ground water. Mr. Clay Carpenter prepared chapters related to over-
view of the exposure assessment process and characterizing and enumerating
exposed populations. Mr. Dan Arrenholz was responsible for preparation of
the chapter on estimation of contaminant concentrations in surface water.
Dr. Arthur Clarke contributed to the chapter on analysis of uncertainties.
Dr. Norman Gabel and Ms. Karen Lisi researched and prepared sections deal-
ing with quantitative assessment of environmental fate. Mr. William Myers
and Mr. Sanjeev Aggarwal contributed to chapters that addressed estimation
of contaminant concentrations in soil and ambient air, respectively.
xiv

-------
REVIEWERS
The following individuals within EPA reviewed an earlier draft of
this document and provided valuable comments:
Mr. Randall Bruins
Environmental Criteria and Assessment Office
Cincinnati, Ohio
Ms. Lynn Del pi re
Office of Toxic Substances
Washington, D.C.
Mr. Charles Ris
Office of Research and Development
Washington, D.C.
Ms. Terri Knonza
Office of Research and Development
Washington, D.C.
The following individuals outside of EPA also reviewed this document
and provided helpful comments and suggestions:
Mr. Mitchell Small
U.S. Army Biomedical Research and Development Laboratory
Fort Dietrick, Maryland
Dr. Thomas McLaughlin
CH2M HILL
Herndon, Virginia
ACKNOWLEDGEMENTS
Quality assurance was provided by Mr. Clay Carpenter, Mr. James Konz,
and Ms. Cindy Lewis. Computer-assisted literature searches and informa-
tion retrieval were provided by Ms. Mimi Faha. Technical editing support
was provided by Ms. Barbara Malczak, Ms. Martha Martin, and Ms. Juliet
Crumrine; word processing was provided by Ms. Lynn Maxfield and Ms. Kammi
Johannsen; and graphics were prepared by Ms. Kathy Bowles. The dedicated
assistance of these Versar Inc. employees allowed the successful comple-
tion of this project and is sincerely appreciated.
xv

-------
1.	INTRODUCTION
1.1 Background
Over the past few years, exposure assessment has played an increasing-
ly important role in many environmental programs at the federal, regional,
and state levels. Exposure assessment has been recognized as an important
component of the risk assessment of chemical releases. Because an expo-
sure assessment requires examination of the entire history of a chemical
in the environment, from release at the site to arrival at the receptor,
a multidisciplinary approach is needed that encompasses the expertise of
engineers, hydrologists, geologists, meteorologists, biologists, chemists,
and toxicologists. Exposure assessment can also be a valuable tool for
examining different options for remediation based on exposure impacts, and
therefore it is also useful in the context of risk management.
Approaches to exposure assessment that have been applied over the past
decade include (USEPA 1988a) the following:
•	Direct measurement methods, which attempt to estimate the
concentrations of pollutants in environmental media to which an
individual or receptor population is exposed;
•	Predictive methods, which attempt to estimate exposure based on
available knowledge of source characterization, the behavior of
pollutants in the environment, and exposure parameters relevant
to receptors (e.g., inhalation rates and frequency and duration
of exposures); and
•	Reconstructive methods, which attempt to reconstruct exposure
information from evidence within an organism such as concentra-
tions in body tissues, toxicological indicators, or other inter-
nal evidence.
All three of these approaches are potentially useful in exposure
assessment. While this Handbook focuses primarily on predictive methods,
it contains elements of the other two approaches as well.
The Exposure Assessment Methods Handbook is intended to serve as a
support document to EPA's Guidelines for Estimating Exposures (USEPA
1986a) and the Proposed Guidelines for Exposure-Related Measurements
(USEPA 1988a) by providing data on standard methods that are used to
1-1

-------
calculate environmental concentrations of toxic chemicals. The guidelines
were developed to promote consistency among the various exposure assess-
ment activities that are carried out by the various EPA program offices.
This Handbook should assist in this goal by providing a consistent frame-
work by which to calculate environmental concentrations. This Handbook
serves as a companion document to the Exposure Factors Handbook (USEPA
1989) which contains the standard factors needed to calculate exposure.
1.2 Purpose and Scope
This Handbook provides a compendium of selected available approaches
to assessment of multimedia exposures from waste disposal facilities.
However, while the release portions of this report are specifically
tailored to waste disposal settings, many of the approaches relating to
fate and exposure are more generic and could be applied to other types of
exposure settings.
The purpose of this Handbook is not only to acquaint exposure
assessors with the release, fate, and exposure algorithms necessary to
carry out an exposure assessment for waste disposal facilities, but also
to provide a step-by-step procedure for performing such an assessment.
This step-by-step procedure is presented in the worksheets that accompany
this report (see Appendix B); these worksheets in a sense represent the
most practical output of this effort. The worksheets and the information
contained within this report do not assume that the user has prior
knowledge of specific models, algorithms, or terminology. However, by
using this document, exposure assessors and others should be able to:
•	Understand the key assumptions, methodology, and limitations
relating to exposure assessment;
•	Understand the basic differences between options for approaching
exposure assessment;
•	Identify sources of information on exposure assessment methodolo-
gies and relevant parameter values; and
•	Comprehend the degree of uncertainty associated with exposure
estimates.
1-2

-------
Although this document contains a step-by-step procedure for perform-
ing exposure assessments of uncontrolled hazardous waste sites, it should
not be used as a substitute for review and consultation with more experi-
enced exposure assessors, environmental fate specialists, modelers, and
other environmental professionals.
1.3 Organization of the Handbook
Following several introductory sections, this Handbook is primarily
divided according to exposure pathways; these sections cover specific
exposure pathways and present the environmental release, fate, and concen-
tration considerations and estimation algorithms relevant to each pathway.
Part I contains media-specific information to be used in the estimation
of environmental concentrations. Part II provides discussions of popula-
tions and uncertainties, and presents the references cited. The specific
sections are as follows for Part I: Section 2 provides an overview of the
exposure assessment process; Section 3 presents methods for quantifying
environmental concentrations in surface water; Section 4 discusses
methodologies for assessing environmental concentrations in ground water;
Section 5 describes quantitative approaches for assessing chemical concen-
trations in fish and shellfish; Section 6 addresses methods for assessing
chemical concentrations in the terrestrial food chain; Section 7 discusses
methodologies for assessing chemical concentrations in ambient air; and
Section 8 presents approaches to determination of chemical concentrations
in soil. In Part II, Section 1 provides information for characterizing
and enumerating exposed populations; Section 2 discusses the sources of
uncertainty in estimates of environmental concentrations and methods for
quantifying uncertainty; and Section 3 presents cited references.
Several appendices are also included in this document. Appendix A
provides a glossary of terms used throughout the report. The format of
the Exposure Assessment Worksheets, which incorporate the environmental
release, fate, concentration, and exposure estimation algorithms is
provided in Appendix B. Completed Exposure Assessment Worksheets for an
1-3

-------
example chemical are presented in Appendix C. Transfer coefficients that
can be used to assess exposures to contaminants in animal-derived food
products are listed in Appendix D.
1-4

-------
2.	OVERVIEW OF THE EXPOSURE ASSESSMENT PROCESS
Planning should be the first step in any exposure assessment; deter-
mining the purpose, scope, depth, and approach will ultimately affect the
quality and usefulness of the resulting exposure estimates. By addressing
these elements in the planning stage, the assessor will focus on the most
critical aspects of the exposure assessment and will conserve resources by
eliminating unnecessary scenarios and inappropriate exposure pathways.
This will allow the assessor to identify the procedures to be used, to
establish the boundaries of the assessment, to facilitate its interfacing
with other assessments (e.g., human health risk assessments), and to
determine and control the final form of the assessment.
2.1	Methodological Framework
While each exposure assessment may be different, certain elements are
common to most exposure assessments. In general, the following major
topics are usually addressed: background information about the chemical;
sources of the chemical's releases; environmental pathways and fate
analysis (transport and transformation); development of the exposure
scenarios that are specific to the individual exposure assessment; summary
of monitoring data and/or prediction of environmental concentrations;
estimation of the exposure/dose levels; evaluation of the exposed human
populations; and integration of the route-specific exposure analyses.
The relationships among these components is shown in Figure 2-1.
2.2	Elements of Exposure Assessment
2.2.1 General Information
The exposure assessment process will usually be initiated by gathering
of information on the chemical of concern and compiling a brief descrip-
tion of these data. Each chemical should be identified by as many of the
following characteristics as appropriate: molecular formula and struc-
ture; synonyms; Chemical Abstracts Service (CAS) Registry Number; and
description of likely contaminants. The available data on chemical and
physical properties should also be tabulated; where data are lacking but
2-1

-------
ro
i
ro
FIGURE 2-1. FRAMEWORK AND COMPONENTS OF AN EXPOSURE ASSESSMENT

-------
are important for characterizing the chemical itself or its potential
transport or tranformation in the environment, the critical properties
should be estimated by the use of appropriate calculations such as those
available in the CHEMEST system (part of GEMS, NUMERICA) and presented by
Lyman et al. (1982). These properties include the molecular weight,
density, melting point, boiling point, vapor pressure, vapor density,
water solubility, partition coefficients, soil sorption coefficient,
chemical half-lives in environmental media, and other properties, as
appropriate.
2.2.2 Estimation of Release
The estimation of the environmental release of a chemical involves the
identification of all known or suspected sources, accounting for specific
releases to each of the environmental media or compartments (e.g., air,
surface water, ground water, soil, biota). The releases should be
characterized in as much detail as possible concerning the amounts of the
chemical and known or estimated rates of the release.
A detailed exposure assessment may include or be based on a previous
"materials balance." The materials balance is a thorough accounting of
the source locations, production, uses, destruction/disposal, and environ-
mental release of the chemical, quantifying the mass that was produced,
incorporated into products, and lost to the environment. This materials
balance may attempt to quantify chemical releases as a function of time,
especially if it is expected that releases are intermittent or sporadic,
as well as to describe the physical characteristics and magnitude of the
source and the physical and chemical form of the chemical being released.
Aqueous releases to receiving waters are often referred to as effluents or
discharges; atmospheric releases are often termed emissions; and releases
to soil may be described as leachate, spills, or leaks. In the absence
of complete mass-balance information, methods for estimating releases of
a given chemical into environmental media are provided in the following
sections.
2-3

-------
2.2.3 Environmental Fate Analysis/Determination of Environmental
Concentrations
This component of the exposure assessment is a description of the
route the chemical follows as it travels from the source to the receptor
population and of how much of the chemical degrades and/or transforms to
other chemicals during transport through the environment. The environmen-
tal fate analysis provides information on the extent and magnitude of
environmental contamination, including the concentrations of the contami-
nant in each of the specific compartments of the environment. In the ab-
sence of direct data on pathways and fate of a given chemical, this infor-
mation can often be predicted based on the physical-chemical properties
of the chemical. In addition, chemical fate can be predicted by the use
of a number of available mathematical models. The degree of specificity
of the environmental pathways and fate estimates is dependent on the level
of detail required of the overall exposure assessment for the individual
chemical. Some environmental routes of exposure or some transformation
processes may be judged to be insignificant and can be excluded from
further consideration.
For a detailed exposure assessment, the fate analysis should describe
each of the transport and transformation processes of the chemical,
because the ultimate environmental distribution and concentrations at the
receptors will differ considerably from concentrations in the initial
releases and levels in the immediate area of the release. The transport
analysis should include hydrological, atmospheric, and/or other physical
processes that describe the routes of the contaminant through each com-
partment and across intercompartmental boundaries from the source to the
receptor. The behavior of the chemical should be assessed at each bound-
ary, including the air/water, the soil/air, and the water/soil interfaces.
This environmental partitioning assessment is important because transport
processes may dilute the chemical or may serve to concentrate the chemi-
cal in certain environmental compartments.
Transformation processes also must be assessed for the chemical in
each of the environmental compartments of interest. Chemical alteration
2-4

-------
may occur via (1) biodegradation; (2) photodegradation; (3) hydrolysis,
oxidation/reduction, and other chemical degradation processes; and
(4) reactions with other chemicals.
Once the detailed transport and transformation pathways have been
identified, one can focus on the individual exposure pathways for more
thorough analyses, depending on the relative importance of different expo-
sure pathways and routes for the chemical.
2.2.4	Development of Exposure Scenarios
Information on the environmental fate and exposure pathways of the
chemical allows the assessor to predict the types of human receptor popu-
lations that could be exposed and the routes by which exposure could
occur. For some chemicals, the exposure routes to be studied may have
been selected at the beginning of the exposure assessment process in order
to focus on specific concerns.
Exposure scenarios should be developed for each relevant major expo-
sure pathway; these may include contact via air, ground water, surface
water, food, and. contaminated soil. Exposure scenarios account for the
route of contact (i.e., dermal, inhalation, and ingestion) and important
exposure parameters such as frequency of contact, duration of each expo-
sure event, and rate of contact (i.e., rate of inhalation of ambient air,
rate of ingestion of drinking water, and rate of ingestion of contaminated
food). In addition, other types of exposure parameters may need to be
specified as part of the development of exposure scenarios. For example,
development of an exposure scenario for inhalation exposure to volatile
organic compounds released from indoor home use of contaminated surface
water would require information or assumptions regarding room volume and
ventilation rate.
2.2.5	Estimation of Exposure/Dose
There are two basic methods of determining concentrations of chemicals
in environmental media, from which an environmental exposure value can be
determined; these are monitoring data and estimates based on models.
2-5

-------
Monitoring is the repeated sampling and analysis of different locations
within selected compartments of the environment to determine the levels of
the chemical and how those levels change with time. At best, most moni-
toring concentration values represent short-term averages. Thus, monitor-
ing data must be carefully reviewed to evaluate their precision and
accuracy and whether the sampling locations and durations used provide
representative values. Because monitoring data are usually neither com-
plete nor comprehensive for each environmental compartment, it is often
desirable to supplement the concentration data with modeling estimates.
In estimating exposure, the exposure assessor should distinguish
between exposure and dose. For some chemicals, the external exposure
resulting from the concentration in the environment to which the receptor
is exposed may be substantially greater than the internal exposure, or
absorbed dose incurred by the receptor. The absorbed dose is a function
of the degree of penetration through and chemical transfer across barriers
(e.g., skin) and membranes (e.g., the lining of the lung) that it con-
tacts; thus, different exposure routes are often associated with different
rates of absorption.
For most exposure assessments, the dose is estimated based on external
exposure, because the human health/animal toxicological data used in com-
bination with exposure data to calculate risk are most frequently based on
administered dose or ambient exposure rather than internal dose. In order
to add exposures incurred from different exposure routes, as discussed in
Section 2.2.6, the absorbed dose is the desired common denominator for
different routes; however, calculation of exposure route-specific absorbed
doses calls for a level of sophistication that is well beyond (1) the
limited available data on pharmacokinetic factors for most chemicals and
(2) the screening-level purpose of most exposure assessors using this
document. Therefore, this document will be directed at external exposure
rather than absorbed dose.
EPA's Exposure Assessment Guidelines (USEPA 1986a; 1988a) define expo-
sure as the contact with a chemical or physical agent. The magnitude of
2-6

-------
the exposure is the amount of the agent available at human exchange bound-
aries (skin, lungs, gut) during some specified time. Starting with a
general integral equation for exposure (USEPA 1988a), several exposure
equations can be derived depending upon boundary assumptions. One of the
more useful of these derived equations for dealing with lifetime exposures
to agents with linear non-threshold responses (i.e., our current assump-
tions about many carcinogens) is the Lifetime Average Daily Exposure
(LADE) discussed below. Exposure assessments are usually done to support
risk assessments; only exposure calculations used to support cancer risk
assessments and repeated and prolonged (chronic) exposures to noncarcino-
gens will be covered in this Handbook. The Proposed Guidelines for
Exposure-Related Measurements (USEPA 1988a) contain an expanded discussion
of some of the other equations that can be used. The Exposure Factors
Handbook (USEPA 1989) discusses exposure parameters and factors used to
estimate exposure.
For cancer risk assessments, exposure is averaged over the body weight
and lifetime:
iAr\c	Total Exposure		(? n
Body Weight x Lifetime	* '
The total exposure can be expanded as follows:
Media
Total ExDosure = Chemical Contact Exposure (?
p	Concentration Rate Duration ' '
The chemical concentration term above is the concentration of the
contaminant in the medium (air, food, soil, etc.) contacting the body and
has units of mass/volume or mass/mass.
The contact rate refers to the rate of inhalation, ingestion, or
dermal contact depending on the route of exposure. For ingestion, the
contact rate is simply the amount of food, soil, or water containing the
chemical of interest that an individual ingests during some specific time
period (units of mass/time).
2-7

-------
The exposure duration is the length of time that contact with the con-
taminated environmental media lasts. The time a person lives in an area,
frequency of activities leading to exposure (e.g., bathing), time spent
indoors vs. outdoors, etc., may all affect the exposure duration. When
the above parameters remain constant over time, duration may be substi-
tuted directly into the exposure equation. When they change with time, a
summation approach is needed to calculate exposure. In either case, the
exposure duration is the length of time exposure occurs at the concentra-
tion and contact rate specified by the other parameters in the equation.
Exposure (sometimes called "administered dose") can be expressed as a
total amount (with units of mass, e.g., mg), as an exposure rate in terms
of mass/time (e.g., mg/day), or as a rate normalized to body mass (e.g.,
with units of mg of chemical per kg of body weight per day (mg/kg-day).
The LADE is usually expressed in terms of mg/kg-day or other mass/mass-
time units..
The lifetime term in Equation (2-1) is the total period of time over
which the administered dose is averaged. For carcinogens, this should
represent the average life expectancy of the exposed population. For
exposure estimates to be used for assessments other than carcinogenic
risk, different averaging periods are frequently used. For acute expo-
sures, the administered doses are usually averaged over a day or single
exposure event. For chronic noncancer effects, the time period used is
often the actual period of exposure (e.g., length of occupational expo-
sure). The objective in selecting the averaging time is to express the
exposure in a way that makes it comparable to the dose-response relation-
ship used in conjunction with the exposure estimate to calculate risk.
The body weight used to calculate the total exposure in the above
equation should reflect the average weight of the exposed population
during the time when the exposure actually occurs. If the exposure occurs
continuously throughout an individual's life or only during the adult
ages, using an adult average weight should provide sufficient accuracy.
2-8

-------
However, when the exposure is limited to childhood, the weight represent-
ing those ages should be used. More detailed information for these expo-
sure parameters is presented in the Exposure Factors Handbook (USEPA
1989).
2.2.6 Estimation of Integrated Exposure
Integration of exposures is the final step in a multimedia exposure
assessment. It usually involves summation of exposures within a given
exposure route (dermal, inhalation, ingestion) across n relevant
exposure pathways (e.g., ambient air, surface water, ground water, food).
In other words,
n
Total route-specific exposure = £ Exposure ^	(2-3)
i=l
Multimedia exposures are frequently not integrated across multiple
exposure routes. This is true for the following reasons:
•	Some combinations of exposure routes are not realistic; for
example, some persons exposed via ambient air near the site may not
be the same individuals who are exposed via drinking water further
away.
•	Different exposure routes for a chemical are often associated
with different absorption rates.
•	Some risk assessments to which exposure estimates provide inputs
must be performed on a route-specific basis (because of differences
in the potency and toxicity endpoint of some compounds according
to exposure route).
•	EPA's regulatory needs often require separate exposure-pathway-
specific estimates to facilitate formulation and evaluation of
control options for specific environmental media.
In other words, the potential application of the exposure numbers dictates
the manner in which exposures are to be integrated.
Assessment of total human exposure is an emerging science that focuses
on the activities of populations and individuals that bring them into
contact with the chemicals of concern. The methodology used in this
2-9

-------
Handbook can be used to estimate the concentrations of chemicals and
resulting exposures at human physical boundaries (e.g., skin, lungs,
gastrointestinal tract), regardless of whether the chemicals arrive at the
receptor via air, water, food, or other exposure pathways.
Total human exposure assessment methodologies are often limited by
the simplistic nature of developed exposure scenarios and frequently-used
default values for parameters. Such integrated approaches may not always
account specifically for variations in exposure parameters or for the
multitude of microenvironments in which people are exposed via a given
exposure pathway and a given exposure route on a daily basis. Prioritiza-
tion of microenvironments for an exposure assessment is dependent upon
human activities (i.e., durations and frequencies of exposures in differ-
ent microenvironments), pollutant concentrations encountered in those
microenvironments, and exposure parameters such as inhalation rate that
vary according to microenvironments and associated activities.
Initially, examination of human activity patterns can identify candi-
date microenvironments. Knowledge of the amount of time individuals spend
performing various activities in specific microenvironments during the
course of a day can provide estimates of the duration and frequency of
exposure. Thus, time-use studies (see Exposure Factors Handbook (USEPA
1989)) provide the type of information needed to initially prioritize
microenvironments for selecting exposure pathways to be integrated. When
activity data are combined with pollutant concentration data, a clear
profile of the relative importance of various exposure pathways becomes
apparent; an illustration of this is shown for inhalation exposure to
nitrogen dioxide (NC^) in Figure 2-2. When such information is not
readily available, a conservative approach that accounts for all potential
exposure routes should be taken.
Equations to be used in calculating exposure for each exposure medium
and route in the final integrated assessment are contained in the Exposure
Factors Handbook (USEPA 1989). The exposure assessor is referred to the
2-10

-------
MICRO ENVIRONMENTS
TOTAL HUMAN EXPOSURE —
TO AIR POLLUTION ~~
POLLUTANT EXPOSURE
IN RESIDENCES
POLLUTANT EXPOSURE
IN OUTDOOR AIR
POLLUTANT EXPOSURE
IN BUILDINGS
(OFFICE. SCHOOL, ETC.)
POLLUTANT EXPOSURE
IN VEHICLES
(CAR, PLANE, TRAIN, ETC )
POLLUTANT
CONCENTRATION
(ppb)
140
I 20- —
100
rv>
i
-120
-100
,4°	POLLUTANT
CONCENTRATION
(PPt>)
0000 0300 0600 0900 1200 1500 1800 2100 2400
TIME (HOUR OF THE DAY)
24-HOUR EXPOSURE PROFILE
SUMMARY OF MICROENVIRONMENT
CONTRIBUTION TO TOTAL EXPOSURE
(SHADED AREA IS TOTAL INDOOR AIR CONTRIBUTION)
Figure 2-2. Representations of microenvironment contributions to total inhalation exposure to nitrogen dioxide.
Source: Adapted from Sexton and Ryan (1987); exposure data are for N02-

-------
Exposure Assessment Worksheets in Appendix B for the step-by-step
procedure for estimating exposures using algorithims contained within
this Handbook and to Appendix C for completed worksheets for an example
chemical.
2-12

-------
3.	ASSESSMENT OF EXPOSURE VIA SURFACE WATER
3.1 Release Analysis - Surface Water
People can be exposed to chemical contaminants in surface water that
result from many sources. Contaminants may be delivered to surface water
in precipitation or through dry deposition (i.e., dust) from the
atmosphere. Contaminated ground water may discharge into the surface
water body of concern. Surface runoff from contaminated sites may carry
contaminated soils or dissolved contaminants to surface water. On-site
treatment processes, where they exist, may result in direct discharge of
contaminants, via the treated effluent, into the surface water body.
This section will emphasize surface runoff releases as the major source
of contaminant entry into surface water.
3.1.1 Screening of Release Potential
In order for chemical contaminants to be available for surface runoff,
they must first be present in the upper layer (1 cm) of soil; if they are
deeper than 1 cm, the contaminants will likely adsorb in various amounts
to the soil or be leached to ground water. Contaminants may be present
in soil because of (1) unintentional releases, such as spills, leaks, or
container failure; (2) intentional surface application of chemical sub-
stances, such as occurs during landfarming of wastes; or (3) deposition
from atmospheric sources. Hydrophobic substances, in particular, can be
expected to readily sorb to soils on site; substances adsorbed to environ-
mental matrices such as soil make up the largest proportion of substances
of concern found at uncontrolled hazardous waste sites.
Once the contaminants are present in the upper soil layer, they are
subject to a variety of fate processes (e.g., volatilization, photolysis)
that may reduce their concentrations in the soil (see Section 8.2). The
remaining contaminants, however, will be available for runoff during
storm events. They will be present in the runoff stream either adsorbed
to suspended soils or dissolved in the runoff, depending on the
individual properties of the chemicals involved.
3-1

-------
3.1.2 Quantitative Release Estimation
Once it has been determined that the contaminants are present in the
upper soil layer and are subject to removal via runoff, it is necessary to
estimate quantitatively the amount of contaminants that will migrate off
site.
The amount of a given hydrophobic compound released in surface runoff
can be estimated using the Modified Universal Soil Loss Equation (MUSLE)
and sorption partition coefficients derived from the compound's
octanol-water partition coefficient. An estimation of the amount of soil
eroded in a storm event of given intensity can be obtained using MUSLE.
By using the sorption partition coefficients, one can estimate the
amounts of contaminants carried along with the soil and the amounts
carried in dissolved form.
(1) Soil loss calculation. The Modified Universal Soil Loss
Equation (Williams 1975), as presented in Mills et al. (1982), is given
in Equation 3-1. Equations 3-2 through 3-5 are provided to guide calcula-
tion of certain input parameters required to utilize Equation 3-1.
Y(S)e = a(Vrqp)°-56KLSCP	(3-1)
where
Y(S)f = sediment yield (tons per event, metric tons per event)
a = conversion constant (95 English, 11.8 metric)*
Vr = volume of runoff (acre-feet, m3)
qc = peak flow rate (cubic feet per second, nr/sec)
K = the soil-erodibility factor; a value of K for a given site
can be obtained from the local Soil Conservation Service
office. This should be treated as a dimensionless number.**
L = the slope-length factor (dimensionless ratio)
S = the slope-steepness factor (dimensionless ratio)
* Metric conversions presented in the following runoff contamination
equations are from Mills et al. (1982).
** In other situations, K is commonly expressed in terms of tons per
acre per dimensionless rainfall erodibility unit. In this equation,
the dimensionless number is treated the same regardless of whether
other parameters are expressed in English or metric units.
3-2

-------
C = the cover factor (dimensionless ratio = 1.0 for bare soil;
see the following discussion for vegetated site "C" values)
P = the erosion control practice factor (dimensionless ratio =
1.0 for uncontrolled hazardous waste sites).
Storm runoff volume is calculated as follows (Mills et al. 1982):
Vr = aAQr	(3-2)
where
Vr = storm runoff volume
a = conversion constant (0.083 English, 100 metric)
A = contaminated area (acres, ha)
Qr = depth of runoff (in, cm).
Depth of runoff, Qr, is determined by (Mockus 1972):
Qr = (Rt - 0.2Sw)2/(Rt + 0.8SW)	(3-3)
where
Qr = the depth of runoff from the watershed area (in, cm)
Rt = the total storm rainfall (in, cm)
Sw = water retention factor (in, cm).
The value of Sw, the water retention factor, is obtained as follows
(Mockus 1972):
Sw = poafi - lOj a	(3-4)
where
Sw = water retention factor (in, cm)
CN = the SCS Runoff Curve Number (dimensionless, see Table 3-1)
a = conversion constant (1.0 English, 2.54 metric).
The CN factor is determined by the type of soil at the site, its
condition, and other parameters that establish a value indicative of the
tendency of the soil to absorb and hold precipitation or to allow precipi-
tation to run off the surface. The analyst can obtain CN values of uncon-
trolled hazardous waste sites from Table 3-1. If further information on
CN values is desired, particularly for agricultural situations, the
analyst can refer to Novotny and Chesters (1981) for additional CN values
3-3

-------
/ uccn
Table 3-1. Runoff Curve Numbers
	Site type	
Overall Road/right of May Meadow	Woods
Soil group	Description	site3
Lowest runoff potential: Includes	59	74	30	45
deep sands with very little silt and
clay; also deep, rapidly permeable
loess (infiltration rate = 8-12 mn/h).
B	Moderately low runoff potential:	74	84	58	66
Mostly sandy soils less deep than A,
and loess less deep or less aggregated
than A. but the group as a whole has
above-average infiltration after
thorough wetting (infiltration
rate = 4-8 mm/h).
C	Moderately high runoff potential:	82	90	71	77
Conprises shallow soils and soils
containing considerable clay and
colloids, though less than those of
group D. The group has below-
average infiltration after presatura-
tion (infiltration rate = 1-4 m/h).
0	Highest runoff potential: Includes	86	92	78	83
mostly clays of high swelling percent,
but the group also includes scrae
shallow soils with nearly impermeable
subhorizons near the surface
(infiltration rate = 0-1 nq/h).
Source: Adapted from Schwab et al. 1966.
aValues taken from farmstead category, which is a conposite including buildings, farmyard, road, etc.
3-4

-------
and more background information. The local Soil Conservation Service
office may also be helpful.
The peak runoff rate, qp, is determined as follows (Haith 1980):
a - aARtQr	<3-5)
p VRt-°-2V
where
qp = the peak runoff rate (ft3/sec, m3/sec)
a = conversion constant (1.01 English, 0.028 metric)
A = contaminated area (acres, ha)
Rj- = the total storm rainfall (in, cm)
Qr = the depth of runoff from the watershed area (in, cm)
Tr = storm duration (hr)
Sw = water retention factor (in, cm).
The soil erodibility factor, K, is indicative of the erosion potential
of given soil types; it is, therefore, highly site specific. Values for
soil erodibility factors are given in Table 3-2. Values for K for sites
under study are available from the local Soil Conservation Service
office. Additional information on determining values for K can also be
obtained from Wischmeier et al. (1971).
The slope length factor, L, and slope steepness factor, S, are common-
ly entered into the MUSLE as a combined topographic factor, LS. This can
be obtained from Figures 3-1 through 3-3.
The vegetative cover factor, C, is determined by the amount and type
of vegetative cover present at the site. For bare soils a value of "1"
is used. Tables 3-3 and 3-4 should be used to obtain C values for sites
with vegetative covers.
The factor, P, refers to any erosion control practices that may be
used on the site. These practices include contouring, compacting, estab-
lishing sediment basins, and other control practices. Some representative
values of P for agricultural land and for construction sites are given in
Tables 3-5 and 3-6, respectively. A conservative, worst-case value of "1"
should be used for abandoned or uncontrolled hazardous waste sites, where
erosion control practices are unlikely to be used.
3-5

-------
7622H
Table 3-2. Magnitude of Soil Erodibility Factor Ka
K for organic matter content (X)
Textural class	<0.5	2	4
Sand	0.05	0.03	0.02
Fine sand	0.16	0.14	0.10
Very fine sand	0.42	0.36	0.28
Loamy sand	0.12	0.10	0.08
Loamy fine sand	0.24	0.20	0.16
Loamy very fine sand	0.44	0.38	0.30
Sandy loam	0.27	0.24	0.19
Fine sandy	loam 0.35	0.30	0.24
Very fine sandy loam	0.47	0.41	0.33
Loam	0.38	0.34	0.29
Silt loam	0.48	0.42	0.33
Silt	0.60	0.52	0.42
Sandy clay	loan 0.27	0.25	0.21
Clay loam	0.28	0.25	0.21
Silty clay	loam 0.37	0.32	0.26
Sandy clay	0.14	0.13	0.12
Silty clay	0.25	0.23	0.19
Clay	0.13-0.2
a The values shown are estimated averages of broad ranges of
specific soil values. When a texture is near the borderline
of two texture classes, use the average of the two K values
(Novotny and Chesters 1981).
3-6

-------

9uti
e J-i
«O0 S/^
6°0
°°
not?.
So,
Urce: he das^
^'°Pe
'**£ »s
Hins
3.7
fo.

shi.
et
^u]
n9t,
("Lbev
on.
*1. '¦ »ey0nd
Or,
'e9on.
<32)
"
-------
Figure 3-2. Soil moisture-soil temperature regimes of the Western United States.
Source: USDA (1974).
3-8

-------
Slope Length, Meters
3.5 6.0 10	20	40 60 100	200 4O0 6O0
Slope Length, Feet
Figure 3-3. Slope effect chart for areas where Figure 3-2 is not applicable.
NOTE: The dashed lines represent estimates for slope dimensions beyond the range of lengths
and steepnesses for which data are available.
Source: USDA (1974).
3-9

-------
Table 3-3. "C" Values for Permanent Pasture. Rangeland, and Idle Land
Vegetation canopy	Canopy 	Cover that contacts the surface
(type and height	coverc 	Percent ground cover
of raised canopy)b	(X) Type 0	20	40	60	80	95-100
No appreciable canopy

G
0.45
0.20
0.10
0.042
0.013
0.003


W
0.45
0.24
0.15
0.090
0.043
0.011
Canopy of tall weeds
25
G
0.36
0.17
0.09
0.038
0.012
0.003
or short brush

W
0.36
0.20
0.13
0.082
0.041
0.011
(0.5 m fall height)
50
G
0.26
0.13
0.07
0.035
0.012
0.003


U
0.26
0.16
0.11
0.075
0.039
0.011

75
G
0.17
0.10
0.06
0.031
0.011
0.003


U
0.17
0.12
0.09
0.067
0.038
0.011
Appreciable brush
25
G
0.40
0.18
0.09
0.040
0.013
0.003
or brushes

W
0.40
0.22
0.14
0.085
0.042
0.011
(2 m fall height)
50
G
0.34
0.16
0.085
0.038
0.012
0.003


W
0.34
0.19
0.13
0.081
0.041
0.011

75
G
0.28
0.14
0.08
0.036
0.012
0.003


W
0.28
0.17
0.12
0.077
0.040
0.011
Trees but no appreciable
25
6
0.42
0.19
0.10
0.041
0.013
0.003
low brush

W
0.42
0.23
0.14
0.087
0.042
0.011
(4 di fall height)
50
6
0.39
0.18
0.09
0.040
0.013
0.003


U
0.39
0.21
0.14
0.085
0.042
0.011

75
G
0.36
0.17
0.09
0.039
0.012
0.003


W
0.36
0.20
0.13
0.083
0.041
0.011
Source: Wisctaeier (1972).
a All values shorn) assme (1) randan distribution of gulch or vegetation and (2) mulch of appreciable
depth where it exists.
^ Average fall height of waterdrops from canopy to soil surface: m = meters.
c Portion of total area surface that would be hidden from view by canopy in a vertical projection
(a bird's-eye view).
^ 6: Cover at surface is grass, grass-like plants, decaying contacted duff, or litter at least 5 am
(2 in) deep.
V: Cover at surface is mostly broadleaf herbaceous plants (as weeds) with little lateral root network
near the surface and/or undecayed residue.
3-10

-------
IbidW
Table 3-4. "C" Values for Woodland
Stand condition
Tree canopy
percent of
area3
Forest
litter
percent of
b
area
Undergrowth0
"C" factor
Uell stocked
100-75
100-90
Managed''
Umanaged*'
0.001
0.003-0.011
Nedinn stocked
70-40
85-75
Managed
Unmanaged
0.002-0.004
0.01-0.04
Poorly stocked
35-20
70-40
Managed
Unmanaged
0.003-0.009
0.02-0.09e
Source: Wischmeier (1972).
a When tree canopy is less than 20 percent, the area will be considered as grassland or
cropland for estimating soil loss.
^ Forest litter is assuoed to be at least 2 inches deep over the percent ground surface
area covered.
c Undergrowth is defined as shrubs, weeds, grasses, vines, etc., on the surface area
not protected by forest litter. Usually found under canopy openings.
^ Managed - grazing and fires are controlled.
Unmanaged - stands that are overgrazed or subjected to repeated burning.
e For unmanaged woodland with litter cover of less than 75 percent. C values should
be derived by taking 0.7 of the appropriate values in Table 3-3. The factor of 0.7
adjusts for much higher soil organic matter on permanent woodland.
3-11

-------
7622H
Table 3-5. Values of P for Agricultural Land
Strip cropping and terracing
Slope (X)
Contouring
Alternate meadows
Close-grown crops
1.1-2.0
0.6
0.30
0.45
2.1-7.0
0.5
0.25
0.40
7.1-12.0
0.6
0.30
0.45
12.1-18.0
0.8
0.40
0.60
18.1-24.0
0.9
0.45
0.70
Source: Novotny and Chesters (1981).
3-12

-------
7622H
Table 3-6. Values of P for Construction Sites
Erosion control practice	P
Surface condition with no cover
Conpact, smooth-scraped with bulldozer or scraper
up and down hill
Sane as above, except raked with bulldozer and
root-raked up and down hill
Contact, smooth-scraped with bulldozer or scraper
across the slope
Sane as above, except raked with bulldozer and
root-raked across slope
Loose as a disked plow layer
Rough irregular surface, equipment tracks in all
directions
Loose with rough surface >0.3 m depth
Loose with smooth surface >0.3 m depth
Structures
Small sediment basins:
0.09 basins/ha	0.50
0.13 basins/ha	0.30
Downstream sediaent basins:
With chemical flocculants	0.10
Without chemical flocculants	0.20
Erosion control structures:
Normal rate usage	0.50
High rate usage	0.40
Strip building	0.75
Source: Novotny and Chesters (1981).
1.30
1.20
1.20
0.90
1.00
0.90
0.B0
0.90
3-13

-------
(2) Dissolved/sorbed contaminant loading. Once the storm event soil
loss has been calculated, the analyst can predict the degree of soil/water
partitioning expected by using the following equations. The amounts of
adsorbed and dissolved substances are first determined, using Equa-
tions 3-6 and 3-7, as adapted from Haith (1980).
Ss = [1/(1 + ec/Kd/j)] (Ci) (A)	(3-6)
and
Ds = [1/(1 + Kd0 /ec)] (Ci) (A)	(3-7)
where
Ss = sorbed substance quantity (kg, lb)
ec = available water capacity of the top cm of soil (difference
between wilting point and field capacity) (dimensionless)
Kd = soil partition coefficient (cm3/g)
j8 = soil bulk density (g/cm3)
= total substance concentration (kg/ha-cm, lb/acre-cm)
A = contaminated area (ha-cm, acre-cm). (Actually a volume; major
assumption is that contamination in upper 1 cm is available for
release.)
Ds = dissolved substance quantity (kg, lb).
The soil partition coefficient, K^, for a given chemical can be
determined from known values of certain other physical/chemical parame-
ters, primarily the chemical's octanol-water partition coefficient,
solubility in water, or bioconcentration factor. Lyman et al. (1982)
present the regression equations that enable the analyst to determine
sorption coefficients for specified groups of chemicals (e.g., herbicides,
polynuclear aromatics). Parameter values required by the appropriate
equations but not available in chemical reference literature can be
estimated according to procedures described in Lyman et al. (1982). In
particular, the octanol-water partition coefficient can be estimated
based on the molecular structure of the substance.
After calculating the amount of sorbed and dissolved contaminant, the
assessor calculates the loading to the receiving water body as follows
(adapted from Haith 1980):
3-14

-------
PX-j = [Y(S)E/(aA/3)] ss	(3-8)
arid
PQi = [Qr/Rt] Ds	(3-9)
where
PXj = sorbed substance loss per event (kg, lb)
Y(S)e = sediment yield (tons per event, metric tons per event)
a = conversion constant (100 metric, 0.714 English)
A = contaminated area (ha-cm, acre-cm)
p = soil bulk density (g/cm3, lb/ft3)
Ss = sorbed substance quantity (kg, lb)
PQ-j = dissolved substance loss per event (kg, lb)
Qr = total storm runoff depth (in, cm)
Rj- = total storm rainfall (in, cm)
Ds = dissolved substance quantity (kg, lb).
PX.j and PQ.j can be converted to mass per volume terms for use in esti-
mating contaminant concentration discharging to the receiving water body
by dividing by the site storm runoff volume (Vf, see Equation 3-2).
(3) Total release. The total contaminant release to a particular
surface water body can be estimated by summing the various inputs of each
source and contaminant to that water body. Surface runoff inputs, both
sorbed and dissolved, can be estimated as explained in the preceding
section. Ground-water discharges into surface water can be quantified
according to the methods outlined in Section 4.2.2. Inputs from on-site
treatment processes, where they exist, that discharge into surface water
can be determined from monitoring data or from the discharger's NPDES
permit.
3.2 Fate Analysis - Surface Water
3.2.1 Screening of Surface Water Fate Processes
Screening of environmental fate processes provides a qualitative
assessment of contaminant transport and transformation processes. The
fate processes of organic contaminants in surface water are discussed
quantitatively in the following section. If the effects of transport
3-15

-------
processes, such as sorption, volatilization, and sedimentation, can be
determined in qualitative terms, one can make a more accurate estimation
of the influence of transformation processes. For example, if the domi-
nant transport process of a contaminant is sorption to suspended or bottom
sediments, it is likely that biodegradation will have a greater effect on
the chemical fate of the contaminant than volatilization and photolysis.
Mills et al. (1982) summarized the relative importance of transport and
transformation processes affecting the aquatic fate of organic priority
pollutants; this information is presented in Table 3-7.
Several generalizations can be made to aid the assessor in a prelimi-
nary screening of surface water fate processes. Sorption of a contaminant
to suspended sediments and subsequent sedimentation of the sorbate-sorbent
(contaminant-sediment) complex can effectively remove the pollutant from
the water column. Resuspension of sediments and dissolution of the
pollutant from bottom sediments are processes that must be considered if
sediments are not transported from the site of contamination.
Hydrophobic organic chemicals tend to sorb to suspended material.
These chemicals typically have a high octanol-water partition coefficient
(Kqw) value and a low solubility in water. Table 3-8 provides
estimates of the fraction of chemicals adsorbed to suspended sediments in
surface waters based on Koc, the organic carbon partition coefficient.
Chemicals with large KQC values will tend to accumulate in the
sediments over time.
Another transport process that can lead to the removal of a chemical
from surface water is volatilization. The chemical parameter that most
significantly influences volatilization of a chemical is Henry's law,
Kh. Henry's law can be defined as:
equilibrium partial pressure of chemical in
=	atmosphere above the water (atm)	(3-10)
equilibrium concentration of chemical in
the water (mole/nr)
3-16

-------
/ oi un
Table 3-7. Relative Importance of Processes Influencing Aquatic
Fate of Organic Priority Pollutants
Compound
Process
PESTICIDES
Acrolein
Aldrin
Chlordane
ODD
DDE
DDT
Dieldrin
Endosulfan and Endosulfan Sulfate
Endrin and Endrin Aldehyde
Heptachlor
Heptachlor Epoxide
Hexachlorocyclohexane	isomers)
-Hexachlorocyclohexane (Lindane)
Isophorone
TCDD
Toxaphene
+
7
+
7
7
E
"S
.1
CD
PCBs and RELATED COHPOUHDS
Polychlorinated Biphenyls
2-Chloronaphthalene
HALOGENATED AND ALIPHATIC HYDROCARBONS
Chloromethane (methyl chloride)
Dichlororaethane (methylene chloride)
Trichloromethane (chloroform)
Tetrachloranethane (carbon tetrachloride)
Chloroethane (ethyl chloride)
1.1-Dichloroethane	(ethylidene chloride)
1.2-Dichloroethane	(ethylene dichloride)
1.1.1-Trichloroethane	(methyl chloroform)
1.1.2-Trichloroethane
1,1,2,2-Tetrach1oroethane
Hexachloroethane
Chloroethene (vinyl chloride)
1.1-Dichloroethene	(vinylidene chloride)
1.2-trans-Dichloroethene
Trichloroethene
+
+
+
3-17

-------
767 OH
Table 3-7. (continued)
Compound
Process
Tetrachloroethene (perchloroethylene)
1.2-Dichloropropane
1.3-Oichloropropene
Hexachlorobutadiene
Hexachlorocyclopentad iene
Branomethane (methyl bromide)
Branod ichloromethane
D i broroochloromethane
Tribromomethane (bromoform)
0ich1orodif1uoronethane
Trichlorof1uoranethane
8
U
o
I

O
L.
HAL06ENATED ETHERS
Bis(chloromethyl) ether
Bis(2-chloroethyl) ether
Bis(2-chloroisopropyl) ether
2-Chloroethyl vinyl ether
4-Chlorophenyl phenyl ether
4-Branophenyl phenyl ether
Bis(2-chloroethoxy) methane
MOHOCfCLIC AR0HAT1CS
Benzene
Chlorobenzene
1.2-Dichlorobenzene	(o-dichlorobenzene)
1.3-Dichlorobenzene	(m-dichlorobenzene)
1.4-Dichlorobenzene	(jg-dichlorotoenzene)
1,2,4-Trichlorobenzene
Hexach1orobenzene
Ethylbenzene
Nitrobenzene
Toluene
2.4-Dinitrotoluene
2,6-Dinitrotoluene
Phenol
2-Chlorophenol
+
+
+
+
+
+
+
7
+
+
3-18

-------
Table 3-7. (continued)
Caipound
Process
2
&
O
4)
>
O
L.
"2
2,4-Dichlorophenol
2,4,6-Tnchlorophenol
Pentachloropheno1
2-Nitrophenol
4-Nitrophenol
2,4-Dinitrophenol
2,4-Dimethyl phenol (2,4-xylenol)
£-Chloro-m-creso1
4,6-Dinitro-o-cresol
PHTHALATE ESTERS
Dimethyl phthalate
Diethyl phthalate
Di-n-butyl phthalate
Di-n-octyl phthalate
Bis(2-ethylhexyl) phthalate
Butyl benzyl phthalate
P0LTCTCL1C AROMATIC HYDROCARBONS
Acenaphthenec
Acenaphthylenec
Fluorene0
Naphthalene
Anthracene
Fluoranthenec
Phenanthrenec
Benzo(a)anthracene
Benzo(b)fluoranthene0
Benzo(k)fluoranthene0
Chrysenec
Pyrenec
Benzo(ghiJperylenec
Benzo(a)pyrene
D i benzo(a,h)anthracenec
Inderto(1.2,3-cd)pyrenec
+
+
+
+
+
~
+
+
+
+
+
+
+
+
+
+
NITR0SAH1NES AND MISC. COMPOUNDS
DimethyInitrosamine
Dipheny1nitrosaaine
Di-n-propyl nitrosaniine
3-19

-------
7670H
Table 3-7. (continued)
Carpound
Process

|
J

*4
m
i
0

u
p
Q
U
19
1
O
I/)
O
>
a
o
£
E
"5
Benzidine
3,3'-Dichlorobenzidine
1,2-Diphenylhydrazine (Hydrazobenzene)
Acrylonitrile
Key to Symbols:
++ Predominant fate determining process.
+ Could be an inportant fate process.
- Not likely to be an important process.
? Inportance of process uncertain or not known.
Notes
a Biodegradation is the only process knom to transform polychlorinated biphenyls under
enviroranenta1 conditions, and only the lighter compounds are measurably biodegraded. There is
experimental evidence that the heavier polychlorinated biphenyls (five chlorine atons or more
per molecule) can be photolyzed by ultraviolet light, but there are no data to indicate that
this process is operative in the enviroraaent.
b Based on infomation for 4-nitrophenol.
c Based on infonutin for PAHs as a group. Little or no information for these canpounds exists.
Source: Mills et al. (1982).
3-20

-------
7671H
Table 3-8. Estimated Adsorption of Chemicals to Suspended
Solids in Surface Waters3
Fraction adsorbed
10	0.00002
100	0.0002
1,000	0.002
10.000	0.02
100,000	0.17
1,000.000	0.67
Assures fraction organic content of suspended solids = 0.01 and mass
fraction of solids in Mater = 0.0002.
Source: USEPA (1985a).
3-21

-------
-3	3
In general, if is >10 (atm-m /mole), the chemical will
probably volatilize from water with a half-life of less than 4 hours. If
-7	3
Kh is <10 (atm-m /mole), the chemical will probably volatilize
only slowly from surface water, with a half-life greater than 1 year
(USEPA 1985a).
Photolysis is a decay process involving the absorption of sunlight by
pollutants and initiation of a chemical reaction. The susceptibility of
a pollutant to direct photolysis can be determined from the chemical's
absorption region. Many references report values for a*, the wave-
lengths of maximum absorption for organics. If [a* < = 270 nm or a* > =
730 nm], then direct photolysis need not be considered an important fate
process (Mills et al. 1982). Lyman et al. (1982) report absorption
spectra and a* values for many organic compounds. In general,
alcohols, ethers, amines, and nitriles (cyanides) are likely to be poor
absorbers of sunlight, and for these groups, photolysis may not be an
important fate process (Mills et al. 1982).
Nonbiological oxidation of organic molecules is not a common fate
determining process in surface waters. The reaction of contaminants with
surface water oxidants may warrant consideration for aromatic amines,
phenols, and electron-rich molecules such as branched olefins, certain
amines, polycyclic aromatics, and sulfides (USEPA 1985a).
Hydrolysis, the reaction of an organic molecule with water to form two
new chemicals, is an important chemical transformation process for many
surface water contaminants. Organic molecules containing functional
groups that can react with the water molecule are susceptible to hydroly-
sis; examples of classes of these functional groups are listed in
Table 3-9. Examples of classes of functional groups that are likely to
be resistant to hydrolysis are shown in Table 3-10.
Biodegradation can be considered a decay process when assessors
screen the chemical fate of an organic chemical in surface waters because
organic chemicals provide a carbon source for microbial populations.
According to Mills et al. (1982), it is not yet possible to predict
3-22

-------
7671H
Table 3-9. Classes of Organic Functional 6roups
Potentially Susceptible to Hydrolysis
Alkyl and benzyl ha 1 ides
Amides
Carbamates
Carboxylic acid esters
Epoxides
Phosphonic acid esters
Phosporic acid esters
Sulfonic acid esters
Sulfuric acid esters
Source: USEPA (1985a).
3-23

-------
7 671H
Table 3-10. Classes of Organic Functional Groups
Potentially Resistant to Hydrolysis
Alkanes
Alkenes
Alkynes
Benzenes/biphenyls
Polycyclic aromatic hydrocarbons
Heterocyclic polycyclic aromatic hydrocarbons
Halogenated aromatics/PCBs
Dieldrin/aldrin and related halogenated hydrocarbon pesticides
Ararat ic nitro confounds
Amines
Alcohols
Phenols
Glycols
Ethers
Aldehydes
Ketones
Carboxylic acids
Sulfonic acids
Nitriles
Source: USEPA (1985a).
3-24

-------
whether an organic chemical will serve as a source of energy and carbon,
based on its chemical structure. However, several generalizations are
helpful in determining the potential influence of decay processes in
aquatic systems. Highly branched compounds may be more resistant to
biodegradation than unbranched compounds, and short alkyl chains are less
easily degraded than are long chains. Polynuclear aromatic hydrocarbons
are resistant to biodegradation resulting from insolubility and branching
of the molecule. Susceptibility to biodegradation appears to decrease as
aliphatic hydrocarbons become large (USEPA 1985a). Biodegradation can be
assumed to occur at only very slow rates for heavily chlorinated or
halogenated compounds (Mills et al. 1982). Introduction of o-methyl,
nitro, and sulfonate groups onto a molecule may also significantly reduce
the rate of biodegradation (USEPA 1985a).
3.2.2 Quantitative Fate Estimation
Once a hazardous substance is introduced into a water body, the com-
pound may be subject to transport or transformation processes. Transport
processes include sorption, volatilization, and sedimentation. These
processes tend to distribute compounds among the atmospheric, aquatic,
and sediment compartments, depending on the chemical characteristics of
the substance and the physical and chemical parameters of the water body.
Transformation processes taking place within each environmental
compartment, or phase, chemically alter contaminants by photolysis,
oxidation, hydrolysis, and biodegradation. The chemical products of
these processes may be of lesser, equivalent, or greater toxicity than
the parent compound.
The three major classifications of surface water bodies are rivers and
streams, impoundments, and estuaries. Algorithms presented below that
apply to a specific type of water body are identified as such. All other
algorithms presented represent general models by which the assessor can
quantify important aquatic fate processes for a contaminant in surface
water.
3-25

-------
It must be stressed that the processes that influence the environmen-
tal fate of pollutants are complex. Some environmental factors that
influence environmental fate, such as temperature, vary greatly from site
to site and season to season, leading to a tremendous range of variability
in the actual rate constants for fate and transformation processes. Thus,
the algorithms below should be thought of as providing only a very approx-
imate depiction of environmental fate of chemicals in surface water.
Their greatest value is in providing a relative view of environmental fate
for different chemicals, rather than an absolute view.
The determination of rate constants to describe transformation
processes, such as photolysis, oxidation, hydrolysis, and biodegradation,
requires more chemical and environmental data. The assessor should con-
sult Table 3-7 for a rough estimate of the relative importance of aquatic
fate processes for a particular chemical before determining rate constants
for all transformation processes, since only one or two fate processes
may "drive" the overall rate of transformation. Rate constants and
related factors can be calculated as described below.
(1) Transport processes.
(a) Sorption. Sorption of nonvolatile, hydrophobic organic
compounds to suspended materials or bottom sediments is an important
removal process in aquatic systems. The key environmental parameters
that affect the degree of sorption of a substance on suspended sediments
are the octanol-water coefficient (KQW)> the solubility of the
substance in water, the suspended sediment concentration, and the organic
carbon content of the suspended sediment (Mills et al. 1982). The
tendency of a dissolved pollutant to be sorbed can be expressed in terms
of soil/ solution partition coefficient, derived from the Freundlich
equation:
(3-11)
3-26

-------
where
x/m = ^g of chemical adsorbed/g of sediment (or comparable units)
Kj = soil/solution partition coefficient
C = concentration of chemical (in units comparable with x/m) in
aqueous phase
1/n = exponential factor; can be assumed to = 1 if measured value not
available (Lyman et al. 1982).
Dividing by the fractional organic carbon content in the sediment
(OC) gives the partition coefficient expressed on an organic carbon basis
(Koc), which is characteristic of the organic contaminant and
independent of the sediment:
Koc = Kd/0C	(3-12)
Methods are also available for estimating K„ from K„,, when Kj
3 oc	ow	d
cannot be determined (see Section 4.2.2).
The relative amount of contaminant sorbed to sediments and dissolved
in the water column depends on the partition coefficient and the
suspended sediment concentration. At equilibrium, the relationship
between dissolved and sorbed phase contaminant concentrations (as
described by Mills et al. 1982) can be approximated by the following
equation:
	1
rs = 1 *
1 + (Kqc . S . OC)
(3-13)
where
rs = fraction of chemical removed from water column because of
sorption onto suspended sediments (unitless)
Koc = °r9anic carbon partition coefficient (unitless)
S = mass fraction of suspended sediments in water (unitless)
OC = fraction of organic carbon in suspended sediments (unitless)
If site-specific values for mass fraction of suspended sediments in water
and fraction of organic carbon in suspended sediments are not available,
default values of 0.002 and 0.01 may be used, respectively.
3-27

-------
(b) Volatilization. Factors that control volatilization of
organic contaminants from surface water include the solubility,
molecular weight, and vapor pressure of the compound, as well as the
turbulence of the air-water interface, wind speed, and mixing depth of
the water body.
The two-film theory approach can be used to describe the rate of
volatilization of an organic compound. The following equations (from
Callahan et al. 1979) provide a means to estimate the volatilization rates
of chemicals from aqueous systems:
d[CJ
Rv —= W	<3"I4>
and
where

i +	BI	
(0.001) kHkg J
and
where
Di
= liquid phase diffusion coefficient (hr)"1
s~£ = liquid phase boundary layer thickness (cm)
Dg = gas phase diffusion coefficient (hr)"1
(3-15)
Rv =	volatilization rate of chemical (moles liter"* hr"1)
Cw =	concentration of C in water (mole liter"1)
kv =	volatilization rate constant (hr"1)
L =	depth (cm)
k^ =	liquid phase mass transport coefficient (cm hr"1)
=	Henry's Law constant (atm-nr/mole)
kg =	gas phase mass transport coefficient (cm hr"1)
R =	gas constant (liter-atm. mole"1 deg"1)
T =	temperature (deg. Kelvin).
For both the gas and liquid phases,
kjg =	(3-16)
kg = Dg/5g	(3-17)
3-28

-------
6g = gas phase boundary layer thickness (cm)
A simplification of the above equations is provided by Callahan et al.
3
(1979). For highly volatile compounds, where > 3.0 atm-m /mole,
Rv is limited by diffusion through the liquid phase boundary layer and
can be determined by the value of k^. The equation presented below
applies to a wide range of environmental conditions, and should be used
to estimate the volatilization rate of highly volatile organics from
surface waters (Callahan et al. 1979):
env - ("X'lab en,	I3"18)
where
r
= volatilization rate constant for the chemical in the laboratory
ky = oxygen reaeration rate constant in the laboratory or the
or environment (hr-1)
oxygen reaeration r;
environment (hr-1).
Delos et al. (1984) note that among substances with a high Henry's law
constant (K^), the volatilization rate is independent of and
dependent on the substance's water diffusivity. Therefore, volatilization
rates of all highly volatile pollutants are essentially the same since
there is relatively little variance between the diffusivities of these
compounds.
For chemicals characterized by low volatility, where <
0.01 atm-m3/mole, Equation 3-15 can be simplified to:
.001 Kh kn
kv - "Trt	<3-19>
In this case, the volatilization rate constant is limited by diffusion
7	3
through the gas phase boundary layer. If K^< 10 atm-m /mole, the
assessor need not consider the turbulence of the water body; surface
transfer will be slow, and the rate will decrease as KH decreases
3-29

-------
(Delos et al. 1984). Both terms in Equation 3-15 are significant for com-
pounds of intermediate volatility where 3.0 > > 0.01 atm-m /mole.
(c) Sedimentation. Chemical concentrations in surface waters can
be reduced by deposition of suspended sediments containing sorbed chemical
and direct sorption onto bottom sediments. Mills et al. (1982) suggest
the use of a simple linear equation to calculate a partition coefficient,
Kp, which is then used to calculate the sedimentation rate in impound-
ments:
Kp = 0.63 (Kow) (OC)	(3-20)
where
Kqw = octanol-water partition coefficient
0C = organic carbon content of sediment.
Then the sedimentation rate, k$, of a toxic compound is computed by the
following equation:
ks = a (D)(Kp)	(3-21)
where
a = fraction of chemical in solution = l/(l+KpS)
D = P x S x Q/V, sedimentation rate constant
S = input suspended organic sediment = OC x S0 (mg/1)
S0 = input of suspended sediment (mg/1)
P = sediment trapping efficiency (default = 1)
Q = mean output flow rate of water from impoundment (liters/day)
V = impoundment volume (liters).
(2) Transformation processes.
(a) Photolysis. Direct and indirect photolysis of hazardous
substances dissolved in surface waters occurs at wavelengths greater than
290 nm. Shorter, or higher energy, wavelengths are filtered out by ozone
in the stratosphere. Photolysis involves the absorption of light and the
subsequent decomposition of a molecule. Mills et al. (1982) note that
products of the photochemical decomposition of a chemical may have
3-30

-------
prolonged environmental persistence relative to the parent compound. An
example given is the irradiation of DDT to form DDE, which is more
persistent in the environment than its parent compound. The assessor
must evaluate both the compound of concern and its possible decomposition
products when determining aquatic fate.
Direct photolysis describes processes in which a chemical absorbs
light and then undergoes one of several transformation reactions, such as
rearrangement, dissociation, or oxidation. The rate of loss of a chemical
(#) is given by the following equation (from Callahan et al. 1979):
dt
- § = kp[C] = ka*[C]	(3-22)
or
kp = katf	(3-23)
where
[C] = concentration of chemical
kp = first order rate constant of photolysis
4> = reaction quantum yield
ka = rate constant for absorption of light by the chemical.
The quantum yield is the fraction of absorbed photons that cause the
reaction(s) and is defined as:
moles of a given species formed or destroyed	(3-2 =1.
Indirect photolysis refers to the absorption of light by other
substances in the environment to form reactive species that interact with
3-31

-------
the compound of interest to produce a photochemical reaction. The rate
equation for indirect photolysis (from Callahan et al. 1979) is presented
below:
" dt = k2P [C][X] = kP'[C]	(3-25)
where
[C] = concentration of chemical
kgp = second order rate constant for photolytic reaction of chemical
[X] = concentration of reactive intermediate
kp' = rate constant, combined term for concentration of excited
state species and quantum yields (or efficiencies) of energy
transfer to and subsequent reaction of the chemical.
(b) Oxidation. Oxidants generated during photochemical processes
in surface waters may cause the oxidation of susceptible hazardous
compounds. Oxidants formed in aquatic systems through photoreactions of
naturally occurring materials include singlet oxygen and alkyl or acyl-
peroxyl radicals (IK^*). The average effective concentrations of
singlet oxygen and free radicals in natural waters have been determined
to be 10~12 and 10~9 moles/liter, respectively (Mill et al. 1979).
The loss of a chemical by reaction with an oxidant (OX) can be
expressed by (from Callahan et al. 1979):
" dt " kox[oxl[c)	(3"26)
where
[C] = concentration of chemical (mole/liter)
kgx = second order rate constant for reaction of the oxidant with
the chemical (moles • sec/liter)"*).
[OX] = concentration of oxidant (moles/liter).
Assuming a constant concentration of oxidant equal to the average
_q
free radical concentration in natural waters cited above (10 M), a
_q
first order rate constant can be approximated equal to 10 times the
second order rate constant:
3-32

-------
K0 - (lCf9)(Kox)	(3-27)
The majority of oxidation reactions involving organic chemicals, where
one or more electrons are lost from the atomic structure, are tnicrobially
mediated (Delos et al. 1984; Mills et al. 1982). Therefore, nonbiolog-
ical oxidation as described in this section need only be considered for
those organics cited in the literature as susceptible to this process.
(c) Hydrolysis. Direct reaction of a compound with water is termed
hydrolysis. The reaction involves the ionization of the water molecule
and the splitting of the chemical hydrolyzed. In general, hydrolysis of
an organic compound results in the introduction of a hydroxyl group (-OH)
into a chemical and the loss of a leaving group, which usually combines
with the remaining hydrogen ion from the water molecule.
Hydrolysis products may be more or less toxic or volatile than the
parent compounds. Hydrolysis of most organic chemicals in aqueous systems
is pH dependent and can be catalyzed by acid and base conditions. Adsorp-
tion of an organic molecule to suspended sediments may also influence
hydrolysis rates. Mills et al. (1982) defined the specific hydrolysis
rate constant, k^, to include the effects of adsorption:
kh = [kn + aw(ka[H+] + kb[0H"])]	(3-28)
where
kn	= neutral hydrolysis rate constant (sec"1)
aw	= decimal fraction of the total amount of compound C, which is
dissolved (default value = 1)
ka	= acid catalyzed hydrolysis rate constant (liter mole"1 sec"1)
[H+]	= molar concentration of hydrogen ion (mole liter -1) ([H+] = 10"pH)
k^	= base catalyzed hydrolysis rate constant (liter mole"1 sec"1)
[OH-]	= concentration of hydroxide ion (mole liter-1 ([0H-] = 10(pH-pKw)
= 10tpH14)).
3-33

-------
To determine rate constants for hydrolysis reactions that are pH
dependent, the assessor must obtain pH values for the water body under
study. Mills et al. (1982) note that in poorly buffered waters
(alkalinity < 50 mg/1 as CaCO^), pH values may show daily fluctuations
of 1 to 2 units resulting from natural processes. If the assessor is
unable to obtain adequate data to fully characterize pH within the water
body, low (conservative) values of must be used when predicting fate
of the chemical.
(d) Biodegradation. Biologically mediated processes that result in
chemical alteration of an organic compound are referred to as biodegrada-
tion. Most biological transformations are carried out by microorganisms,
since a carbon source is one requirement for the growth and maintenance of
cells. Assessors should consider biodegradation a decay process when
estimating chemical concentrations in surface waters. The rate of
biodegradation is a function of the microbial biomass in the water body
and the concentration of the chemical in surface water under given
environmental conditions (Callahan et al. 1977).
Because of the potential for the presence of a large variety of
carbon sources in an aquatic system and the probability that a large
microbial population will not change significantly when a relatively low
concentration of pollutant is consumed, Callahan et al. (1979) and Mills
et al. (1982) suggest the use of a first order kinetics equation to
describe the biodegradation rate:
- f - kb [C]	(3-29)
where
kk = first-order biodegradation rate constant (time-*)
[C] = pollutant concentration (mass/volume).
Another consideration is the microbial degradation of compounds not
used as a growth or nutrient substrate. The process is termed cometabo-
3-34

-------
lism, and in this case the rate of biodegradation is directly proportional
to the size of the microbial population. Since the products of cometabo-
lism do not provide energy to microbes, however, the size of the popula-
tion is unaffected by the presence of degradation products.
A second order rate law can also be used to describe microbially
catalyzed reactions (from Mills et al. 1982):
"dt " k2b	(3-30)
where
[C] = pollutant concentration (mass/volume)
k2b = second-order biodegradation rate constant
(volume/organism/time)
[B] = magnitude of bacteria (count or biomass/volume).
In the absence of bacterial cell count data, a low-range value for [B] of
100 cells/ml can be assumed as a default for natural surface waters. The
product K2b[B] then becomes equivalent to the first-order rate constant.
According to Callahan et al. (1979), the half-life of the chemical,
t^» a* a given cell concentration, B, will be:
tl/2 = kgjj [B] •	(3"31)
The biodegradation rate constant (Equation 3-31) is used to describe
biodegradation in aquatic systems where microorganisms are acclimated to
a pollutant and can produce enzymes to break down the chemical. Lag time
is associated with newly introduced pollutants, however, and this
initially may be due to the small population of microorganisms capable of
degrading the chemical, or a period of acclimation during which microbes
are induced to synthesize the necessary enzymes. When a pollutant is
released into an uncontaminated water body, Callahan et al. (1979) stress
that the lag period must be accounted for in the following manner:
3-35

-------
where
t1/2 = *o + tl/2
(3-32)
T1/2 = time required to reduce 50 percent of the original concentra-
tion of pollutant
t0 = time required to reach acclimation
tj/2 = half-life of transformation (Equation 3-31).
A quantitative generic description of tQ is not available since
many site- and chemical-specific factors will influence the value.
(3) Integration of fate processes. There are four approaches to
obtaining on overall fate rate constant for organic chemicals in surface
waters. The approach used by the exposure assessor will depend on data
availability and appropriateness to the level of detail of the
assessment. In the simplest approach, it can be assumed that the
chemical behaves conservatively and the concentration is not reduced by
transport or transformation processes. This approach requires no
chemical or environmental data and yields the highest levels of surface
water contamination. Its main limitation is that transport of the
chemical to other environmental compartments (i.e., air and bottom
sediments) is ignored, and this may lead to an underestimation of
exposure resulting from these pathways and an overestimation of exposure
via surface water-related pathways.
The second approach (from Mills et al. 1982) considers the processes
that lead to removal only by transport of the pollutant from the water
column. It is assumed that no chemical transformation processes take
place. The first order rate constant, ky, describing transport
processes is given below:
kT = ks + ky	(3-33)
where
k$ = specific rate constant for removal to bottom sediment
ky = volatilization rate constant.
3-36

-------
The rate constant, k^, incorporates specific chemical data including
acid-base equilibria, sediment sorption, and solubility. The advantage of
assessing only transport processes is the availability of chemical data.
The assessor may choose this method if data needed to determine transfor-
mation rate constants are unreliable or unavailable.
The third approach focuses on environmental transformation processes
only. First-order decay coefficients for transformation processes are
also additive, and an aggregate degradation coefficient, kg, can be
determined (adapted from Delos et al. 1984):
kg = kg + k^ + kp + K0	(3-34)
where
kg = first order biodegradation rate constant
k^ = first order hydrolysis rate constant
kp = first order photolysis rate constant
k0 = first order oxidation rate constant.
Since nonbiological oxidation is often not an important process for most
organic chemicals, the assessor may opt to drop it from the above
equation.
The fourth and most complex approach incorporates speciation,
transport, and transformation processes. The result is an addition of
all first-order rate coefficients expressed in Equations 3-33 and 3-34:
ky = k$ + ky + kg + + kp + k0	(3-35)
The sorption/sedimentation term can be dropped from the above equation if
the surface water concentration is adjusted for this process in other
ways, e.g., using the output of Equation 3-13. As mentioned already,
prior to calculating all rate constants needed for Equation 3-35, the
assessor should determine the most significant fate processes for the com-
pound. In particular, photolysis and oxidation are frequently not signif-
icant fate processes for many chemicals when compared to other fate proc-
esses such as volatilization, biodegradation, and sorption/sedimentation.
3-37

-------
Table 3-7 provided an assessment of the relative importance of processes
influencing the aquatic fate of organic priority pollutants; Tables 3-9
and 3-10 indicated those classes of chemicals that are generally
susceptible or resistant, respectively, to hydrolytic attack.
3.3 Estimation of Concentrations - Surface Water
The simplified equation presented below (adapted from Delos et al.
1984) provides an estimation of the concentration of a chemical substance
downstream from a point source release into a flowing water body, after
dilution by the receiving river or stream:
C Q
C -	(3-36)
where
C = concentration of substance in stream (mass/volume)
Ce = concentration of substance in effluent (mass/volume)
Qe = effluent flow rate (volume/time)
Qt = combined effluent and stream flow rate (volume/time).
The chemical concentration can also be estimated if introduced into a
flowing water body from a nonpoint source, if the release rate is known
in terms of mass per unit time, with use of the following equation:
where
C = concentration of chemical in stream (mass/volume)
Tr = intermedia transfer rate (mass/time)
Qt = stream flow rate (volume/time).
These equations assume that (1) the chemical is conserved (i.e., that its
environmental concentration is not reduced by fate processes such as
volatilization, biodegradation, and sedimentation); (2) complete and
3-38

-------
instantaneous mixing takes place; and (3) stream flow and rate of
chemical release are both constant. It must be noted that the equations
provide in-stream chemical concentrations from site releases only. To
estimate total in-stream chemical concentrations, the assessor should add
background in-stream chemical concentrations to those estimated by
Equations 3-36 and 3-37.
Because the dilution equations (Equations 3-36, 3-37) assume that the
chemical introduced into the flowing water body is conserved, the
resulting prediction is an estimated stream/river concentration that
remains constant throughout a segment of the stream. The concentration
only decreases with further dilution from tributary stream flow. The
equations are useful as a basic model to determine the fate of conserved
hazardous substances. In the case of hazardous substances that are
subject to a variety of loss processes, these equations provide a
worst-case estimate of contaminant concentration in the water body.
To estimate the concentration of a nonconserved substance at selected
exposure points downstream (below the mixing zone), Equation 3-38 below
incorporates an overall decay coefficient. This coefficient, K,
represents a combination of all decay and loss rates affecting the removal
of a substance from a water body (from Delos et al. 1984):
-K*
W(x) = W(0)e u	(3-38)
where
W(x) = concentration at downstream distance x (mass/volume)
W(0) = concentration immediately below point of introduction (from
Equations 3-11, 3-12)
e = 2.71828
K = overall decay coefficient (time)"*
x = distance downstream from point of introduction (length)
u = stream velocity (length/time).
Assumptions incorporated into this equation are as follows: steady-state
conditions are present, mixing is complete, the substance travels
3-39

-------
downstream at stream velocity, and decay rates are a direct function of
chemical concentration.
The overall decay coefficient, K, can be derived from monitoring data
or estimated through the use of literature decay rate constants. Data
from at least two monitoring points, one directly below the point of
hazardous substance release into the water body and one downstream from
the mixing zone, are required to estimate K for a particular stream. If
possible, the assessor should take into account seasonal variations that
may affect chemical and physical parameters in the stream body. If a
worst-case concentration estimate is desired, the lowest decay rate for
the substance of concern in a particular stream/river, or worst-case K
value, should be used.
The most important removal process affecting the chemical of interest
present in the water body must be identified to estimate K from published
decay rate constants. The user may refer to Callahan et al. (1979),
Mabey et al. (1982), or Table 3-7 for information about the relative
importance of organic contaminant aquatic fate processes. Decay rate
constants for many organic compounds are provided in Verschueren (1984)
and Dawson et al. (1980). An overall rate constant can be calculated as
described above in Section 3.2.2.(3), based on the single process rate
constants developed using the algorithms presented previously.
3.4 Exposure Estimation - Surface Water
The information presented in Section 3 allows the assessor to estimate
concentrations of chemicals in surface water, which can then provide input
to estimate exposures by using the equations and parameters presented in
the Exposure Factors Handbook (USEPA 1989).
3-40

-------
4.	ASSESSMENT OF EXPOSURE VIA GROUND WATER
4.1 Release Analysis - Ground Water
When analyzing release of a chemical to ground water, the exposure
assessor should conduct an initial evaluation of the potential release
mechanisms. As a first step, this evaluation involves a qualitative
screening approach aimed at determining the sources of release. A more
detailed analysis of the factors and mechanisms that may significantly
affect the release potential is then undertaken; this may require the use
of modeling or monitoring.
4.1.1 Screening of Release Potential
Contaminant release to ground water may result from a number of mech-
anisms. These mechanisms can affect ground water directly or indirectly,
and they include the following:
•	Overland flow (e.g., from impoundment overflow or failure, drum
leakage);
•	Direct discharge (e.g., on-site release from treatment process
or incineration);
•	Leachate generation (e.g., from buried wastes, lagoons, leaks,
land application); and
•	Generation of surface runoff.
Leachate generation is the only mechanism that directly affects ground-
water contamination. The remaining three mechanisms affect ground water
indirectly and have been addressed in the surface water section (see
Section 3).
The assessor must consider several factors when determining leachate
release potential. Information about the physical/chemical properties and
toxicity of the contaminant of concern may help to determine the disposal
practices associated with the chemical, which will likewise help to deter-
mine the potential for release. For example, the potential for release
of wastes disposed of in a RCRA landfill facility can be expected to be
lower than from an industrial or municipal landfill.
4-1

-------
Ground-water contamination results from the leaching of contaminants
from surface or subsurface soils or from treatment/storage/disposal areas.
Contaminants can be leached from a landfill site or from a land applica-
tion site by rainwater that percolates through the soil or subsoil. Un-
lined lagoons and surface impoundments may introduce contaminants directly
into ground water. Contaminants stored above ground may also leak from
storage containers and percolate to ground water. The potential for
release from any of these sources must be evaluated; any that are signifi-
cant must then be subjected to quantitative analysis as discussed below.
4.1.2 Quantitative Release Estimation
Release estimation involves two aspects: (1) quantification of con-
taminant concentrations in waste and/or leachate and (2) quantification
of volumetric leachate release rates. The procedures used vary depending
upon the characteristics of the release site. In all cases, the proce-
dures reported are for the determination of the steady-state release rate
or annual average release rate.
(1)	Estimating contaminant release concentration. Determining the
concentration of a contaminant depends upon site characteristics. For
lagoons or impoundments, the concentration of the contaminant in the
lagoon or impoundment is considered the concentration of the leachate.
For landfills, the equilibrium solubility of the solid waste is used,
with the assumption that the contaminants will have equilibrated with the
percolating rainwater. This may not be the case for landfarms, since
equilibrium conditions may not be in effect. Therefore, it is necessary
to estimate contaminant concentration as a fraction of the equilibrium
solubility for landfarms.
(2)	Estimating volumetric release rate. The volume of leachate is
calculated in two ways, one for solid wastes and one for liquid wastes.
For solid wastes, percolating water (from direct precipitation and/or
stormwater runoff onto the site) is the only source of liquid. The load-
ing rate to ground water for this situation can thus be calculated using
the following equation:
4-2

-------
Lc - q A C0
(4-1)
where
Lc = contaminant loading rate (mg/sec)
q = percolation rate (see Equation 4-17 for calculation of rate
(cm/sec))
A = area of landfill (cnr)
C0 = water solubility of solid chemical (mg/cm3).
For liquid wastes, precipitation has a minimal effect, since the liquid
wastes will percolate to ground water under the influence of gravity. In
this case, the rate of percolation is dependent upon the permeability of
the liner or the underlying soil at the contamination site. The volumet-
ric release rate for liquid wastes can be estimated using the following
equation (Bowers et al. 1979):
Q = Ks i A	(4-2)
where
Q = volume loading rate (cm3/sec)
Ks = Darcy's coefficient; for unlined lagoons use native soil
hydraulic conductivity; conductivity (cm/sec) (see Section 4.3
for sources of hydraulic conductivity)
i = hydraulic gradient (dimensionless). Equations 4-1 and 4-2 will
handle situations where the liquids in the lagoon have a free
depth. In many cases the depth of the free liquids is small, or
it is small with respect to the distance between the lagoon and
the water table (when the K$ is for native soil). In these
cases, the term "i" can be taken as 1.
A = area of lagoon (cm2).
This Q is then used to estimate mass loadings with the following
equation:
Lc = Mc Q	(4-3)
where
Lc = contaminant loading rate (mg/sec)
Mc = contaminant concentration in lagoon fluid (mg/cm3)
Q = volume loading rate (cm3/sec).
4-3

-------
Equations 4-2 and 4-3 model the release rate from a lagoon whether
the flow through the vadose zone* is saturated or unsaturated. For
unlined active lagoons, the flow is typically saturated all the way to
the water table. For clay-lined lagoons, the flow is saturated through
the liner and unsaturated between the liner and the water table (assuming
no breaches in the liner). Equations 4-2 and 4-3 are appropriate when
lagoon releases are analyzed, but should not be used for spills or other
conditions where there is no ponding of the chemicals on the surface for a
long period of time. Under these conditions, the assumption of saturated
flow (through the liner or soil) may be violated.
Equations 4-2 and 4-3 apply to liquids that are mostly water. For
lagoons containing organic fluids, however, the equations may have to be
corrected. For liquids with a density or viscosity that differs from
water, K$ can be corrected for this viscosity and density by calculating
the term Kc, using the following:
Kc = Kfl (Dc/Dw) (uw/uc)	(4-4)
where
Kc = corrected Ks term = hydraulic conductivity of contaminant
(cm/sec)
Kw = hydraulic conductivity of ground water (cm/sec)
D = density of liquids; c = contaminant; w = water (mg/liter)
u = dynamic viscosity of liquids; c = contaminant; w = water;
(mg-cm/sec).
and then substituting Kc for Ks in Equation 4-2.
For waste sites that are lined with flexible membrane liners (FML), the
release rate is dependent upon the characteristics of the contaminant
(Steingiser et al. 1978). The permeability equation that can be used to
The vadose zone is the soil zone not saturated with water (i.e.,
unsaturated area) lying between the soil surface and the water table.
4-4

-------
estimate permeation rates is as follows (Salame 1961, 1973, 1985;
Steingiser et al. 1978):
"SH
Ps = ApU)e H	(4-5)
where
Ps = permeation rate (g-mi1/100 in^-day-cmHg)
Ap = constant solely dependent on the type of polymers used
(g-mil/100-inz-day cm/Hg)
su = constant solely dependent on the type of polymers used
(cnr/cal)
 = the polymer "permachor" calculated for each polymer-permeant
pair (cal/cirr).
Salame provides a list of values for these parameters obtained from
his extensive experimental work. These values are shown in Tables 4-1
through 4-4.
For permeation of water through FMLs, polymers are categorized into
five groups based on the values of the solubility parameter as shown in
Table 4-2. This grouping was achieved after examination of experimental
data for about 70 different polymers (Salame 1985). The solubility
parameter provides an indication of polymer interaction with water, with
more interaction occurring at higher values of the solubility parameter.
Examples of hydrogen bonding for polymer group 5 include hydroxyl (OH)
and amid (NHCO) radicals as in nylon and polyvinyl alcohol. The polymer
with hydrogen bonding but with the value of "delta" less than 11 does not
belong to group 5. Permachor values for some selected organic liquids
and for water are shown in Tables 4-3 and 4-4, respectively. It should
be noted that the water "permachor" values for various polymers given in
Table 4-4 apply under dry conditions only. For water permeation under wet
conditions, the permachor values should be reduced by about 20 percent.
The term P$ can be used to calculate the release rate in grams per
day. P$ is multiplied by the area of the liner and then divided by its
thickness. This assumes a normal water vapor pressure of 1 cm Hg at
ambient temperature. The equation is:
4-5

-------
7632H
Table 4-1. Parameter Values for Permeation Equation (at 25*C)
Parameter
Waste liquid organics permeating
through given liner materials
PE
PVC
Contaminated leachate (water) permeating
	through various liner categories
A ( g-mil
	
100 in 'day'cmHg
1x10
1x10*
11.5	10.2	5.4xl02 25
i
a~.
s (cc/cal)
0.506
0.23
0.16	0.135 0.115	0.035 0.099
^ (cal/cc)
Table 8-3c
Table 8-4c
Source: Salame, no date; Salame 1961.
^See Table 4-2 regarding polymer category. Source: Salame 1985.
cSee the table indicated for these values.
= 0.33 exp (0.064 x 42), where t is the solubility parameter, (cal/cc)"2.

-------
7628H
Table 4-2. Polymer Categorization for Peraeation of Water
Polymer group
Categorization
1
Any polymer with t < 8.9a
Z
Any polymer with 8.9 < S < 10
3
Any polymer with 10 < « < 11
4
Polymer containing nitrile (CN) group

with 10.5 < i
5
Polymer with H bonding and 11 < t
a	1 /?
6 = Solubility parameter (cal/cc) .
Source: Salame 1985.
4-7

-------
7628H
Table 4-3. Pennachor Values of Sane Organic Liquids in Polyethylene and PVCa
In nonpolar polvwer	In polar polymer
Liquid	#	4
Acetic acid
13.0
44.0
Benzaldehyde
15.9
4.0
Benzene
5.4
7.0
2-Butoxy ethanol
24.4
75.0
Butyl acetate
13.0
5.0
Butyl alcohol
18.0
50.0
Butyl ether
10.4
46.0
Butyra ldehyde
13.5
0.0
Caprylic acid
19.0
50.0
Carbon tetrachloride
5.8
22.0
p-Chlorotoluene
7.6
7.5
Cyclohexane
7.0
45.0
D i but y 1 phtha 1 ate
31.4
17.0
Diethylamine
10.0
5.7
Ethanol
16.0
48.0
Heptane
7.0
44.0
Hexane
6.0
43.0
Methyl ethyl ketone
12.5
1.0
Methanol
15.0
47.0
Hitroethane
15.4
7.0
i-Pentyl propionate
15.0
7.0
i-Propyl amine
11.0
6.7
Trichloroethylene
5.4
3.0
o-Xylene
9.4
11.0
p-Xylene
7.4
9.0
a Polyethylene and PVC are nonpolar and polar polymers, respectively.
Sources: Salaae, no date; Steingiser et al. 1978.
4-8

-------
7628H
Table 4-4. Water Permachor Value for Dry Polymers
Permachor
value
Polyraer	4
Polyvinyl alcohol	160
Polyacrylonitrile	109
Cellulose (dry)	97
Polyvinylidene chloride	87
Polycaprolactam (dry)	80
Polyacrylonitrile styrene	76
(70/30) (Lopac)
Polyacrylonitrile styrene/butadiene	75
(70/23/7) (Cycopac 930)
Polychlorotrifluorethylene	71
Polyethylene terephthalate	68
Polyvinyl idene fluoride (ICynar)	67
Polyacrylonitrile styrene/=a/butadiene	65
(56/27/4/13) (Cycopac 920)
Polyvinyl chloride	62
Polyoxymethylene (Delrin)	57
Polymethyl methacrylate	55
Polyvinyl acetate (dry)	45
Polystyrene/acryIonitrile (74/26)	45
Polyethylene (HO)	40
Polysulfone	34
Polypropylene	33
Polycarbonate (Lexan^)	33
Polystyrene	28
Polyethylene (LD)	26
Polyisobutylene	17
Polyethylene/vinyl acetate (85/15)	15
Polybutadiene	8
Polymethyl pentene (TPX)	8
Polydimethyl siloxane (dry)	-4
Sources: Salame 1961; Salame, no date; Steingiser et al. 1978.
4-9

-------
Lc - Ps A (p/d-,)
(4-6)
where
Lc = contaminant loading rate (g/day)
Ps = permeation rate (g-mi1/100 in^*dav*cmHg)
A = area of liner (in units of 100 in2)
p = vapor pressure (cmHg)
d-j = thickness of the liner (mils).
Total leachate release for a given contaminant from a particular site
can thus be estimated by selecting the appropriate equation for the con-
taminant. These equations represent very simplified approaches for esti-
mating leachate release rates. More in-depth analyses can be performed
using computerized models when appropriate. Models that may be useful are
the Hydrologic Evaluation of the Landfill Performance (HELP) model and the
Seasonal Soil Compartment (SESOIL) model.
4.2 Fate Analysis - Ground Water
4.2.1 Screening of Fate Processes
The nature of the ground-water environment restricts the number of
fate processes that chemical contaminants can undergo as they are trans-
ported from their source to the receptor area in which they are detected.
These processes can be classified by their effects as either transport-
related or degradation-related. Transport-related processes (i.e., sorp-
tion, ion exchange, and precipitation) can have a facilitating or retard-
ing effect on the movement of ground-water contaminants, but degradation-
related processes (i.e., hydrolysis, biodegradation, and reduction) can
only retard the spread of those contaminants that are subject to them.
(1) Characteristics of the subsurface environment. The ground-water
environment can be divided into unsaturated and saturated (vadose zone)
regimes (Freeze and Cherry 1979). The irregular interface that forms the
boundary between these two regimes is generally known as the water table.
Below the water table, the pores and spaces within and between individual
soil particles are filled with water. Above the water table, these pores
and spaces are partially occupied by gases derived primarily from the
4-10

-------
atmosphere and from water. Some of the gases are absorbed directly from
the atmosphere, some are released from solution as water percolates down-
ward, and some are generated by microbial or chemical processes in the
soil. Water from the soil/atmosphere interface percolates downward
through the unsaturated zone until it reaches the saturated zone of the
aquifer.
The loose, solid material of both zones consists primarily of sili-
cates (sand and silt), metallosilicates (clay), and gravel originating
from sedimentary and igneous rocks. The saturated zone also extends into
fractured bedrock that is in contact with the unconsolidated (soil) zone.
The upper part of the unsaturated zone usually contains up to 8 percent
organic matter, which is leached by the water percolating downward
(Lindsay 1979). This organic matter is of biologic origin and consists
primarily of humic and fulvic acids, with smaller amounts of carbohy-
drates, lipids, and protein-derived material (Thurman 1984). Organic
matter reacts with the oxidizing species of the soil, such as gaseous and
dissolved oxygen, ferric ion, and nitrate, to produce an environment that
becomes progressively less oxidizing and less acidic as water proceeds
from the unsaturated zone to the saturated zone. The anoxygenic saturated
zone, in its uncontaminated state, contains only small amounts of organic
matter (including microorganisms) and the reduced valence states of
inorganic matter.
(2) Chemical identity of contaminants. The chemical identity (or
integrity) of some contaminants can be affected by acid/base interactions,
oxidation/reduction reactions (redox), metal speciation, or
complexation. These phenomena may alter the chemical and physical
properties of the contaminants that have entered the ground water. A
description of acid/base effects is presented for organic contaminants.
Speciation and complexation are discussed under inorganic contaminants
(Thurman 1984).
In the saturated and unsaturated zones of an aquifer, naturally occur-
ring materials can participate in acid/base reactions with contaminants.
Some examples are:
4-11

-------
CH3COOH + [humate-0"]~n " CH3COO" + [humate-OH]"^"1^
CH3COOH + [clay-0"]"n " CH3COO" + [clay-OH]~(n~1}
ch3cooh + co3"2	1 ch3coo" + hco3"
The humates and clays are polyelectrolytes that are partially ionized and
therefore can enter into acid/base reactions as acids or bases. In this
manner, they exert a buffering effect on the acidity of the ground water
as long as the ion .exchange capacity of their ionizable groups is not
exceeded. As a result of these acid/base reactions, the ionic charge of
an acidic or basic contaminant can be changed, thus altering its identity
and its properties as an adsorbate.
(3) Transport-related processes. Versar (1987a) has developed a
screening level procedure for transport of chemicals in ground water as
the output of a ground-water screening-level assessment methods working
group at the Office of Toxic Substances. The object of this screening
procedure is to identify those chemicals that, based on physical or
chemical properties, would absorb strongly to soil and clay particles and
not migrate significantly via ground water. For a chemical to fall into
this category it must meet at least two of the following criteria:
•	Net charge at neutral pH > +2
•	Molecular weight >. 5,000 g/mole
•	Solubility <0.1 mg/1
•	Koc > 20,000.
In evaluation of net charge at neutral pH, metals and metal complexes are
excluded from consideration. In evaluation of molecular weight above, the
average molecular weight should be used for a given chemical in the case
of mixed chemical species (e.g., polymers of different chain lengths).
Specific transport- related processes affecting chemical fate in the
ground-water environment are discussed in the following subsections.
(a) Sorption. The tendency of a chemical dissolved in ground water
to be sorbed by the aquifer's solid phase can be expressed in terms of the
4-12

-------
soil/solution partition coefficient, K^, derived from the Freundlich
isotherm equation
where x/m is the amount adsorbed (ug chemical per gram of soil) and C is
the concentration of chemical (ug/ml) in the aqueous phase. The value of
1/n depends on the sorbate and sorbent being studied and is usually close
to one (Lyman et al. 1982).
Sorption to soil by neutral organic chemicals is generally correlated
with the content of organic matter in the soils (Lyman et al. 1982, Leifer
et al. 1983). If the soil partition coefficient, K^, is divided by the
fractional amount of organic carbon (OC) in a soil, a new coefficient is
obtained that is characteristic of the organic chemical under study. This
value is called the organic carbon partition coefficient, KQC, and it
can be estimated from KQW, the octanol-water partition coefficient.
Sorption of neutral organic chemicals in the unsaturated zone of an
aquifer will be affected by the presence of organic matter, but sorption
of these same chemicals in the saturated zone will be due primarily to the
high surface area (per mass) of minerals such as clays. Lipophilic sub-
stances tend to form films at water/solid interfaces just as they do at
the air/water interface. Thus, if the saturated zone contains clays or
other high surface area minerals, the ground water is presented with a
large water/ solid interface on which contaminants can form a surface
film. Adsorption isotherms in which sorption can be correlated with the
surface area of the adsorbent are called Langmuir isotherms. Adsorption
phenomena of this type are not linear and can reach a saturation limit
after which further adsorption will not occur, even from water with higher
levels of contaminant. Several variations of the adsorption isotherm
equation are available for the fitting of empirical data from experimental
sorption studies (Kinniburgh 1986). In order to provide a simplified,
conservative approach, however, this analysis assumes that the magnitude
of such surface area-related sorption approximately parallels that associ-
ated with sorption to a nominal organic carbon content in the soil (i.e.,
(4-7)
4-13

-------
0.1 percent soil organic carbon content). The means for calculating sorp-
tion retardation is thus the same for sorption to both soil organic carbon
and the surface of mineral matter particles.
ENPART (Environmental Partitioning Model) uses an organic matter con-
tent of 2 percent in soils and 4 percent in sediments for generic assess-
ments. Liefer et al. (1983) discuss the relationship that has been
established between Kd and chemical mobility in soils. The results are
summarized in Table 4-5. An EPA test guideline and a support document
(USEPA 1982a, Leifer et al. 1983) provide the details of this relation-
ship.
(b) Ion exchange phenomena. Ionic contaminants, both organic and
inorganic, are retarded within an aquifer primarily by ion exchange
phenomena and precipitation. Ion exchange capability within an aquifer
is almost exclusively limited to colloidal clay and silica particles
-3	-6
(diameters in the range 10 to 10 mm), because these particles
have a large ionic charge relative to their surface areas. This charge
is the result of (1) cationic substitutions within the crystal lattice
and (2) ionic dissociation at the surface. To neutralize this charge,
an adsorbed layer of cations and anions forms a zone adjacent to the
hydroxylated layer (Parks 1967). The net charge of this zone can be nega-
tive or positive, depending on the pH of the immediate environment. At
low pH, a positively charged surface prevails; at a high pH, a negatively
charged surface develops (Freeze and Cherry 1979). The tendency for sorp-
tion of either cations or anions therefore depends on the pH.
Neutralized colloidal particles may be transported in the ground water
with organic and inorganic contaminants on their surfaces. Additionally,
humic substances can exist as colloidal particles and also serve as ion
exchangers. Some contaminants that might otherwise be sorbed to station-
ary material in the aquifer could be transported in the sorbed layers of
these mobile colloids. Sorption in this case has facilitated transport.
The cation exchange capacity (CEC) of soils and other geological materials
is usually expressed as the number of milliequivalents of cations that can
be exchanged in a sample with dry mass of 100 grams.
4-14

-------
7628H
Table 4-5. The General Relationship Between the Soil/Solution
Partition Coefficient and Soil Mobility
Nobility class	Distance surface-applied chemical may leach
0.1
1
10
102
10
10°
10M
2.5
Very nobile
Very mobile
Hobile
Low mob)1ity
lnroobi )e
Innobi le
Imnobile
Much of the chemical leaches through top
20 cm of soil into subsoil.
Much of chemical leaches into soil, but peak
concentration in top Z0 cat of soil.
Only small anount of leaching and peak
concentration normally in top 5 cm of soil.
No significant leaching.
Source: Leifer et al. (19B3).
4-15

-------
(c) Precipitation (phase separation from ground water). After a
chemical has contaminated the ground water, changing conditions in the
aquifer of temperature, pH, and other chemical constituents may bring
about precipitation (i.e., phase separation) of the intruding chemical.
Some sparingly soluble organic chemicals (e.g., diethylhexyl phthalate)
are more soluble in landfill leachate than they are in natural environmen-
tal waters (USEPA 1979a). This effect may be due to the presence of
enough soluble organic material in the leachate to alter its solvent
properties, or it may be due to complexation of the contaminant with
natural colloids. As a landfill leachate (or seepage from other sources
of organic chemicals) becomes diluted in the ground water, the solubility
of lipophilic contaminants may decrease, thereby initiating their phase
separation as a liquid or solid. A similar effect can occur if the
concentration of ionic species in the ground water increases through
dissolution of aquifer minerals. This latter effect is referred to as
salting-out.
Other variable conditions in the aquifer that may bring about the
phase separation of dissolved organic contaminants are decreases in tem-
perature and changes in acidity. The solubility of most chemicals varies
directly with the temperature, and the ground water's level of acidity can
affect the solubility of some organic acids and bases. Decreases in pH
can lessen the solubility of weak acids (e.g., carboxylic acids and
phenols) and increase the solubility of amines (e.g., aniline).
(4) Degradation-related processes.
(a) Hydrolysis. Although hydrolysis affects the fate of only a
small number of contaminating organic chemicals (e.g., some esters,
amides, epoxides, phosphorus pesticides, and nonaromatic halides), this
process may be the only operative process of degradation for some of these
chemicals following their introduction to ground water. A convenient
chart to aid in determining whether hydrolysis will affect the fate of a
chemical contaminant is given in Figure 4-1. The decreased temperature
of the ground-water environment, compared to surface water, will increase
4-16

-------
Alky!
Hat ides
n = 13













Allyl and
Benzyl Halides
n = 9












Polyhalo
Methanes
n = 10


1	0
•
	1



Epoxides
n = 14




to
•	
	1

Aliphatic
Acid Esters
n = 18












Aromatic
Acid Esters
n = 21








Amides
n = 12


-t> m

—H



Carbamates
n = 18















Phosphonic Acid Esters,
Dialkylphosphonates
n = 6












Phosphoric Acid,
Thiophosphoric Acid
Esters n = 6



-
>—•	

	1

Phosphoric Acid Halides,
Dialkylphosphonohalidates,
Dialkylphosphorohalides
n = 13









» u.
II
X

\ 1 ——V
X
» <
o
Pesticides and Misc.
Compounds
n = 13




H—
•










2-2x10* 2.2x104 2.2x10* 2.2 yr 8 days 1.9 hr 1.15min 0.69 s
yr	yr	yr
Key:	Half-Life
% Average
> Median
n No. of Compounds Represented
Source: Mabey and Mill 1978, as adapted by Lyman et al. 1982.
Figure 4-1. Examples of the range of hydrolysis half-lives for various
types of organic compounds in water at pH 7 and 25°C.
4-17

-------
the hydrolytic half-lives of chemical contaminants several times over what
is depicted in Figure 4-1. If the hydrolytic half-lives of the chemical
group into which the contaminant fits are relevant to fate in ground
water, the rate constant can be obtained from one of the literature
sources cited in the following paragraph.
Rates of hydrolysis are difficult to predict for chemicals whose
degradation processes have not been studied (Lyman et al. 1982) However,
rate constants for a large number of hydrolyzable chemicals have been
estimated over the range of environmental conditions from existing data
(Mabey and Mill 1978, Mill 1979). Estimates of this type are usually
based on extrapolations from data obtained at higher reaction tempera-
tures. Note that they are only valid if the molecular mechanisms for
hydrolysis at the two temperatures are the same. Some hydrolysis studies
of polyhalogenated compounds do not differentiate between nucleophilic
displacement of a covalently bound halide, elimination of hydrogen halide,
or addition of water to a double bond followed by elimination of hydrogen
halide (Callahan et al. 1979). In such studies these three different
reactions have been collectively measured and noted as aqueous degradation
(Dilling et al. 1975).
(b)	Biodegradation. The level of both microorganisms and organic
contaminants in ground water is often low compared to surface water. This
condition makes it unlikely that much microbial degradation will occur in
the flowing water itself, because acclimation is usually necessary for the
degradation of anthropogenic chemicals. However, since the water is flow-
ing through porous, solid material with a large surface area per volume,
both microorganisms and contaminants can become sorbed in concentrated
stationary films that would allow acclimation and biodegradation (Bitton
and Gerba 1984). As a consequence, biodegradation of contaminants that
are strongly sorbed may be facilitated.
(c)	Reduction. In general, investigations of reductive transforma-
tions do not effectively distinguish between biotic and abiotic processes.
4-18

-------
A recent review by Macalady et al. (1986) defines abiotic reduction proc-
esses as all those not directly involving the participation of metabol-
ically active microorganisms. Under this definition, biologically derived
materials (including extracellular enzymes) can enter into processes that
would still be termed abiotic.
The most important reductive reaction in ground water is alkyl dehalo-
genation, which can take place in the presence of reduced iron porphyrins
(derived from plant chlorophyll). The reduction of organic nitro groups
to amines has also been studied (Southwick and Willis 1979). These reac-
tions can occur abiotically and are affected by the redox conditions of
the system. The interconversion of sulfones, sulfoxides, and sulfides is
also a redox process, and it is implicated in the degradation of several
pesticides (Walter-Echols and Lichtenstein 1977). The quinone-
hydroquinone structure of humic acids may have a role in mediating abiotic
reduction reactions in ground water.
4.2.2 Quantitative Fate Estimation
(1) Transport and degradation. After the fate processes have been
screened to determine those that are relevant for the quantitative fate
estimation of the chemical contaminant, the physical/chemical properties
required for this estimation should be tabulated or themselves estimated
(if not available).
(a) Estimation of mobility. Sorption of an organic contaminant in
the unsaturated zone of an aquifer can usually be correlated with the
content of organic matter in the soil (Lyman et al. 1982, Leifer et al.
1983). Consequently, the estimated sorption is expressed as the organic
carbon adsorption coefficient, KQC, which is defined by the relationship
Koc - Kd/0C	(4-8)
where
ICj = soil partition coefficient
OC = the fractional mass of organic carbon in the soil.
Estimated KQC values are multiplied by the fractional amount of soil
organic carbon to obtain K^, which can then be related to soil mobility
4-19

-------
(Leifer et al. 1983). Calculations for sorption (or retardation) in the
saturated zone usually assume an organic matter content of 0.1 percent
for that environment.
Several equations for the estimation of KQC are given in Lyman
et al. (1982). All have the form of regression equations in which the log
Koc is related to either log S, log KQW, or log BCF (S = solubility,
Kqw = octanol-water partition coefficient, BCF = bioconcentration fac-
tor). The equations were developed from experimental data that correlated
log Kqc with the other three parameters for a particular set of organic
compounds. Selection of an appropriate equation depends on the physical -
chemical property data available for the chemical and also on how closely
the subject chemical matches the set of organic compounds from which the
equation was developed. While all the equations are considered useful,
they are not intended for use with organic chemicals that will ionize in
water.
In examining the regression equations that relate KQC to KQW, the
analyst should select the equation that is based on a chemical class
closest to the subject contaminant. If the contaminant does not fit into
a specific class, the first regression equation should be used, since it
is based on the largest sample. Six of the better-known regression equa-
tions are the following (Lyman et al. 1982):
Log Koc = 0.544 log Kow + 1.377	(4-9)
based on a wide variety of contaminants, mostly pesticides;
log Koc = 0.937 log Kow - 0.006	(4-10)
based on aromatics, polynuclear aromatics, triazones, and dinitroaniline
herbicides;
log Koc = 1.00 log Kow - 0.21	(4-11)
based on mostly aromatic or polynuclear aromatics;
log Koc = 0.94 log Kow + 0.02	(4-12)
based on s-triazines and dinitroaniline herbicides;
4-20

-------
log Koc = 1.029 log Kow - 0.18	(4-13)
based on a variety of insecticides, herbicides, and fungicides; and,
log Koc = 0.524 log Kow + 0.855	(4-14)
based on substituted phenylureas and alkyl-N-phenylcarbamates.
EPA test recommendations (as cited in Lyman et al. 1982) list organic
matter content of soils to be between 1 and 8 percent for valid use of
Koc estimated from these regression equations. Measured property values
are preferred as input data, but estimated values can also be used.
Measured values should be used in the following order of preference:
K > S > BCF.
ow
If the chemical is polymeric, organometal1ic, or an organic salt,
estimation methods given in Lyman et al. (1982) cannot be applied.
Moreover, there are several phenomena, other than lipophilic partitioning
into soil organic matter, that affect the sorption and mobility of organic
chemicals in soil. These are (1) ion exchange binding of amines and qua-
ternary ammonium compounds to natural polyanions, such as clays, silica,
and humic acids; (2) coordination of organic compounds with mobile or
immobile metal cations; (3) humic acid binding of aromatic compounds;
(4) hydrophobic adsorption of sparingly soluble substances to inorganic
surfaces in the saturated zone of an aquifer; and (5) vaporization/
condensation of the contaminant in the unsaturated zone.
The following suggestions may be appropriate, but experienced profes-
sional judgment is recommended:
• Organic salts. If the chemical decomposes in water or exists in
water as a metallic cation and an organic anion, one can determine
from a handbook such as Merck Index the solubility of the cationic
species as a carbonate, phosphate, aluminate, or silicate. Movement
of sparingly soluble substances in an aquifer will be retarded.
Some organic cations will be quaternary ammonium compounds; these
should be retarded by anionic mineral surfaces. Many organic
anions, however, will be mobile. Their solubility as calcium or
magnesium salts may indicate whether retardation by precipitation
is possible.
4-21

-------
• Polymers. Most polymers should be immobile in a landfill or sur-
face impoundment and will not enter the aquifer. Moreover, soluble
polyamines, polyacids, and polyalcohols should be retarded by inter-
action with mineral surfaces. Monomeric chemicals may be released
from the polymers, however, and it may be necessary to assess their
fate.
(b)	Estimation of hydrolytic stability. Methods for estimating the
rates of hydrolysis of organic compounds are limited to ring-substituted
benzamides and ethyl benzoates, benzyl halides, and paratoluenesulfonates
and some aromatic esters (Lyman et al. 1982). Since correlation data are
based on rates of hydrolysis in mixed organic-aqueous solvents, even these
estimated rate constants must be regarded as lower limits of values that
might be observed in ground water. The exposure assessor should consult
Figure 4-1 and the references given in Section 4.2.1(4) for hydrolysis
rate data. If values for hydrolytic stability cannot be estimated, it
can be assumed that the contaminant does not hydrolyze for a worst-case
analysis.
(c)	Estimation of biodegradability. Biodegradation and volatiliza-
tion are the two main mechanisms by which organic chemical wastes are
removed from a landfill; however, the rate at which biodegradation occurs
cannot be predicted under such circumstances with any reliability (Lyman
et al. 1982). At present, the biodegradability of chemicals is best esti-
mated by comparing the chemical structure to other organic chemicals that
have been studied. This structure/activity approach has been summarized
in tabular form in Lyman et al. (1982). When developing a worst-case
scenario, the assessor should consider the chemical resistant to biodegra-
dation.
(2) Vertical transport. The most significant contaminant movement
in soils is a function of liquid movement. Dry, soluble contaminants
dissolved in rainwater, surface run-off onto the site, or water applied
through human activity will percolate into the soil. Liquid contaminants
may percolate downward either as a solute in water or as a separated
phase. After rainwater or an organic liquid infiltrates the surface of
the ground, it travels vertically downward through the unsaturated zone
4-22

-------
(vadose zone) until it reaches the water table. Thereafter, its flow in
the saturated zone is primarily horizontal.
Darcy's law can be used to calculate the velocity of percolating rain-
water flowing through the vadose zone. However, the hydraulic conductivi-
ty must be corrected to reflect the effect of partially-filled pore spaces
when the hydraulic loading is below that necessary to support saturated
flow.
Interstitial pore water velocity for unsaturated transport through
the vadose zone can be calculated as follows (Enfield 1982):
vpw = q/e	(4-15)
where
Vpw = interstitial ground-water (pore water) velocity (length per
unit time)
q = average percolation or recharge rate (depth per unit time)
8 = volumetric water content of the unsaturated zone (decimal
fraction, representing volume of water per volume of soil).
This equation applies to steady-state conditions or those that can be
assumed to be steady state. For hydraulic loading under non-steady-state
conditions, q and e will vary with time and depth. Additionally, the dis-
tribution of q and e will vary as the water migrates downward, making
determination of the average transport velocity burdensome. For situa-
tions where steady-state conditions cannot be assumed, the analyst should
use a computer model; for example, SESOIL (one component of EPA's GEMS
computer system) calculates the time of travel for seasonally varying
rainfall rates.
The volumetric water content, e, in the unsaturated zone can be
estimated using the following equation (Clapp and Hornberger 1978):
e = (es) (q/Ks)1/(2b+3)	(4-16)
where
0 = volumetric water content in the unsaturated zone (volume/
volume or unitless)
es = volumetric water content of soil under saturated conditions
(volume/volume or unitless)
4-23

-------
q = percolation rate (assumed to be equal to the unsaturated hydraul-
ic conductivity term in original Clapp and Hornberger equation)
(depth per unit time)
Ks = saturated hydraulic conductivity (depth per unit time)
b = soil-specific exponential parameter (unitless).
Representative values of e, b, and the term l/(2b+3) are listed
in Tables 4-6 and 4-7.
The saturated volumetric water content, e$, saturated hydraulic
conductivity, K$, and the exponential coefficient, b, are all related
to soil properties. The most reliable values for these parameters are
empirical values (if available) measured during site investigation. Where
empirical values are unavailable, the values in Tables 4-6 and 4-7 provide
guides for the rough estimation of e$ and the term (l/2b+3). Representa-
tive values of Kg from two different sources are presented in Tables 4-8
and 4-9. These tables demonstrate the variability in estimates for these
values.
Theoretically, the value of e cannot exceed es, which is the saturated
soil moisture content. Therefore, when e calculated by Equation 4-16
equals or exceeds e$, it must be assumed that saturated conditions exist.
In such cases, Equations 4-19 and 4-20 from the next subsection should be
used to calculate saturated flow velocities.
Similarly, the minimum value for e that is applicable to Equation 4-16
is the field capacity of the soil. This value represents the volumetric
moisture content remaining in the soil following complete gravity drainage
and is the moisture content below which downward flow of water caused by
gravity through unsaturated soil ceases. Field capacity is a function of
soil type; the most reliable values are those measured empirically.
Where measured values are not available, default values can be taken from
Table 4-6. Wherever Equation 4-16 results in a value for e that is
less than the specific retention of the soil, it should be assumed that
no downward movement of moisture (and dissolved contaminant) occurred for
the associated time increment, and that V is equal to zero.
pw
4-24

-------
762BH
Table 4-6. Representative Values for Saturated Moisture Contents and
Field Capacities of Various Soil Types
Saturated noisture content	Field capacity
(0s)a	(cm /cm r
n^/cm^)^
Nunber of soils	Mean + 1 standard deviation Mean + 1 standard deviation
Sand	762	0.437	0.347 - 0.500	0.091	0.018 - 0.164
Loamy sand	338	0.437	0.368 - 0.506	0.1Z5	0.060 - 0.190
Sandy loam	666	0.453	0.351 - 0.555	0.207	0.126 - 0.288
Loam	3B3	0.463	0.375 - 0.551	0.270	0.195 - 0.345
Silt loam	1.206	0.501	0.420 - 0.582	0.330	0.258 - 0.402
Sandy clay	loam 498	0.398	0.332 - 0.464	0.255	0.186 - 0.324
Clay loam	366	0.464	0.409 - 0.519	0.318	0.250 - 0.386
Silty clay	loam 689	0.471	0.418 - 0.524	0.366	0.304 - 0.428
Sandy clay	45	0.430	0.370 - 0.490	0.339	0.245 - 0.433
Silty clay	127	0.479	0.425-0.533	0.387	0.332-0.442
Clay	291	0.475	0.427 - 0.523	0.396	0.326 - 0.466
aFram total soil porosity measurenents con^iled by Rawls et al. (1982) from nunerous sources.
''Water retained at -0.33 bar tension; values predicted based on coopiled soil property
measurenents.
Source: Rawls et al. 1982.
4-25

-------
7628H
Table 4-7. Representative Values of Hydraulic Parameters
(Standard Deviation in Parentheses)

No. of


1

c
Soil texture
soils3
bb

2b*3
0
5
Sand
13
4.05
(1.78)
0.090
0.395
(0.056)
Loamy sand
30
4.38
(1.47)
0.085
0.410
(0.068)
Sandy loan
204
4.90
(1.75)
0.080
0.435
(0.086)
Silt loam
384
5.30
(1-87)
0.074
0.485
(0.059)
Loam
125
5.39
(1.87)
0.073
0.451
(0.078)
Sandy clay loam
80
7.12
(2.43)
0.058
0.420
(0.059)
Silt clay loan
147
7.75
(2.77)
0.054
0.477
(0.057)
Clay loam
262
8.52
(3.44)
0.050
0.476
(0.053)
Sandy clay
19
10.40
(1.64)
0.042
0.426
(0.057)
Silt clay
441
10.40
(4.45)
0.042
0.492
(0.064)
Clay
140
11.40
(3.70)
0.039
0.482
(0.050)
a Nunber of individual soil sables included in data ccnpiled by Clapp and
Homberger (1978).
^ Empirical parameter relating soil matrix potential and noisture content;
shown to be strongly dependent on soil texture.
c Volmetric soil moisture content (voluae of water per voluae of soil).
Source: Adapted from Clapp and Homberger 1978.
4-26

-------
7628H
lable 4-8. Representative Values of Saturated
Hydraulic Conductivity
Hydraulic conductivity
Soil texture	Ninber of soils9	(Ks; aa/sec)b
Sand
762
5.8
X
10'3
Loamy sand
338
1.7
X
Hf3
Sandy loam
666
7.2
X
io~4
Loam
383
3.7
X
io"4
Silt loam
1,206
1.9
X
io"4
Sandy clay loan
498
1.2
X
io~4
Silt clay loam
366
4.2
X
10'5
Clay loam
689
6.4
X
Hf5
Sandy clay
45
3.3
X
io"5
Si It clay
127
2.5
X
io"5
Clay
291
1.7
X
10~5
a Nunber of individual soil sanples included in data compiled by Rawls
et al. (1982).
^ Predicted values based on ccnpiled soil properties.
Source: Adapted from Rawls et al. 1982.
4-27

-------
7628H
Table 4-9. Saturated Hydraulic Conductivity Ranges for
Selected Rock and Soil Types
Saturated hydraulic
conductivity (a*/sec)
Soi Is
linweathered marine clay
Glacial till
Silt, loess
Silty sand
Clean sand
Gravel
5 x 10
-10
-11
10
10
-7
10
-5
10
-4
10
-1
-	10
-	10
-	10*
-	10
-	1
-	102
-7
-4
-1
Rocks
Unfractured metamorphic
and igneous rock
Shale
Sandstone
Limestone and dolomite
Fractored igneous and
metamorphic rock
Permeable basalt
ICarst limestone
10
8
5 x 10
lo8
5 x 10
-12
-8
10
10
10
-5
— 10
-2
-7
- 10
- 5 x 10
- 5 x 10
-4
-2
Sources: Adapted frcn Freeze and Cherry 1979.
4-28

-------
Note that the percolation rate, q, cannot exceed the saturated
hydraulic conductivity, K$, for the site soil. Whenever q > Kg (and
therefore e as calculated by Equation 4-16 >. es) for the duration of the
study period, it must be assumed that saturated conditions exist and that
saturated flow prevails. Equations 4-19 and 4-20 in the next subsection
provide a means of estimating saturated flow velocities.
The following equation can be used to estimate the term q (Enfield
1982):
q = HL + Pr - ET - Qr	(4-17)
where
q = percolation rate
HL = hydraulic loading from manmade sources (depth per unit time)
Pj. = precipitation (depth per unit time)
ET = evapotranspiration (depth per unit time)
Qr = runoff (depth per unit time).
Records of estimated percolation rates for the site locality during
the time period in question (or annual average percolation rate estimates)
are often available from local climate or soil authorities, including
regional U.S. Geological Survey (USGS) and U.S. Soil Conservation Service
offices.
An estimation procedure can be used to evaluate percolation rates, q,
at sites where the sources listed above cannot provide them directly.
This estimation procedure requires data for precipitation, evaporation,
and runoff rates. In addition to the above sources, the National Weather
Service, Forest Service offices, National Oceanic and Atmospheric Adminis-
tration (NOAA) gauging stations, or other first order weather stations
(e.g., at local airports) are possible sources for these three types of
data.
The average precipitation rate per unit time, Pr, for the study period
can be obtained from various local weather authorities such as those
listed above.
4-29

-------
A value of ET for substitution into Equation 4-17 is estimated by
using measured Class A pan evaporation rates (a measure of local evapora-
tion rates under standardized conditions, available from the nearest NOAA
gauging station) in the equation:
ET = EVAP x Cet x Cveg	(4-18)
where
EVAP = region-specific or site-specific measured evaporation rates
(depth per unit time)
Cet = correction factor for converting measured pan evaporation
rates to evapotranspiration rates from turf grass (unitless)
Cveg = correction factor for converting evapotranspiration from turf
grass to evapotranspiration from other vegetative cover types
(unitless).
Values for Cgt are taken from Table 4-10, which requires climatological
and pan descriptive information.
The term Cyeg is used mainly for agricultural crops (see Table 4-11)
and varies with the thickness, depth, and characteristics of vegetative
cover. Typical values are 0.87 for shorter broadleaf plants (alfalfa) to
0.6 for taller broadleaf plants (potatoes, sugar beets) and 0.6 for taller
grains and grasses. Where crop-specific data are unavailable, a conserva-
tive default value for this term is the smallest reasonable value, or 0.6.
A value of Qr, or the average runoff over the study period, can be
estimated for Equation 4-17 using the method of Mockus (1972). A more
reliable value for this term, however, can be obtained from local USGS
gauging stations. For relatively level sites, a reasonable conservative
default value is Qr = 0, where site-specific data are unavailable or
cannot be estimated.
The above method for predicting the velocity of water migrating
through the vadose zone is the best approximation available; however, real
world heterogeneities, such as root holes and macropores, can result in
faster velocities than predicted. The analyst is not expected to correct
for this, yet it is important to be aware of the limitations of the
method.
4-30

-------
/ UCOli
Table 4-10. Suggested Value for Relating Evaporation from a U.S. Class A Pan to
Evapotranspiratlon from 8- to 15-cm Tall. Well-Watered Grass Turf
Pan surrounded by a short green crop	Pan surrounded by a dry surface ground

Upwind



Upwind




fetch of
Average regional

fetch of dry

Average regional

Wind
crop
relative tumidity.
X*
fallow
relative huniditv. X*


(m from pan)
20-40
40-70
>70
(m from pan)
20-40
40-70
>70

0
0.55
0.65
0.75
0
0.7
0.8
0.85
Light
10
0.65
0.75
0.85
10
0.6
0.7
0.8
<170 km/day
100
0.7
0.8
0.85
100
0.55
0.65
0.75

1000
0.7
0.85
0.85
1000
0.5
0.6
0.7

0
0.5
0.6
0.65
0
0.65
0.75
0.8
Moderate
10
0.6
0.7
0.75
10
0.55
0.65
0.7
170-425 km/day
100
0.65
0.75
0.8
100
0.5
0.6
0.65

1000
0.7
0.8
0.8
1000
0.45
0.55
0.6

0
0.45
0.5
0.6
0
0.6
0.65
0.7
Strong
10
0.55
0.6
0.65
10
0.5
0.55
0.65
425-700 km/day
100
0.6
0.65
0.7
100
0.45
0.5
0.6

1000
0.65
0.7
0.75
1000
0.4
0.45
0.55

0
0.4
0.45
0.5
0
0.5
0.6
0.65
Very strong
10
0.45
0.55
0.6
10
0.45
0.5
0.55
>700 km/day
100
0.5
0.6
0.65
100
0.4
0.45
0.5

1000
0.55
0.6
0.65
1000
0.3
0.4
0.45
*Hean of maxinun and minimun relative humidities.
Source: Jensen 1973, as presented by Enfield et al. 1982.

-------
7628H
Table 4-11. Crop Coefficients for Estimating
Evapotranspi rat ion
Crop
Period
Coefficient
^eg'
Alfalfa
April 1 - October 10
0.B7
Potatoes
May 10 - Septenber 15
0.65
Snail grains
April 1 - July 20
0.6
Sugar beets
April 10 - October 15
0.6
Source: Jensen 1973, as presented by Enfield et al. 1982.
4-32

-------
(3) Horizontal transport. Darcy's law is used to describe the volu-
metric flow of water through a porous medium under saturated conditions.
The volumetric flow (or discharge) is proportional to the product of the
driving force, the soil's ability to transmit water, and the cross-
sectional area perpendicular to the flow direction. The driving force is
the difference in the energy (hydraulic head) between two points in the
aquifer divided by the distance between the two points. This driving
force is called the hydraulic gradient. A soil's ability to transmit
water is represented by an empirically determined coefficient of hydraulic
conductivity. The hydraulic conductivity is determined by the properties
of the liquid (water or contaminant) and the permeability of the porous
medium. The soil has an intrinsic property of permeability, which is
determined by the size, orientation, and connectedness of the pore spaces.
Estimating contaminant velocity is based on estimating the velocity
of water. For those contaminants that flow as water flows, contaminant
velocity equals water velocity (vertical or horizontal). For those that
flow at rates that differ from water, the estimated water velocity must
be adjusted to approximate that of the contaminant.
Ground-water velocity can be determined for both the saturated zone
and the vadose (unsaturated) zone. Saturated zone velocity is discussed
in this section, while vadose zone velocity is examined in the next
section.
Ground-water velocity in the saturated zone is calculated using
Darcy's law, which is as follows (Bouwer 1978):
v = Ksi	(4-19)
where
v = Darcy velocity of water, also termed superficial velocity, or
specific discharge (length/time)
Ks = hydraulic conductivity of soil or aquifer material (length/
time)
i = hydraulic gradient (length/length).
4-33

-------
However, v, the Darcy velocity, is not the real macroscopic velocity of
the water, but the velocity as if the water were moving through the entire
cross-sectional area normal to the flow, solids as well as pores (Bouwer
1978). The ground-water velocity is calculated from the Darcy velocity
by dividing it by soil porosity, or, for precise modeling, by effective
porosity. (This approach takes into account the fact that the entire
cross-section of the pore is not flowing because of boundary layer
effects.) For clay soils, the effective porosity also corrects for the
effect of electroosmotic counterflow and the development of electrokinetic
streaming potentials (Bouwer 1978). The equation for calculating ground-
water velocity from Darcy velocity using effective porosity is as follows
(Bouwer 1978):
vpw = V/Pe	(4-20)
where
Vpw = ground-water (pore water) velocity (length/time)
v = Darcy velocity (superficial velocity, specific discharge)
(length/time)
Pe = effective porosity (dimensionless fraction).
The above terms should be determined for the site being studied. If
this is not possible for all parameters, then literature values can be
used for those parameters that are not available. Literature values for
saturated hydraulic conductivity are presented in Table 4-8 (Rawls 1982)
and Table 4-9 (Freeze and Cherry 1979).
The hydraulic gradient (the change in the elevation of the water table
over distance from the site) should also be taken from field data devel-
oped during the site investigation. Water levels in existing nearby wells
can provide an indication of hydraulic gradient. Table 4-6 lists values
for saturated moisture content, which is roughly equal to the effective
porosity, or Pg, for several soil types.
If site-specific information on effective porosity is available, it
should be used; however, literature values for soils with the same hy-
draulic conductivity do provide sufficient accuracy. Effective porosity,
Pg, can be approximated by the difference between the moisture content
4-34

-------
~
at saturation and at the "wilting point" (-15 bar) . The equation is
as follows (Rawls 1986):
^e = nsat " n(-15)	(4-21)
where
Pe = effective porosity (fraction, dimensionless)
nsat = water content when the pores are fully saturated (fraction,
dimensionless)
n(-15) = wiping point moisture content (fraction, dimensionless).
This estimation procedure addresses the fraction of the pore spaces
that contributes to flow, but does not address the effect of electro-
osmotic counterflow and the development of electrokinetic streaming poten-
tials. For clays, this can be a significant difference. Literature
values listed in Table 4-6 should be used for clay solids (these values
incorporate the effects of the clay's ionic double layer (Rawls et al.
1982); either technique can be used for sand or loam soil.
The above method for predicting the average velocity of ground water
is the most widely accepted approximation; however, it is only an approxi-
mation, and further refinement can be made to this approach to improve
its accuracy. Corrections for the path length difference between the
straight line distance versus the tortuous path that ground water flows
through can improve the precision (Freeze and Cherry 1979). This refine-
ment is not included for Superfund exposure assessments because the
literature does not provide a consistent correction factor to apply.
(4) Net dilution. Dilution (mixing) in ground water is different
from dilution in air and in surface water. In both air and surface
water, dilution is a major phenomenon. Flow in both air and surface
The wilting point is determined by drawing a suction of -15 bar to
draw water out of the soil in a manner similar to the suction of a
plant root; the bar is a measure of pressure (dynes/cnr).
4-35

-------
water can be turbulent. Turbulent flow means that all the flow paths are
not essentially parallel to the gross direction of motion. Flow
components that are perpendicular to the bulk fluid motion cause the
plume to spread laterally, thus reducing the concentration in the plume,
while making the plume contaminate a larger volume of air or surface
water.
However, in ground water, the magnitude of dilution is much smaller
and turbulent flow rarely exists. The slow speed of ground water coupled
with the straightening effect of many soil particles keeps the flow smooth
and laminar. In an idealized conceptual model, the interconnecting pore
spaces can be thought of as forming flow channels or tubes; any tendency
for the flow to eddy is resisted by the sides of the flow channel. How-
ever, since the interconnecting pore spaces do not make a continuous flow
channel, in real soil there is some lateral mixing. Dispersion in ground
water is not affected by eddy currents. Dispersion (neglecting diffusion
for the moment) is caused by four principal phenomena: varying pore
sizes, varying path length, variation in velocity gradient across pore
space, and flow splitting around soil particles with mixing within the
pore space. The first three phenomena contribute to longitudinal disper-
sion; the last phenomenon causes lateral dispersion. In ground water, the
magnitude of the mixing is much larger for longitudinal mixing than for
lateral mixing. Researchers have reported longitudinal dispersivity
values ranging from 2 to 25 times higher than transverse dispersivity
values (Gelhar et al. 1985).
In ground water, dilution occurs at a much slower rate than it does
in air or surface water. For short-term releases (spills), longitudinal
mixing and resulting dilution of plume concentrations can be significant.
This is because the plume can effectively mix with the uncontaminated
water in front of and behind the slug of contamination, whereas continuous
sources result in a plume of such length that its middle section cannot
effectively mix with clean water in front or behind it.
(5) Adjustments for hydrology. Use of monitoring data to assess the
depth of contamination for the purposes of calculating volume or mass in
4-36

-------
the plume must be done carefully. The difference between monitoring and
pumping wells (i.e., those intended to provide a water supply) will affect
the interpretation of the concentrations found in the wells. Monitoring
wells are more desirable, but since most existing wells will be pumping
wells, monitoring wells typically will have to be installed. The cost
burden associated with drilling monitoring wells may cause the analyst to
rely on existing pumping wells.
Monitoring wells extract a small quantity of water (a sample); this
minimizes the well's influence on the flow of the ground water. Since
these wells do not induce a large vertical component in the ground-water
flow, they sample a horizontal slice of the aquifer. The concentration
in a sample removed from a monitoring well represents a concentration at
the depth of the well screen. Thus, monitoring wells at various depths
can be used to assess the depth of contamination.
Pumping wells draw large quantities of water from an aquifer. This
causes a cone of depression to form on the water table and influences the
flow direction above and beneath the well screen. Pumping wells induce
vertical flow in the aquifer near the well. This vertical movement causes
the concentration in the well to reflect the average concentration for a
depth range that is substantially greater than that intended, because
water will be drawn from above and below the well screen. Thus, the well
water does not reflect the concentration of a particular depth but rather
an average concentration from a range of depths.
(6) Integration of fate processes.
(a) Retardation effects. Hydrophobic or cationic contaminants that
are migrating as a dilute solute are subject to retardation effects.
Concentrated plumes are not subject to this phenomenon. Contaminant
migration as a dilute solute in mixed solvents will show some retardation,
although not as much as when present in dilute concentrations in ground
water.
Algorithms describing retardation are based on the assumption that ad-
sorption of hydrophobic contaminants to soil particles is due to sorption
4-37

-------
to organic carbon in the soil. Basing the adsorption coefficients on soil
organic carbon rather than total mass eliminates much, but not all, of the
variation in sorption coefficients between different soils. The remaining
variation may be due to other characteristics such as surface area of soil
particles per mass of soil, which is a function of particle size. Numer-
ous studies of the correlation of Kd (the factor describing distribution
of a chemical between interstitial water and soil particles) with various
soil variables have found that the organic carbon content usually gives
the most significant correlation. Furthermore, this correlation often
extends over a wide range of organic carbon content — from 0.1 percent to
nearly 20 percent of the soil in some cases (Lyman et al. 1982).
This does not imply that hydrophobic contaminants will not adsorb on
minerals free of organic matter. Some adsorption will always take place,
and it may be significant under certain conditions, such as clay soils
(high surface area per mass of soil) with very low organic carbon content.
Unfortunately, methods for estimating adsorption coefficients under these
conditions are not available at present (Lyman et al. 1982).
The equation used to calculate the retardation is as follows
(Kent et al. 1985):
Rd = 1 + (0 Kd)/Pt	(4-22)
where
R(j = retardation factor (unitless)
p = bulk density (g/ml)
pt = total porosity (unitless)
K(j = distribution factor for sorption on aquifer medium (from
sorption isotherm column studies, or from regression equation
based on the octanol/water partition coefficient, in ml/g).
The use of the retardation factor is described in the following
equation (Kent et al. 1985):
Rd = Vpw/Vc	(4-23)
where
Rj = retardation factor (unitless)
Vpw = velocity of ground water (length/time)
Vc = velocity of contaminant (length/time).
4-38

-------
The term is based on sorption isotherm column studies. While
this approach is precise, the analyst will typically have to work with
estimated parameters. For hydrophobic contaminants, the term can be
estimated from the term Koc by using Equation 4-8 (Lyman et al. 1982),
previously presented. The term "fraction of organic carbon" (OC) in
Equation 4-8 is precisely calculated when taken from empirical measure-
ments of the soil in the study area; if this is not possible, estimates
can be made. For the vadose zone velocity, a value of OC from Rawls
(1986) provides a good estimate. Rawls' work focused on soils near the
surface, the area of interest to agriculture. For saturated zone veloci-
ty, the analyst has two choices. If the subsoil comes from igneous or
metamorphic rock, the value of OC decreases with depth. The actual value
may be quite low; however, the model to predict retardation is only useful
down to 0.1 percent. For this situation, the analyst should use 0.1 per-
cent for the OC or assume no retardation and set equal to 1. If the
subsoil comes from sedimentary rock, the value of OC distribution may be
similar to the distribution for agricultural soils done by Rawls. The
variation of OC with depth may be relatively constant. The carbon, which
was at the surface at one time, has been buried over geological time.
Hence, the analyst should use a value of OC from the Rawls (1986) distri-
bution for the saturated zone velocity determination (Trask 1942). Soil-
water partition coefficients have been developed for many important
contaminants (Callahan et al. 1979, Mabey et al. 1982). If Koc is not
known, it can be estimated from the regression equations that relate KQC
to Kqw (octanol-water partition coefficient).
The retardation effects computed from the octanol-water partition
coefficient (K ) relate the concentration in polar solvent (water) to
uw
the concentration in hydrophobic solvent (usually octanol, which simulates
the soil organic carbon). If the contaminated plume has a large concen-
tration of organic chemicals dissolved in the ground water, the actual
partitioning will be from a solvent-organic chemical system. This will
raise the concentration in the fluid and lower the concentration on the
soil organic carbon. This shift in partitioning will lower R^, (i.e.,
4-39

-------
the contaminant will migrate at a speed closer to that of ground water).
The analyst should be aware of the general influence of this effect, but
should not model the precise numerical difference. For dilute plumes, the
analyst should model full retardation; however, for concentrated plumes,
no retardation should be modeled.
Retardation can be modeled for complex leachates and separated phases
of immiscible liquids, but the methods are not presented in this report.
The analyst should refer to Nkedi-Kizza et al. 1985, Rao et al. 1985, and
Woodburn et al. 1986, for guidance on performing these calculations.
(b) Degradation effects. Data for rates of degradation in ground
water are very limited. Both hydrolysis and biodegradation are important
processes in the vadose (unsaturated) zone, but in the saturated zone, the
microbial population may not be sufficient for biodegradation to occur.
If data are available, this type of estimation can be performed by using
an overall decay coefficient, which represents a combination of all decay
and loss processes affecting the removal of a substance being transported
by flowing water (Delos et al. 1984). Generally, the rate of the princi-
pal fate process is sufficient for the estimation.
4.3 Estimation of Concentration - Ground Water
There are several different approaches to estimating the concentration
of a chemical of interest at the receptor. One approach is based on the
proportionality of volume and concentration of the waste versus those of
ground water.
For liquid contaminants, the following proportionality exists:
^1 ^1 = Vgw CgW	(4-24)
where
Vj = volume of liquid chemical released (liters)
VqW = volume of contaminated ground water (liters)
Cj = average concentration of chemical contaminant in the
released liquid (mg/1)
Cgw = average concentration of contaminant in ground water (mg/1).
Both volumes and concentrations should be in the same units.
4-40

-------
For solid contaminants, the following proportionality exists:
«c = ^gw ^gw
(4-25)
where
Mc = mass of solid waste (mg)
Cc = concentration expressed as mass fraction, fraction of
contaminant in waste (dimensionless)
Cgw = concentration of contaminant in ground water (mg/1)
VgW = volume of contaminated ground water (liters).
If at least limited ground-water monitoring data at the release point
are available and sufficient environmental fate data are available to
calculate an overall decay rate (see Section 4.2.2) the concentration at
the receptor well can be calculated.
The concentration of the degrading substance at a selected point
downgradient from the release point is given by the following equation:
W(x) = concentration at downgradient distance x (mass/volume)
W(0) = concentration immediately below point of introduction
e = 2.71828
K = overall decay coefficient (time)"1
x = distance downgradient from point of introduction (length)
Vc = contaminant velocity (length/time) from Equation 4-23.
In the absence of ground-water monitoring data, mathematical models
are often used to estimate concentrations of contaminants in ground water
at the receptor wells. The Office of Solid Waste (OWS) has developed a
simplified model for its delisting program that relates the leachate con-
centration from a landfill to the concentration at a hypothetical receptor
well 500 feet downgradient from the landfill site. This procedure (the
VHS model) involves back-calculation starting with a health-based ground-
water concentration at the receptor well. The only data required by the
VHS model are the leachate concentration, the annual volume of wastes dis-
posed, and the constituent concentrations of toxicants in the wastes (to
W(x) = W(0)e Vc
(4-26)
where
4-41

-------
ensure that they are present in sufficient mass to sustain leaching).
The model calculates a different dilution factor depending on the annual
volume of waste disposed of at the site. All other input parameters to
the model are fixed at reasonable worst-case values. While recognizing
the oversimplifications built into this model, the assessor can consider
the VHS model useful for screening purposes under CERCLA; is not recom-
mended for site-specific in-depth analysis.
Several references are available that provide detailed derivations and
outline the application of more sophisticated models for the analysis of
contaminant migration in the saturated and unsaturated zones. The analyst
is referred to the following documents: Van Genuchten and Alves (1982);
Walton (1984); Javendel et al. (1984); USEPA (1986b), Geotrans (1986); and
van der Heijde et al. (1985, 1987). Tables 4-12 and 4-13 present informa-
tion regarding the features and input data requirements, respectively, of
some of the currently available modeling procedures for in-depth assess-
ment, of contaminant fate in ground water. The outputs from these models
are ground-water concentrations of contaminants at specified distances
from the source. Two of the models addressed in Tables 4-12 and 4-13 are
SESOIL and AT123D, which are part of GEMS and which have received wide
usage both inside and outside of the Agency. AT123D is of particular
note, because it is versatile and is applicable to a wide range.of situa-
tions; specifically, it is capable of simulating contaminant transport and
fate under 300 different user-selected situations (Yeh 1981). This model
determines contaminant concentraiton at any downgradient and lateral
distance and depth specified by the user, as a function of time from the
beginning of release at the source. Information on other ground-water
models is contained in USEPA (1988b).
Versar (1987a) has developed a screening level procedure for estimat-
ing contaminant concentrations in ground water at the receptor; this is
the result of an Office of Toxic Substances working group on developing
screening level procedures for ground water exposure assessment. This
procedure, which is shown in Worksheet VI of Appendix B, allows applica-
tion of a unit concentration in ground water (mg/1 per KKg/year released)
4-42

-------
Table 4-12. Features of Unsaturated Zone and Ground-Water Models

-------
Table 4-12 (continued)
LPMM
Kant M al. 1985
ONE-D
van Qanuchtan and Alvaa 1082
Plum* (3D)
Codail at al. 1S82
Rataq
Javandal at al. 1884
SLM
HuyaKorn at al. 1887
VHS
Demanleo and Palolauakaa 1882
RW-Analyt.
van dar Hal|da and Srlnlvaaan 1986
Baavar Soft
Baar and Vamilft 1887
Klnzalbaoh
Klnzalbaoh 1888
Walton 35
Walton 1885
Soluta
Bal)ln 1885
Famtran
Martlnaz 1885
Famwatar/Famwaata
Yah 1887, Yah, and Ward 1881
Porllo
Runehal at al. 1885
QS2/Q91
Oavto and Sagol 1885
PRZM
Caraal at al. 1884
SESOIL
Bonazountaa and Wagnar 1884
SUMATRA-1
van Qanuehtan 1878
SUTRA
Voaa 1884
TRIPM
Quraghlan 1883
CFEST
Gupta at al. 1887
FEWA/FEMA
Yah and Huff 1885
HST3D
Ktpp 1887
MOC(USaS 2-D)
KonBiow and Badahoaft 1878
MOC DEN9E
Sanford and Konlkowr 1885
MOC NRC
Tracy 1882
RW3TM
Prlckatt at al. 1881
SWENT
Intara 1881
SWIFT II
Raavaa at al. 1888
TRACR 3D
Travla 1884
TRAFRAP/WT
Huyakom at al. 1887
BIOPLUME-II
PFATiN
Rlfal at al. 1887

-------
Table 4-12 (continued)
Domanteo and Palelauakaa 1082
RW.Analyt.
van dar- Haljda and Srtntvaaan 1086
Baaver Soft
ndVarruift 1087
Kinzalbach
Ktnzalbaeh 1086
Walton 35
Walton IMS
Baijfn 1965
Uartlnaz 1085
Famwatar/Famwaata
Yah 1067, Yah, and Ward 1661
Porflo
Runehal at al. 1065
and Sagol 1065
PRZM
Caraal at al. 1084
SESOIL
Bonaiountaa and Wagnar 1064
SUMATRA-1
van Qanuehtan 1078
SUTRA
TRIPM
Guraghlan 1083
CFEST
Gupta at aJ. 1087
FEWA/FEMA
Yah and Huff 1085
HST3D
Klpp 1087
MOC(USGS 2-D)
Konlkow and B*dahoaft 1978
MOC DENSE
Sanford and Konlkow 1965
MOC NRC
Tracy 1082
RWSTM
Prickatt at aJ. 1081
SWENT
tntara 1083
SWIFT II
naavaa at al. 1886
TRACR 30
Travto 1084
TRAFRAP/WT
Huyafcom at al. 1087
BIOPLUME-II
RHai at al. 10ST
PESTAN
EnftaM at al. 1082
HELP
•NON-AQUEOUS PHASE LIQUIDS 1) FOR UNSATURATED ZONE ONLY.
4-45

-------
Table 4-13. Data Requirements for Unsaturated Zone and Ground-Water Models

-------
to a release rate at the site to yield the ground-water concentration at
a hypothetical receptor. These unit concentrations, which are derived
using the SESOIL/ATD123 models, are selected from a table based on the
Henry's law constant and Koc values. The values in the table in
Worksheet VI of Appendix B are 70-year average concentrations at a well
200 meters from the edge of a 1-hectare site at which occurs a release of
1 KKg per year for a period of 10 years. Two geographical locations rep-
resenting sandy loam and silt loam soils are represented in these numbers.
The total distance from the soil surface to the water table is assumed to
be 8 meters for this screening level procedure.
4.4 Exposure Estimation - Ground Water
This section has presented methods for estimating ground-water concen-
tration at the receptor. Use of these estimated concentrations allows
the assessor to estimate exposure based on the equations and parameter
values presented in the Exposure Factors Handbook (USEPA 1989).
4-47

-------
5.	ASSESSMENT OF EXPOSURE VIA FISH/SHELLFISH
5.1	Release Analysis - Fish/Shellfish
Water contamination with subsequent uptake into the food chain pro-
vides an indirect route of human exposure to contaminants. Foods subject
to direct or indirect water contamination are those derived from fish and
shellfish.
5.1.1	Screening of Release Potential
Chemical contaminants may reach fish and shellfish habitats through
several release mechanisms, which include (1) precipitation or dry deposi-
tion (i.e., dust) into surface water from the atmosphere; (2) discharge
of contaminated ground water into surface water; (3) surface runoff from
contaminated sites; and (4) on-site treatment processes leading to direct
discharge of the contaminant into the surface water. Determination of
the potential for releases of a contaminant via these mechanisms will
indicate whether an assessment of this exposure pathway is necessary.
Section 3 discusses these mechanisms in more detail.
5.1.2	Quantitative Release Estimation
Releases from surface water are relevant in estimating potential for
aquatic food contamination. Section 3.1.2 (surface water) describes the
quantitative release estimation techniques that may be used for this
estimation depending on site characteristics.
5.2	Fate Analysis - Fish/Shellfish
5.2.1 Screening of Fate Processes
Fate analysis for this exposure pathway is conducted using the meas-
ured or estiamted concentration data developed for surface water, air,
and ground-water pathways. The human populations that may be potentially
exposed are assessed in terms of their likelihood of experiencing food
chain exposure. Any pathways that are identified as potential sources of
human exposure are then assessed more quantitatively to determine the rate
at which contaminants may be introduced into the food chain.
5-1

-------
5.2.2 Quantitative Fate Estimation
In estimating the fate of a chemical in biotic populations, two major
processes must be considered: bioavailability and bioaccumulation.
Bioavailability is a measure of the tendency of a chemical to partition
between the environmental media and terrestrial organisms. The octanol-
water partition coefficient is the parameter generally used to indicate
the bioavailability of organic chemicals to aquatic organisms. Bioaccumu-
lation refers to the uptake of contaminants by organisms; this uptake is
an indication of the contaminant's lipophilicity. The bioconcentration
factor is the parameter used as a measure of bioaccumulation potential for
both organic and inorganic chemicals.
(1) Bioavailability
(a) Organic chemicals. Aquatic organisms, both pelagic and
benthic, are exposed to organic chemicals almost entirely via contact with
contaminated water. Moreover, an organic chemical usually must be uncom-
plexed and dissolved in the water for it to be absorbed by an aquatic
organism (Hamelink and Spacie 1977, McCarthy et al. 1985, Landrum et al.
1985). The reason for this limitation is that sorption or complexation of
the chemical compound with particulates or dissolved macromolecules pre-
vents transport of the chemical across cell membrances and consequently
into the body of the organism.
Most current investigations stress the importance of sorption to
sediments, suspended particulates, and dissolved humic substances as the
principal process controlling the bioavailability of organic chemicals in
aquatic ecosystems. For chemicals that are considered to be persistent in
water, an evaluation of sorptive processes is probably sufficient. Most
chemicals that are introduced into the environment, however, vary greatly
in their persistence in water as a function of degradation or mobility
characteristics.
Estimation of the bioavailability of an organic compound in the
aquatic environment requires information on the ambient concentration of
5-2

-------
the chemical in the water and the extent of its sorption to dissolved or-
ganic macromolecules (i.e., humic matter). Ambient concentration levels
can be obtained from monitoring data, or they can be estimated from in-
dustrial and consumer releases and an analysis of the chemical's environ-
mental fate and transport. After ambient concentration levels have been
established, the amount of the chemical in the water column that is bio-
available, i.e., neither complexed nor bound by dissolved organic macro-
molecules, can be estimated through use of the following relationships:
where
Cdiss = concentration of chemical that is neither complexed nor
bound by dissolved macromolecules
fdom = fraction of contaminant chemical bound to dissolved
organic macromolecules
Ct = ambient concentration of chemical in surface water
Kqc = organic carbon adsorption coefficient
(mgC/L),jom = dissolved organic macromolecules expressed as
milligrams of carbon per liter.
The term	"in Equation 5-1 above represents the fraction of chem-
ical that is unbound and bioavailable. The concentration of dissolved
organic macromolecules (i.e., humic matter) ranges from 0.5 to 4.0 mg of
carbon per liter (mgC/L) in most lakes and rivers, but can reach 10 to
50 mgC/L in wetlands and marshes (Thurman 1984). Equation 5-2 was devel-
oped by McCarthy and Black (1987) from studies with lipophilic aromatic
hydrocarbons.
dependent on their speciation and on their sorption to both soluble and
insoluble components of aquatic systems. Metals are present in water,
not only as hydrated cations and oxyanions but also in association with
particulates, colloids, and soluble inorganic and organic complexing
materials, such as clays and humic acids. With respect to the types of
components involved in their aquatic distribution, the metals are somewhat
(5-2)
(5-1)
(b) Metals. Bioavailability of metals in surface water is
5-3

-------
similar to organic pollutants. They differ from organic pollutants, hoiH
ever, in the physical and chemical nature and binding strength of their
interactions with these components. No single physical property, such as
1ipophilicity, serves conveniently as a criterion for estimating sorption
in the case of metal cations. Also important in the aquatic distribution
of metals is the hardness of the water, which is a function of the concen-
trations of calcium and magnesium ions, and the alkalinity of the water.
Several studies have indicated that uncomplexed metal cations or their
low molecular weight complexes may be more readily incorporated into
aquatic organisms (especially those with gill structures) than the corre-
sponding high molecular weight metal complexes (Jackson and Morgan 1978,
Hung 1982, and Winner 1985). In test aquaria, for example, oysters have
shown decreased cadmium uptake in the presence of dissolved organic matter
that can form complexes with metals (Zamuda and Sunda 1982). This behav-
ior appears similar to observations of decreased uptake of lipophilic com-
pounds due to their sorption by humic materials. However, the mechanism
for sorption would not be the same. There are few studies from which an
empirical relationship can be derived between metal complexing and bio-
availability. Analytical methods for determining speciation under natural
conditions at low metal concentrations are not readily available (Campbell
and Evans 1987).
From studies with freshwater mussels, Campbell and Evans (1987) have
concluded that several natural complexing materials affect the bioavail-
ability of cadmium and lead in particular, and presumably affect the bio-
availability of other metals as well. For example, most of the variabili-
ty in bioconcentration of cadmium and lead could be explained by varia-
tions in pH, humic content, and water hardness. These conclusions were
based on data from laboratory and field studies, including measurements
on metal concentrations in water and mussels from ten lakes with broad
variations in characteristics.
These studies also illustrate the metal-specific differences in avail-
ability. For example, the amount of available cadmium cation (or its low
molecular weight complexes) is higher than it is for lead, given the sa|
5-4

-------
total metal concentration and similar water chemistry. Lead availability
is markedly decreased by even small quantities of humic acids; in most
surface water, very little lead is available except in waters that are
very low in pH and humic substances or very high in calcium carbonate. On
the other hand, available cadmium decreases with increasing humic sub-
stances; however, the effect is not substantial. Unfortunately, no single
environmental factor accounts for sufficient variation in the bioavail-
ability of most metals to be useful as a predictive model (Campbell and
Evans 1987).
Models for predicting humic acid binding of metal cations (without
reference to bioavailability) have been evaluated by Giesy and Alberts
(1984). Humic materials bind metals by coordinate covalence and chelation
through phenolic and carboxylic acid groups. Humic materials also occlude
metals and their sparingly soluble compounds in cavities within the macro-
molecular structure. Stability constants for these interactions as well
as the acid dissociation constants of the phenolic and carboxylic acid
groups are required for models of humic-metal equilibria. Techniques for
estimation of these constants are given by Giesy and Alberts (1984);
stability constants have been used in the GEOCHEM model (Sposito and
Mattigod 1979). Model predictions in conjunction with field observations
and laboratory data indicate that trivalent cations have higher stability
constants for binding with natural organic matter than do most divalent
trace metals which, in turn, form stronger bonds than do calcium and
magnesium (the natural hardness metals).
The relative ranking of metal cation binding by humic acids has been
provided by several investigators. Beveridge and Pickering (1980)
reported that binding constants decrease in the order Pb > Cu > Cd > Zn >
Ca > Mg, with the individual values for lead and copper being an order of
magnitude greater than those of calcium and magnesium. This relative
ranking, however, varied somewhat within the pH range of natural waters.
Schnitzer and Kerndorff (1980) developed the following relative ranking
of 11 metals for the formation of water-insoluble complexes at pH 6:
5-5

-------
Fe(111), Cr(111), A1 > Pb, Cu > Hg(ll) > Zn, Ni, Co, Cd, Mn.
There is an obstacle to the use of these relative rankings as a
measure of relative bioavailability. Some of these metals are present in
natural waters as oxyanions as well as cations (Callahan et al. 1979);
these oxyanions could be available for absorption by aquatic animals be-
cause they do not form the same type of complexes as their cation counter-
parts. Therefore, while empirical data exist on the bioavailability of
metals, there do not appear to be any general methods for estimating the
bioavailability of metals in the aquatic environment.
(2) Bioaccumulation
(a) Organic chemicals. Bioaccumulation refers to the uptake of
contaminants by organisms; this uptake is an indication of the contami-
nant's 1ipophilicity and persistence. The aquatic bioconcentration
factor, BCF, for an organic chemical indicates the extent to which the
chemical may accumulate in aquatic organisms. It is an equilibrium parti-
tioning ratio of the concentration in the organism divided by the concen-
tration in the water, as noted by the following equation:
BCF = ^organism	(5-3)
Cwater
Values of BCF range from less than 1 to over 106. The concentration
units of the numerator and the denominator must be expressed on a weight
per weight basis, and the organism's weight must be its wet weight. Lyman
et al. (1982) recommended three equations for estimating the log BCF:
log BCF = 0.76 log KQW - 0.23	(5-4)
(r2 = 0.823 n = 84)
log BCF = 2.791 - 0.564 log solubility*	(5-5)
(r2 = 0.49 n = 36)
log BCF = 1.119 log KQC - 1.579	(5-6)
(r2 = 0.757 n = 13)
* Units for solubility are in parts per million (ppm).
5-6

-------
where
r^ = the coefficient of determination (proportionate reduction in
error)
n = the number of chemicals from which the regression is developed.
If measured values for KQW or aqueous solubility are unavailable,
the techniques described in Lyman et al. (1982) should be employed to
estimate KQW from the fragment constants of the chemical in question.
The soil adsorption coefficient, KQC, can be used when it has been
experimentally determined. It is most useful when the determination has
been site specific.
Estimations of BCF from these regression equations should be used with
the understanding that the calculated values may be valid only to within
an order of magnitude (Lyman et al. 1982). Moreover, many compounds with
high estimated BCF values (i.e., BCF > 10^) move across cell membranes
very slowly, and may not reach equilibrium levels in the aquatic organism
for at least 1 month. The lipid content of aquatic organisms also varies
among species, stage of growth, and position in the reproductive cycle.
Therefore, experimental attempts to define the bioconcentration potential
of an organic compound without field verification could introduce signifi-
cant error.
Furthermore, calculated BCFs should be corrected for reduction in
bioavailability of the chemical due to binding to dissolved organic macro-
molecules, using the following equation:
Observed BCF = (1 - F^) • expected BCF	(5-7)
where
fdom = fraction of contaminant bound to dissolved organic molecules,
and consequently unavailable for uptake.
(b) Netals. The potential for bioaccumulation of metals has
not been related successfully to any chemical or physical property that
can then be used for predictive purposes. This is, of course, unlike the
5-7

-------
bioaccumulation of organic pollutants, which is thought to be a partition-
ing process directly related to environmental concentration and lipophil-
icity. Bioconcentration factors for many metals in aquatic animals are
available (Phillips and Russo 1978, Callahan et al. 1979); some are pre-
sented in Table 5-1. The level of bioconcentration appears to vary with
the aquatic conditions and the biological species of interest. Therefore,
it is recommended that estimations of metal bioaccumulation in food chain
species should be based on bioconcentration factors specifically deter-
mined for that species or a closely related one.
Bioaccumulation of metals by plants and animals may exert a major
influence on the concentrations of trace elements in aquatic environments.
Studies indicate that the uptake of a metal increases with the metal's
ability to be complexed by ligand molecules and anions (Martin 1970).
The chemical interactions responsible for metal bioaccumulation within a
living organism are somewhat similar to the chemical interactions in the
environment that regulate metal bioavailability, i.e., those metals
exhibiting the least availability in surface water as cations (or their
low molecular weight complexes) will also tend to become the most biocon-
centrated. Thus, the alkaline earth elements (e.g., barium, strontium,
magnesium, and calcium) tend to become considerably less bioconcentrated
than the heavy metal transition elements (e.g., copper, nickel, cobalt,
zinc, and manganese).
Several mechanisms for bioconcentration of metals have been suggested,
including:
•	Complexation by biological chelating molecules,
•	Replacement for similar cations with specific physiological
functions in organisms, and
•	Ion exchange and sorption on membrane surfaces (Martin 1970).
Usually, metal residue levels are higher in animals than in plants and
also higher in carnivores than in herbivores.
A useful compendium on metal bioaccumulation in fishes and aquatic
invertebrates has been prepared by Phillips and Russo (1978). Twenty-one
5-8

-------
7984H
Table 5-1. Reasonable Worst-Case Bioconcentration
Factors (BFCs) for Metals in Fish/Shellfish
Hetal	Bioconcentration Factor (BCF)a
Ant imony
40
Arsenic
333
Berylliua
100
Cadniun
3,000
ChraniiiB
440
Copper
30,000
Lead
60
Mercury
1,670
Nickel
100
Selenius
400
Silver
3.330
Thai liun
100,000
Zinc
24,000
a Based on highest value for marine or freshwater fish or
shellfish.
Source: Adapted from Callahan et al. (1979).
5-9

-------
metals of concern are considered in individual sections of this review.
The authors conclude that most metals do not accumulate in the edible
portions of fish; however, mercury, arsenic, and radioactive cesium may
reach hazardous concentrations in both fish and shellfish. In addition,
cadmium, lead, silver, and other radioactive isotopes may also exceed safe
levels for human consumption in shellfish.
5.3	Estimation of Concentrations - Fish/Shellfish
Data on the concentrations of chemical contaminants in fish and shell-
fish are quite limited. At the present time there is no mandatory fish
testing program at the federal level. Of the approximately 2.8 billion
lbs. of commercial fish consumed in the U.S. in 1982 the FDA only tested
approximately 920 fish, or 0.00058% (Public Voice 1986). A report by
Public Voice (1986) gives an overview of the status of fish testing
performed in the U.S. Because very little testing is conducted, actual
contaminant tissue concentrations in fish are scarce.
5.4	Exposure Estimation - Fish/Shellfish
The information presented in Section 5 allows the assessor to estimate
concentrations of hazardous chemicals in fish/shellfish. This information
can then be used to estimate exposures by using the equations and parame-
ters presented in the Exposure Factors Handbook (USEPA 1989).
5-10

-------
6.	ASSESSMENT OF EXPOSURE VIA TERRESTRIAL PLANT AND ANIMAL FOOD
SOURCES
6.1	Release Analysis - Terrestrial Food Sources
Soil contamination with subsequent uptake into the food chain provides
an indirect route of human exposure to chemical contaminants . Foods sub-
ject to direct soil contamination are those derived directly from plants
and those derived from animals that ingest the contaminated soil and
plants. Since most crops consumed directly by humans are washed to remove
as much adherent soil and dust as possible, the following discussion will
focus primarily on foods derived from domestic and grazing animals that
ingest soil and deposited dust in combination with uptake of contaminated
feed material.
6.1.1	Screening of Release Potential
Chemical contaminants may reach terrestrial plant and animal food
sources through several release mechanisms: (1) deposition of particu-
lates onto plants and soils with subsequent consumption by food animals;
(2) contamination of soil by runoff from the site; (3) contamination of
plants from ground water via root systems; (4) contamination of soils,
plants, and livestock from irrigation; and (5) livestock watering.
Sections 3.1.1, 4.1.1, and 7.1.1 (surface water, ground water, and air)
discuss these mechanisms in more detail.
6.1.2	Quantitative Release Estimation
Releases from surface water, air, and ground water are relevant in
estimating potential for food contamination. Sections 3.1.2 (surface
water), 4.1.2 (ground water), and 7.1.2 (air) describe the quantitative
release estimation techniques that may be used for this estimation depend-
ing upon site characteristics.
6.2	Fate Analysis - Terrestrial Food Sources
6.2.1 Screening of Fate Processes
Fate analysis in biotic populations is conducted using the ambient
concentration data developed for surface water, air, and ground water
6-1

-------
pathways. The populations that may be exposed via the relevant environ-
mental pathways are assessed in terms of their likelihood of experiencing
food chain exposure. Any pathways that are identified as potential
sources of human exposure are then assessed more quantitatively to deter-
mine the rate at which contaminants may be introduced into the food chain.
Chemicals that demonstrate unusually high bioavailability to plants and
high bioaccumulation potential in animals are of particular concern.
6.2.2 Quantitative Fate Estimation
(1) Bioavailability.
(a) Organic Chemicals. Soil invertebrates and plants require
an organic chemical to be uncomplexed in aqueous solution for direct
absorption, although invertebrates that ingest soil organic matter may be
able to take up soil-complexed organic chemicals during digestion of that
organic matter.
Most current investigations stress the importance of sorption to
sediments, suspended particulates, and dissolved humic substances as the
principal process controlling the bioavailability of organic chemicals.
For chemicals that are considered to be persistent in water or soil, an
evaluation of sorptive processes is probably sufficient. Most chemicals
that are introduced into the environment, however, vary greatly, in their
persistence in water and soil as a function of degradation or mobility
characteristics. Therefore, bioavailability should be related more
broadly to the persistence of a chemical in specific environmental media,
as determined by an analysis of its environmental fate and transport. In
estimating the fate of a contaminant in biotic populations, the assessor
should consider two major processes: bioavailability and bioaccumulation.
Bioavailability is a measure of the tendency of an organic chemical
to partition between environmental media and terrestrial organisms.
Within the soil environment, the availability of an organic chemical for
uptake by plants depends on how much of the chemical can be dissolved in
water and remain unbound by organic macromolecules (i.e., humic matter),
particulates, or clay. If the organic chemical is strongly sorbed or
6-2

-------
becomes covalently bonded to dissolved organic macromolecules, it will be
less available for transport across the outer layer of cells making up the
exterior membrane of a plant's root system. Thus, although an organic
chemical may persist in soil, it could be available to plants in only
limited quantities, depending on its ability or tendency to bind to both
dissolved and solid macromolecular matter. Polar organic chemicals
(especially ionic compounds) can become sorbed to clays as well as to
humic matter through electrostatic attraction and ion exchange reactions.
However, because soil binding is a reversible phenomenon, soil can act to
some extent as a reservoir, leading to slow release of small amounts of
compound over time.
The amount of a specific chemical that is bioavailable to plants
(i.e., not sorbed or complexed by the organic matter of soil) can be esti-
mated through use of the organic carbon partition coefficient, KQC. Thus,
the soil partition coefficient (K^) that is estimated from a chemical's
Koc value and the soil's fractional organic carbon content (OC) should
permit estimation of the unbound dissolved chemical in the water associ-
ated with soil.
Thus,
(6-1)
(6-2)
^soil " ^diss + ^sorbed ~ ^diss^ + ^d)
(6-3)
then
^diss " Csoil/(l + Kj) .
Substituting Equation 6-1 into Equation 6-4, one obtains the
following expression:
^diss = Csoil/(* + Koc • OC)
(6-4)
(6-5)
6-3

-------
where for Equations (6-1) through (6-5)
Kj = soil partition coefficient
Kqc = or9an^c carbon adsorption coefficient
OC = fractional organic carbon content of soil (default value =
0.02)
Csorbed = concentration of chemical sorbed to soil, mobile particu-
lates, and dissolved organic macromolecules
Cdiss = concentration of chemical dissolved in the water associ-
associated with soil (this is the concentration of chemical
that can be considered bioavailable)
Csoii = total concentration of chemical in soil obtained from
monitoring data or from a source and environmental fate
assessment.
The fraction of chemical in dissolved form in soil, and thus available
from soil, can be determined by rearranging Equation (6-5):
^diss/^soil = V(1 + Koc • OC)	(6-6)
(b) Metals. With regard to the soil environment, bioavailabil-
ity of metals obviously is dependent on their sorption to soil components
and the effect of this sorption on mobility. This is a problem involving
more variables than the bioavailiability of metals in aquatic systems.
An investigation of this type has been conducted by Romero and Elejalde
(1985), but their results are species- and site-specific and may be
applicable only to the plants that were studied.
(2) Bioaccumulation.
(a) Organic chemicals. Bioaccumulation refers to the uptake
of contaminants by organisms. This uptake is a measure of the contami-
nants' 1ipophilicity and persistence. The root systems of terrestrial
plants preferentially transport soil constituents that are necessary for
growth. Most of a soil xenobiotic, however, remains sorbed to the outer
layer of a plant's root system and can enter into the food chain if these
roots are ingested (USEPA 1986c). A small fraction of soil xenobiotic
will reach the cytoplasm via the root system, but direct foliar absorption
may also be significant (Russell et al. 1971, Gillett 1980). The roots of
plants generally accumulate higher concentrations of soil-applied organic
6-4

-------
chemicals than do aerial plant parts. This observation has been demon-
strated with sugar beets, carrots, turnips, wheat, and pasture grass
(Smelt and Leistra 1974, Schuenert et al. 1983).
Lipophilic partitioning does not have the same role in bioaccumulation
by plants as it does in animals. Transport of xenobiotics in plants has
also been shown to differ significantly (Christ 1978, Price 1978). Fur-
thermore, plants lack the excretory mechanisms of animals and must rely
primarily on metabolic degradation or conjugation, followed by deposition
in vacuoles, to remove xenobiotics from the cytoplasm (Matile 1976, Menzie
1980).
At present, the most useful indication of potential bioaccumulation
of organic chemicals in plants may be the chemical's persistence over the
growing season. The root system of a plant increases its volume and sur-
face area over the growing season, while the soil organic matter and other
constituents that complex the xenobiotic remain relatively constant.
Thus, since persistence in soil is usually dependent on sorption to soil
constituents, an increasing amount of xenobiotic will become sorbed to
the root system as it increases its mass over the growing season. The
persistence and bioaccumulation of the chemical in question can be com-
pared to the persistence and bioaccumulation of better-studied soil con-
taminants, such as hexachlorobenzene (USEPA 1986c). The persistence would
be determined from the fate and transport analysis and from available
monitoring data.
The bioaccumulation factor (BF) for terrestrial birds and mammals is
defined by Trabalka and Garten (1982) as:
BF = (mg/kg contaminant in fat)/(mg/kg contaminant in diet) (6-7)
where both concentrations have achieved a steady state (i.e., dynamic
equilibrium over a prolonged feeding period). This definition does not
account for biotransformation of the contaminant and assumes that accumu-
lation occurs only through lipid partitioning. Although these are ideal
conditions, a BF defined in this way could be useful in assessing dietary
exposure to humans, since xenobiotics that do not bioaccumulate in animal
6-5

-------
fat tend to be excreted quickly and present little potential for human
exposure.
Kenaga (1980a,b) has correlated bioaccumulation in ruminant fat with
log Kqw (positive association) and with log solubility (negative associ-
ation) as follows:
log BF = -3.457 + 0.500 log KQW	(6-8)
(r2 = 0.62 n = 23)
and
log BF = -1.476 - 0.495 log solubility	(6-9)
(r2 = 0.67 n= 23)
Trabalka and Garten (1982) found, however, that when the data from a
larger number of examples were included in the development of these
regression equations, the amount of variance precluded their usefulness
for accurately predicting BF based on K or aqueous solubility. Kenaga
UW
(1980 a,b) and Trabalka and Garten (1982) also attempted to correlate BF
for animal fat with BF for fish (Garten and Trabalka 1983). Garten and
Trabalka (1983) have shown that fish tests are inadequate for predicting
bioaccumulation in birds and mammals and that any regression equations
developed to relate bioaccumulation between these classes of organisms
had such high variance that they could not be considered reliable.
Furthermore, because of the problems in variability of log BF values
about a given value for log Kni. or log solubility, Trabalka and Garten
UW
(1982) have suggested the use of a screening cutoff level to determine
whether organic chemicals bioaccumulate in animal fat. This screening
level is a value of log KQW or log solubility beyond which no compounds
are observed to significantly bioaccumulate. A value for log BF of -0.5
has been empirically chosen by these researchers as the value below which
bioaccumulation of an organic chemical is considered insignificant. This
means that for each mg/kg of chemical in the diet, only 0.3 mg/kg will be
deposited in the body fat for a chemical meeting this BF cutoff criterion.
Trabalka and Garten (1982) base the screening cutoff on a frequency
analysis of log BF and log K performed on 123 paired observations for
Un
6-6

-------
68 different chemicals. Since al_L compounds with log KQW values less than
3.5 had log BF values less than -0.5, a log KQW of >3.5 was proposed as
the level for identifying xenobiotics that could potentially bioaccumulate
in mammals. In addition, a frequency analysis of log BF and log solubili-
ty, based on 90 paired observations for 47 different chemicals, indicated
that no compounds with a solubility greater than 10 mg/L bioaccumulated
in tested mammals. In a study performed on birds, it was found that com-
pounds with log Kftl. <3.5, or a water solubility greater than 10 mg/L,
uw
also did not exhibit a potential for bi©accumulation (Trabalka and Garten
1982). Regression equations were developed from these data but, since
2
their coefficients of determination are poor (r = 0.37 and 0.60,
respectively), the equations themselves are not recommended for estimating
bioaccumulation levels as part of an exposure assessment.
(b) Metals. The uptake of most heavy metals by edible plants
*
usually occurs via their root systems (Kloke et al. 1984). Most
heavy metals exhibit low mobilities once they enter soil, unless the soil
is acidic. Lead, chromium, and mercury are usually immobile in soils, but
cadmium and zinc are readily transported to plant roots. Kloke et al.
(1984) consider lead, cadmium, thallium, and copper to have the greatest
potential for consumption by humans via edible plants. Other heavy
metals, such as zinc, nickel, chromium, and mercury, reach a toxic level
in plants before their level of concentration becomes a hazard to human
consumers.
The most metabolically active plant parts, especially the leaves,
exhibit higher heavy metal accumulation than stems and storage organs or
fruits and seeds (Kloke et al. 1984). In cereal grains, most heavy metals
Uptake resulting from airborne deposition on plant foliage can be
significant for some heavy metals in particular (e.g., lead) and
for situations involving heavy foliar deposition (e.g., waste
application or deposition from other nearby pollutant sources).
6-7

-------
concentrate in the outer layers, so that the processed flour contains
little of the total metal content compared to the bran (Oberlander and
Roth 1982). Table 6-1 gives the relative accumulation of heavy metals by
edible plant parts by different crops (USEPA/USFDA/ USDA 1981).
6.3 Estimation of Concentrations - Terrestrial Food Sources
Domestic livestock may ingest contaminated soil, leading to human
exposure via the food chain. For grazing animals, soil ingestion rates
have been estimated to represent between 1 and 18 percent of dry matter
for cattle and up to 30 percent for sheep (Thornton and Abrahams 1983).
If it is assumed that all chemicals deposited on the soil could be
ingested, the amount of contaminant ingested by grazing animals can be
estimated according to the following equation:
IR = C x S	(6-10)
where
IR = contaminant ingestion rate (mg/day)
C = contaminant concentration in soil (mg/g)
S = soil ingested by livestock per day (g/day).
The amount of soil ingested per day (S) can be estimated by multiply-
ing the forage intake amount by the soil intake fraction. According to
Fries et al. (1973), the forage intake rate for cattle is 16.5 kg/day.
Using a typical soil ingestion rate of 2 percent for cattle (Thorton and
Abrahams 1983), a daily soil ingestion rate of 330 g/day can be assumed.
Based on this intake rate, tissue concentrations can then be calculated.
Methods for estimating contaminant concentrations in terrestrial plant
and animal tissues have been identified in Section 6.2.2. In general, the
methods rely on partitioning coefficients for animals and chemical persis-
tence for plants. Since these estimation techniques are imprecise, actual
See Section 8 for soil concentration estimation methods.
6-8

-------
7695H
Table 6-1. Relative Accunulation of Heavy Metals into Edible
Plant Parts of Vegetable Crops
High	Moderate	Low	Very low
uptake	uptake	uptake	uptake
Lettuce	Kale	Cabbage	Snapbean family
Spinach	Col lards	Sweet com	Pea
Chard	Beet	Broccoli	Melon family
Esearole	Turnip roots	Cauliflower	Tomato
Endive	Radish globes	Brussel sprouts Pepper
Cress	Mustard	Celery	Eggplant
Turnip greens Potato	Berry fruits	Tree fruits
Beet greens
Carrot
Source: USEPA/USFDA/USDA (1981).
6-9

-------
measurements of plant and animal tissue levels are recommended for all
sites where there is a high probability of exposure through consumption ot
terrestrial plants and animals.
Concentrations in animal tissues (e.g., meat) and animal products
(e.g., eggs, milk) can be determined for certain contaminants using trans-
fer coefficients. A transfer coefficient represents the fraction of a
contaminant ingested daily through feed by a domestic animal that is found
in one kilogram of meat, one liter of milk, or one kilogram of eggs.
Transfer coefficients were initially designed for use in U.S. Nuclear
Regulatory Commission (NRC) models to evaluate the dose from ingestion of
terrestrial foods contaminated by radionuclides that are released to the
atmosphere and hydrosphere by light-water reactors (Ng et al. 1982).
Because the biological half-life in muscle for most elements is
relatively short, the difference between isotopic and elemental transfer
coefficient values in meat is usually negligible for radionuclides with
long chemical half-lives. Some of the elements for which transfer coeffi-
cients have been estimated based on stable-element concentrations in m?3*
and the associated diet are trace elements essential in animals. Thest
include manganese, iron, cobalt, copper, zinc, selenium, and iodine.
Hemostatic regulation of these elements maintains optimal tissue concen-
trations over a wide range of dietary exposures (Ng et al. 1982). Trans-
fer coefficients that are greater than one suggest that some bioaccumula-
tion occurs for the associated element.
Collateral data are available for transfer coefficients for some ele-
ments. Collateral data are defined as data derived from experiments on
related species and are used when data on the tissue concentration versus
time relationship for a given species of interest are incomplete (Ng
et al. 1982). Known transfer coefficients are presented in Appendix D.
Table D-l presents the feed-to-milk transfer coefficients and Table D-2
presents the feed-to-meat and feed-to-eggs transfer coefficients.
Selected data for transfer coefficient values have been excluded from
these tables if (1) the data are associated with diets deficient in the
6-10

-------
element under study; (2) the diet is supplemented with the element at
levels much greater than that in the normal or basal diet; or (3) the
data are based on experiments in which radiobiological effects due to
administration of radiosiotopes were observed.
The concentration of a chemical contaminant in animal tissues or prod-
ucts is a function of contribution from ingestion of contaminated soil
during grazing, ingestion of contaminated water and feed, and inhalation
of contaminants in air. Normally, the air route will be a minor source of
contamination compared to soil, feed, and water.
The animal tissue concentration due to ingestion of soil is calculated
as follows:
^ts = x P$ x Cs x fsa	(6-11)
where
C^s = animal tissue/product concentration of chemical due to inges-
tion of soil (mg/kg)
Gf = feed consumption rate (kg feed dry weight/day)
Ps = fraction of diet as soil (kg soil consumed/kg feed consumed)
Cs = soil concentration (mg chemical/kg soil)
fsa = soil-to-animal product transfer coefficient (fraction/kg
animal product/day).
The animal tissue concentration due to ingestion of feed is calculated
as follows:
Ctf = Gf x Cf x ffa	(6-12)
where
C^s = animal tissue/product concentration of chemical due to inges-
tion of feed (mg/kg)
Gf = feed consumption rate (kg feed dry weight/day)
Cf = concentration of chemical in feed (mg/kg feed)
ffa = feed-to-animal product transfer coefficient (fraction/kg
animal product/day).
The animal tissue concentration due to ingestion of contaminated
drinking water is calculated as follows:
^tw = x Cw x *wa	(6-13)
6-11

-------
where
Ctw = animal tissue/product concentration of chemical due to inges-
tion of water (mg/kg)
Gw = consumption rate of water (1/day)
Cw = concentration of chemical in water (mg/1)
fwa = water-to-animal product transfer coefficient (fraction/kg
animal product/day).
The concentrations of chemical contaminant in surface water and ground
water are obtained using methods in Sections 3.2 and 4.2, respectively.
The tissue concentration due to inhalation of chemical contaminants in
air is calculated as follows:
Cta = ^a x ^a x ^aa	(6-14)
where
Cta = animal tissue/product concentration of chemical due to inhala-
tion of air (mg/kg)
Ga = inhalation rate (myday)
Ca = chemical concentration in air (mg/nr)
faa = air-to-animal product transfer coefficient (fraction/kg
animal product/day).
The concentration of chemical contaminant in air is obtained using the
methods described in Section 7.2.
Thus, the total tissue concentration Cj is:
Ct = C^s + ^tf + C^w + C^a .	(6-15)
Parameter values needed to calculate animal tissue/product concentrations
are presented in Table 6-2.
6.4 Exposure Estimation - Terrestrial Food Sources
Exposure to contaminated food may occur as a result of the consumption
of homegrown fruits and vegetables, homegrown beef and dairy products, or
game residing or grazing in contaminated areas. Each potential exposure
pathway should be investigated on a site-specific basis.
Use of the concentration estimates developed employing the methodolo-
gies noted above allows the assessor to estimate exposure based on
consumption rates and other parameters presented in the Exposure Factors
Handbook (USEPA 1989).
6-12

-------
7695H
Table 6-2. Parameter Values for Estimation of Concentrations in Animal Tissues/Products Using Transfer Coefficients
Animal
Feed consult ion
(kg dry weight/day)
Fraction of
diet as soil
Drinking water
intake (1/day)
Inhalation
rate (m^/day)
Animal
Type
Tissue/Product
Mass (kg/animal)
Cattle
7.6
0.02
19
150
Beef
Milk
250
Pigs
Z.3
0.05
5.8
45
Pork/Bacon
58
Sheep
1.0
0.05
2.5
20
Laitto
18
Goats
1.4
0.05
3.5
27
Neat
26
Poultry
0.14
0.10
0.35
3
Chicken Neat
Egg
1.5
.05/egg
Source: Based on Thome (1984).

-------
7.	ASSESSMENT OF EXPOSURE VIA AMBIENT AIR
7.1 Release Analysis - Ambient Air
Some mechanisms that may release chemical contaminants into ambient
air are (1) volatilization of pollutants from soil and water, (2) fugitive
dust generation, and (3) combustion. Contaminant release analysis is
conducted in two stages: screening of contaminant release mechanisms and
quantitative analysis.
7.1.1 Screening of Release Potential
The screening, which is a qualitative evaluation, identifies each
potential release source, determines the environmental media affected by
each release, and broadly defines the possible extent of the release.
In order to assess the potential for a chemical release, the physical
and chemical properties of the chemical and the medium in which it is
dispersed are important. Typically, the chemical can be released into
the air from the soil or from a liquid surface by volatilization or can
be released from combustion during incineration of waste. The volatility
and concentration of a chemical will determine the amount that will be
released.
In some cases, unintentional releases (e.g., spills, leaks, container
failure) may take place and release the pollutants into the ambient air.
Significant releases may occur at uncontrolled hazardous waste sites.
Since in such cases high concentrations of the chemical may result in
contaminated surface soil, leaks and spills from these sites may pose a
substantial degree of risk to the environment.
Chemicals bound to particulate matter may also be released to ambi-
ent air through the following mechanisms: fugitive dust resulting from
wind erosion of contaminated soils, vehicular travel over contaminated
unpaved roadways, excavation and transfer of contaminated soils, and
particulate releases from on-site waste incineration.
7-1

-------
After the contaminants are released to ambient air, they are subject
to a variety of fate processes, as will be discussed in Section 7.2.
Once the potential sources of on-site contaminant release have been
screened, those requiring further evaluation are quantitatively analyzed.
7.1.2 Quantitative Release Estimation
One can estimate releases by listing the factors on which the mode of
release depends and deriving an empirical relationship between the factors
and the quantity of release. These equations will be discussed for each
mode of release.
(1) Particulate emissions analysis. Emissions of contaminated soil
particles can result from a combination of the following factors:
•	Wind erosion of wastes and contaminated soils;
•	Vehicular traffic movement over contaminated, unpaved roads;
•	Heavy equipment activity at the site (excavation, loading); and
•	Incineration of wastes during remediation.
(a) Wind erosion.* Wind erosion of disturbed soils depends
upon a number of factors, including surface roughness and cloddiness;
surface soil moisture content; kind and amount (and orientation, if
applicable) of vegetative cover; wind velocity; and the amount of soil
surface (field length) exposed to the eroding wind force. A wind erosion
estimation algorithm that takes this range of variables into account
(Skidmore and Woodruff 1968) is a function of the following variables:
This applies to nonadhering, noncompacted contaminated soil or
waste materials.
7-2

-------
E = f(I', C', K', L', V)	(7-1)
where
E = potential annual average wind erosion soil loss (kg/m2/sec)
I' = soil erodibility index (dimensionless)
C' = climatic factor (dimensionless)
K' = soil ridge roughness factor (dimensionless)
L' = field length along the prevailing wind direction (feet)
V = vegetative cover factor (dimensionless).
An investigator applies the wind erosion equation via the use of well-
developed tables, equations, and graphs that are region specific; the
equation is easy to use in evaluating given locations across the country.
As noted in USEPA (1983a), however, it does exhibit one significant
drawback—it computes the total wind erosion soil loss caused by the
combination of surface creep, saltation, and suspension. Although this
is entirely appropriate for studies of agricultural soil loss, for which
the equation was developed, in risk evaluations the analyst is generally
concerned only with that fraction of the total soil loss that consists of
particles of suspendable, wind transportable, and inhalable size. Thus,
when the wind erosion equation is used to estimate contaminated fugitive
dust exposure situations, the total soil loss results obtained must be
adjusted (reduced) to reflect only that portion of the total soil loss
that is suspendable and transportable over significant distances by wind.
Considerable discussion of the cut-off point for suspendable soil particle
size exists in the literature (Sehmel 1980, Smith et al. 1982, and USEPA
1983a,b). Table 7-1 presents environmental variables and model parameters
for the wind erosion equation.
As a group, particles <100 ^m in diameter include all wind suspendable
and transportable particles, and thus encompass inhalable particle sizes
(see Miller et al. (1979) for a discussion of the extent to which various
particle sizes penetrate the human respiratory system). Of particles
within this broad range, those in the 30 to 100 diameter range will
be susceptible to impeded settling within a few hundred feet of the source
(USEPA 1983b), while those particles <30 /im in diameter can be transported
for considerable distances downwind. To estimate inhalation exposure,
7-3

-------
7630H
Table 7-1. Environoental Variables and Model Parameters for the Wind Erosion Equation
Equivalent SCS wind erosion equation
priaary wind erosion variables
Parameters
Soil erodibility index, I (function of soil
particle size distribution; read from a table)
Knoll erodibility,
steepness; read from a graph)
Is (function of knoll slope
Soil and knoll erodibility, I' (equal to
I x Is)
Surface crust stability, F
Disregarded—crust is transient
Soil ridge roughness, 1^. (function of height,
width, and spacing of clods and furrows)
Annual average wind velocity, v (read fron nap)
Surface soil moisture, M (estimated using
Thomthwaite's (1931) precipitation-evaporation
index)
Soil ridge roughness factor, K* (estimated by
Caspar i son to a set of standard photographs
included in SCS wind erosion equation users'
manuals)
Local wind erosion climatic factor, C' (may
be calculated but cannonly read fron maps
of C)
Distance across field, Df (field width in
direction of primary erosive wind)
Sheltered distance, D^ (calculated fron barrier
.height upwind of field)
Field length, L' (the difference between
Df and Dfa)
Quantity of vegetative cover, R' (mass of standing
or fallen vegetative residue per unit area)
Kind of vegetative cover, S (factor related to
erosion-reducing effectiveness of residues fron
different crops)
Orientation of vegetative cover. Kg (factor
relating erosion reduction to standing vs.
fallen crop residues)
Equivalent vegetative cover, V (the product of
R', S, and Kg) - can often be assuned = 0
for abandoned waste sites (see text)
Source: Smith et al. (1982).
7-4

-------
however, only the inhalable fraction of suspended particulates needs to be
considered. The respirable fraction consists of particles less than or
•k
equal to 10 /im in diameter. Therefore, the 10 nm cut-off is used in the
following estimation methods.
(b) Excavation and transfer of contaminated soils. One can use
the following equation (USEPA 1983b) to estimate fugitive dust releases
resulting from on-site excavation and dumping of contaminated soils.
E	(f) (2^) ('lf)	(7.2)
"HE = k(0.00090) 	2	OIF
r m
where
= emission factor for heavy equipment (batch dump) operations
(kg/kkg of material dumped)	^
k = 0.36 = particle size multiplier for particles <10 ug
s = silt content (of contaminated material) (%)
U = mean wind speed (m/s)
H = drop height (m)
M = material moisture content /%)
Y = dumping device capacity (m3).
The equation is designed for application to batch dump material trans-
fer and does not take into account material released to air during excava-
tion. The equation is valid for situations that comply with the following
boundary conditions (which must be compared with existing on-site condi-
tions):
Personal communication between Gary Whitmyre (Versar Inc.) and
Kent Berry (Strategy and Air Standards Division, Office of Air
Quality Planning and Standards, U.S. Environmental Protection
Agency, Research Triangle Park, N.C.), July 23, 1985.
See USEPA (1983b) for "k" values used when release of other
particle size groups is to be estimated.
7-5

-------
•	Silt content = 1.3% to 7.3%
•	Moisture content = 0.25% to 0.70%
•	Dumping capacity = 2.10 to 7.6 m3 (2.75 to 10 yd3).
(c) Travel on contaminated unpaved roads. Travel of loaded and un-
loaded vehicles at the site of interest can be a very significant source
of fugitive dust emissions. The following equation (USEPA 1983a) can be
used to estimate fugitive dust releases associated with vehicles traveling
on contaminated unpaved roads on site:
Eyj = emission factor for vehicular traffic, (lb/vehicle mile
traveled; kg/vehicle kilometer traveled)
k = 0.45 = particle size multiplier for particles < 10 ^m (i.e.,
particles that may remain suspended once they become airborne
and which can be inhaled into the respiratory systerj^.
s = silt content (of road surface material), (percent).
Sp = mean vehicle speed, (mph; kph).
W = mean vehicle weight, (tons; Mg).
w = mean number of wheels.
Dp = number of days with at least 0.254 mm (0.01 in) of precipitation
per year (see Figure 7-1).
See EPA (1983a) for "k" values used when release of specific
particle size groups other than < 10 is desired.
Soil silt content can be estimated from SCS Soils 5 File data by
subtracting the "percent clay" value from the "percent material
passing No. 200 sieve" value. (Personal communication between
Lee Schultz (Versar Inc.), and Keith Young (U.S. Department of
Agriculture, Soil Conservation Service), Washington, D.C., May 1,
or in metric form
(7-3)
where
1984.)
7-6

-------
.140
Source: USDC 1968.
Figure 7-1. Mean number of days per year with >0.01 inches of precipitation (I.e., "wet days").

-------
The emission factor (Eyy) can be multiplied by a "vehicle kilometers
traveled per time" value to generate a "dust release per time value."
Short-term (maximum release) estimates can be made by using a reduced
value of "Dp" in the equation to reflect drought conditions at the site.
Consultation with the local National Weather Service office may provide
locale-specific insight into what "Dp" values should be used to represent
dry years at the site. Figure 7-1 reflects the range of average "Dp"
values for locations in the United States. Long-term (average) releases
can be estimated by using the annual average value for "Dp." USEPA
(1983a) states that this equation is valid for situations that are consis-
tent with the following source conditions:
•	Road surface silt content = 4.3 - 20 percent;
•	Mean vehicle weight = 3-157 tons (2.7-142 Mg);
•	Mean vehicle speed = 13-40 mph (21-64 kph); and
•	Mean number of wheels = 4-13.
For an overview of the utility and limitations associated with the appli-
cation of emission factors to particulate release estimation problems, the
user of this manual can refer to USEPA (1983a,b), Farino et al. (1983),
Sehmel (1980), and Smith et al. (1982).
(d) Incineration. Incineration of toxic solid wastes may prove a
feasible alternative to land disposal. Although incineration can be
effective in destroying toxic substances, it can also result in the
release of contaminated particulates. Such a release can occur as combus-
tion emissions, or as fugitive emissions from the waste feed system or the
incinerator itself. Combustion releases are a function of the waste feed
rate and the system's destruction and removal efficiency (DRE) for speci-
fied toxic chemicals in the waste materials. In the absence of emissions
data, a combustion release of 0.01 percent can be assumed based on the
requirements of 40 CFR 264.343 (USEPA Regulations for Owners and
Operators of Permitted Hazardous Waste Facilities; Subpart 0 -
Incinerators). In order to simplify analysis, fugitive emissions can be
7-8

-------
assumed to mirror the relative proportions of toxic components in the
waste feed. To estimate total particulate emissions from a given
incineration option, the analyst should review the proposed system design
parameters, as well as the permit operating requirements and pertinent
performance standards (i.e., 40 CFR 264.340-345).
(e) In-depth analysis. For contaminated fugitive dust emissions,
in-depth analysis can consist of monitoring and modeling activities to
back-calculate releases. Generally, air sampling will be conducted down-
wind and upwind of the uncontrolled hazardous waste site. The difference
in particulate loading obtained at the two (or more) sampling locations
will provide a quantification of the particulate mass loading attributable
to the site alone (assuming that air sampling stations can be sited so as
to eliminate interference from other sources). Either simple dispersion
*
equations or computerized air dispersion modeling can then be used to
delineate the emission level at a "virtual point source" (i.e., the
uncontrolled hazardous waste site) necessary to result in the ambient
particulate concentrations observed downwind.
(2) Volatilization emission analysis. Volatilization of contami-
nants from uncontrolled hazardous waste sites can occur as a result of the
following sources:
•	Covered landfills - without internal gas generation;
•	Covered landfills - with internal gas generation;
Although computerized dispersion modeling can be used to obtain
contaminant release rates, it is primarily considered to be a tool
for determining contaminant atmospheric fate. Thus, refer to
Section 7.2.2, Environmental Fate Analysis, for detailed discus-
sions of air dispersion models applicable to uncontrolled hazardous
waste facilities.
7-9

-------
•	Spills, leaks, landfarms - concentrated wastes on the surface or
adhered to soil particles below the surface; and
•	Lagoons - wastes dissolved in or mixed with water.
Exposure occurs at or near contaminated sites through inhalation of
ambient air contaminated with toxic vapors or particulate matter on which
a contaminant is absorbed. Volatile constituents emitted from contami-
nated soil or water will be diluted by the ambient air before they are
inhaled. When a contaminant adheres to soil or is absorbed in water, the
vapor pressure of the contaminant above the soil/water surface will always
be smaller than the pure component vapor pressure. In other words, the
adsorption/absorption phenomenon depresses the vapor pressure that can
exist under saturated conditions.
In calculating concentrations of contaminants in ambient air, the
assessor first has the task of estimating emission rates. The emission
rates for steady-state conditions can be calculated using the methods
presented by Farmer et al. (1978), and subsequently summarized by Hwang
(1982). In addition, Thibodeaux and Hwang (1982) presented the transient
emission rate from land-treated facilities.
The volatilization rate of a given substance is determined primarily
by its chemical properties. The equations presented to estimate the vola-
tilization rate requires inputs of chemical property values. These data
are available for many chemicals in the literature. In addition, readily
accessible computerized systems are available to predict numerous perti-
nent chemical properties. For example, the computerized Graphic Exposure
Modeling System (GEMS), and its subsystem CHEMEST, provides one such tool.
This system has been developed and is managed by the EPA Office of Toxic
Substances in Washington, D.C.
(a) Landfills without internal gas generation. An equation can be
developed to estimate volatilization releases from covered landfills con-
taining toxic materials alone, or toxic materials segregated from other
landfilled nonhazardous wastes. The starting point for this equation is
Fick's First Law of steady-state diffusion, which assumes that diffusion
7-10

-------
into the atmosphere occurs at a plane surface where concentrations remain
constant. It ignores biodegradation, transport in water, adsorption, and
production of landfill gas. Thus, diffusion of the toxic vapors through
the soil cover is the controlling factor.
The Fick's Law equation was simplified by Farmer et al. (USEPA 1980a)
by incorporation of a number of assumptions, such as completely dry soil
(worst case) and zero concentration of volatilizing material at the soil
surface. Farmer's equation as modified by Shen (1981) is:
= emission rate of component i (g/sec)
= diffusion coefficient of component i (cnr/sec)
Csj = saturation vapor concentration of component i (g/cnr)
A = exposed area (cm2)
Pt = soil porosity (dimensionless)
djC = effective depth of soil cover (cm)
H-j = mole fraction of toxic component i in the waste (g/g).
The presence of water in a soil cover will tend to decrease the flux
rate of a volatile compound by effectively decreasing the porosity.
Water will also increase the geometric complexity of the soil pore system
(because water will adhere to soil particles), thereby effectively extend-
ing the length of the path that vapor must traverse to be released (USEPA
1980b).
If not provided in the literature, Di, a compound's diffusion
coefficient (required for the above equation), can be calculated by
Fuller's Method (Perry and Chilton 1973):
(7-4)
where
PjaVj)170 + (IVa)1/J]
7-11

-------
where
T = absolute temperature (°K)
MW-j; MWa = molecular weights of toxic substance and air (28.8),
respectively (g/mole)
Pa = absolute pressure (atm)
EVj; XVa = molecular diffusion volumes of toxic substance
and air, respectively; this is the sum of the atomic diffu-
sion volumes of the compound's components (cnr/mole)•
To estimate short-term (maximum) release rates, the analyst should use a
value for temperature that reflects expected summer maximum temperatures.
Annual average temperatures should be used to estimate long-term
(average) release rates. Relevant atomic diffusion volumes for use in
estimating are (Perry and Chilton 1973):
C = 16.5	CI = 19.5	Aromatic ring = -20.2
H = 1.98	Br = 35.0	Heterocyclic ring = -20.2
0 = 5.48	F = 25.0
N = 5.69	S = 17.0
Table 7-2 presents diffusion coefficients that have been calculated for a
variety of compounds.
An alternative method (Shen 1981) for approximating involves the
identification of a compound listed in Table 7-2 that has a molecular
weight and molecular diffusion volume (calculated) similar to those of the
toxic substance under evaluation. The unknown diffusion coefficient can
then be calculated using:
1/2
Di= D'
MW
MWi
(7-6)
This value is from Shen (1981).
7-12

-------
763OH
Table 7-2. Diffusion Coefficients of Selected Organic Conpounds
Molecular
Compound	Formula weight
Atonic Diffusion coefficients tan /seel
diffusion at	at	at
volune 10*C	30*C	50'C
Acetaldehyde
w
44
46.
.40
0.11758
0.13249
0.14816
Acetic acid
w2
60
51.
.88
0.10655
0.12007
0.13427
Acetone
C3H6°
58
66.
.86
0.09699
0.10930
0.12223
Aniline
C6H7N
93
118.
.55
0.07157
0.08065
0.09019
Benzene
CH
6 6
78
90.
.68
0.08195
0.09234
0.10327
Branoethane
CH Br
3
95
57.
.44
0.09611
0.10830
0.12111
Broraoforn
CtSr
3
118
53.
.48
0.09655
0.10880
0.12167
Carbon tetrachloride
CC1
4
154
94.
.50
0.07500
0.08451
0.09451
Chlorobenzene
C6H5C1
113
128
40
0.06769
0.07627
0.08530
Chloroethane
C2H5C1
65
62.
.40
0.09789
0.11031
0.12336
ChloroforHi
CHC!3
120
76.
.89
0.08345
0.09404
0.10517
Chloronethane
CH CI
3
51
57.
.94
0.10496
0.11827
0.13226
Cyclohexane
C H
6 12
84
122.
.76
0.07139
0.08045
0.08996
Dichloroethane
C H CI
2 4 2
99
75.
.96
0.08557
0.09643
0.10784
Di chloroethylene
C HC1
2 2 2
97
106
.96
0.07442
0.08386
0.09377
Dichloropropalene
C3H6C12
113
100.
.38
0.07519
0.08473.
0.09475
Dimethylamine
C2H7N
45
52.
.55
0.11161
0.12577
0.14065
Ethanol
C H 0
2 6
46
50.
.36
0.11297
0.12730
0.14236
Ethyl acetate
Wz
88
92.
.80
0.07991
0.09005
0.10070
Ethylamine
C2H7N
45
52.
.55
0.11161
0.12577
0.14065
Ethylbenzene
VlO
116
151.
.80
0.06274
0.07070
0.07906
Fluorotoluene
C7H7F
110
154.
.36
0.06262
0.07056
0.07891
Heptane
C7H16
100
146.
.86
0.06467
0.07287
0.08149
Hexane
C H
6 14
86
126.
.72
0.07021
0.07912
0.08848
Isopropanol
c,"0°
3 8
60
37
.82
0.12004
0.13526
0.15126
Methanol
CH 0
4
32
29.
.90
0.14808
0.16686
0.18660
Methyl acetate
C H 0
3 6 2
74
72.
.34
0.09054
0.10203
0.11410
Methyl chloride
CH CI
2 2
85
59
.46
0.09610
0.10830
0.12111
7-13

-------
7630H
Table 7-2. (continued)
Ccnpound
Foroula
Molecular
weight
Atomic
diffusion
volune
Diffusion coefficients (
at at
10*C 30"C
an /sec)
at
50*C
Methylethyl ketone
C4H8°
72
87.32
0.08417
0.09485
0.10607
PCB (1 CI)
C12H9C1
189
235.32
0.04944
0.05571
0.06230
Pentane
C5H12
72
106.26
0.07753
0.08737
0.09770
Phenol
C HO
6 6
84
96.16
0.07919
0.08924
0.09980
Styrene
CH
8 8
104
137.84
0.06620
0.07460
0.08343
Tetrachloroethane

168
114.96
0.06858
0.07729
0.08643
Tetrachloroethylene
C2C14
166
111.00
0.06968
0.07852
0.08781
Toluene
C7H8
92
111.14
0.07367
0.08301
0.09283
Trichloroethane
C H CI,
2 3 3
133
97.44
0.07496
0.08447
0.09446
Trichloroethylene
CHC1
2 3
131
93.48
0.07638
0.08606
0.09625
Trichlorof1uororoethane
CC1 F
3
138
100.00
0.07391
0.08329
0.09314
Vinyl chloride
W1
63
58.44
0.10094
0.11375
0.12720
Xylene
C8H10
106
131.60
0.06742
0.07597
0.08495
Source: Shen 1961.
7-14

-------
where
Di = diffusion coefficient of the compound to be estimated from the
known D'
D'= diffusion coefficient of a compound that can be found in the
table, the molecular weight, and molecular diffusion
MW = molecular weight of the selected compound D'
MW^ = molecular weight of the compound to be estimated, the volume
of which is close to that of the unknown.
Total soil porosity, Pt, can be calculated as follows (USEPA
1980a,b):
= total soil porosity (dimensionless)
p = soil bulk density (g/cm3); generally between 1.0 and
2.0 g/cm3
p = particle density (g/cnr); usually 2.65 g/cnr is used for
most mineral materials.
For estimation purposes, Pt can be assumed to be approximately 0.55 for
dry, noncompacted soils and about 0.35 for compacted soils. This same
value (0.35) is also appropriate for use as a generic air-filled soil
porosity (P,) when analyzing volatilization release from soils with a
a
high moisture content (Shen 1981).
Saturation vapor concentration, Cs^, can be determined by (USEPA
1980c):
Values for soil bulk density for specified locations can be
obtained from the U.S. Soil Conservation Service, Soils 5 File data
base.
(7-7)
where
7-15

-------
where
Csi = saturation vapor concentration^ component i (g/cm3)
p = vapor pressure of the chemical (mm Hg)
MWj = molecular weight (g/mole)
R = molar gas constant (62,361 mm Hg-cnr/mole-°K)
T = absolute temperature (K).
(b)	Landfills with internal gas generation. Thibodeaux (1981)
developed a method for estimating toxic vapor releases from co-disposal
landfills. These facilities contain toxic wastes in combination with
municipal or sanitary wastes which, because of their considerable organic
content, generate landfill gases (e.g., ^ CH^, C02). In such
cases, the upward movement (convective sweep) of the landfill gas becomes
the significant controlling factor, greatly accelerating the upward migra-
tion and subsequent release to the atmosphere of the co-disposed toxic
substances. In fact, review of Thibodeaux's work indicates that the
accelerating effect of the toxic gas migration of the landfill gas is so
great that soil and gas phase diffusion becomes essentially insignificant.
Thus, the following simplified equation is recommended for estimating
volatilization of toxic substances from co-disposal landfills:
Ei = C -j *VyA	(7-9)
where
Ei = emission rate (g/sec)
C-j* = concentration of compound i in the soil pore spaces (g/cm3)
Vy = mean landfill gas velocity in the soil pore spaces
(Thibodeaux (1981) provides an average value of 1.63 x
10~3 cm/sec for this factor.)
A = area (cm2).
(c)	Spills, leaks. Equations 7-10 and 7-12 estimate the volatili-
zation release from new and old (respectively) chemical spills on soil.
If the vapor pressure of a chemical under consideration is not
available in standard reference texts, it can be estimated as
described in Lyman et al. (1982).
7-16

-------
Equations 7-11 and 7-13 through 7-15 provide means of estimating certain
input variables required to solve Equations 7-10 and 7-12.
As discussed in Farino et al. (1983), one can apply Equation 7-11
(adapted from Thibodeaux and Hwang 1982) to estimate volatilization re-
leases resulting from spills or leaks where a contaminant pool is visible
on the soil surface, or where soil is contaminated (saturated) from the
surface down. The equation does not consider soil phase mass transfer
resistance, and therefore is not appropriate for use when spilled contami-
nants have seeped into surface soils (in this case, use the landfarming
equation that follows). Similarly, because it does not consider liquid
phase resistance, it is only useful for estimating releases of pure com-
pounds.
Ei = kiG C-j*A	(7-10)
where
E^ = emission rate of chemical i (g/s)
kiG = 9as Phase mass transfer coefficient of chemical i (cm/s)
C^* = vapor concentration of chemical i (g/cm3)
A = area (cm2).
Hwang (1982) has developed the following simplified means of estimat-
ing a compound's gas phase mass transfer coefficient.
kiG
mwh2o
MWi
0.335
r t i
298
1.005
CiG'V)	(7"1"
where
K,G = gas phase mass transfer coefficient of chemical i (cm/s)
MWh q = molecular weight of water (g/mole)
= molecular weight of compound i (g/mole)
7 = temperature (°C)
kiG*H2O = gas phase mass transfer coefficient for water vapor at
at 25°C (0.58 cm/sec).
In cases where past spills, leaks, or intentional disposal directly
onto or into surface soils (landfarming) have resulted in contaminated
surface soils, Equation 7-12 can be used to estimate volatilization
7-17

-------
releases. This equation assumes that soil pore spaces connect with the
soil surface, that soil conditions are isothermal, and that there is no
capillary rise of contaminant. Farino et al. (1983) have determined that
it is preferable to other approaches when estimating the volatilization
release of chemicals spilled or incorporated into soils, because it takes
into account the contaminant loss over time. It describes vapor diffusion
as being soil-phase controlled, and essentially assumes that contaminant
concentrations in the soil will remain constant (until all contaminant is
lost to the air), and that contaminant release will occur by "peeling
away" successive unimolecular layers of contaminant from the surface of
the "wet" (contaminated) zone. Thus, over time, this process results in
a "dry zone" of increasing depth at the soil surface, and a wet zone of
decreasing depth below the dry zone.
E. =			(7-12)
d + / 2D C t + d2
\hr
where
Ej = average emission rate of component i over time t (g/sec)
D = phase transfer coefficient (cmz/sec)
Cs = the liquid-phase concentration of component i in the soil
(g/cm2)
Cr = bulk contaminant concentration in soil (g/cnr)
A = contaminated surface area (cm2)
d = depth of dry zone at sampling time (cm)
t = time measured from sampling time (seconds).
2
D (cm /sec) is related to the amount of contaminant i that goes from
liquid to gas phase and then from gas phase to diffusion in air. It can
be estimated as follows:
D = D-j (Pt)4/3 H/	(7-13)
7-18

-------
where
D = phase transfer coefficient (cm2/sec)
D.j = diffusion coefficient of component i in air (cm2/sec)
P^ = total soil porosity (dimensionless). Again, use of total soil
porosity in this equation results in a worst-case (dry soil)
estimate for D. In some cases (i.e., where soils are wet more
often than dry), it may be more appropriate to use air filled
soil porosity (Pa) in place of Pt. See text addressing
Equation 7-4 for a discussion of the application of and values
for these two terms.
H-j' = Henry's law constant in concentration form (dimensionless).
H.j', the Henry's law constant in concentration form (ratio of the
boundary layer concentration of contaminant in air to the boundary layer
concentration of contaminant in "wet" soil) can be determined as follows
(Lyman et al. 1982):
V = Hi	(7-14)
RT
where
Hj = Henry's law constant of contaminant i (atm-m^/mol)
R = gas constant (8.2 x 10"5 atm-nr/mol-°K)
T = absolute temperature (°K).
Note that Equation 7-12 assumes that the contaminant concentration in
the liquid and gas phases in the soil remains constant until all of the
contaminant has been released to air. Also, the equation holds from time
zero (the time at which the soil was sampled) to t^ (the time at which
the soil becomes dry, i.e., all contaminant has volatilized and the
release process stops). The formula for calculating td (in seconds) is:
.2
ld =
2D
• CB	(7-15)
C
7-19

-------
where
tj = the time at which all contaminant has volatilized from the
soil (sec)
h = depth from soil surface to the bottom of the contaminated region
(cm)
d = depth of dry zone at sampling time (cm)
D = phase transfer coefficient (cnr/sec)
Cg = bulk contaminant concentration in soil (g/cnr)
Cs = contaminant liquid phase concentration (g/cm3).
(d) Lagoons. Mackay and Leinonen (1975) have developed an equation
for estimating volatilization releases of low solubility compounds from
water bodies such as hazardous waste lagoons. This approach assumes that
conditions are steady state (i.e., no constant addition of contaminant),
that diffusion is liquid state controlled, and that it occurs from a well-
mixed water phase to a well-mixed air phase across a stagnant water/air
interface. If one can assume that the atmospheric background levels of
the contaminant of concern are negligible, then Mackay and Leinonen's
basic equation can be simplified to the following form (which includes an
area term to convert flux rate to emission rate):
E-j = K^A	(7-16)
where
Ei = emission rate (g/sec)
K.j = overall mass transfer coefficient (cm/sec)
C$ = contaminant liquid phase concentration (g/cm3)
A = area (cmz).
The overall mass transfer coefficient, , can be calculated via the
following relationship:
+	(7-17)
K, kjL Hlkis
7-20

-------
where
= overall mass transfer coefficient (cm/sec)
k^i = liquid phase mass transfer coefficient (cm/sec) (see
Equation 7-17)
R = ideal gas law constant (8.2 x lO'^ atm-nr/mol-°K)
T = temperature (°K)
H.j = Henry's law constant of compound ^ (atm-nr/mol)
kiG = 9as Phase mass transfer coefficient (cm/sec).
Hwang (1982) provides a simplified method for determining a compound's
liquid phase and gas phase mass transfer coefficient for use in the above
equation. To estimate k^, use the following equation:
0.5
fMW
kiL "
0,
MW.

(7-18)
where
= liquid phase mass transfer coefficient (cm/sec)
MW2; MW-j = molecular weight of oxygen and of compound i, respectively
(g/mole)
T = temperature (°K)
k|_,O2 = liquid phase mass transfer coefficient for oxygen at 25°C
(0.0022 cm/sec in rivers; 0.0005 cm/sec in lakes).
(3) Long-term release calculations. Long-term release values
(70 years) for wind erosion of contaminated particulates and for volatili-
zation from landfills with and without internal gas generation, spills,
and lagoons can be estimated as follows:
"Ai
!c£i
70
E
1 - e
VcCi
(2.2x10 )
(7-19)
where
f:
e =
E =
average annual release of contaminant i (g/yr)
volume of contaminated region (cm3)
concentration of contaminant i (g/cm3)
2.71828
total release rate of contaminant i (g/sec) obtained by summing
all above-listed releases of the contaminant at the site. For
particulates, convert the average annual release from Equation 7-
to mass per second by dividing 3.16 x 10' seconds.
7-21

-------
Note that Vc and must be based on the same value. The product of
these two parameters (Vc*C^) is equal to the total mass of contam-
inant; it can be the total mass of contaminant in a lagoon, contaminated
soil, or landfill.
To estimate long-term release from contaminated surface soils,
Equation 7-15 (converted to years by dividing by 3.16 x 107) is first
used to determine the dry-out time. If no contaminant is expected to
remain after 70 years (i.e., 70 > t^), the total amount of contami-
nant present at the time of site investigation is determined, and is
divided by 70 years (in seconds) to get a conservative long-term release
Q
value (i.e., AC$(h - d)/2.21 x 10 ). If some contaminant is expected
to remain after 70 years (i.e., 70 < td), the following equation
should be used to estimate long-term release:
E = ACs
Ai —*¦
M1 70
(d + 4.4 x 109 Cs D)1/2 - d
CB
(7-20)
where
= average long-term release of contaminant i (g/yr)
A = contaminated area (cm2)
C5 = liquid phase concentration of contaminant i (g/cm3)
a = depth of dry zone at sampling time (cm)
D = relates the amount of contaminant that goes from liquid to gas
phase, and then from gas phase to diffusion in air (see
Equation 7-13)
CB = bulk contaminant concentration in soil (g/cm3).
Note that this approach does not include consideration of contaminant loss
resulting from chemical half-life and is thus a conservative approach.
7.2 Fate Analysis - Ambient Air
After an airborne pollutant has been released to the atmosphere, its
fate is determined by degradation and transport. Degradation occurs
through photochemical transformations of direct photolysis and indirect
photooxidation. A discussion of both processes is given in Section 7.2.1.
The rates of degradation that are estimated by the methods suggested in
Section 7.2.1 are then used as part of the input data for atmospheric
7-22

-------
transport modeling. Models for the transport of airborne pollutants are
discussed in Section 7.2.2.
7.2.1 Atmospheric Degradation
Photochemical transformations are those reactions in which the com-
pound absorbs energy to form an electronically excited intermediate that
undergoes further reactions. This may occur by direct photolysis in which
the compound itself absorbs solar radiation or by sensitized photolysis
in which energy is transferred to the compound from some other chemical
substance.
Photooxidation reactions involve the interaction of a compound in its
electronic ground state with a reactive species such as ozone or hydroxyl
radical. In this case, the rate determining step is the bimolecular reac-
tion of the compound and the reactive species. Atmospheric photooxidation
is often called indirect photolysis.
(1) Direct photolysis. Chemical reactions such as oxidation, reduc-
tion, hydrolysis, and various rearrangements may occur as a consequence of
direct absorption of solar radiation. The upper atmosphere filters out
solar radiation of those wavelengths of less than about 290 nm. Thus,
direct photochemical transformations of a chemical below the upper atmos-
phere can only occur if it absorbs photons at 290 nm or longer wave-
lengths. In the earth's troposphere, photochemically relevant solar radi-
ation extends from about 290 nm to 730 nm. The distribution in intensity
is not uniform in this spectral range and drops sharply at wavelengths
shorter than about 400 nm.
A convenient method for determining environmental photolysis rates is
to monitor the rate of disappearance of the chemical in an aqueous solu-
tion, vapor phase, or thin surface layer when it is exposed to outdoor
sunlight. At the same time, the photolysis of another chemical having a
similar absorption spectrum and well-characterized quantum yield also
should be measured. (The quantum yield is the efficiency for converting
the absorbed light into a chemical reaction.) The variations in sunlight
7-23

-------
intensity are thereby taken into account, and the need for determining
the detailed spectrum or quantum yield for the new chemical is avoided.
Environmental photolysis rates may also be estimated by performing
laboratory measurements of  at a single wavelength, A. The
literature contains average sunlight intensity (1^) data as a func-
tion of time, day, season, and latitude (Mabey et al. 1979, Zepp and Cline
1977, Davenport and Mill 1985). Since 4> generally does not vary sig-
nificantly with wavelength, the rate constant in sunlight	is
Kp(S) = *	(7"21)
where
eA = Molar absorbance of the chemical at wavelength (determined from
its absorption spectrum)
IA = sunlight intensity
and the half-life in sunlight is
(ti/2)(S) - M93	(7-22)
'	kp(S)
Computer and hand calculator methods are available to sum the products
of £aIa over a range of wavelengths. The data will yield a plot of the
rate constant or half-life of the chemical with respect to photolysis as
a function of the month of the year and latitude (Mabey et al. 1979,
Hendry et al. 1979). An upper limit for kp may be calculated from
Equation (7-21) by assuming  = 1. If the rate constant is small com-
pared to the rate constant for atmospheric photooxidation, then no further
photolysis calculations are needed. This method can, therefore, be used
as a screening tool.
Table 7-3 lists classes of chemicals that have significant light
absorption at wavelengths of greater than 290 nm. This list is not inclu-
sive. Some generalizations can provide a basis for a crude separation of
chemical structures into the solar-active and solar-inactive categories.
Carbon atoms singly bonded to carbon or to other more electronegative
elements absorb weakly or not at all in the solar spectrum. Carbon atoms
7-24

-------
763OH
Table 7-3. Approximate Absorption Regions for Organic Molecules
Class
Exanple
Region
(ran)
Reference
Aliphatics
hydrocarbons
n-c.H
4 10
120-170

a
fluorides, chlorides, bromides
CHBR . CHCL
3 3
180-260

a, b
iodides
w
180-320

b
ethers, alcohols
W" VWb
150-200

a
aldehydes
CH CHO
3
240-340

a
ketones
C2H5C(0,CH3
240-320

a, b
acids
CH3C(0)0
200-230

a. b
esters
CH3C(0)0CH3
200-240

a
amides
C^H^C(0)HHt^
180-220

a, b
amines
(C H ) N
2 5 3
180-240

a. b
azines
(ch3)2c=h-n=c(ch3)2
200-290

b
azo
CH -N=N-CH
3 3
200-230,
320-400
a
nitro
CH NO
3 2
200-220,
260-320
a. b
nitroso
(ch3)3cno
200-300,
540-740
a
nitrite
(ch3)3cono
200-240,
320-400
a, b
nitrate
C2H50N02
200-300

a
sulfide
C2H5SC2H5
200-223

a, b
disulfide
CH SSCH
3 3
200-240

a, b
c
Non-conjugated
Olefinics and Acetylenes




hydrocarbons
(CH3)2c=CtCH3)2
170-240

a, b
aldehydes
CH =CHCH0
2
200-250,
290-400
b
ketones
CH CH=CH-C(0)CH
3 3
200-240,
260-330
b
nitro
CH C{N0 )=C(CH )
180-380

b
7-25

-------
7630H
Table 7-3. (continued)
Region
Class	Example	(ran)	Reference
Aranatics
benzene
Benzene
200-280
a
alkylbenzenes
Toluene
200->280
a
ha lobenzertes
CI- or Br-benzene
200-280
a
nitrobenzenes
Nitrobenzene
200-330
a

Trinitrotoluene
200-450
d
aminobenzenes
Aniline
200-3300
a
azobenzenes
Azobenzene
<200->500
a
phenols
Phenol
200-305
a
vinylbenzenes
Styrene
200-300
a

Stilbene
200-330
a
acids
4-Nitrostilbene
200-390
a

Benzoic acid
200-300
a
carbonyls
Benzaldehyde
200-375
a

Benzophenone
200-410
a
polycyclic aranatics
Napthalene
200-325
a
Anthracene
ZOO->400
a

Phenanthrene
200-380
a


Benzo[a]pyrene
200-410
e
heteroaranatics
Quinoline
200-360
e

Benzoquinoline
200-380
e

9H-Carbazole
200-390
e

Benzo[b]thiophene
200-312
e
a Calvert and Pitts (1967).
^ Perkanpus et al. (1971).
c From Hendry and Kenley (1979).
d Spanggord et al. (1979).
e Smith and Bamberger (1979).
7-26

-------
that are multiple bonded with electronegative elements exhibit weak-to-
strong absorption between 290 and 350 nm, and conjugated structures absorb
more strongly at higher wavelengths. Singly bonded heteroatoms such as
0-0, S-0, and N-N also exhibit weak to moderate absorption bands in the
solar spectrum.
(2) Indirect photooxidation. An important transformation process in
the atmosphere is photooxidation. It can occur when significant concen-
trations of oxidants are formed by photochemical reactions with either
natural or anthropogenic light-absorbing chemicals. In polluted urban
atmospheres, HO* radicals and 0^ are generated in a complex photo-
chemical cycle involving NO2, O2, and organic chemicals. Only a few
environmental oxidants are important in either air or water. Table 7-4
presents these oxidants and their average diurnal concentrations.
Tables 7-5 and 7-6 list half-lives for a variety of organic chemical
classes in reactions with HO* and 0^. These half-lives are based
upon the assumption that the rate of a specific oxidation process follows
the relation:
-d[C]/dt = k2[0X][C].	(7-23)
If [OX] is not significantly depleted, Equation (7-23) becomes:
-d[C]/dt = k0X[C],	(7-24)
and the half-life for chemical C is:
11/2 = M93.	(7-25)
^ox
The most important oxidant in the atmosphere is the HO* radical.
It can react by H-atom abstraction and addition to double bonds and aro-
matic rings. Tables 7-5 and 7-6 list rate constants and half-lives for
reaction of the HO* radical with a variety of organic structures.
More detailed lists of rate constants are found in Hendry and Kenley
(1979) and Hampson and Garvin (1978). Most of the organic molecules with
7-27

-------
763OH
Table 7-4. Oxidant Concentrations in Water and Air
Oxidant	Concentration, H
Water8
R02	1 x 10"9
^	1 x 10"12
Airb
HO	3.4 x 10~15
(8.2 x 10"8 ppn)
03	1.7 x 10"9
(4.1 x 10~2 ppm)
a Fran Mill et al. (1979).
** Fran Hendry et al. (1979).
7-28

-------
7630H
Table 7-5. Rate Constants for Oxidation by HO Radical
in the Atmosphere at 25 Ca
Class
'"""ho
(h"1 s"1)
Half-life.
V
(Days)
n-Alkanes
(C3 " C8>
iso-Alkanes
tc6 ~ ^10^
Cycloalkanes
(c4 - c6)
Haloraehtanes
(1-3 fluorines or chlorines)
1.3 - 5.0
0.6 - 2.9
0.7 - 4.1
1.3 - 4.3
1.9 - 9.4
1.4 - 8
1.2(-4) - 0.065	87 - 47.000
Haloethanes
(1-3 chlorines and 3-4 fluorines)
6(-3) - 0.23	24 - 950
Butanone
1.9
2.9
p-Alcohols
(c2 - c3)
sec-Alcohols
(c3 - c4)
Ethers
(c2 - c6)
Terminal olefins
(c2 ~ C^)
Internal olefins
(C2 - c5)
Anna tics
benzene
toluene
xylenes
4.1
2.5 - 1.0
4.6 - 34
29-90
0.82
3.5
5.9 - 12
2.8
1.3
0.6 - 2.2
0.2 - 1.2
0.06 - 0.2
6.8
1.6
0.47 - 1.0
7-29

-------
7630H
Table 7-5. (continued)
Half-life.
-i -i
Class	(Ms)	(Days)
Trimethylbenzenes	15-30	0.2-0.4
Ethylbenzene	4.4	1.3
Propylbenzene	3.5	1.6
Cunene	4.6	1.2
o-Cresol	20	0.3
Benzaldehyde	7.6	0.74
a Frora Hendry and Ken ley (1979).
^ Assuses [HO ] is 3.4 x 10~^ H and 10 hrs solar days:
t1/2 = 0.69/k(3.4 x 10"15) x (3600) x (10).
7-30

-------
763OH
Table 7-6. Rates of Oxidation by Ozone in
the Atmosphere at 25 Ca
Class
(H_1 s"1)
Half-life, t
(Days)
1/2
Alkanes
Terminal Olefins
Internal Olefins
Branched Internal Olefins
Chloroethylenes
Alkylaroaatics
Alkynes
<0.005
5.8(3)
110(3)
(300-900)(3)
<60
<60
<60
>2(6)
1.4
0.7
0.2 - 0.6 hr
>130
>130
>130
Fran Hendry and Kenley (1979).
D
Per nolecule.
Based on 1 * 10 N 0^; t = Pn/8.6 x 10
k in days.
7-31

-------
-CH- or -CHg- bonds and the aromatic compounds react rapidly with half-
lives of less than a day. Compounds that are highly halogenated such as
the freons are much less reactive or are unreactive.
Rate constants and half-lives of atmospheric oxidation by hydroxyl
radicals for chemicals not listed in Tables 7-5 and 7-6 can be estimated
using a computer program referred to as the Fate of Atmospheric Pollutants
(FAP), which is based on chemical structure. The rate constant k^g for
oxidation of a specific structure can be estimated using the additivity
procedure of Hendry and Kenley (1979) which, depending on the structure,
requires estimation of the individual rate constants for H-atom transfer
and addition. The probable error in values of k^Q estimated in this
way is about a factor of two (+ 100%). Davenport and Mill have updated
and revised kHQ values (1985).
Ozone is a selective oxidant that reacts with electron-rich molecules
such as olefins, eneamines, some phenols, and polycyclic aromatics. Rate
constants summarized in Table 7-6 show that, of the structures listed,
only branched alkanes react fast enough with ozone for this reaction to
compete with oxidation by HO* radical. Hendry and Kenley (1979) have
written a useful summary of specific rate constants for ozone oxidations
in the gas-phase. At present, there are insufficient data to allow for
reliable estimates of ozone rate constants.
7.3 Estimation of Concentrations - Ambient Air
Atmospheric transport models are used in the exposure assessment
process to estimate the pollutant concentrations that occur at the recep-
tor, i.e., the exposed population. Transport models use estimated or
measured emission rates as input data and produce as output a matrix of
airborne concentrations as a function of distance (and sometimes direc-
tion) from the sources. If atmospheric fate processes are not accounted
for in the model, the concentration output values should be adjusted for
them, where possible (see Section 7.2).
Table 7-7 summarizes the key features of the EPA preferred models that
are considered most appropriate for in-depth analysis of the atmospheric
7-32

-------
Table 7-7. Capabilities of Reviewed Models
PT	COMPLEX
MODEL FUNCTION
ISCLT
ISCST
CRSTER
MPTER
RAM
MODELS!
VALLEY
LONGZ
SHORTZ
1
RTDM
Poinl Source





#



.
•
Area Source




•






Volume Source











Short-Term


•
•
•
•



•
*
Long-Term


•
•
•




•

Hourly Meteorological Data


•
•
•
•



•
•
Annual Meteorological Dala










•
Rums Trapping


•

•
•



•

Half-Lile


•
•
•




•

Gravitational Settling











Building Wake Ellects











Buoyant Plume Fiise


•
•
•
•
•



•
Momentum Plume Rise


•
.
•






Slack-Tip Downwash


•
•
•
•





Terrain Ad|ustmenl


•
•


•



•
Complex Terrain






~



•
Urban/Rural Dispersion


•
•
•
•
•




Rural Dispersion Only











Windspeed Extrapolation


•
•
•
•

•
•

•
Em. Rates Fund, ol Met Condl'ns










•
Em. Rales Fund. Ol Time ol Day



•
•





•
Buoyancy Induced Dispersion


•
•
•
•

*
•

•
Deposition







•
•


Discrete Receptors

•

•
•
•

•
•
•
•
STAR Data






•
•



Maximum Number ol Sources
100ft
lOOtt
19
250P
250P100A
25(PTMTP)
50
14.000
300
250
35
Maximum Number ol Receptors
400| f
400 (t
180
180
<50
30
-------
fate of substances released from point and area sources. The applicabili-
ty of each of these models is discussed briefly in the following pages.
Figure 7-2 presents a decision tree to aid in selecting models for air
transport of pollutants.
To apply air dispersion modeling to point and area sources, one should
first examine the terrain where the receptors lie. Terrain can be classi-
fied as either a simple or complex. Model evaluation exercises conducted
to date to determine the best, most appropriate point source model for use
in a simple terrain have found no specific model to be clearly superior.
Thus, based on past use, public familiarity, and availability, CRTSTER re-
mains the recommended model for rural, simple terrain, single point source
applications. For urban, simple terrain, single point source applica-
tions, RAM is highly recommended. For area sources in simple terrain,
ISC is the most appropriate model. These recommendations are presented in
Table 7-8.
For complex terrain, initial screening is usually done using a VALLEY
model. In cases where more detailed information is desired, COMPLEX-1 is
the model recommended for rural areas. In urban areas, SHORTZ is the
model recommended for short-term modeling and LONGZ is preferred for long-
term modeling.
(1)	Some other preferred models. For simple terrain having multiple
sources, MPTER is recommended for short-term modeling in rural areas, and
RAM is recommended for urban areas. For long-term dispersion modeling, if
there are only a few sources, RAM can be used; in other cases, however,
CDM is the preferred model. For line sources, the more appropriate model
in rural areas is BLP. The resource requirements and information sources
for these models are presented in Table 7-9; additional features and data
requirements of these models are presented in Tables 7-10 and 7-11,
respectively.
(2)	Other possible models. Certain other models have advantages in
particular situations. In some instances, estimation of short-term dis-
persion and assessment of the impact on air quality of portions of urban
7-34

-------
Yes

Consult with Regional

Meterologlst or Other


Expert on Proper


Modeling Techniques
Yas

Short


Term

i
r

Long
Term

l
'

Rural

<

Point

Area

Point

Area

Point

Area

Point

Area
Source

Source

Source

Source

Source

Source

Source

Source
ISCST
RAM
SHORTZ
PAL
IMPACT
RPM-11
TEM-8
CRSTER
MPTER
ISCST
ISCST
BLP
SHORTZ
PAL
IMPACT
RPM-11
TEM-8
ISCLT
LONQZ
IMPACT
AODM
ERTAQ
RTM-11
TEM-2
CRSTER
MPTER
ISCLT
ISCLT
BLP
LONQZ
IMPACT
ERTAQ
RTM-11
TCM-2

Short


Term

1
'

Long
Term


'
Urban

Rural

Urban

Rural
SHORTZ
IMPACT
RTDM
COMPLEX-1
SHORTZ
IMPACT
RTDM
LONOZ
IMPACT
RTM-11
COMPLEX-1
LONGZ
IMPACT
RTM-11
rigure 7-2. Decision tree for selection of air transport models.
Source: Versar 1988.

-------
7630H
Table 7-8. Preferred Models for Selected Applications in Single Terrain
Short-Term (1-24 hours)	Land Use	Model*
Single Source	Rural	CRSTER
Urban	RAM
Multiple Source	Rural	MPTER
Urban	RAM
Ccnplicated Sources**	Rural/Urban	ISCST
Buoyant Industrial Line Sources	Rural	BLP
Long-Term (monthly, seasonal, or annual)
Single Source	Rural	CRSTER
Urban	RAM
Multiple Source	Rural	MPTER
Urban	CDM 2.0 or RAM***
Complicated Sources**	Rural/Urban	ISCLT
Buoyant Industrial Line Sources	Rural	BLP
* Several of these models contain options that allow them to be
interchanged. For exanple, ISCST can be substituted for CRSTER, and
equivalent, if not identical, concentration estimates can be obtained.
Similarly, for a point source application, MPTER with an urban option
can be substituted for RAM. Uhere a substitution is convenient to
the user and equivalent estimates are assured, it can be made. The
models as listed here reflect the applications for which they were
originally intended.
** Complicated sources are sources with special problems such as aero-
dynamic downwash, particle deposition, volune and area sources, etc.
*** If only a few sources in an urban area are to be modeled.
Source: Versar 1988.
7-36

-------
Table 7-9. Resource Requirements and Information Sources: Atmospheric Fate Models
References, sources of docunent.
Model	Description	Resource requirements	software
PTMAX
Steady-state 6aussian plume model
Point source
Short-term
Assures conservative pollutant
Produces maxinur hourly concentrations
for each stability and Mind speed
location of maximm concentration
Available through GEMS
FORTRAN IV Program, applicable
to wide range of con^iuters;
has been Inplemented on UNIVAC
1110
Approximately 12 K bytes memory
required
Software available as part of
UNAMAP package for $420
References: SSC 1982
Software: Computer Products. National
Technical Information Service (NTIS).
Springfield. Va. 22161
PTDIS
Steady-state 6aussian pliane model
Point source
Accomodates limit to upward vertical
mixing
Short-term
Assures conservative pollutant
Produces estimates of hourly
concentrations at a user-selected
downwind distance
Available through GEMS
FORTRAN IV Program, applicable
to wide range of confiuters;
has been Inplemented on UN I VAC
1110
Approximately 12 K bytes memory
required
Software available as part of
UNAMAP package for $420
References: GSC 1982
Software: Conputer Products, NTIS,
Springfield, VA 22161
Texas Episodic Model
(TEMJ*
Steady-state model
Point or area sources
Short-term - 10 minutes to 24 hours
Produces maxinun and average concentrations
over time periods selected by user
User can select up to.2500 downwind
receptor points, according to an automatic
or specified grid array
Handles nonconservative pollutants
Up to 24 meteorologic scenarios can be
input for a single run
FORTRAN program applicable to
a wide range of ccnjiuter types;
has been implemented on
Burroughs 6810/11
Requires approximately 26 K bytes
memory
Engineering, meteorology,
atmospheric transport background
useful
Reference: Christiansen 1976

-------
7630H
Table 7-9. (continued)
References, sources of docunent.
Model	Description	Resource requirements	software
Box Model	• Area Source	• Available through GEMS
•	Vertical dispersion or no
vertical dispersion option
•	Basic box model
Texas Climatological Model
Control	(TCM)*
Long-tern (seasonal or annual)
Gaussian dispersion
Two pollutants per run
Includes option for simulation of urban
area turbulence classes
Handles nonconservative pollutants
Point or area sources
Up to 2500 receptor locations on downwind
user-specific grid
Outputs average concentration data
Requires stability array data Documentation:
FORTRAN program language; has Board 1980
been inplemented on Burroughs
6810/11
Batch mode
17 K bytes memory required
Technical background in meteorology,
air pollution useful
Texas Air Control
Climatological Dispersion
Model (COM)
Long-term seasonal or annual
Point or area sources
6aussian plime model
Simulates nonconservative pollutants
Can simulate turbulence over urban
areas
Outputs long-term average concentrations
at user-specified receptors
Requires stability array data
FORTRAN V program language; has
been inplemented on the
UN1VAC 1110
22 K bytes storage required
Software available as part of
UNAMAP package.
Docunentation: Busse and Zirnnerman
1976
Software: Conputer Products, HTIS.
Springfield, Va. 22161
VALLEY*
Short- or long-term
Simulates plume inpact in complex
terrain
Provides screening estimates of worst-
case short-term concentrations
Provides annual average concentrations
12-receptor grid
Hay require careful analysis of
output by experienced air
quality modeler
FORTRAN V program, applicable
to wide range of confiuters
Approximately 13 K bytes memory
required
Reference: Burt 1977
Software: Conputer Products, NTIS,
Springfield, Va. 23161

-------
Table 7-9. (continued)
References, sources of document.
Model	Description	Resource requirements	software
•	Accomodates nonconservative pollutants • Software available as part of
•	Requires stability array data for	UNAMAP.
long-term option
•	Requires user-input worst-case
meteorological data for short-term
screening option
Industrial Source Caiplex
(ISC)
Operates in both long-term and short-
term modes
Accounts for settling and dry deposition
of particles; downwash, area, line, and
volune sources; plune rise as a function
of downwind distance; separation of point
sources; and limited terrain adjustments.
Appropriate for industrial source com-
plexes. rural or urban areas, flat or
rolling terrain, transport distances less
than 50 kilometers, and one hour to annual
averaging times.
Integrated into GEMS
Source data: location, onisison
rate, physical stack height,
stack gas exit velocity, stack
inside diameter, and stack gas
tenperature. Optional inputs
include source elevation, build-
ing dimensions, particle size,
distribution with corresponding
settling velocities, and surface
reflection.
Meteorological data: includes
stability wind rose (STAR deck),
average afternoon mixing height,
average morning mixing height,
and average air tenjierature.
Doc mentation: Bowers et al. 1979
Software: Conputer Products, NTIS,
Springfield. VA 22161
Ram
Steady-state 6aussian plune model.
Appropriate for point and area sources,
urban areas, flat terrain transport dis-
tances less than 50 kilometers, and one
hour to one year averaging times.
May be used to model primary pollutants,
however settling and deposition are not
treated.
Available code on UNIMAP
(Version 6)
Source data: point sources
require location, emission rate,
physical stack height, stack gas
exit velocity, stack inside
diameter and stack gas tenpera-
ture. Area sources require
location, size, emission rate,
and height of emission.
Reference: Turner and Novak, 1978.

-------
7630H
Table 7-9. (continued)
Model
Description
Resource requirements
References, sources of docunent,
software
Meteorological data: hourly
surface weather data from the
preprocessor RAfMET. Actual
anemometer height is also
required.
CRSTER
Steady-state 6aussian dispersion model.
Designed to calculate concentrations
from point sources at a single location.
Highest and second highest concentra-
tions are calculated at each receptor.
Appropriate for single point sources,
rural or urban areas, transport distances
less than 50 kilometers, and flat or roll-
ing terrain.
Available on UNINAP (Version 6).
Source data: emission rate,
physical stack height, stack
exit velocity, stack inside
diameter and stack gas tempera-
ture.
Meteorological data: hourly
surface weather data from the
preprocessor RAMMET. Actual
anemometer height is also re-
quired.
Reference: USEPA 1977a.
MPTER
Multiple point source algorithn useful
for estimating air quality concentration
of relatively nonreactive pollutants.
Apropriate for point sources, rural or
urban areas, flat or rolling terrain,
transport distances less than 50 kilome-
ters, and one hour to one year averaging
times.
Source data: location, emission
rate, physical stack height,
stack gas exit velocity, stack
inside diameter, stack gas tem-
perature, and optional ground
level elevation.
Meteorological data: hourly
surface weather data fran the
preprocessor RAtWET. Actual
anerameter height Is also
required.
Docunentatlon: Pierce and Turner 1980.
Chico and Catalano 19B6.
* These models are not EPA-preferred models. They can. however, be used if it can be demonstrated that they estimate concentrations equivalent to those
provided by the preferred models (e.g., COM, RAM, ISC, MPTER, CRSTER, for a given application).
** This model is reconmended for screening applications only.
3	Bonazountas et al. 1982; USEPA 1982a; USEPA 1979b.

-------
Table 7-10. Features of Atmospheric Models
* This model is recommended for screening applications only.
** These models are not EPA preferred models. These models can be used if they can be demonstrated to estimate concentra-
tions equivalent to those provided by the preferred models, e.g., CPM, RAM, ISC, MPTER, CRSTER, for a given application.
Source: Bonazountas et al. 1982; USEPA 1979b; USEPA 1982a

-------
Table 7-11. Data Requirements for Atmospheric Models
* This model is recommended for screening applications only.
** These models are not EPA preferred models. These models can be used if they can be demonstrated to estimate concentra-
tes equivalent to those pnmded by the preferred	e.g., CRM, RAM, ISC, MPTER, CRSTER, for a given applicati^
Source: Bonazountas et al. 1982; USEPA 1979b; USEPA 19

-------
areas such as shopping centers, large parking areas, and airports, can be
accomplished using PAL, if there is justification. Similarly, AQDM can be
used to estimate dispersion of sulfur dioxide and particulate. For cases
of multiple reflections involving fugitive dust and accounting for
entrainment of particulates from ground level sources, ERTAQ can be used
to model deposition. IMPACT can be used with less than 20 point sources
for inert or reactive pollutants when one needs to consider the effects of
vertical temperature stratifications on wind and diffusion fields, shear
flows caused by the atmospheric boundary layer or by terrain effects, and
chemical transformations. For compounds that undergo photochemical
changes, RPM-11 can be used. For applications at distances up to
15 kilometers in a complex terrain for one or more sources, one may use
RTDM. However, this model does not consider deposition effects or chemi-
cal transformations within the plume.
7.4 Exposure Estimation - Ambient Air
This section has provided guidance for calculating environmental
concentrations resulting from contaminant release to air. The presence
of contaminants in ambient air can pose a potential health risk to humans
through the following exposure routes:
•	Inhalation of air containing contaminants in vapor or particulate
form;
•	Absorption of contaminants through the skin from the ambient air
(unlikely to be a major route of exposure for most pollutants);
•	Ingestion of a portion of inhaled particulate-associated pollut-
ants as a result of lung clearance mechanisms; and
•	Ingestion of foods contaminated with pollutants as a result of
dry and wet deposition processes, uptake, and bioaccumulation in
the food chain.
Equations and factors to be used in calculating exposures to chemicals
in ambient air are contained in the Exposure Factors Handbook (USEPA
1989).
7-43

-------
8.	ASSESSMENT OF EXPOSURE VIA DIRECT CONTACT WITH CONTAMINATED
SOILS
In this section, exposure to chemically-contaminated soils and the
resulting effects on human health are discussed. Contaminated soils can
be found at controlled as well as uncontrolled hazardous waste treatment,
storage, and disposal sites. Chemical contaminants in these soils (or
soil environment) can be released into a variety of environmental media,
including air and water; routes for release are discussed in Sections
3.1, 4.1, and 7.1 of this report.
The soil environment is defined as a stratum extending from the top
of the soil layer to the capillary fringe of the ground water. This
environment is a dynamic, interdependent system of air spaces, bulk soil,
soil water, and soil biota that are linked by physical, chemical, and
biological processes. Hazardous substances introduced into soil use
these linkages to migrate within and between the soil systems. Hazardous
substances in soils can potentially volatilize, transform, adsorb,
degrade, or percolate depending upon the soil and waste characteristics.
8.1 Release Analysis - Soil
Release of a given contaminant from the soil occurs primarily as a
result of volatilization, transformation or degradation, plant uptake, or
leaching. These processes determine the ultimate disposition of
hazardous components in the soil.
One of the major release pathways from soil is to ground water. As a
general rule, those substances that sorb to or form complexes with soil
particles are less likely to be mobile than those that do not. On the
other hand, care should be taken in considering the release potential of
various chemicals with regard to their potential health impact. For
example, a substance that has a low solubility (i.e., low release
potential), but a high health hazard at low concentrations, should be
8-1

-------
examined very carefully before deciding to exclude it from a list of
chemicals of concern. Furthermore, chemicals that are strongly sorbed to
soil particles may still be of concern for human exposure via other
routes, e.g., direct dermal contact with contaminated soil particles and
inhalation of contaminated particulate matter emitted by the site (see
Section 7.1).
Release of the chemical in vapor form from soil to the atmosphere may
also occur. The volatility of a chemical and its concentration in soil
control the extent to which it will exist in the vapor phase within the
soil. The porosity, chemical activity, and general water content of the
soil determine how fast and how far the gaseous chemical will migrate
both laterally and vertically. For near surface cases, the barometric
pressure and ambient temperature will also play a part in determining
release rates from soil to the atmosphere.
8.1.1 Screening of Release Potential
Depending upon where the hazardous wastes are deposited in or on the
land, chemical releases may occur from the soil environment by any one or
all of the processes noted above. It is safe to assume that chemical
releases from uncontrolled hazardous waste sites will occur as long as
the wastes remain in the soil regime. On the other hand, contaminant
releases from controlled sites are expected to be minimal under nominal
conditions; however releases may occur due to unanticipated events (e.g.,
failure of the liner).
Table 8-1 provides a comprehensive list of soil, waste, and site
factors and characteristics that determine the degree and extent of
hazardous material releases from contaminated soils. The parameters
listed in Table 8-1 are applicable in determining contaminant release and
transport, and in identifying potential populations that may be receptors
of releases from the site.
(1) Air releases. Release of particulate-bound chemicals to the
air from contaminated soils is likely to occur by wind erosion from
unpaved roads and land and through excavation and transport of soil. The
8-2

-------
7596H
Table 8-1. Factors Necessary to Determine Releases
from Contaminated Soils
Release
nedia
Area of
concern
Form of
release
Factors
Air
Particulate
release
Uind erosion
Soil erodibility index
Soil ridge roughness factor
Field length along prevailing Mind direction
Vegetative cover factor
Concentration of contaminants
Volune of contaminated region
Unpaved roads	- Silt and clay content
- Mean nirtoer of wheels speed, and mean weight of
vehicles traversing contaminated area
Excavation and
transfer of soil
Volatilization General
Silt and clay content
Mean wind speed
Drop height
Naterial moisture content
Dwping device capacity
Depth from soil surface to bottom of contaminated
region
Area of contamination
Depth of uncontaminated zone
Concentrations of contaminants in soil and in liquid
phase
Soil porosity
Absolute tenperature
Baronetric pressure
Concentration of contaminants in soil
Effective depth of soil cover
Mole fraction of contaminants in waste
Without internal
gas generation
With internal gas
generation
Area of contamination
Soil porosity
Absolute ambient tenperature
Absolute ambient pressure
Soil bulk density
Vol line of contaminated region
Vapor concentration of contaminants in soil pore
spaces
Area of contamination
Concentration of contaminants in soil
Effective depth of soil cover
Mole fraction of contaminants in waste
8-3

-------
7S96H
Table 8-1. (continued)
Release
media
Area of
concern
Form of
release
Soil and waste characteristics
Surface Water Runoff
- Soil erodibility factor
-	Slope-length factor
-	Vegetative cover factor
-	Erosion control practice factor
-	Area of contamination
-	Soil bulk density
-	Total area concentrations of contaminants
-	Solubility
-	Adsorption
-	Degradation
-	Concentration and loading
-	Density, viscosity
-	Soil porosity
-	Infiltration, net recharge, and volunetric Mater
content
-	Bulk density
-	Hydraulic conductivity
-	Chemical characteristics of mediun
-	Dispersion
-	Depth to ground aater
-	Structural and geologic features
Ground Water Release to
ground water
-	Contaminant identity and physical state
-	Extent of the contamination
8-4

-------
quantities of contaminated particulate released into the air are
dependent on the factors identified in Table 8-1. Volatilization of
contaminants from soil into air is a physiochemical process that is
dependent on soil and chemical-specific factors. Human activities such
as excavation and site remediation may increase the volatilization rate
of some contaminants from the site.
(2)	Surface water releases. Release of chemical contaminants from
soils to surface waters occurs primarily via runoff. Runoff may be
either sporadic or long term in nature. During sporadic storm events,
large quantities of contaminants can be removed at a rapid rate and
transported long distances. In addition, soil erosion by wind and
previous runoff cause exposure of lower soil layers, which can contribute
to the future release of contaminants into the surface waters.
(3)	Ground-water releases. Precipitation or rainfall intensity and
pattern distribution have a critical influence on the extent of chemical
transport through soils. Chemicals bound to the soil may become
mobilized by infiltration of water and can then move towards the ground-
water table. Also important in this scenario are the pH of the soil and
rainwater, the type of soil (clay type, sand, silt, etc.), ion exchange
capacity, organic content, and the physical state of the chemical. An
example of the importance of the chemical's physical state would be the
recent recognition by environmental scientists that dioxin can travel
some distance through the subsurface without being bound to the soil as
would normally be expected. This phenomenon is attributed to the fact
that in some instances the dioxin is preferentially dissolved into tiny
hydrophobic micelles that are capable of moving through the soil pores
with percolating water.
8.1.2 Quantitative Release Estimation
Once potential chemical release sources have been identified and
screened, those requiring further evaluation are quantitatively
analyzed. This may involve the application of simplified estimation
8-5

-------
approaches or more rigorous and resource-intensive in-depth computerizeu
modeling, which requires considerably more site-specific data. The goal
of this analysis is the generation of quantitative release rate estimates
for each release source. These analyses can then be used as a basis for
exposure estimates.
(1)	Air releases from dust generation. Emissions of chemically-
contaminated fugitive dusts are attributable to wind erosion, vehicular
traffic at the site, and excavation and loading of contaminated soils.
Using the release equations given in Section 7.1.2 of this report,
assessors can estimate suspended dust levels and quantities of chemical
contaminants expected to enter the air by multiplying the amount of dust
generated by the weight percent of the chemical either in the contaminated
soil or in actual fugitive dust monitoring samples. The total particulate
emissions can be estimated by summing the values for wind erosion,
vehicular traffic, and excavation and transfer of contaminated soils.
(2)	Air releases from volatilization. One other form of air release
is via volatilization of chemicals present in the soils. Volatilization
is dependent on a wide variety of site and climatologic characteristics
as well as time (see Table 8-1). For estimating emissions from volatili-
zation from sites with and without internal gas generation, assessors can
use the appropriate equations in Section 7.1.2.
Most of the equations mentioned above and discussed in Section 7.1.2
are generally used for calculating short-term releases. When calculating
long-term releases (arbitrarily assumed to be 70 years) from wind erosion
of contaminated particulate and for volatilization from landfills both
with and without internal gas generation, assessors can use Equation 7-19
presented in Section 7 of this report for a more appropriate estimate.
If there exists a potential for a chemical to remain in soils beyond
70 years, Equation 7-20 may be more appropriate to use. Neither of these
equations takes into consideration chemical loss via transformation
processes, which is measured by chemical half-life.
8-6

-------
(3)	Surface water releases. Release of chemicals into surface
waters from soil depends primarily on adsorption/desorption mechanisms
and how hydrophobic and/or polar the material of concern is. The release
that is soil-related occurs through three routes: (1) runoff that carries
dissolved chemicals as well as soil particulates into the receiving body;
(2) wind deposition that is the result of particulates being blown into
the receiving body; and (3) bank seepage where, after a precipitation
event, water drains into a receiving body through ordinarily unsaturated
soils. Of the three, runoff is by far the most important release
mechanism. For a detailed discussion of dissolved and sorbed contaminant
release, see Section 3.1 of this Handbook.
(4)	Ground-water releases. Release of toxic substances from contam-
inated soils to ground water occurs through leaching and percolation.
Methods for estimating releases from soil to ground water are described
in Section 4.1.
8.2 Fate Analysis - Soil
Using quantitative release estimation methods, the characteristics of
the hazardous contaminants, and various environmental transport, transfor-
mation, removal, and migration mechanisms, fate analysis can be performed
and used to estimate exposure.
8.2.1 Screening of Soil Fate Processes
The fate of hazardous contaminants in the soil medium is assessed
whenever significant quantities and concentrations of contaminants are
shown to be present. Contaminants dissolved in water and infiltrated
liquids will migrate or seep through percolation into the subsurface
layers of the soil. Migration rates for waterborne contaminants depend
on net water recharge rates, chemical characteristics, and permeability
of the soil, and to a lesser extent water solubility of the chemical(s)
of interest, while migration of gaseous materials also takes into account
volatility, temperature, barometric pressure, and initial water content
of the soil.
8-7

-------
The environmental fate of land-disposed industrial wastes and other
chemicals introduced to soil can be assessed by a consideration of the
following fate and transport processes: (1) sorption, (2) hydrolysis,
(3) biodegradation, (4) volatilization, and (5) photolysis/oxidation. A
brief statement about the role of each process in a landfill or
contaminated soil is presented in the following subsections. Directions
are also given for estimating the significance of each process in
determining the fate of specific chemicals.
(1)	Sorption. The most important property for predicting a
chemical's potential to persist in soil systems is its potential for
sorption to the solid phase of soil. Sorption should be estimated (if
not available) and considered along with any available information on
other fate processes. Sorption to soil by organic chemicals can usually
be correlated with the content of organic matter in the soil (Lyman et
al. 1982, Leifer et al. 1983). Consequently, the estimated sorption is
generally calculated as the organic carbon partition coefficient, KQC,
which is defined by the relationship
Koc - Kd/0C	(8-1)
where
ICj = the soil partition coefficient
OC = the fractional mass of organic carbon in the soil.
Estimated KQC values can be multiplied by an assumed fractional amount
of soil organic carbon to obtain K^, which can then be related to soil
mobility (Leifer et al. 1983). See Section 4.2.1 for a discussion on
sorption and mobility.
(2)	Hydrolysis. Since hydrolysis in soil is obviously dependent on
the presence of soil moisture, the efficacy of this degradation process
will be diminished in arid regions or dry seasons. For a discussion on
estimating the rate of hydrolysis for a specific chemical, see
Section 4.2.1.(4). Figure 4-1 in particular provides a useful reference
for hydrolytic potentials and rates.
8-8

-------
(3)	Biodeqradation. Although biodegration is one of the principal
mechanisms by which organic chemical pollutants are removed from soil,
the rate at which biodegradation occurs in soil cannot be predicted with
any reliability (Lyman et al. 1982). Even the more general property of
susceptibility to biodegradation is not readily predicted, and
predictions regarding it must be approached with considerable
professional judgment. At present, the biodegradability of chemicals is
best estimated by comparing the chemical structure to that of other
organic chemicals that have been studied (Lyman et al. 1982).
(4)	Volatilization. Rates of volatilization from soil are
difficult to predict (Lyman et al. 1982). An estimate of relative soil
volatility (RSV) for arid regions or dry seasons may be obtained from the
following relationship (USEPA 1985b)
RSV - Pvp(MW)"1/4	(8-2)
where
RSV = Relative soil volatility (atm)
PVp = Vapor pressure at 25°C (atm)
MW = Molecular weight.
The equation was developed from the Hamaker Method for determining vola-
tilization from dry soil (Lyman et al. 1982). The relative importance of
different RSV values is shown in Table 8-2.
To determine the importance of volatilization from wet soils and
surface impoundments, the Henry's law constant is derived from the
following relationship (Lyman et al. 1982):
H = Pvp/S	(8-3)
where
PVp = vapor pressure in atmospheres
S = solubility (mol/m3)
H = Henry's law constant (atm-nr/mol).
If H<3 x 10-7, volatilization from a wet soil or landfill can be
considered unimportant. If H>3 x 10~7, the chemical can be considered
8-9

-------
7596H
Table 8-2. Inportance of Volatility from Dry Soils
Relative soil volatility

(RSV)a
Inportance for transport
> 1
Highly volatile
10"3 < RSV < 1
Moderately volatile
10"6 < RSV < 10"3
Slightly volatile
< 10"6
Nonvolatile
'Calculated from RSV = P (MY)"1/4 in USEPA (1985b). P is the
vp	vp
vapor pressure at 20 C, and NU is the nolecular weight.
8-10

-------
volatile from water and volatilization from wet soils may be an important
transport process. If solubility data are unavailable, this parameter
can be estimated using the methods of Lyman et al. (1982).
(5) Photolvsis/Oxidation. Photolysis and oxidation generally do
not contribute significantly to the destruction of most chemicals in soil
or to their potential for contamination of ground water. However, some
chemicals are unstable with respect to atmospheric oxidation (e.g., some
aromatic amines) and may decompose appreciably in the aerated surface
soil of a landfill. Thus, oxidation may be responsible for the formation
of some chemical degradation products. Photolytic/oxidative destruction
of chemicals can also occur in the aqueous surface film of wet soil.
Section 7.2.1 can be consulted to identify which chemicals are susceptible
to destruction by this process.
8.2.2 Quantitative Fate Estimation
Appropriate subsections on fate analysis in surface water
(Section 3.2), ground water (Section 4.2), and air (Section 7.2) should
be consulted for quantitative treatment of the fate processes discussed
in Section 8.2.1.
8.3 Estimation of Concentrations - Soil
The simplified equations presented below employ previously developed
information and will usually produce conservative estimates for final
ambient concentrations and the extent of chemical contaminant migration.
Since these simplifications may, on occasion, lead to an overstatement of
human exposure, it may be necessary to use more in-depth environmental
fate methods, such as those methods discussed in Sections 3.2.3, 4.2.3,
and 7.2.3, to arrive at an accurate exposure estimate.
8.3.1 Estimation of Airborne Concentrations Resulting from Contaminated
Soil
For assessing airborne concentrations, the following equation can be
used:
* °z m
8-11

-------
where
C/x\ = concentration of substance at distance x from site
Q = release rate of substance for site (mass/time)
7t = 3.14159
ay = lateral dispersion coefficient
oz = vertical dispersion coefficient
n = mean wind speed (distance/time)
Values for oz, and n are obtained from Figures 8-1 and 8-2.
Equation 8-4 takes into consideration the two predominant atmospheric fate
mechanisms--advection and dispersion, and estimates ground-level atmos-
pheric concentrations of chemicals at selected points directly downwind
from a ground-level source. Wind direction should be obtained from site
specific meteorological data (or the nearest wind rose collection sta-
tion). As a generic default procedure, Equation 8-4 can be solved assum-
ing steady state wind conditions, mean wind speed of 3 m/sec, negligible
lateral dispersion, and lack of removal or decay processes. Additional
default assumptions are shown in Table 8-3.
The loss of chemicals in vapor form is limited by the chemical vapor
concentration that is maintained at the soil surface and by the rate at
which this vapor is carried away from the soil surface to the atmosphere.
The volatilization from soil depends on the balance between two rate
processes: the flux of chemical from the soil body to the soil surface
and the flux of chemical away from the soil surface to the atmosphere.
The total quantity of chemical (Cy) per unit soil volume is given by
Cy = <7|jC^ + flC|_ + aCg	(8-5)
where
Cfl = adsorbed chemical concentration (ug/g soil)
= dissolved technical concentration (ug/cnr solution)
Cq = vapor density (ug/cnrsoil air)
, = soil bulk density (g/cm3)
e = volumetric water content
a = volumetric air content.
If the results of the above equation indicate that significant ambient
atmospheric concentrations may occur, then additional in-depth analysis
8-12

-------
10,000 -
1,000-
E
>
100-





















T" 7
1












































!.•

•























•





















•
•
•



*
•



















•*
•
•

~*
•



















•
•
•
•
•
•
*
•
•
*
•
•

•
•
•

















•
•
•
•
•
•
-•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
*
•















•
•
4
•
4
•
»
4
«
~ •
• •
•
•
•
•
•
•
•
•
•
•
•
~
»
•
•
•
%
•
•
•
•
«
•
•
«
•
•
•
»•
•
•
«
•
*
•
•
•
•
•<
•
«













•

•


•
•

















•

•


I
•
•
•

















•
•
»<
•


*
•
* *1
•

















•
•
•


•
<#

• •
• •
•
•
















#
•

»

•

•
u ~
« •
•
















•
• *1
~
•
•
«
•
»

•
•
•
y
















AV*"
«
*
•
~
•
•
•
•
•

~
•
•
¦»
w
•
#
•
•
~
•













•


B.-'l
•
•
~
#
Li*
•
•
~
0
•
*
•
•
.*
0
i#
~
•
•
(.1













•
•
•
•
*
•
«k
•
•
L) .
•
*
•
~
•
•
•
•
•
•
•
»
F .
•
•
•
•
*
•
•
•
•



















	^—r2
•























• ~*
•























. • A
























• V
• s
• s





























































































































































































































































i










10
0.1
1	10
DISTANCE DOWNWIND, km
100
Figure 8-1. Horizontal dispersion coefficient as a function of downwind distance from the source.
* Lines designated A through F represent dispersion coefficient functions for atmospheric stability
classes A through F.
Source: Turner (1970).
8-13

-------
vo	1 I I I I I |	r——l I l I 111	——I I ll I
0.1	1	10	100
DISTANCE DOWNWIND, km
Figure 8-2. Vertical dispersion coefficient as a function of downwind distance from the source.
* Curves designated A through F represent dispersion coefficient functions for atmospheric
stability classes A through F.
Source: Turner (1970).
8-14

-------
7596H
Table 8-3. Default Assunptions for Calculation of
Short-tern Maxinun Concentrations in Air*
Stability	Percent
Duration Hind speed	class	towards exposure point
1 hour	1 m/sec	F	100
24 hour	2 m/sec	E	50
7 days	3 */sec	D	30
* These asswptions are provided for general guidance only. For seme
sites, such as deep valleys, these input assmptions nay be
inappropriate and detailed site-specific data nay be required
(see Section 7.2).
8-15

-------
of concentrations at the receptor may be required, as described in
Section 7.3, incorporating the fate considerations noted in Section 7.2.
8.3.2	Estimation of Concentrations in Surface Water Resulting from
Contaminated Soil
Estimation of concentrations of chemical contaminants in surface
water resulting from contaminated soil is discussed in Section 3.3
8.3.3	Estimation of Concentrations 1n Ground Water Resulting from
Contaminated Soil
The ground-water concentration resulting from chemicals adsorbed in
the soil matrix (i.e., unsaturated zone) can be determined by mainly
site-specific factors. The equation used to estimate the mean rate of
downward travel is as described below:
where
Vpw = interstitial pore water velocity (length per unit time)
q = average percolation or recharge rate (depth per unit
time) obtained for the site data
e = volumetric water content of unsaturated zone.
When average percolation rate information is not available, it can be
estimated by subtracting values given for the precipitation, evaporation,
and runoff rates from hydraulic loading values. In-depth analysis of con-
taminant migration in the unsaturated zone may be performed using models
such as SESOIL, PESTAN, or HELP, which are described in Section 4.3.
For assessing the adsorption potential of organic chemicals onto soil,
it is essential to understand the diversity and complexity of the adsorp-
tion surfaces. Most organic chemicals are attracted to both clay mineral
surfaces and organic matter surfaces. For a given chemical, the nature
of the bonding mechanisms to the accessible adsorption sites depends on a
variety of factors such as clay composition, organic water content, soil
water content, soil bulk density, pH, solution concentration, chemical
polarity and soil temperature. The bonding of chemicals on the soil
8-16

-------
surface occurs because of the formation of ionic, ligand, charge dipole,
dipole-dipole, hydrogen, charge transfer, and van der Waal's-London bonds.
Using the organic carbon partition coefficient (Koc), it is
possible to determine the adsorption potential of a chemical, as noted in
previous sections:
Koc = Kd/0C	(8-7)
where OC is organic carbon fraction and is the soil partition
coefficient. The Kd can be related to other physical-chemical properties
such as the octanol-water partition coefficient, K (Lyman et al.
uw
1982). The amount of a chemical adsorbed on the soil particle is deter-
mined by the Freudlich isotherm equation:
x = KdC1/n	(8-8)
m
where
^ = the amount adsorbed (ug chemical/gm of soil)
C = the concentration of chemical (ug/mL) in the aqueous phase
= soil/solution partition coefficient
n = a constant.
Following the assessment of the transport, fate, and loss mechanisms,
including the dissolution, volatilization, and dispersion losses, the
amount of contaminants remaining on the soil can be calculated by
rearranging Equation 8-5 as follows:
CA = CT " gCL" aCG	(8-9)
ab
If the results of this equation indicate that the concentrations of
chemical contaminants are significant, then additional in-depth analysis
of the concentrations resulting from releases from soil may be necessary,
using the methods described in Section 4.3 and incorporating the fate
consideration noted in Section 4.2.
8-17

-------
8.4 Exposure Estimation - Soil
In order to assess human exposure to chemical contaminants, a compari-
son between environmental contamination data and population data is
needed, which requires the essential steps of screening potential exposed
populations and identifying exposure pathways. Primary exposure to
contaminated soils is through direct dermal contact. Secondary indirect
routes include inhalation of dust particles and ingestion of contaminated
soils and foods grown in the contaminated soils. Another route of
exposure is contaminated surface water or ground water intake.
Once the exposure potential of the site is identified, demographics
(e.g., age, sex) of the potentially exposed populations is recommended.
This step will allow the identification of high-risk groups such as
children playing on the contaminated soils, children eating dirt, and
children and adults inhaling the fugitive dust, volatiles, or both.
Equations and parameters for estimating these exposures are presented in
the Exposure Factors Handbook (USEPA 1989).
8-18

-------
1.	EXPOSED POPULATIONS
Detailed information regarding the methods for enumerating and charac-
terizing populations is contained in Methods for Assessing Exposure to
Chemical Substances. Volume 4 - Methods for Enumerating and Characterizing
Populations Exposed to Chemical Substances (USEPA 1985c). This report
provides guidance on choosing methods and presents a series of steps to be
performed.
Populations that are potentially exposed to contaminants via inhala-
tion of ambient air will be identified on the basis of their geographic
location relative to the site or source of chemical contaminants. Choos-
ing the most appropriate method for enumerating a population exposed in
the ambient environment depends not only on the route of exposure but also
on the characteristics of the chemical source. Different methods apply to
point sources, area sources, and line sources and the methods may differ
according to the number of similar sources to be dealt with. A point
source release is easily defined, and its position can be accurately
determined; the population consequently exposed is described in terms of
distance from the source. The report discusses the use of a program
called SECPOP, part of EPA's computerized GEMS system. SECPOP uses data
from the Census Bureau to accurately perform population enumeration,
along with an atmospheric dispersion/exposure model (ATM) if desired.
These sources may be updated with detailed population data from Census
Bureau documents or through contact with the appropriate local and county
governments.
The methods discussed above focus on inhalation exposures. Other
types of exposures that can be important are ingestion of drinking water,
dermal contact with waterborne pollutants during showering and recreation,
and more incidental exposures such as soil ingestion.
Some exposure assessments require that the exposed populations be ex-
ceptionally well-defined. The methods report on populations (USEPA 1985c)
presents age and sex distributions for the U.S. as a whole and lists
1-1

-------
Census publications that go to various levels of geographic resolution
with such data.
Populations exposed via food ingestion are of interest if food chain
exposure is deemed to be important for the site of interest. The methods
are keyed to the source of contaminant: agricultural practices, process-
ing or packaging, other sources, and unknown sources (when contamination
is demonstrated by monitoring data). Data from national surveys of food
consumption, such as those performed by the USDA, can be used to estimate
the fraction of households consuming certain foodstuffs and are the basis
of the methods to characterize the exposed populations. Key information
from USDA also includes data on consumption of home-grown vegetables.
The presence of possible contamination through surface water and
ground water will necessitate characterization and enumeration of popula-
tions exposed via the consumption of drinking water. Data maintained by
EPA are the basis for most of the methods. One method relies on EPA files
such as the Hydrologically-Linked Data File (HLDF) and the Water Supply
Data Base (WSDB). HLDF is a group of data bases on surface water systems
in the United States that contains information on discharges, drinking
water supplies (including the number of persons served by each private or
public utility in a given locality), stream/river flow rates, and other
data. This data base is maintained by the Office of Water Regulations and
Standards at EPA. Populations that are potentially exposed to contami-
nants via ingestion or have use of drinking water from specific contami-
nated surface water sources and contaminated ground water can be obtained
from EPA's Federal Reporting Data System (FRDS).
In addition to the guidance provided in the population methodology
report (USEPA 1985g), the Exposure Factors Handbook (USEPA 1989) discusses
the activity patterns and mobility of various populations. Use of activi-
ty pattern data in conjunction with population size data will provide
important information regarding exposure duration for specific subpopula-
tions.
1-2

-------
2.	CHARACTERIZATION AND ANALYSIS OF UNCERTAINTIES IN THE SAMPLING,
MONITORING, AND MODELING OF ENVIRONMENTAL CONCENTRATIONS OF
CONTAMINANTS
The goal of an analysis of uncertainties is to provide decision makers
with the complete spectrum of information concerning the quality of a con-
centration estimate, including the potential variability in the estimated
concentration, the inherent variability in the input parameters, data
gaps, and the effect these data gaps have on the accuracy or reasonable-
ness of the concentration estimates developed. Analysis and presentation
of uncertainties will allow the user(s) or decision maker(s) to better
evaluate the concentration results in the context of other factors being
considered.
2.1 Causes of Uncertainty
The basic cause of uncertainty as discussed by Cox and Baybutt (1981)
is a lack of knowledge on the part of the analyst due to inadequate, or
even nonexistent, data on parameter values. Morgan et al. (1984) and
Henrion (1985) have identified and further explained the causes for a lack
of knowledge in risk-policy analysis that are directly related to concen-
tration estimates among other estimates. A modification of the causes of
uncertainty as identified by Morgan and Henrion with examples relevant to
the the concentration estimate process are as follows:
1.	Measurement error (uncertainty in true value) - Uncertainty in
this case arises from random and systematic error in the measure-
ment technique for direct empirical data. Random or indeterminate
error in a measured value results from imprecision of the measure-
ment process. Systematic error in the measurement technique is a
tendency or bias to measure something other than what was
intended. Examples relevant to the concentration estimate process
include experimental error in measured physical- chemical proper-
ties (e.g., solubility, vapor pressure, boiling point) and experi-
mental errors in environmental monitoring data.
2.	Indirect empirical or generic data - When data of direct rele-
vance are not available, indirect empirical data may be used to
fill the data gap. Uncertainty in indirect data arises from both
the applicability of the indirect data and the measurement error
in the data; however, application is generally of greater concern.
In concentration estimation, the use of indirect or generic data
can be significant due to frequently encountered data gaps. Data
2-1

-------
for structurally similar chemicals are also frequently used to
evaluate physical-chemical properties and the environmental fate
of a chemical substance where chemical-specific information is
lacking.
3.	Variability - The natural variability in environmental and
concentration related parameters can frequently be a major source
of uncertainty. Inherent random variability in physical phenomena
can lead to unavoidable unpredictability in the quantity of
interest. Examples of variability include meteorological (e.g.,
wind speed, temperature), hydrological (e.g., river flow, water
depth), and chemical intake related parameters (e.g., inhalation
rate, ingestion rate, bioaccumulation). Concentration estimation
involves the process of estimating concentrations for settings and
scenarios that represent actual conditions of interest. Selection
of values for inherently random input parameters, therefore, can
create a significant state of uncertainty.
4.	Modeling - In the context of this report, modeling implies
the use of some tool, algorithm, or framework that is designed to
simulate some form of reality. Models can be simple algorithms
(e.g., a x b = c), complex computer based algorithms, and even the
framework or process for conducting an analysis such as a concen-
tration estimate. Modeling approximations can be and usually are
a major source of uncertainty in concentration estimation, the
basic question being "How close to reality is the model function
and output?" Modeling uncertainties are compounded because they
also overlap with other sources of uncertainty and can propagate
them to the confusion of the analyst. For example, not only is
the analyst concerned about the model form, he is also concerned
with the selection of suitable input parameters that can be
affected by inherent variability as well as measurement error.
5.	Sampling error - Sampling error as a source of uncertainty is
intimately related to natural variability, and it can be argued
that sampling error is a factor of variability and not a separate
source. However, because the use of environmental monitoring data
and other types of sampling data is extremely important in the
concentration estimation process, sampling error as a source of
uncertainty has been listed separately. The major question
related to the quality of sampling data (and, therefore, a measure
of uncertainty) is whether the data are truly representative of
the population sampled. Assessment of sampling error involves
evaluation of the randomness of sample selection, the adequacy of
the number of samples and sampling frequency, and sampling
accuracy and precision.
2-2

-------
6. Professional judgment/disagreement - The use of professional
judgment is an inherent factor in concentration estimation from
the planning of the estimation scope and approach, to the imple-
mentation of the approach, evaluation of information, and inter-
pretation of results. It is often thought, however, that profes-
sional judgment is only a factor when it is used to explicitly
fill gaps in required information. Even when one thinks that an
analysis is truly objective, professional judgment is present in
the selection of the analysis procedure. As will become apparent
later in this report, selection of a quantitative procedure to
determine variability in an output parameter is also dependent
upon the professional judgment of the analyst.
2.2 Qualitative Analysis of Uncertainties
Once the causes of the uncertainties in a concentration estimation
have been identified, a qualitative analysis of the impact these uncer-
tainties have on the estimation results should be determined. Where
uncertainty exists, data or estimates of a range of plausible values
should be gathered. The effects of the range of assessment input values
on the estimated results (or output value) can be determined numerically
through the substitution of plausible alternative values for each input
parameter. This procedure is similar to a sensitivity analysis, which is
discussed in the next subsection. The variation in output attributable
to variations in input values or parameters can thus be evaluated. It is
recommended that the concentration estimator construct a table that
includes a listing of possible variations in each parameter that would
encompass a reasonable range of actual expected concentration conditions.
Text or footnotes should accompany this table explaining the basis of each
assumption.
A qualitative review of uncertainties should include a qualitative
statement of the impact of pertinent assumptions in light of reasonable
variations that could be encountered for actual concentration conditions.
If possible, special emphasis should be placed on evaluating extremes in
each assumption. For example, have parameters been chosen to evaluate
extreme or average concentration conditions? Are the concentration esti-
mates likely to be overestimates, underestimates, or typical values? This
2-3

-------
type of analysis will transmit the level of confidence in the results to
the decision maker and will aid in determining future actions.
2.3 Quantitative Analysis
For detailed concentration estimates, uncertainty can be characterized
using one or several of the following methods: (1) sensitivity analysis
techniques, (2) mean and variance of concentration, (3) the Monto Carlo
simulation procedure, or (4) confidence interval for concentration charac-
teristics.
(1) Sensitivity Analysis
Sensitivity analysis is a technique that tests the sensitivity of an
output variable to the possible variation in the input variables of a
given model. The purpose of the sensitivity analysis is to identify the
influential input variables, and develop bounds on the model output (in
this case, concentration estimates). By identifying the influential input
variable(s), more resources can be directed to reduce their uncertainties
and hence reduce the output uncertainty. The sensitivity of the output
variable with respect to each input variable is computed and the sensitiv-
ities of all input variables are compared. When computing the sensitivity
with respect to a given input variable, all other input variables are held
fixed at their nominal or middle values.
A sensitivity analysis should be performed when limited information or
data are available about the input variables. Information on input data
can be in the form of either best estimates of the centralities of the
input variable or estimates of the ranges (minimum and maximum) of the
input variables. Accordingly, one of two types of sensitivity analysis
can be pursued: point sensitivity or range sensitivity.
Point sensitivity is the sensitivity of the output to a given input
variable at a given value of that input variable with the other input
variables held at best estimates of their centralities. Range sensitivity
analysis estimates the range of output that would result as individual
input variables are varied from their minimum to their maximum possible
2-4

-------
value with other input variables held at fixed values, e.g., their mid-
ranges.
Point Sensitivity
Assume that the general model
y = f(xj» X2>•••> *k)	(13-1)
yields the concentration y of an environmental contaminant given the
values of the input variables Xj, X2,..., x^. Define
e = Ay/y = the relative error in y due to
e^ = ax^/x^ = the relative error in x-, i = 1 to k.
The total differential of y with respect to Xj, X2>..., x^ is
Ay = Ei = 1k(af/axi) Axi?
from which
e = Si = 1ksiei,	(13-2)
where s^ = (x^/f)(af/ax^), and af/ax^ is the usual partial derivative of f
with respect to x^. The quantities s^, i = 1 to k, are termed the first-
order sensitivity coefficients of y with respect to x-, i = 1 to k.
Equation 13-2 states that the relative error in concentration, e, is
determined by the relative errors or uncertainties in the input variables,
e^, multiplied by the associated sensitivity coefficients, s^. Even
if the input errors ei are small, the overall output error e may still
be large if the sensitivity coefficients s^ are large. In general, the
sensitivity coefficients will be a function of the input variables, x^.
That is, the values of the sensitivity coefficients may change as the
values of the input variables change. To compute representative values
of the sensitivity coefficients, one might take an average value over the
range of the input variables. Alternatively, one can evaluate the sensi-
tivity coefficients at typical or mid-range values of the input variables.
Example 1. Consider the simple Gaussian plume model for ground level
air concentrations of a pollutant at a specified downwind distance from an
elevated pollutant emitting source:
2-5

-------
C = (KQ/xua)exp-1/2(H2/£72j
(13-3)
where
C = Ground level air concentration (10"l®g/m3)
K = 8.65 x 10' = normalization constant
Q = Emission rate (10"18g/sec)
x = 1000m = downwind distance
u = Source height wind speed (m/sec)
a = Atmospheric dispersion coefficient (m)
H = Effective source height (m)
To specify input variable uncertainties, assume:
Q... is lognormally distributed with mean 13.2 and standard deviation
11.5:Q~Lognormal (13.2, 11.5). Let the minimum and maximum values
of Q be those at the boundaries of the symmetric 95 percent
confidence interval surrounding the mode of the associated normal
distribution of mean 2.3 and standard deviation 0.75. The minimum,
median, and maximum values of Q are then 2.29, 10.0, and 43.4,
respectively.
u... is lognormally distributed with mean 6.11 and standard deviation
2.80:U~Lognormal (6.11, 2.80). Let the minimum and maximum values
of u be those at the boundaries of the symmetric 95 percent
confidence interval surrounding the mode of the associated normal
distribution of mean 1.715 and standard deviation 0.437. The
minimum, median, and maximum values of u are then 2.36, 5.55, and
13.1, respectively.
H... is normally distributed with mean 100 and standard deviation
10:H~Normal (100, 10). let the minimum and maximum values of H be
those at the boundaries of the symmetric 95 percent confidence
interval. The minimum, median, and maximum values of H are then
80.4, 100, and 119.6, respectively.
a... is equally likely to assume one of the three values 73, 120, and
200: a-Duniform (73, 120, 200). The three values represent the
x = 1,000 m values of the Briggs-modified Pasquill-Gifford dispersion
curves for stability classes C, B, and A, respectively.
With constant x = 1,000 m, the sensitivity coefficients of equation
13-3 are:
Sq = 1, su = -1, sH = -(H/a)1 = -(100/120)^ = -0.694,
Sa = (H/o)2-1 = (100/120)2-1 = -0.306, 73< a <200
= (100/92.5)2-l » 0.169, 73< a <120
= (100/156.2)2-l = -0.590, 120 < a <200
2-6

-------
The latter two central estimates of a appearing in the calculation
of sa were obtained by interpolation.
The central difference estimates for the input variable errors are:
eQ =	(43.4-2.29)/(2 x 10) = 2.05
eu =	(13.1-2.36)/(2 x 5.55) = 0.964
eH =	(119.6-80.4)/(2 x 100) = 0.196
ea =	(200-73)/(2 x 120) = 0.529, 73 
-------
8495H
Table 2-1. Relative Errors and Sensitivity Coefficients for the Gaussian Plume Concentration Model
of Examples 1 & 2: C = (8.65xl07Q/1000u<7)exp~1/2(H2/fZ)
Input variables
-18 3
Output concentrations. CflO q/m 1
Values
Variable name, symbol
Min. Med. Max
Relative
error
=	Sensitivity
(Max - Min)/ coefficient
(2 Med.)	sj
Error
component
siei
Values
Lo Mid Hi
Relat ive
error
(Hi - Lo)/
(2 Mid)
-18
Emission rate. Q (10 g/sec)
2.29
10
43.4
2.05
1
2.05
210
917
3980
2.05
Source height wind speed, u (m/sec)
2.36
5.55
13.1
0.964
-1
-0.964
2160
917
390
-0.964
Effective source height, H (m)
80.4
100
119.6
0.196
-0.694
-0.136
1037
917
790
-0.135
Atmospheric dispersion coefficient, a (m)
73
120
200
0.529
-0.306
-0.162
835
917
687
-0.081

73
120

0.254
0.169
0.043
835
917

0.047


120
200
0.256
-0.590
-0.151

917
687
-0.143

-------
maximum possible values with the other input variables held at fixed
values, e.g., their midranges.
Example 2: consider the Gaussian plume air dispersion model of
Example 1. The x = 1000m range sensitivies of the model are given by
CLo - C(Q = Qm1n) = (8.65 x 104 x 2.29/5.55 x 120) exp~1/2 (1002/1202) = 210
CMid E C(Q = Qmed) = (8'65 x 104 x 10/5-55 x 12°) exp"1/2 (1002/1202) = 917
CU4 - C(Q = Qmav) = (8.65 x 104 x 43.4/5.55 x 120) exp"1/2 (1002/1202) = 3980
n 1	ilia X
Range sensitivity of C with respect to Q is (3980-210)/917 = 4.11
CLo - C(u = umin) = (8.65 x 104 x 10/2.36 x 120) exp"1/2 (1002/1202) = 2160
CMid B C(u = umed} = (8'65 x 104 x 10/5-55 x 12°) exp~1/2 (1002/1202) = 917
CHi - C(u = umax) = (8.65 x 104 x 10/13.1 x 120) exp"1/2 (1002/1202) = 390
Range sensitivity of C with respect to u is (390-2160)/917 = -1.93
CLq - C(H = Hmin) = (8.65 x 104 x 10/5.55	x	120) exp"1/2 (80.42/1202) = 1040
CMid - C(H = Hme(J) = (8.65 x 104 x 10/5.55	x	120) exp"1/2 (1002/1202) = 917
CHi = C(H = Hmax) = (8.65 x 104 x 10/5.55	x	120) exp"1/2 (119.62/1202) = 790
Range sensitivity of C with respect to	H	is (790-1040)/917 = -0.27
CLq b C(a = amin) = (8.65	x 104 x 10/5.55	x	73) exp"1/2 (1002/732) = 835
CMid s C(a = ame(J) = (8.65	x 104 x 10/5.55	x 120) exp~1/2 (1002/1202) = 917
CHi . C(a = amax) = (8.65	x 104 x 10/5.55	x 200) exp"1/2 (1002/2002) = 687
Range sensitivity of C	with respect to	a is (687-835)/917 = -0.16
The central difference output errors associated with these ranges of
values can be estimated from
(3980-210)/(2 x 917)	= 2.05
(390-2160)/(2 x 917)	= -0.964
(790-1037)/(2 x 917)	= -0.135
(687-835)/(2 x 917)	= -0.081, 73 
-------
(2)	Mean and Variance of the Output Variable for Nonadditive Models
The uncertainty associated with a given input variable or concentra-
tion estimate can also be expressed in terms of measures of centrality
(arithmetic means, geometric means, medians) and measures of variability
(variances, covariances) associated with multiple determinations. These
statistics can be determined directly for an actual sample or simulated
sample. Calculation of these common measures of quantitative variability
is addressed in standard statistic textbooks and is not discussed here.
(3)	Monte Carlo Simulation Method
In the Monte Carlo approach, the distribution of the concentration
output y is simulated directly from the known or assumed probability dis-
tributions of the input variables. The first step is to assign a joint
probability distribution type (e.g., normal, lognormal) to the input
variables Xj, ..., x^ of the concentration estimation algorithm. Most
often the input variables are considered independent, but independence is
not an absolute requirement for using the Monte Carlo method. Next, a
large number (n>=l,000 iterations) of independent samples (x^, X2^, •••>
Xkj) i - 1, 2, ..., n from the assigned joint distribution are simulated
(usually employing a computer) and the corresponding outputs y-, i =1,
..., n are calculated. This is accomplished by repeated computer runs
through the concentration estimation algorithm using random numbers to
assign values to the input parameters based on their known or suspected
ranges, means and probability distributions). The resulting output repre-
sents a sample from the true output distribution.
Methods of statistical inference are used to estimate, from the con-
centration output sample, some parameters of the concentration distribu-
tion such as percentiles, mean, variance or to test a hypothesized proba-
bility distribution of output. Confidence intervals for parameters of the
output distribution can also be calculated using the simulated output
data.
When a specific probability distribution is used to express uncertain-
ty in one or more input factors of an concentration model, a distribution
2-10

-------
of concentration is generated. The decision maker may investigate the
effects of using different probability distributions for input variables
and/or the effects of using different functional forms of the concentra-
tion estimation algorithm on the type of concentration distribution out-
put. The Monte Carlo simulation procedure yields a concentration distri-
bution which is strictly a consequence of the assumed distributions of the
model inputs and the assumed functional form of the model.
The Monte Carlo method is also commonly used approach to derive the
probability distribution of output for other models. If more than one
functional form of the model can be used to express the relationship of
the exposure with the input variables, then the Monte Carlo method would
be used to study the consequence of using each functional form on the con-
centration probability distribution; the model which produces a probabili-
ty distribution with smaller variance (smaller uncertainty in the output)
may be chosen for further use.
In general, the selection of a probability distribution to represent
an input factor in a concentration model should be based upon any gathered
information about that factor, theoretical arguments, and/or expert opin-
ion. A probability distribution can be ascertained from such information
as the following: the general type of the distribution, minimum, maximum,
mode, mean, medium, midrange and other percentiles. A matching between
the properties of these distributions and the properties of input factors
would justify the use of a particular distribution. The following consid-
erations are relevant to the process of selection of a distribution:
•	A uniform distribution would be used to represent a factor when
nothing is known about the factor except its finite range. The use
of a uniform distribution assumes that all possible values within
the range are equally likely.
•	If the range of the factor and its mode are known, then a
triangular distribution would be used.
•	If the factor has a finite range of possible values and a smooth
probability function is desired, a Beta distribution (scaled to the
desired range) may be most appropriate. The Beta distribution can
be fit from the mode and the range that defines a middle specific
percent (e.g. 95 percent) of the distribution.
2-11

-------
•	If the factor only assumes positive values, then a Gamma,
Lognormal or a Weibull distribution may be an appropriate choice.
The Gamma distribution is probably the most flexible of these; its
probability function can assume a variety of shapes by varying its
parameters, and it is mathematically tractable. These distribu-
tions also can be fit from the knowledge of the mode and the
percentiles that capture the middle 95 percent of the distribution.
•	If the factor has an unrestricted range of possible values and is
symmetrically distributed around its mode, then a normal distribu-
tion may be an appropriate distribution.
Once the probability distributions of all concentration factors are
developed or assigned, the Monte Carlo simulation can be performed.
Normally, this simulation is time-consuming and requires the use of a
computer. Numerous commercial personal computer based programs are avail-
able to perform the simulation. A simulation can be performed in little
more than the time required to enter the algorithm, range, and probability
distribution of each input value. The output of the simulation is a
frequency distribution and cumulative frequency distribution from which
the mean, median, variance, and percentile concentration levels can be
extracted.
The following is an example demonstration of the use of the Monte
Carlo technique to simulate the probability distribution of a given
concentration. The analysis was performed using commercially available
simulation software.
Example 3. Consider yet again the Gaussian model of Example 1.
Assume the following independent probability distributions for Q, H, and
a'.
Q - Lognormal (13.2, 11.5)
H - Normal (100, 10)
a - Duniform (73, 120, 200)
For illustrative purposes add the further complexity that the source
height wind speed is determined from a reference height wind speed and
actual source height according to
2-12

-------
u = u0(h/z0)P
where
u = source height wind speed (m/sec)
u0 = reference height wind speed (m/sec)
h = 50m = actual source height
z0 = 10m = reference height
p = wind profile power law exponent
The distribution of uQ is assumed to be Lognormal (4.80, 2.20). The
distribution of p is assumed to be Duniform (0.1, 0.15, 0.2) and perfectly
inversely correlated with the distribution of a; that is, p = 0.2 if
a = 73, p = 0.15 if a = 120, and p = 0.1 if a = 200. If p = 0.15, the
resulting distribution of u is Lognormal (6.11, 2.80), the same as in
Example 1.
The mean, range, and half-decade percentile values for the output
concentration resulting from an n = 1000 iteration Monte Carlo simulation
of the model are as follows:
0% .
.. 55
35% ..
565
70% .
. 1242
5% .
.. 188
40% ..
648
75% .
. 1390
10% .
.. 268
45% ..
722
80% .
. 1616
15% .
.. 319
50% ..
808
85% .
. 2000
20% .
.. 393
55% ..
884
90% .
. 2522
25% .
.. 439
60% ..
972
95% .
. 3492
30% .
.. 504
65% ..
1101
100% .
. 11674
mean ... 1178 range ... 11619
The variability in simulated concentrations is large and caused primarily
by the large variability in the emission rate which is, in effect, assumed
to be known only to within about two orders of magnitude. The simulated
concentrations are approximately lognormally distributed with an associ-
ated normal distribution of mean 6.68 and standard diviation of 0.78.
Assume now that the emission rate is constant and equal to 10; that
is, Q - Normal (10, 0). The mean, range, and half-decade percentile
values for the output concentration are then ( n = 1000, still):
2-13

-------
0 % .
. 157
35% ..
663
70% ..
1018
5% .
. 385
40% ..
709
75% ..
1109
10% .
. 438
45% ..
763
80% ..
1192
15% .
. 495
50% ..
816
85% ..
1308
20% .
. 548
55% ..
862
90% ..
1435
25% .
. 593
60% ..
910
95% ..
1690
30% .
. 626
65% ..
967
100% ..
3364

mean .
. 895
range .
. 3207

The spread is simulated concentrations is now much smaller than previous-
ly. The simulated concentrations are approximately lognormally distrib-
uted with an associated normal distribution of mean 6.69, the same as
previously, and standard deviation 0.41, about one-half as much as
previously.
Example 4. In Example 3 the downwind distance probability distribu-
tion was specified as an input. In this example we simulate the
distribution of downwind distances for which the concentration given by
Equation 13-3 is a maximum. It can be shown that this distance is
approximated by
xra - J(H2/2c2)	(13-4)
where
xm = Downwind distance to maximum concentration (m),
H - Normal (100,10),
c - Duniform (0.065,0.073,0.085,0.100,0.120,0.143,0.170,0.200,0.234),
and the distribution of c is correlated with the distribution of a
such that
c - Duniform (0.065,0.073,0.085) if a = 73,
Duniform (0.100,0.120,0.143) if a = 120,
Duniform (0.170,0.200,0.234) if a = 200.
The maximum concentration (at x = xm) is determined from Equation 13-3
with x given by Equation 13-4 and
a = 0.073xm if 0.065 < c < 0.085,
= 0.120xm if 0.100 =s c =s 0.143,
= 0.200xm if 0.170 < c < 0.234.
2-14

-------
The mean, range, and half-decade percentile values for the distance
to maximum concentration resulting from a n = 1000 iteration Monte Carlo
simulation of Equation 13-4 are:
0% .
. 233
35% ..
469
70% .
. 800
5% .
. 243
40% ..
505
75% .
. 856
10% .
. 318
45% ..
539
80% .
. 908
15% .
. 345
50% ..
588
85% .
. 970
20% .
. 368
55% ..
635
90% .
. 1032
25% .
. 400
60% ..
688
95% .
. 1137
30% .
. 429
65% ..
740
100% .
. 1338

mean .
. 639
range .
. 1105

The 50th percentile value for xm, 588, is virtually the same as that
computed from Equation 13-4 using the input variable central values
H = 100 and c = 0.12:
J(1002/2x0.122) = 589.
The Monte Carlo simulated values for the distance to maximum concentration
are approximately lognormally distributed with an associated normal dis-
tribution of mean 6.35 and standard deviation 0.49.
The mean, range, and half-decade percentile values for the maximum
concentration (at x = xm) resulting from a n = 1000 iteration Monte
simulation of Equation
13-3 are:



0% .
. 191
35% ...
1036
70% .
. 2017
5% .
. 423
40% ...
1139
75% .
. 2200
10% .
. 537
45% ...
1235
80% .
. 2432
15% .
. 647
50% ...
1353
85% .
. 2733
20% .
. 750
55% ...
1479
90% .
. 3175
25% .
. 835
60% ...
1608
95% .
. 3993
30% .
. 933
65% ...
1794
100% .
. 9397

mean ..
. 1693
range .
. 9206

The 50th percentile value for the maximum concentration, 1353, is approxi-
mately equal to the concentration computed from Equation 13-3 using the
input parameter values xm = 589, Q = 10, u = 5.55, a = 0.12x589 =
70.7, H - 100:
(8.65xl08/589 x 5.55 x 70.7)exp'1/2(i0o2/70.72) = 1375.
2-15

-------
The simulated values for the maximum concentration are approximately
lognormally distributed with an associated normal distribution of mean
7.20 and standard deviation 0.62.
(4) Confidence Intervals for some Characteristics of the
Concentration Distribution
A major reduction in the uncertainty associated with a concentration
estimation can be achieved by obtaining confidence intervals for some
characteristics of the concentration distributions such as concentration
percentiles, mean concentration or the range of the concentration distri-
bution. This approach of "statistical inference" can be made on the
basis of the results obtained from a sample drawn from the population of
concentrations. The sample information about concentration can be
obtained indirectly (e.g., via Monte Carlo simulation; see (3) above) or
directly. The population may be thought of as a set of N subjects that
each experience their individual values of the model input variables and
resulting concentration. If the functional relationship (model) of con-
centration output to its input factors is validated then it may be much
easier and less costly to collect data on the model input variables,
rather than the concentration itself. Having collected the input variable
sample data, the concentration model can be used to compute the predicted
concentration for each sampled member. Methodologies for calculation of
confidence intervals can be found in standard statistics textbooks and,
therefore, are not presented here.
2.4 Presentation of Uncertainty Analysis Results
Comprehensive qualitative analysis and rigorous quantitative analysis
are of little value if the analysis results are not clearly presented for
use in the decision-making process. The following procedures can be used
to present the results of the uncertainty analysis:
• Concentration estimates should clearly identify the input parame-
ters to which they are most sensitive, including a quantitative
statement of the magnitude of sensitivity (e.g., percent variability
of the output across the range of possible input values). Where
Monte Carlo simulations are performed, concentration estimates could
2-16

-------
be presented as percentile intervals (or cumulative probability per-
centile intervals). For each interval, the associated concentration
conditions should be described.
•	Each concentration estimate should conclude with statements of
qualitative and quantitative uncertainties, including the major
data gaps or other causes of uncertainty. This information can be
presented in tabular format.
•	Uncertainties for the overall concentration estimation should be
summarized and presented as a section that qualifies all other re-
port sections or, alternatively, the results could be incorporated
into an Executive Summary.
An example of an uncertainty analysis report that would ideally be
associated with a hypothetical concentration estimation is shown in
Table 2-2.
2-17

-------
8495H
Table 2-2. Uncertainty Analysis Report for Example 3 Concentration Estimation
Concentration model: Ground level air concentration from Gaussian model
C= (KQ/xu
-------
8495H
Table 2-2. (continued)
Actual source height -- very high confidence based upon easily performed and very accurate measurement tech-
niques.
Wind profile exponent -- medium confidence in values attaching to uncertainty in atmospheric stability class
applicable to site-specific conditions.
Source height wind speed -- same confidence as for reference height wind speed. Uncertain but small values of
the wind profile exponent cause little change in the confidence between the reference and source height wind
speeds.
Atmospheric dispersion coefficient -- same confidence as for wind profile exponent and for the same reasons.
Effective source height -- medium/high confidence because of additional uncertainty introduced by plume rise
above actual source height.
Concentration -- medium confidence in Gaussian model. The model may not accurately reflect certain site-
specific conditions.
(2) Sensitivity Analysis:
Indicates that estimated concentration is most sensitive to the emission rate and source height wind
speed, less sensitive to the effective source height, and least sensitive to the atmospheric dispersion
coefficient. The point and range sensitivities of concentration with respect to the four input variables are:
Point	Range
Sensitivity	Sensitivity
(% change concentration/	(% change
Variable % change input variable)	concentration)
Emission rate 1	411
Source height wind speed	-1	-193
Effective source height	-0.694	- 27
Atmospheric dispersion coefficient	-0.306	- 16
(3) Probabilistic Analysis:
Monte Carlo simulated median, mean, range, and selected percentile concentrations are:
OX...55	25%...439	90%...2,522
5%...188	50%...808	95%...3,492
10%...'268	75%...1,390	100%.. .11,674
median...808	mean...1,178 range...11,619
The simulated concentrations are lognormally distributed with an associated normal distribution of mean
6.68 and standard deviation 0.78.
2-19

-------
3.	REFERENCES
Bartha R, You IS, Saxena A. 1982. Humus-bound residues of phenylamide
herbicides: their nature, persistence and monitoring. In: Pesticide
Chemistry. Vol. 3. Mode of Action, Metabolism and Toxicology, J.
Miyamoto and P.C. Kearney, eds. New York: Pergamon Press, pp 345-350.
Bear J, Verruijt A. 1987. Modeling groundwater flow and pollution.
Dordrecht, The Netherlands: D. Reidel Publishing Company. As cited in
USEPA (1988b.)
Beljin MS. 1985. A program package of analytical models for solute
transport in groundwater 'SOLUTE'. Indianapolis, IN: International
Ground Water Modeling Center, Holcomb Research Institute, Butler
University. As cited in USEPA (1988b).
Berglund R. 1984. Personal communication to G.W. Dawson. National
Cattleman's Association, Denver, CO.
Beveridge A, Pickering WF. 1980. Influence of humate-solute interactions
on aqueous heavy metal ion levels. Water Air Soil Pollut. 14:171-185.
Bitton G, Gerba CP. 1984. Ground-water pollution microbiology.
New York: John Wiley and Sons.
Bonazountas M, Fiksel J, et al. 1982. Environmental mathematical
pollutant fate modeling handbook/catalogue. Draft. Washington, DC:
U.S. Environmental Protection Agency, Office of Policy and Resource
Management. EPA Contract No. 68-01-5146. As cited in USEPA (1988b).
Bonazountas M, Wagner J. 1981. SESOIL, a seasonal soil compartment
model. Cambridge, MA: A.D. Little Inc., for U.S. Environmental
Protection Agency. Contract No. 68-01-6271. As cited in USEPA (1988b).
Bouwer H. 1978. Groundwater hydrology. New York: McGraw-Hill Pub.
Co. As cited in USEPA (1988b).
Bowers JF, et al. 1979. Industrial source complex (ISC) dispersion model
user's guide, volumes I and II. Washington, DC: U.S. Environmental
Protection Agency. EPA 450/4-79-030. As reviewed in: Bonazountas M,
Fiksel J, et al. 1982. Environmental mathematical pollutant fate
modeling handbook/catalog. Draft. Washington, D.C.: U.S. Environmental
Protection Agency, Office of Policy and Resource Management. Contract
No. 68-01-5146. As cited in USEPA (1988b).
Brown DS, Allison JD. 1987. MINTEQ1, equilibrium metals speciation
model: A user's manual. Athens, GA: U.S. Environmental Protection
Agency. As cited in USEPA (1988b).
3-1

-------
Brown SM, Boutwell SH, Roberts BR. 1983. Selection and use of models
for remedial action evaluation at uncontrolled hazardous waste sites.
Palo Alto, CA: Anderson-Nichols & Co., Inc. Draft report. EPA Contract
No. 68-03-3116. As cited in USEPA (1988b).
Bureau of the Census. 1986. Current population reports. Series P-25,
No. 1000. Washington, DC: Department of Commerce.
Burt E. 1977. VALLEY model user's guide. Washington, DC: U.S.
Environmental Protection Agency. EPA-450/2-77-018. As reviewed in:
U.S. EPA. 1982. Environmental Modeling Catalogue. Washington, DC:
U.S. Environmental Protection Agency. Available from Information
Clearing House as PM-211A. As cited in USEPA (1988b).
Busse AD, Zimmerman JR. 1976. User's guide for the climatological
dispersion model. Research Triangle Park, NC: U.S. Environmental
Protection Agency. EPA-R4-73-024. As reviewed in: Bonazountas M,
Fiksel J, et al. 1982. Environmental mathematical pollutant fate
modeling handbook/catalogue. Draft. Washington, DC: U.S. Environmental
Protection Agency, Office of Policy and Resource Management. EPA
Contract No. 68-01-5146. As cited in USEPA (1988b).
Callahan MA, Slimak MW, Gabel NW, May IP, Fowler CF et al. 1979. Water-
related environmental fate of 129 priority pollutants. Washington, DC:
U.S. Environmental Protection Agency, Office of Water Planning and
Standards. EPA 440/4-79-029ab.
Calvert JG, Pitts JN. 1967. Photochemistry. New York: John Wiley.
Campbell JH, Evans RD. 1987. Inorganic and organic ligand binding of
lead and cadmium and resultant implications for bioavailability. Sci.
Total Environ. 62:219-227.
Carsel RF, Smith CN, Mulkey LA, Dean JD, Jowise P. 1984. User's manual
for the pesticide root zone model (PRZM): release 1. Athens, GA:
Environmental Research Laboratory, U.S. Environmental Protection Agency,
EPA-600/3-84-109. As cited in USEPA (1988b).
Chico T, Catalano JA. 1986. Addendum to the user's guide for MPTER.
Research Triangle Park, NC: U.S. Environmental Protection Agency. As
cited in USEPA (1988b).
Christ RA. 1978. Physiological and physicochemical requisites for the
transport of xenobiotics in plants. In: Advances in pesticide science,
H. Geissbuhler, ed. New York: Pergamon.
3-2

-------
Christiansen JH. 1976. Design and application of the Texas episodic
model. Proceedings of the Conference on Environmental Modeling and
Simulation. Washington, DC: U.S. Environmental Protection Agency.
EPA-600/9-76-016. As reviewed in: USEPA (1982). Environmental Modeling
Catalogue. Washington, DC: U.S. Environmental Protection Agency.
Available from Information Clearing House as PM-211A. As cited in USEPA
(1988b).
Clapp RB, Hornberger GM. 1978. Empirical equations for some soil
hydraulic properties. Water Resources Research 14:601-604. As cited in
USEPA (1988b).
Codell RB, Key KT, Whelan G. 1982. A collection of mathematical models
for dispersion in surface water and ground water. Washington, DC: U.S.
Nuclear Regulatory Commission. NUREG/CR-0868. As cited in USEPA (1988b).
Cox DC, Baybutt P. 1981. Methods for uncertainty analysis: a
comparative survey. Risk Analysis 1 (4):251-258.
Davenport J, Mill T. 1985. Summary report: structure activity
relationships. Data generation and development of structure/reactivity
relationships for environmental fate processes/properties. EPA Contract
68-02-3968.
Davis LA, Segol G. 1985. Documentation and user's guide: GS2 and
GS3-variably saturated flow and mass transport models. Washington,
D.C.: U.S. Nuclear Regulatory Commission. NUREG/CR-3901. As cited in
USEPA (1988b).
Dawson GW, English CJ, Petty SE. 1980. Physical chemical properties of
hazardous waste constituents. Athens, GA: U.S. Environmental Protection
Agency, Environmental Research Laboratory. As cited in USEPA (1988b).
Delos CG, Richardson WL, DePinto JV, et al. 1984. Technical guidance
manual for performing wasteload allocations, Book II: streams and rivers.
Draft final report. Washington, DC: U.S. Environmental Protection
Agency, Office of Water Regulations and Standards, Water Quality Analysis
Branch. As cited in USEPA (1988b).
Dilling WL, Tefertiller NB, Kallos GJ. 1975. Evaporation rates of
methylene chloride, chloroform, 1,1,1-trichloroethane, trichloroethylene,
tetrachloroethylene, and other chlorinated compounds in dilute aqueous
solutions. Environ. Sci. Technol. 9:833-838. As cited in Callahan et
al. (1979).
Domenico PA, Palciauskas VV. 1982. Alternative boundaries in solid waste
management. Ground Water 20(3):303. As cited in USEPA (1988b).
3-3

-------
Enfield CG, Carsel RF, Cohen SZ, Phan T, Walters DM. 1982. Approximat!
pollutant transport to ground water. Ground Water 20(6): 711-722. As
cited in USEPA (1988b).
Enfield CG. 1984. Chemical transport facilitated by multiphase flow
systems. Presented at seminar on Degradation, Retention and Dispersion
of Pollutants in Ground Water, Copenhagen, Denmark, September 12-14.
Ada, OK: U.S. Environmental Protection Agency, Robert S. Kerr.
Environmental Research Laboratory. As cited in USEPA (J988b).
Farino W, Spawn P, Jasinski M, Murphy B. 1983. Evaluation and selection
of models for estimating air emissions from hazardous waste treatment,
storage and disposal facilities. Revised draft final report prepared for
U.S. Environmental Protection Agency, Office of Solid Waste, Land
Disposal Branch. Contract No. 68-02-3168. Bedford, Massachusetts: GCA
Corporation, GCA/Technology Division. As cited in USEPA (1988b).
Farmer WJ, Yang MS, Letey J, Spencer WF, Roulier MH. 1978. Land disposal
of hexachlorobenzene waste: controlling vapor movement in soils. San
Antonio, TX. 4th Annual Symposium on Land Disposal. As cited in USEPA
(1988b).
Freeze RA, Cherry JA. 1979. Groundwater. Englewood Cliffs, NJ:
Prentice-Hall, Inc. As cited in USEPA (1988b).
Fries GF, Marrow GS Jr, Gordon CH. 1973. Long-term studies of residue^
retention and excretion by cows fed a polychlorinated biphenyl (Arochlol
1254). J. Agric. Food Chem. 21:117-120.
Fries GF. 1986. Assessment of potential residues in foods derived from
animals exposed to TCDD-contaminated soil. Presented at Dioxin 86: 6th
International Symposium on Chlorinated Dioxins and Related Compounds,
September, Fukuoka, Japan.
Fuhrman TW, Lichtenstein EP. 1978. J. Agric. Food Chem. 26:605-610.
GSC. 1982. Graphical exposure modeling system (GEMS) user's guide.
General Software Corporation. Washington, DC: U.S. Environmental
Protection Agency, Office of Pesticides and Toxic Substances. Contract
No. 68-01-6618. As cited in USEPA (1988b).
Garten CT, Trabalka JR. 1983. Evaluation of models for predicting
terrestrial food chain behavior of xenobiotics. Environ. Sci. Teochnol.
17:590-595.
Gelhar LW, Mantoglou A, Welty C, and Rehfeldt KR. 1985. A review of
field-scale physical solute transport processes in saturated and
unsaturated porous media. Palo Alto, CA: Electric Power Research
Institute. EPRI EA-190. As cited in USEPA (1988b).
3-4

-------
Geotrans. 1986. Analytical models for evaluating leachate migration in
groundwater systems. Washington, DC: U.S. Environmental Protection
Agency, Office of Solid Waste. As cited in USEPA (1988b).
Giesy JP, Alberts JJ. 1984. Trace metal speciation: the interaction of
metals with organic constituents of surface waters. In: Workshop
Proceedings. The Effects of Trace Elements on Aquatic Ecosystems. Palo
Alto, CA: Electric Power Research Institute. EA-3329.
Gillett JW. 1980. Terrestrial microcosm technology in assessing fate,
transport and effects of toxic chemicals, in Dynamics, Exposure and
Hazard Assessment of Toxic Chemicals, R. Hague, editor. Ann Arbor
Science, Ann Arbor, MI. As cited in Trabalka and Garten (1982).
Gupta SK, Cole CR, Kincaid CT, Monti AM. 1987. Coupled fluid, energy,
and solute transport (CFEST) model: formulation and user's manual.
Columbus, OH: Battelle Memorial Institute, Office of Nuclear Waste
Isolation. As cited in USEPA (1988b).
Gureghian AB. 1983. TRIPM: a two-dimensional finite-element model for
the simultaneous transport of water and reacting solutes through saturated
and unsaturated porous media. Columbia, OH: Battelle Memorial Institute,
Office of Nuclear Waste Isolation. BMI/ONWI-465. As cited in USEPA
(1988b).
Haith DA. 1980. A mathematical model for estimating pesticide losses in
runoff. J. Environ. Quality 9(3):428-433. As cited in USEPA (1988b).
Hamelink JL, Spacie A. 1977. Fish and chemicals: The process of
accumulation. Ann. Rev. Pharmacol. Toxicol. 17:167-177.
Hamelink JL, Waybrant RC, Ball RC. 1971. A proposal: Exchange
equilibria control the degree chlorinated hydrocarbons are biologically
magnified in lentic environments. Trans. Am. Fish. Soc. 100:207-214.
Hamelink JL. 1980. Bioavailability of chemicals in aquatic environments.
In: Biotransformation and Fate of Chemicals in the Environment. A.W.
Maki, K.L. Dickson and J. Cairns, Jr., eds. Washington, DC: Am. Soc.
Microbiol.
Hampson RF Jr., Garvin D. 1978. Reaction rate and photochemical data
for atmospheric chemistry-1977. Gaithersburg, MD: National Bureau of
Standards. NBS Special Publication 513.
Hansch C, Leo A. 1979. Substituent constants for correlation analysis
in chemistry and biology. New York,: John Wiley and Sons.
Hawley J. 1985. Assessment of health risk from exposure to contaminated
soil. Risk Analysis 5:289.
3-5

-------
Helling CS, Krivonak AE. 1978. J. Agric. Food Chem. 26:1164-1172.
Hendry D6, Kenley R. 1979. Atmospheric reaction products of organic
compounds. EPA Final Report. EPA Publication No. 560/12-79-001. As
cited in USEPA (1988b).
Hendry DG, Mill T, Mabey W. 1979. Test protocols for environmental
processes: Oxidation and photolysis in air. EPA draft report.
Washington, DC: U.S. Environmental Protection Agency. EPA Contract No.
68-03-2227.
Henrion M. 1985. Issues in uncertainty analysis. Paper presented at
Versar Inc., Springfield, VA, December 10, 1985. Pittsburgh, PA:
Carnegie-Mellon University, Department of Engineering and Public Policy.
Hung Y.W. 1982. Effects of temperature and chelating agents on cadmium
uptake in the American oyster. Bull. Environ. Contam. Toxicol.
28:546-551.
Huyakorn PS, Ungs MJ, Mulkey LA, Sudicky SA. 1987. A three-dimensional
analytical method for predicting leachate migration. Ground water, Vol.
25, No. 5. As cited in USEPA (1988b).
Hwang ST. 1982. Toxic emissions from land disposal facilities.
Environ. Progress 1 (l):46-52. As cited in USEPA (1988b).
Intera. 1983. SWENT: A three-dimensional finite-difference code for the
simulation of fluid, energy, and solute radionuclide transport. Intera,
Inc. Columbus, OH: Battelle Memorial Institute, Office of Nuclear Waste
Isolation. BMI/ONWI-457. As cited in USEPA (1988b).
Jackson G, Morgan J. 1978. Trace metal-chelator interactions and
phytoplankton growth in seawater media: theoretical analysis and
comparisons with reported observations. Limnol. Oceonogr. 23:268-282.
Javandel I, Doughty C, Tsang CF. 1984. Groundwater transport: handbook
of mathematical models. AGU Water Resources Monograph 10:240. As cited
in USEPA (1988b).
Jenson ME, ed. 1973. Consumptive use of water and irrigation water
requirements. New York, NY: American Society of Civil Engineers. As
presented by Enfield et al. 1982. Approximating pollutant transport to
ground water. Ground Water 20(6): 711-722. As cited in USEPA (1988b).
Karickhoff SW, Brown DS, Scott TA. 1979. Sorption of hydrophobic
pollutants on natural sediments. Water Res. 13:241-248.
3-6

-------
Karickhoff SW, Morris KR. 1985. Pollutant sorption: relationship to
bioavailability. Athens, GA: U.S. Environmental Protection Agency,
Office of Research and Development, Environmental Research Laboratory.
EPA 600/D-85/104.
Kenaga EE. 1980a. Predicted bioconcentration factors and soil sorption
coefficients of pesticides and other chemicals. Ecotoxicol. Environ.
Safety 4:26-38.
Kenaga EE. 1980b. Correlations of bioconcentration factors of chemicals
in aquatic and terrestrial organisms with their physical and chemical
properties. Environ. Sci. Technol. 14:553-556.
Kent DC, Pettyjohn WA, Prickett TA. 1985. Analytical methods for the
prediction of leachate plume migration. Ground Water Monitoring Review.
Spring 1985. Vol. 5 #2 pp. 46-59. As cited in USEPA (1988b).
Kent DC, Pettyjohn WA, Witz F, Prickett T. 1982. Prediction of leachate
plume migration and mixing in ground water. Solid and Hazardous Water
Research and Development Annual Symposium proceedings. Columbus, OH:
National Water Well Association. As reviewed in Versar Inc. (1983).
Theoretical evaluation of sites located in the zone of saturation. Draft
final report. Chicago, IL: U.S. Environmental Protection Agency.
Contract No. 68-01-6438. As cited in USEPA (1988b).
Kinniburgh DG. 1986. General purpose adsorption isotherms. Environ.
Sci. Technol. 20:895-904.
Kinzelback W. 1986. Groundwater modeling: An introduction with sample
programs in BASIC. Amsterdam, The Netherlands: Elsevier Science
Publishers. As cited in USEPA (1988b).
Kipp KL Jr. 1987. HST3D: A computer code for simulation of heat and
solute transport in three-dimensional groundwater flow systems. Lakewood,
CO: U.S. Geological Survey. WRI 86-4095. As cited in USEPA (1988b).
Kloke A, Sauerbeck DR, Vetter H. 1984. The contamination of plants and
soils with heavy metals and the transport of metals in terrestrial food
chains. In: Changing Metal Cycles and Human Health, J.C. Nriagu, ed.,
New York: Springer-Verlag.
Konikow LF, Bredehoeft JD. 1978. Computer model of two-dimensional
transport and dispersion in ground water. Techniques of water resource
investigation, Book 7, Chapter 2. Washington, DC: U.S. Geological
Survey. As reviewed in: Versar Inc. 1983. Theoretical evaluation of
sites located in the zone of saturation. Draft final report. Chicago,
IL: U.S. Environmental Protection Agency. Contract No. 68-01-6438. As
cited in USEPA (1988b).

-------
Kufs C, Repa E, Rogoshewski P, et al. 1983. Leachate plume migration
control. Unpublished draft. Cincinnati, OH: U.S. Environmental
Protection Agency. Contract No. 68-03-3113. As cited in USEPA (1988b).
Landrum PF, Nihart SR, Eadie BJ, Herche LR. 1987. Reduction in
bioavailability of organic contaminants to the amphipod Pontoporeia hovi
by dissolved organic matter of sediment interstitial waters. Environ.
Toxicol. Chem. 6:11-20.
Landrum PF, Reinhold MD, Nihart SR, Eadie BJ. 1985. Predicting the
bioavailability of organic xenobiotics to Pontoporeia hovi in the presence
of humic and fulvic materials and natural dissolved organic matter.
Environ. Toxicol. Chem. 4:459-467.
Leifer A, Brink RH, Thorn GC, Partymiller KG. 1983. Environmental
transport and transformation of polychlorinated biphenyls. Washington,
DC: U.S. Environmental Protection Agency, Office of Toxic Substances.
EPA 560/5-83-025.
Lindsay WL. 1979. Chemical equilibrium in soils. New York: John
Wiley & Sons.
Lyman WJ, Reehl WF, Rosenblatt DH. 1982. Handbook of Chemical Property
Estimation Methods. New York: McGraw-Hill Book Company.
Mabey WR, Mill T, Hendry DG. 1979. Test protocols for environmental
processes: Photolysis in water. EPA draft report. EPA Contract
68-03-2227.
Mabey WR, Mill T. 1978. Critical review of hydrolysis of organic
compounds in water under environmental conditions. J. Phys. Chem. Ref.
Data 7L383. As quoted in Callahan et al. (1979).
Mabey WR, Smith JH, Podoll RT, et al. 1982. Aquatic fate process data
for organic priority pollutants. Final draft report. Washington, DC:
U.S. Environmental Protection Agency, Office of Water Regulations and
Standards. As cited in USEPA (1988b).
Macalady DL, Tratnyek PG, Grundl TJ. 1986. Abiotic reduction reactions
of anthropogenic organic chemicals in anaerobic systems: A critical
review. J. Contain. Hydrol. 1:1-28.
Mackay D, Leinonen PJ. 1975. Rate of evaporation of low-solubility
contaminants form water bodies to atmosphere. Environ. Sci. Technol.
Vol. 9(13):1178-1180. As cited in USEPA (1988b).
Martin OF. 1970. Marine Chemistry, Vol. 2. New York: Marcel Dekker.
3-8

-------
Martinez MJ. 1985. A finite element computer program for simulating
radionuclide transport through porous media. Albuquerque, NM: Sandia
National Laboratory. SAND 84-0747. As cited in USEPA (1988b).
Mathur SP, Morley HV. 1975. Biodegradation approach for investigating
pesticide incorporation into soil humus. Soil Sci. 120:238-240.
Matile P. 1976. Plant cell: Substructures and subfunctions. 8.
Vacuoles. In: Plant Biochemistry, J.F. Bonner and J.E. Varner, eds.
New York: Academic Press.
McCarthy JF, Black MC. 1987. Partitioning between dissolved organic
macromolecules and suspended particulates: Effects on bioavailability
and transport of hydrophobic organic chemicals in aquatic systems. Tenth
Symposium on Aquatic Toxicology and Hazard Assessment. Philadelphia, PA:
Am. Soc. for Testing Materials.
McCarthy JF, Jimenez BD, Barbee T. 1985. Effect of dissolved humic
material on accumulation of polycyclic aromatic hydrocarbons: Structure-
activity relationships. Aquatic Toxicol. 7:15-24.
McCarthy JF, Jimenez BD. 1985. Reduction in bioavailability to bluegills
of polycyclic aromatic hydrocarbons bound to dissolved humic material.
Environ. Toxicol. Chem. 4:511-521.
Menzie CM. 1980. Metabolism of Pesticides: Update III. Special
Scientific Report - Wildlife No. 232. Office of Environmental Assistance,
U.S. Dept. of Interior, Fish and Wildlife Service, Washington, DC. As
cited in Trabalka and Garten (1982).
Mercer JW, Silka LR, Faust CR. 1983. Modeling ground-water flow at Love
Canal, New York. J. Environ. Eng. 109(4):924-942.
Mill T, Mabey WR, Hendry DG. 1979. Test protocols for environmental
processes: Oxidation in water. EPA Report (Draft). EPA Contract
68-03-2227.
Mill T. 1979. Structure reactivity conditions for environmental
reactions. EPA560/11-79-012. As cited in Callahan et al. (1979).
Miller FJ, Gardner DE, Graham JA, Lee RE Jr., Wilson WE, Bachmann JD.
1979. Size considerations for establishing a standard for inhalable
particles. J. Air Pollut. Control Assoc. 29(6):610. As cited in USEPA
(1988b).
Mills WB, Dean JD, Porcella DB, et al. 1982. Water quality assessment:
A screening procedure for toxic and conventional pollutants: Parts 1, 2,
and 3. Athens, GA: U.S. Environmental Protection Agency. Environment
Research Laboratory. Office of Research and Development. EPA Publication
No. 600/6-82/004a,b,c. As cited in USEPA (1988b).
3-9

-------
Mockus J. 1972. Estimation of direct runoff from storm rainfall. In
National engineering handbook. Section 4: hydrology. Washington, DC:
U.S. Department of Agriculture, Soil Conservation Service. As cited in
USEPA (1988b).
Morgan M, Henrion M, Nair I, Mull in T, Wiecha C. 1984. A generic
"pre-HEED" on characterizing and dealing with uncertainty in health and
environmental risk assessment. Pittsburgh, PA: Carnegie-Mellon
University, Department of Engineering and Public Policy.
NAS. 1977. National Academy of Sciences. Drinking water and health,
Volume 1. Washington, DC: National Research Council.
Ng YC, Colsher CS, Quinn DJ, Thompson, SE. 1977. Transfer coefficients
for the prediction of the dose to man via the forage-cow-milk
pathway from radionuclides released to the biosphere. Lawrence Livermore
National Laboratory. Report No. UCRL - 51939. U.S. Department of Energ)
Contract No. W-7405-Eng-48.
Ng YC, Colsher CS, Thompson SE. 1982. Transfer coefficients for
assessing the dose from radionuclides in meat and eggs. Final Report.
Lawrence Livermore National Laboratory. Washington, DC: U.S. Nuclear
Regulatory Commission. NUREG/CR-2976/UCID-19464.
Niimi AJ, Cho CY. 1981. Elimination of hexachlorobenzene (HCB) by
rainbow trout (Salmo aairdneri) and the examination of its kinetics in
Lake Ontario salmonids. Can. J. Fish. Aquat. Sci. 38:1350-1356.
Nkedi-Kizza P, Rao PSC, Hornsby AG. 1985. Influence of organic
cosolvents on sorption of hydrophobic organic chemicals by soils.
Environ. Sci. and Technol. 19 (10): 975-979. As cited in USEPA (1988b),
Novotny V, Chesters G. 1981. Handbook of nonpoint pollution, sources
and management. New York: Van Nostrand Reinhold Company
Oberlander HE, Roth K. 1982. Uptake and distribution of soil-applied
labelled heavy metals in cereal plants and products. FAO/IAEA Research
Coordination Meeting, "Agrochemical Biota Interactions in Soil and Water
Using Nuclear Techniques," Vienna: International Atomic Energy Agency.
Technical Document No. 283. As quoted in Kloke et al. 1984.
Parks GA. 1967. Aqueous surface chemistry of oxides and complex oxide
minerals. Equilibrium concepts in natural water systems. Washington,
DC: American Chemical Society, pp. 121-160. As quoted in Freeze and
Cherry (1979).
Perkampus et al. 1971. UV atlas of organic compounds. Vol. I-V. New
York: Plenum Press.
3-10

-------
Perry R, Chilton C. 1973. Chemical engineers handbook. 5th edition.
New York: McGraw-Hill. As cited in USEPA (1988b).
Phillips GR and Russo RC. 1978. Metal bioaccumulation in fishes and
aquatic invertebrates: A literature review. Environmental Research
Laboratory. Office of Research and Development. U.S. Environmental
Protection Agency. Duluth, Minnesota. EPA Publication No. 600/3-78-103.
Pierce D, Turner DB. 1980. User's guide for MPTER. Research Triangle
Park, NC: U.S. Environmental Protection Agency. EPA Publication
No. 600/8-80-016. NTIS Document No. PB-80-176361. As cited in USEPA
(1988b).
Price CE. 1978. Movement of xenobiotics in plants - perspectives. In:
Advances in Pesticide Science, H. Geissbuhler, ed. New York: Pergamon.
Prickett TA, Naymik TG, Lonnquist CG. 1981. A "random-walk" solute
transport model for selected groundwater quality evaluations. Champaign,
IL: Illinois Department of Energy and Natural Resources. ISWS/BUL-65/81.
As reviewed in: Versar Inc. 1983. Theoretical evaluation of sites
located in the zone of saturation. Draft final report. Chicago, IL:
U.S. Environmental Protection Agency. Contract No. 68-01-6438. As cited
in USEPA (1988b).
Public Voice. 1986. Public voice for food and health policy. The great
American fish scandal: Health risks unchecked. Washington, DC: Public
Voice.
Rao PSC, Hornsby AG, Kilcrease PP, Nkedi-Kizza P. 1985. Sorption and
transport of hydrophobic organic chemicals in aqueous and mixed solvent
systems: model development and preliminary evaluation. J. Environ.
Qual. 14 (3):376-383. As cited in USEPA (1988b).
Rawls WJ, Brakensiek DL, Saxton KE. 1982. Estimation of soil water
properties. Trans. Am. Soc. Agri. Eng. 25(5):1316-1320 and 1328. As
cited in USEPA (1988b).
Rawls WJ. 1986. Computer printouts from the soils data base, dated
August 28, 1986. From WJ Rawls. Beltsville, MD: Beltsville
Agricultural Research Center. As cited in USEPA (1988b).
Reeves M, Ward DS, Johns NJ, Cranwell RM. 1986. The Sandia waste-
isolation flow and transport model for fractured media; release 4.84:
Theory and implementation. Washington, DC: U.S. Nuclear Regulatory
Commission. NUREG/CR-3328. As cited in USEPA (1988b).
3-11

-------
Rifai HS, Bedient PB, Borden RC, Haasbeek JF. 1987. BIOPLUME II;
Computer model of two-dimensional contaminant transport under the
influence of oxygen limited biodegradation in ground water. Houston,
TX: National Center for Ground Water Research, Rice University. As
cited in USEPA (1988b).
Romero F, Elejalde C. 1985. The uptake of metallic elements by plants
from polluted soils. Toxicol. Environ. Chem. 10:247-255.
Runchal A, Sagar B, Baca RG, Kline NW. 1985. PORFLO - A continuum model
for fluid flow, heat transfer, and mass transport in porous media.
Richland, WA: Rockwell Hanford Operations. RHO-BW-CR-150P. As cited in
USEPA (1988b).
Russell RS, Bartlett BO, Bruce RS. 1971. The significance of long-lived
nuclides after a nuclear war. In: Survival of food crops and livestock
in the event of nuclear war. D.W. Bensen and A.H. Sparrow, eds. Oak
Ridge, TN: USAEC. As cited in USEPA (1988b) and Trabalka and Garten
(1982).
Salame M. (no date) Permeability-structure relationships of high
polymers. Obtained by private communication. Bloomfield, CT: Monsanto
Co. As cited in USEPA (1988b).
Salame M. 1961. The prediction of liquid permeation in polyethylene and
related polymers. SPE Trans. 1(4):153. As cited in USEPA (1988b).
Salame M. 1973. Transport properties of nitrile polymers. J. Polymer
Sci. 41:1-15. As cited in USEPA (1988b).
Salame M. 1985. Personal communication. Bloomfield, CT: Monsanto Co.
As cited in USEPA (1988b).
Sanford WE, Konikow LF. 1985. A two-constituent solute transport model
for groundwater having variable density. Reston, VA: U.S. Geological
Survey. WRI 85-4279. As cited in USEPA (1988b).
Scheunert I, Marra C. Viswanathan R, Klein W, Korte F. 1983. Fate of
hexachlorobenzene - *4C in wheat plants and soil under outdoor
conditions. Chemosphere 12(7/8):843-858.
Schnitzer M and Kerndorff H. 1980. Reactions of fulvic acid with metal
ions. Water Air Soil Pollut. 14:97-108.
Schwab GO, Frevert, RK, Edminster, T.W., Barnes KK. 1966. Soil and
water conservation engineering. 2nd ed. New York: John Wiley and
Sons. As cited in USEPA (1988b).
3-12

-------
Sehmel GA. 1980. Particle resuspension: A review. In Environment
International. Vol. 4. London, England: Pergamon Press, Ltd. As cited
in USEPA (1988b).
Sexton K, Ryan PB. 1987. Assessment of human exposure to air pollution:
methods, measurements, and models. In Watson A, Bates RR, Kennedy D
(eds). Air pollution, the automobile and public health: research
opportunities for quantifying risk. Washington, DC: National Academy
Press (In press).
Shen T. 1981. Estimating hazardous air emissions from disposal sites.
Pollut. Eng. 13(8):31-34. As cited in USEPA (1988b).
Skidmore EL, Woodruff NP. 1968. Wind erosion forces in the United
States and their use in predicting soil loss. Agriculture Handbook No.
346. Washington, DC: U.S. Department of Agriculture, Agricultural
Research Service. As cited in USEPA (1988b).
Smelt JH, Leistra M. 1974. Hexachlorobenzene in soils and crops after
soil treatment with pentachloronitrobenzene. Agric. Environ. 1:65-71.
Smith WJ, Whicker FW, Meyer HR. 1982. Review and categorization of
saltation, suspension and resuspension models. Nuclear Safety
23(6):685-699. As cited in USEPA (1988b).
Southwick LM, Willis GH. 1979. Interaction of trifluralin with soil
components under flooded anaerobic conditions. Paper presented in the
Division of Pesticide Chemistry at the 176th ACS National Meeting,
Washington, DC. As cited in Macalady et al. (1986).
Spanggord RF, Mill T, Chou TW, Mabey WR, Smith JH, Lee S. 1979.
Environmental fate studies on certain munition wastewater constituents.
Phase I. Literature review. Final report. Fort Detrick, MD: U.S. Army
Medical Bioengineering Research and Development Lab.
Sposito G, Mattigod SV. 1979. GEOCHEM: A computer program for the
calculation of chemical equilibria in soil solutions and other natural
water systems. Riverside, CA: Kearney Foundations of Soil Science,
University of California.
Steingiser S, Nemphos SP, Salame M. 1978. Barrier polymers. In:
Kirk-Othmer encyclopedia of chemical technology, 3rd ed. New York: John
Wiley and Sons. As cited in USEPA (1988b).
Texas Air Control Board. 1980. User's guide to the Texas climatological
model. Austin, TX: Texas Air Control Board. As reviewed in: USEPA.
1982. Environmental Modeling Catalogues. Washington, DC: U.S.
Environmental Protection Agency. Available from Information Clearing
House as PM-211A. As cited in USEPA (1988b).
3-13

-------
Thibodeaux LJ, Hwang ST. 1982. Landfarming of petroleum wastes -
modeling the air emission problem. In Environmental Progress 1(1):	.
As cited in USEPA (1988b).
Thibodeaux LJ. 1981. Estimating the air emissions of chemicals from
hazardous waste landfills. J. Haz. Mat. 4 (3):235-244. As cited in
USEPA (1988b).
Thorne MC. 1984. ECOS: Values of parameters to be used for domestic
animals. Associated Nuclear Services Final Report No. 372. Epsom,
Surrey, England; Department of the Environment Contract
No. PECD7/9/94-128/82.
Thornton D, Abrahams P. 1983. Soil Ingestion: A major pathway of heavy
metals into livestock grazing contaminated land. Sci. Total Environ.
28:287-294.
Thurman EM. 1984. Organic geochemistry of natural waters. The Hague,
Netherlands: Martinus Nijhoff. As quoted in McCarthy et al. (1985).
Trabalka JR, Garten, CT, Jr. 1982. Development of predictive models for
xenobiotic bioaccumulation in terrestrial ecosystems. ORNL-5869. Oak
Ridge, TN: Environmental Sciences Division, Oak Ridge National
Laboratory. NTIS Document No. DE83 003171.
Tracy JV. 1982. User's guide and documentation for adsorption and decay
modifications to the USGS solute transport model. Washington, DC: U.S.
Nuclear Regulatory Commission. NUREG/CR-2502. As cited in USEPA (1988b).
Trask PD, Patnode HW. 1942. Source beds of petroleum. Tulsa, OK: The
American Association of Petroleum Geologists. As cited in USEPA (1988b).
Travis B. 1984. TRACR3D: A model of flow and transport in porous/
fractured media. Los Alamos, NM: Los Alamos National Laboratory.
LA-0667-MS. As cited in USEPA (1988b).
Turner DB, Novak JH. 1978. User's guide for RAM. Vols, a and b.
Research Triangle Park, NC: U.S. Environmental Protection Agency.
EPA-600/8-78-016. As cited in USEPA (1988b).
Turner DB. 1970. Workbook of atmospheric dispersion estimates. Research
Triangle Park, NC: U.S. Environmental Protection AGency, Office of Air
Programs. AP-26.
USDA. 1974. Universal soil loss equation. Agronomy Technical Note No.
32. Portland, OR: U.S. Department of Agriculture, Soil Conservation
Service, West Technical Service Center. As cited in USEPA (1988b).
3-14

-------
USDC. 1968. Climatic atlas of the United States. Washington, D.C.:
U.S. Department of Commerce. Environmental Sciences Services
Administration. Environmental Data Service. As cited in USEPA (1988b).
USEPA. 1977a. User's manual for single source (CRSTER) model. Research
Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air
Quality Planning and Standards, EPA-45012-77-013. As cited in USEPA
(1988b).
USEPA. 1979a. Nonaquatic fate of priority pollutants. Versar Inc.
Washington, DC: U.S. Environmental Protection Agency, Office of Water
Planning and Standards. Contract No. 68-01-3852.
USEPA. 1979b. Environmental modeling catalogue. Washington, DC: U.S.
Environmental Protection Agency. Contract No. 68-01-4723. As cited in
USEPA (1988b).
USEPA. 1980a. Hazardous waste background document for the control of
hazardous waste leachate. Cincinnati, OH: Municipal Environmental
Research Laboratory, Solid and Hazardous Waste Research Division, Briefing
document for Review Committee Meeting held at the U.S. Environmental
Protection Agency, Washington, DC. January 23, 1980. As cited in USEPA
(1988b).
USEPA. 1980b. Land disposal of hexachlorobenzene wastes: controlling
vapor movement in soil. Cincinnati, OH: U.S. Environmental Protection
Agency, Municipal Environmental Research Laboratory, Office of Research
and Development. EPA-600/2-80-119. As cited in USEPA (1988b).
USEPA. 1982a. Chemical fate test guidelines. Test Guideline CG-1700.
Washington, DC: U.S. Environmental Protection Agency. EPA
560/6-82-003. Available from NTIS as PB 82-233008. As quoted in Leifer
et al. (1983).
USEPA. 1982b. Establishment of guidelines for modeling groundwater
contamination from hazardous waste facilities. Preliminary groundwater
modeling profile. Discussion draft. Washington, DC: U.S. Environmental
Protection Agency, Office of Solid Waste.
USEPA. 1983a. Compilation of air pollutant emission factors: AP-42.
Research Triangle Park, NC: U.S. Environmental Protection Agency.
Office of Air, Noise and Radiation, Office of Air Quality Planning and
Standards. As cited in USEPA (1988b).
USEPA. 1983b. Methods for assessing exposure to windblown particulates.
Washington, DC: U.S. Environmental Protection Agency, Office of Health
and Environmental Assessment, Office of Research and Development.
EPA-600/4-83-007. As cited in USEPA (1988b).
3-15

-------
USEPA. 1985a. Predicting the fate of organic chemicals. Handout
materials for EPA's risk assessment workshop, October 28-29, 1985,
Denver, Colorado. Washington, DC: U.S. Environmental Protection Agency,
Office of Toxic Substances, Exposure Evaluation Division, Exposure
Assessment Branch.
USEPA. 1985b. Physical-chemical properties and categorization of RCRA
wastes according to volatility. Research Triangle Park, NC: U.S.
Environmental Protection Agency, Office of Air Quality Planning and
Standards. EPA Publication No. 450/3-85-007.
USEPA. 1985c. Methods for assessing exposure to chemical substances,
Volume 4-Methods for enumerating and characterizing populations exposed
to chemical substances. Washington, D.C.: U.S. Environmental Protection
Agency, Office of Toxic Substances. EPA No. 560/5-85-004.
USEPA. 1986a. U.S. Environmental Protection Agency. Guidelines for
estimating exposures. Federal Register 51:34042-34054.
USEPA. 1986b. Mathematical model	selection criteria for performing
exposure assessments: Groundwater	contaminants from hazardous waste
facilities. Washington, DC: U.S.	Environmental Protection Agency,
Office of Health and Environmental	Assessment. As cited in USEPA (1988b).
USEPA. 1986c. Exposure assessment for hexachlorobenzene. Final
report. Washington, D.C.: U.S. Environmental Protection Agency, Office
of Toxic Substances. EPA Publication No. 560/5-86-019.
USEPA. 1986d. Methods for assessing exposure to chemical substances.
Vol. 8. Methods for assessing environmental pathways of food
contamination. Washington, DC: Environmental Protection Agency. EPA
560/5-85-008.
USEPA. 1987. Cornerstone paper on exposure assessment. First draft.
Office of Research and Development, Washington, D.C. U.S. Environmental
Protection Agency, Office of Research and Development.
USEPA. 1988a. U.S. Environmental Protection Agency. Proposed
guidelines for exposure-related measurements. Federal Register
53:48830-48853.
USEPA. 1988b. Superfund Exposure Assessment manual. Washington, D.C.:
U.S. Environmental Protection Agency, Office of Emergency and Remedial
Response. EPA Publication No. 540/1-88/001.
USEPA. 1989. Exposure factors handbook. Washington, DC: U.S.
Environmental Protection Agency, Office of Health and Environmental
Assessment. EPA No. 600/8-89/043.
3-16

-------
USEPA/USFDA/USDA. 1981. Land application of municipal sewage sludge for
the production of fruits and vegetables. A statement of federal policy
and guidance. Joint Policy Statement SW-905. Washington, DC: U.S.
Environmental Protection Agency.
van Genuchten MT, Alves WJ. 1982. Analytical solutions of the one-
dimensional convective-dispersive solute transport equation. Riverside,
CA: U.S. Department of Agriculture. Technical Bulletin 1661. As cited
in USEPA (1988b).
van Genuchten MT. 1978. Mass transport in saturated-unsaturated media:
one-dimensional solutions. Princeton, NJ: Department of Civil
Engineering. Princeton University. 78/WR-ll. As cited in USEPA (1988b).
van der Heijde PKM, Paul KM, Bachmat Y, Bredehoeft J, Andrews B, Holtz D,
Sebastian S. 1985. Groundwater management: The use of numerical
models. Washington, DC: American Geophysical Union, Water Resources
Monograph #5, 2nd Edition. As cited in USEPA (1988b).
van der Heijde PKM, Beljin JS. 1987. Model assessment for delineating
well head protection areas. Washington, DC: Office of Groundwater
Protection, U.S. Environmental Protection Agency. International
Groundwater Modeling Center, Holcomb Research Institute. Butler
University, Indianapolis, Indiana. As cited in USEPA (1988b).
van der Heijde PKM, Srinivasan P. 1986. A "random-walk" model for
solute transport in groundwater 'RANDOM WALK/RWH'. Indianapolis, IN:
International Ground Water Modeling Center, Holcomb Research Institute,
Butler University. As cited in USEPA (1988b).
Veith GD, Macek KJ, Petrocelli SR, Carroll J. 1980. An evaluation of
using partition coefficients and water solubility to estimate
bioconcentration factors for organic chemicals in fish. In: Aquatic
Toxicology, ASTM STP 707, Philadelphia, PA: Am. Soc. for Testing
Materials. As cited in McCarthy and Black (1987).
Versar, Inc. 1983. Theoretical evaluation of sites located in the zone
of saturation. Draft final report. Chicago, IL: U.S. Environmental
Protection Agency. Contract No. 68-01-6438. As cited in USEPA (1988b).
Versar, Inc. 1986. Methods for assessing exposure to chemical
substances. Volume 10. Methods for the identification,
characterization, and analysis of uncertainties. Washington, D.C.: U.S.
Environmental Protection Agency, Office of Toxic Substances, Exposure
Evaluation Division. EPA Contract No. 68-02-3968.
Versar, Inc. 1987. Groundwater assessment methods for screening level
exposure assessments. Final report. Washington, DC: U.S. Environmental
Protection Agency, Office of Toxic Substances. EPA Contract
No. 68-02-4254.
3-17

-------
Versar, Inc. 1988. Air dispersion modeling as applied to hazardous
waste incinerator evaluations: An introduction for the permit writer.
Washington, DC: U.S. Environmental Protection Agency, Office of Solid
Waste.
Verschueren K. 1984. Handbook of environmental data on organic
chemicals. New York: Van Nostrand/Reinhold Press. As cited in USEPA
(1988b).
Voss CI. 1984. SUTRA: A finite element simulation model for saturated-
unsaturated fluid density-dependent ground water flow with energy
transport or chemical reactive single species solute transport. Reston,
VA: U.S. Geological Survey, Water Resources Investigation 84-4369. As
cited in USEPA (1988b).
Walter-Echols G, Lichtenstein EP. 1977. Microbial reduction of phorate
sulfoxide to phorate in soil-lake mud-water microcosms. J. Econ. Entomol.
70:505-509. As quoted in Macalady et al. (1986).
Walton WC. 1984. Handbook of analytical ground water models.
Indianapolis, IN: International Ground Water Modeling Center, Holcomb
Research Institute, Butler University. As cited in USEPA (1988b).
Walton WC. 1985. Thirty-five BASIC groundwater programs for desktop
microcomputers 'WALT0N84-35BASIC'. Indianapolis, IN: International
Ground Water Modeling Center, Holcomb Research Institute, Butler
University. As cited in USEPA (1988b).
Williams, Jr. 1975. Sediment yield prediction with the universal
equation using runoff energy factor. In: Present and prospective
technology for predicting sediment yields and sources. Washington, DC:
U.S. Department of Agriculture. ARS-S-40. As cited in USEPA (1988b).
Winner R. 1985. Bioaccumulation and toxicity of copper as affected by
interactions between humic acid and water hardness. Water Res.
19:449-455.
Wischmeier WH. 1972. Estimating the cover and management factor on
undisturbed areas. Proceedings of the USDA Sediment Yield Workshop.
Oxford, MS: U.S. Department of Agriculture. As cited in USEPA (1988b).
Wischmeier, WH, Johnson, CD, Cross, BV. 1971. A soil erodibility
nomograph for farmland and construction sites. J. Soil Water Cons.
26:189-193.
Woodburn KB, Rao PSC, Fukui M, Nkedi-Kizza P. 1986. Solvophobic
approach for predicting sorption of hydrophobic organic chemicals on
synthetic sorbents and soils. J. Contam. Hydrology 1:227-241. As cited
in USEPA (1988b).
3-18

-------
Yeh GT, Huff DD. 1985. FEMA: A finite element model of material
transport through aquifers. Oak Ridge, TN: Oak Ridge National
Laboratory, 0RNL-6063. As cited in USEPA (1988b).
Yeh GT, Ward DS. 1981. FEMWASTE: A finite element model of waste
transport through saturated-unsaturated porous media. Environmental
Services Division Publication No. 1462, Oak Ridge National Laboratory.
0RNL-5602. 137 p. As reviewed in: Versar Inc. 1983. Theoretical
evaluation of sites located in the zone of saturation. Draft final
report. Chicago, IL: U.S. Environmental Protection Agency. Contract
No. 68-01-6438. As cited in USEPA (1988b).
Yeh GT. 1981. AT123D. Analytical transient one-, two-, and three-
dimensional simulation of waste transport in the aquifer system.
Environmental Sciences Division Publication No. 1439. Oak Ridge, TN:
Oak Ridge National Laboratory. ORNL-5601. As cited in USEPA (1988b).
Yeh GT. 1982. CHNTRN: a chemical transport model for simulating
sediment and chemical distribution in a stream/river network. Washington,
DC: U.S. Environmental Protection Agency, Office of Pesticides and Toxic
Substances. Contract No. W-7405-eng-26. As reviewed in: Versar 1983.
Methodology for assessing exposures to chemical substances via the
ingestion of drinking water. Washington, DC: U.S. Environmental
Protection Agency. Contract No. 68-01-6271. As cited in USEPA (1988b).
Yeh GT. 1987. FEMWATER: A finite element model of water flow through
saturated-unsaturated porous media. First revision. Oak Ridge, TN: Oak
Ridge National Laboratory. ORNL-5567/RI. As cited in USEPA (1988b).
Zamuda C, Sunda W. 1982. Bioavailability of dissolved copper to the
American oyster, Crassostrea virqinica. I. Importance of chemical
speciation. Mar. Biol. 66:77-82.
Zepp RG, Cline DM. 1977. Rates of direct photolysis in aquatic
environments. Environ. Sci. Technol. 11:359-366.
3-19

-------
APPENDIX A
Glossary of Terms

-------
GLOSSARY OF TERMS
Absorbed Dose - The amount of a substance penetrating across the
exchange boundaries of an organism, via either physical or biological
processes, after contact (exposure).
Accuracy - The measure of the correctness of data, as given by the
difference between the measured value and the true or standard value.
Administered Dose - The amount of a substance given to a human or test
animal in determining dose-response relationships, especially through
ingestion or inhalation (see applied dose). Even though this term is
frequently encountered in the literature, administered dose is actually a
measure of exposure, because even though the substance is "inside" the
organism once ingested or inhaled, administered dose does not account for
absorption (see absorbed dose).
Agent - A chemical, radiological, mineralogical, or biological entity
that may cause deleterious effects in an organism after the organism is
exposed to it.
Ambient - surrounding conditions. "Indoor ambient" and "outdoor
ambient" are sometimes used to differentiate between indoor and outdoor
surroundings.
Ambient Measurement - A measurement (usually of the concentration of a
chemical or pollutant) taken in an ambient medium, normally with the
intent of relating the measured value to the exposure of an organism which
contacts that medium.
Ambient Medium - One of the basic categories of material surrounding or
contacting an organism, e.g., outdoor air, indoor air, water, or soil,
through which chemicals or pollutants can move and reach the organism (see
biological medium, environmental medium).
Applied Dose - The amount of a substance given to a human or test animal
in determining dose-response relationships, especially through dermal
contact (see administered dose). Even though this term is encountered in
the literature, applied dose is actually a measure of exposure, since it
does not take absorption into account.
Arithmetic Mean - The sum of all the measurements in a data set divided
by the number of measurements in the data set.
Bias - A systematic error inherent in a method or caused by some feature
of the measurement system.
A-l

-------
Biological Medium - One of the major categories of material within an
organism, e.g., blood, adipose tissue, or breath, through which chemicals
can move, be stored, or be biologically, physically, or chemically
transformed (see ambient medium, environmental medium).
Data Quality Objectives (DQO) - Statements of the level of uncertainty
an assessor is willing to accept in results derived from environmental
data.
Dose - The amount of a substance available for interaction with
metabolic processes of an organism following exposure and absorption into
the organism. The amount of a substance crossing the exchange boundaries
of skin, lungs, or digestive tract is termed absorbed dose, while the
amount available for interaction by any particular organ or cell is termed
the delivered dose for that organ or cell. Theoretically, the sum of the
delivered doses plus the metabolic transformations should equal absorbed
dose. (The terms administered dose and applied dose refer to amounts of
a substance made available for absorption, and therefore are measures of
exposure rather than dose. As such, these terms, sometimes found in the
literature, are somewhat confusing and should be avoided if possible by
exposure assessors. The term exposure dose, a common radiological term,
refers to exposure but carries the assumption that the absorption fraction
is one, making exposure equal to absorbed dose. Again, this term is some-
what confusing and should be avoided if possible.)
Dosimeter - Instrument to measure dose; many so-called dosimeters
actually measure exposure rather than dose.
Dosimetry - Process of measuring dose.
Ecological Exposure - Exposure of a nonhuman receptor or organism to a
chemical, radiological, or biological agent.
Effluent - (Waste) material being discharged into the environment,
either treated or untreated. Effluent generally is used to describe water
discharges to the environment, although it can refer to stack emissions or
other material flowing into the environment.
Environmental Fate - The destiny of a chemical or biological pollutant
after release into the environment. Environmental fate involves temporal
and spatial considerations of transport, transfer, storage, and transfor-
mation.
Environmental Fate Model - In the context of exposure assessment, any
mathematical abstraction of a physical system used to predict the concen-
tration of specific chemicals as a function of space and time subject to
transport, intermedia transfer, storage, and degradation in the environ-
ment.
A-2

-------
Environmental Medium - One of the major categories of material found
in the outdoor natural physical environment that surrounds or contacts
organisms, e.g., surface water, ground water, soil, or air, and through
which chemicals or pollutants can move and reach the organisms (see
ambient medium, biological medium).
Exposure - Contact of an organism with a chemical, physical, or biologi-
cal agent. Exposure is quantified as the amount of the agent available
at the exchange boundaries of the organism (e.g., skin, lungs, digestive
tract) and available for absorption.
Exposure Assessment - The determination or estimation (qualitative or
quantitative) of the magnitude, frequency, duration, and route of expo-
sure.
Exposure Pathway - The course a chemical or pollutant takes from the
source to the organism exposed.
Exposure Rate - Exposure per unit time (compare exposure, dose),
yielding units of amount (mass, fibers, etc.)/time, for example, mg/day.
Exposure rates are often normalized to body weight, yielding units such
as mg/kg/day.
Exposure Route - The way a chemical or pollutant enters an organism
after contact, e.g., by ingestion, inhalation, or dermal absorption.
Exposure Scenario - A set of assumptions about how exposure takes place
(including assumptions/conditions concerning sources, exposure pathways,
concentrations of pollutants, individual or population habits and charac-
teristics), which aid the exposure assessor in evaluating, estimating, or
quantifying exposures.
Fixed-Location Monitoring - Sampling of an environmental or ambient
medium for pollutant concentration at one location continuously or
repeatedly over some length of time.
Geometric Mean - The nth root of the product of n values.
Guidelines - Principles and procedures to set basic requirements for
general limits of acceptability for assessments.
Limit of Detection (LOD) - The smallest concentration or amount of a
substance that can be differentiated from background by a given measure-
ment process (e.g., analytical instrument/sample matrix combination) (see
1imit of quantitation).
A-3

-------
Limit of Quantitation (LOQ) (also "limit of quantification") - The
lower limit of concentration or amount of substance for which quantitative
results may be obtained with a specified degree of confidence (i.e., the
amount that must be present before a particular method is considered to
provide reliable and reproducible quantitative results) (see limit of
detection).
Maximally Exposed Individual (ME1) - The single individual with the
highest exposure in a given population.
Median Value - The value in a measurement data set such that half the
measured values are greater and half are less.
Microenvironment Method - A method used in predictive exposure assess-
ments to estimate exposures by sequentially assessing exposure for a
series of areas (microenvironments) that can be approximated by constant
or well-characterized concentrations of a chemical or other agent.
Microenvironments - Well-defined areas such as the home, office, auto-
mobile, kitchen, store, etc. that can be treated as homogeneous (or well
characterized) in the concentrations of a chemical or other agent.
Mode - The value in the data set that occurs most frequently.
Monte Carlo Method - A method used in predictive exposure assessments
that uses the Monte Carlo technique, a repeated random sampling from the
distribution of values for each of the parameters in a generic exposure
equation, to derive an estimate of the distribution of exposures in the
population.
Non-Parametric Methods - Statistical methods that do not assume a
particular statistical distribution for the statistical population(s) of
interest ("distribution-free methods").
Parametric Methods - Statistical methods that assume a particular
statistical distribution for the statistical populations of interest
("distribution-specific methods").
Personal Measurement - A measurement collected from an individual's
immediate environment using direct methods such as personal air pumps.
Pharmacokinetics - The study of the time course of absorption, distribu-
tion, metabolism, and excretion of a foreign substance (e.g., a drug or
pollutant) in an organism's body.
A-4

-------
Predictive Exposure Assessment - An approach to quantifying exposure
by measurement or estimation of both the amount of a substance contacted,
and the frequency/duration of contact, and subsequently linking these
together to estimate exposure. Predictive exposure assessments commonly
use the scenario method, the microenvironment method, or the Monte Carlo
method (or combinations of these) to make the link between amount
(intensity of contact) and duration of contact.
Probability Samples - Samples selected from a statistical population
such that each sample has a known probability of being selected.
Quality Assurance - The system of activities whose purpose is to provide
to the user of the exposure assessment the assurance that the data used
as a basis for the assessment meet defined standards of quality.
Quality Control - The overall system of activities whose purpose is to
control the quality of the measurement data so that they meet the needs of
the user.
Random Samples - Samples selected from a statistical population such
that each sample has an equal probability of being selected.
Range - The difference between the largest and smallest values in a
measurement data set.
Reconstructive Exposure Assessment - An approach to quantifying exposure
by reconstructing absorbed dose (and exposure) after exposure has
occurred, from evidence within an organism such as chemical levels in
tissues or fluids, or from evidence of other biomarkers of exposure.
Representativeness - The degree to which a sample is, or samples are,
characteristic of the whole medium, exposure, or dose for which the
samples are being used to make inferences.
Sample - A small part of something designed to show the nature or
quality of the whole. Exposure-related measurements are usually samples
of environmental or ambient media, exposures of a small subset of a
population for a short time, or biological samples, all for the purpose
of inferring the nature and quality of parameters important to evaluating
exposure.
Sampling Frequency - The time interval between the collection of
successive samples.
Sampling Plan - A set of rules or procedures specifying how a sample is
to be selected and handled.
A-5

-------
Scenario Method - A method used in predictive exposure assessment'to
account for variability in exposure parameters by setting up one or more
alternative sets of assumptions (scenarios), evaluating each for the
resulting exposures, and then using the resulting exposure estimate(s) as
being illustrative of the exposures in the actual populations being
evaluated. In the case of a single scenario, a sensitivity analysis is
usually performed on the parameters to evaluate the variability.
Source Characterization Measurements - Measurements made to characterize
the rate of release of agents into the environment from a source such as
an incinerator, landfill, industrial or municipal facility, consumer
product, etc.
Standard Operating Procedure (SOP) - A procedure adopted for repeti-
tive use when performing a specific measurement or sampling operation.
Statistical Control - The process by which the variability of measure-
ments or of data outputs of a system is controlled to the extent necessary
to produce stable and reproducible results.
Statistical Significance - An inference that the probability is low that
the observed difference in quantities being measured could be due to vari-
ability in the data rather than an actual difference in the quantities
themselves. The inference that an observed difference is statistically
significant is typically based on a test to reject one hypothesis and
accept another.
Surrogate Data - Substitute data or measurements on one substance used
to estimate corresponding values of another substance.
A-6

-------
APPENDIX B
Exposure Assessment Worksheets

-------
APPENDIX C
Completed Exposure Assessment Worksheets for
An Example Chemical

-------
Appendix D
Transfer Coefficients for Assessing
Exposures to Contaminants in
Animal Tissues and Animal Products

-------
IU.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Table D-l. Elemental Feed-to-Milk Transfer Coefficients
Transfer coefficient
(fraction of
ingested element
Element	per liter of Bilk) Code3	Chemical form''
H	1.40E-02	1	Tritiated water
He	2.00E-02	9
Li	2.00E-02	6
Be	9.10E-07	1	Berylliun chloride
B	1.50E-03	2
C	1.50E-02	4
N	2.30E-02	3
0	2.00E-02	9
F	1.10E-03	4
He	2.00E-02	9
Ha	3.50E-02	3
Mg	3.90E-03	2
A1	2.00E-04	4
Si	2.00E-05	4
P	1.60E-02	2
S	1.60E-02	2
CI	1.70E-02	2
Ar	2.00E-02	9
K	7.20E-03	2
Ca	1.10E-02	3
Sc	5.00E-06	6
Ti	1.00E-02	4
V	2.00E-05	4
Cr	2.00E-03	4
Hn	8.40E-05	2
Fe	5.90E-05	3
Co	2.00E-03	3
D-l

-------
8085H
Table D-l. (continued)
Atonic
no.
Transfer coefficient
(fraction of
ingested element
Element per liter of milk)
Code
Chemical form
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
Ni
Cu
Zn
6a
Ge
As
Se
Br
Kr
Rb
Sr
1
Zr
Nb
Ho
Tc
Ru
Rh
Pd
Ag
Cd
In
Sn
Sb
Te
1.00E-02
1.70E-03
1.00E-02
5.00E-05
7.00E-02
6.20E-05
4.00E-03
2.00E-02
2.00E-02
1.20E-02
1.40E-03
2.00E-05
8.00E-Q2
2.00E-02
1.40E-03
9.90E-03
6.10E-07
1.00E-02
1.00E-02
3.00E-02
1.00E-03
1.00E-04
1.20E-03
2.00E-05
2.00E-04
Sodinn arsenate
Selenous acid
Rutheniun chloride,
nitrosyl rutheniun
trinitrate
Antimony trichloride
D-2

-------
tu 0
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
Table D-l. (continued)
Transfer coefficient
(fraction of
ingested elaoent
Elanent per liter of milk)
Code"
Chemical form''
I
Xe
Cs
Ba
La
Ce
Pr
Nd
Pm
So
Eu
Gd
Tb
Dy
Ho
Er
Tm
Yb
Lu
Hf
Ta
W
Re
Os
Ir
9.90E-03
2.00E-02
7.10E-03
3.50E-04
2.00E-05
2.00E-05
2.00E-05
2.00E-05
2.00E-05
2.00E-05
2.00E-05
2.00E-05
2.00E-05
2.00E-05
2.00E-05
2.00E-05
2.00E-05
2.00E-05
2.00E-05
5.00E-06
2.80E-06
2.90E-04
1.30E-03
5.00E-03
2.00E-06
Cerous chloride
Tantalun oxalate
Sodiun and potass inn
tungstate
Sodiun perrhenate
AmoniiiD iridiun
hexachloride
D-3

-------
8085H
Table D-l. (continued)
Atonic
no.
Transfer coefficient
(fraction of
ingested element
Element per liter of milk)
Code"
Chemical form
78
Pt
5.00E-03
5
79
Ail
5.30E-06
1
80
Hg
9.70E-06
1
81
T1
1.90E-03
1
82
Pb
2.60E-04
2
83
Bi
5.00E-04
5
84
Po
1.40E-04
1
85
At
1.00E-02
6
86
Rn
2.00E-02
9
87
Fr
2.00E-02
6
88
Ra
4.50E-04
3
89
Ac
2.00E-05
7
90
Th
5.00E-06
5
91
Pa
5.00E-06
6
92
U
6.10E-04
2
93
Np
5.00E-06
6
94
Pu
1.00E-07
8
95
Am
2.00E-05
7
96
Cm
2.00E-05
7
97
Bk
2.00E-05
7
98
Cf
2.00E-05
7
99
Es
2.00E-05
7
100
Fm
2.00E-05
7
Auric chloride
Mercuric chloride
Thallous nitrate
Poloniun dioxide
Plutoniun citrate
The codes designating the procedures used for estimating the transfer coefficient
(f||) are as folloMs:
D-4

-------
8085H
Table D-l. (continued)
Code
1	The f|j is based on tracer data.
2	Thef, is based on concentrations in milk and related forage.
3	The f|j is based on a combination of tracer data and concentrations in milk and
related forage.
4	Thef„ is based on concentrations in milk and forage.
5	In the absence of tracer data or data on concentrations in milk and forage, the fH
is assumed to be 0.05 x fraction absorbed per liter. The fraction absorbed was
assuoed to be the ICRP value.
6	The fj| is based on collateral data.
7	The fg is based on that for ceriun and the related rare earths elenents.
8	The f|j is based on experimental data and collateral data.
9	The fp| of oxygen is estimated on the basis of the daily water consumption
(50 liters) and the daily secretion of water in nilk (10 liters), i.e., f^ =
0.02/liter. The fM of the inert gases is similarly based on water balance in the
lactating cow.
b
If the transfer coefficient is based on a particular chemical form of the element,
the name of the compound is given in this colum.
Source: Ng et al. 1977.
D-5

-------
8085H
Table D-2. Suimary of Feed to Meat and Feed to Egg Transfer Coefficients
Transfer Coefficient
ffraction of inoested element per kilogram)
Element or	Animal	No. of
isotope	product observations Average8	Range	Approach''
Na
Beef
1
8.3xl0~2

4

Egg (contents)
1
6.1

3
Mg
Beef
1
l.BxlO"2

3

Egg (contents)
1
1.6

3
P
Beef
1
4.9xl0"2
4.1xl0"2 - 5.7xl0"2
3,4
S
Lamb
2
1.4
1.3 - 1.4
3
K
Beef
1
1.8xlO"Z

4

Egg (contents)
1
1.1

3
Ca
Beef
2
1.6xl0~3
7.2xl0"4 - 2.5xl0~3
3,4

Chicken
2
4.4xl0"2
2.4x10"2 - 4.4xl0"2
3.5

Egg (contents)
2
4.4X10"1
4.4xl0_1 - 6-OxlO"1
3.5

Egg (whole)
2
1.3X101
l.lxlO1 - 1.3x1a1
3,5
Cr
Beef
1
9.2xl0"3

4
Mn
Beef
5
5xl0~4
3.8xlO"4 - 7xl0"4
3.4,5

Pork
2
3.6xl0~3
1.9xl0"3 - 5.3xl0~3
3

Lamb
3
5.9xl0~3
2.0xl0_3 - 9.0xl0~3
3

Chicken
2
5.1xl0"2
3.3xl0"2 - 6.8x10"2
1.3

Egg (contents)
4
6.5xl0"2
3.1xl0"2 - 1.3x10"*
3

Egg (whole)
1
6.0xl0~2

3
Fe
Beef
7
2.1xl0"2
2xl0~3 - 4.7xl0"2
3,4,5

Pork
1
2.6xl0~2

3

Lanb
1
7.3xl0~2

3

Chicken
1
1.5

3

Egg (contents)
3
1.3
6.2X10"1 - 2.0
2.3

Egg (whole)
3
1.2
5.6xl0_1 - 1.9
2,3
Co
Beef
3

2xl0"3 - 6.9xl0~2
3,4,5

Pork
1
1.7X10"1

3

Lant)
1
6.2xl0"2
5.3xl0_1 - l.OxlO1
3

Chicken
3

1.3,5

Egg (contents)
2

2xl0"2 - 4.9
1,3

Egg (whole)
2

2xl0'2 - 1.3x10*

D-6

-------
8085H
Table D-2. (continued)
Transfer Coefficient
(fraction of ingested element per kilogram)
Eleoent or	Animal	No. of
a	b
isotope	product observations Average	Range	Approach
Ni
Beef
1
2.0xl0~3

4
Cu
Beef
3
9.0xl0~3
4.7xl0~3 - 1.3xl0"2
3.4

Pork
2
2.2xl0-2
1.4x10"2 - 2.9xl0"2
3

Laab
1
3.9xl0"2

3

Chicken
1
5-lxlO"1

3

Egg (contents)
2
4.9x10"*
2.5x10"1 - 7.2x10"1
3

Egg (whole)
1
5.3xl0-1

3
Zn
Beef
5
9.8xl0"2
3.5xl0~2 - 2-OxlO"1
3.4

Pork
2
1.5xl0-1
5.4xl0'2 - 2.4xl0_1
3

Lamb
1
4.1

3

Chicken
1
6.5

3

Egg (contents)
3
2.6
1.2 - 4.6
3

Egg (whole)
2
2.4
1.1 - 4.4
3
Se
Port
1
3.2X10"1

3

Chicken
1
8.5

3

Egg (contents)
1
9.3

3
Rb
Beef
1
l.lxlO"2

4
Sr
Beef
6
8.1xl0~4
7.8xl0"5 - 1.8xl0"3
2.4

Pork
5
3.9xl0~2
4xl0"3 - 6.9xlO"Z
2

Laid)
7
2.2xl(f3
l.lxlO"3 - 3.7xl0~3
2

Chicken
8
3.5xlO~2
l.BxlO'2 - B.OxlO"2
2

Egg (contents)
4
3.0xl0_1
2.7X10"1 - 3.5xl0_1
2

Egg (whole)
4
7.1
6.6 - 8.6
2
Y 88
Beef
1
lxlO"3

5
Y 91
Chicken
1
lxl0"2

5

Egg (contents)
1
2xl0~3

5

Egg (whole)
1
2xl0"3

5
Zr
Beef
1
2x10"Z

4
Nb
Beef
1
2.5X10"1

4
Nb 95
Chicken
1
2xl0~3

5

Egg (contents)
1
3x10"3

5

Egg (whole)
1
3xl0"3

5
D-7

-------
8085H
Table 0-2. (continued)
Transfer Coefficient
(fraction of ingested element per kilogram)
Element or	Animal	No. of
isotope	product observations Average8	Range	Approach'*
Ho
Mo 99
Beef
Chicken
Egg (contents)
Egg (whole)
6.8x10
-3
5x10
5x10
4x10
r2
fl
-1
t„ 95
IC D
pertechnetate
in feed0
pertechnetate
in feed0
pertechnetate
in feed0
Chicken
Chicken
Egg (contents)
Egg (contents
Egg (Dhole)
Egg (whole)
2.0x10
6.3x10
1.9
1.9
1.5
1.5
fl
-2
6.3xl0"2 - 6.6x10 2
Ru
Ag
Cd
Sb
Te 132
Beef
Chicken
Egg (contents)
Egg (whole)
Beef
Beef
Pork
Chicken
Egg (contents)
Beef
Chicken
Egg (contents)
Egg (whole)
Beef
Pork
Lamb
Chicken
Egg (contents)
Egg (whole)
2.0x10
7x10"3
-3
6x10"
6x10
-3
-3
2x10
3.5x10"
3.0x10
8.4x10
lxlO"1
"3
-1
9.2x10"
2.7x10
-1
8x10
7x10
-1
-1
3.6x10
3.3x10
3.0x10
2.8
2.5
f3
r3
-2
7 xlO"4 - 2.0xl0"3
1.9xl0"3 - 5.5xl0"3
1.8xl0"4 - 3.3xl0"3
4xl0"3 - 9.4xl0"2
1.8 - 3.8
1.5 - 3.4
2.5
5
5
5
4.5
3
3
5
5
5
5
5
2
2.3
2
2
1.2
1.2
D-8

-------
8085H
Table D-2. (continued)
Transfer Coefficient
(fraction of ingested element per kilogram)
Elenent or	Animal	No. of
isotope	product observations Average8	Range	Approach'*
Cs
Beef
16
2.6xl0"2
7.2xl0"3 - 9.2x10"2
2

Pork
2
2.5x10"'
2.4x10"1 - 2.6x10"1
2

Lanb
5
2.9x10"'
7.Bxl0"2 - 5.0x10"'
2

Chicken
2
4.4
4.2 - 4.6
1,2

Egg (contents)
4
4.9x10"'
4.1x10"' - 5.7xl0_1
1.2

Egg (whole)
4
4.4x10"'
3.6xl0_1 - 5.2xl0_1
1.2
Ba
Beef
1
9.7xl0"5

4

Chicken
2
2xl0~2
lxlO"2 - 2.9xlO"Z
5

Egg (contents)
2
9xl0-1
5x10"' - 1.2
5

Egg (whole)
2
2
1.0 - 2.7
5
La 140
Chicken
1
lxlO"1

5

Egg (contents)
1
9xl0~3

5

Egg (whole)
1
lxlO"2

5
Ce
Beef
1
2xl0~3

5

Chicken
1
lxlO"2

5

Egg (contents)
I
5xl0"3

5

Egg (whole)
1
5xl0~3

5
Pr 142
Chicken
1
3xl0~2

5

Egg (contents)
1
5xl0~3

5

Egg (whole)
1
5xl0~3

5
Nd 147
Chicken
1
9xl0"2

5

Egg (contents)
1
3xl0~4

5

Egg (whole)
1
5xl0"4

5
PB
Chicken
1
2xl0"3

S

Egg (contents)
1
2x10"Z

5

Egg (whole)
1
2xl0~2

5
W
Beef
1
3.7xl0"2

4
Hg
Chicken
1
2.7xl0"Z

1
Pb
Beef
S
4xl0"4
l.OxlO"4 - 7xl0"4
3.5
Po 210
Beef
3
4.5xl0"3
5.5xl0~4 - 4.6xl0~3
5
D-9

-------
8085H
Table D-2. (continued)
Transfer Coefficient
(fraction of ingested element per kilogram!
Element or
Animal
No. of



isotope
product
observat ions
Average3
Range
Approach''
Ra 226
Beef
1
9xl0~4

5
U
Pork
1
4xl0"2

3

Chicken
1
1.2

3

Egg (contents)
1
9.9X10"1

3

Egg (whole)
1
1.3

3
Pu
Beef
4
2x10"6
1.7x10"7 - 2xl0~5
2.5
(dioxide)
Chicken
1
1.9xl0"5

2
(citrate)
Chicken
1
1.6xl(f4

2
(dioxide)
Egg (contents)
1
3.3xl0"5

2
(citrate)
Egg (contents)
1
7.6xl0"3

2
(dioxide)
Egg (whole)
1
2.9xl0"5

2
(citrate)
Egg (whole)
1
6.8xl0~3

2
An





(citrate)
Chicken
1
l.BxlO"4

2
(citrate)
Egg (contents)
1
8.5xl0~3

2
(citrate)
Egg (whole)
1
7.6xl(T3

2
a When the indivicfcial estimates were disparate, no average Mas listed.
b The transfer coefficient (Ff) values are estimated by the following approaches:
(1)	The Ff	is based on the recovery of a single actainistered dose of a radioisotope.
(2)	The Ff	is based on the recovery of a radioisotope fed repeatedly or continuously.
(3)	The Ff	is based on stable-element concentrations in neat or eggs and the associated feed.
(4)	The Ff	is based on stable-elenent concentrations in associated beef and feed.
(5)	The Ff	is based on collateral information and data.
0 Isotope incorporated into feed.
Source: Ng et al. 1982.
D-10

-------
I'? V?
[)5
0
EPA Library Region 4
IRHIlllllll
1011510
r


DATE DUE
-




J

-------