-------
EPA/540/1-881001
OSWER Directive 9285.5-1
April 1988
Superfund Exposure Assessment
Manual
U.S. Environmental Protection Agency
Office of Remedial Response
Washington, DC 20460
-------
Notice
This report was prepared under contract to an agency of the United States
Government. Neither the United States Government nor any of its employees,
contractors, subcontractors, or their employees makes any warranty,
expressed or implied, or assumes any legal liability or responsibility for any
third party's use of or the results of such use of any information, apparatus,
product, or process disclosed in this report, or represents that its use by such
third party would not infringe on privately owned rights.
-------
Table of Contents
Chapter Page
List of Tables vi
List of Figures VIM
Foreword ix
Executive Summary xi
Acknowledgments xii
1 INTRODUCTION 1
1.1 Purpose 1
1.2 Background 1
1.3 Scope 1
1.4 Use of the Manual 2
1.5 Timeframe of Analysis 4
1.6 Analysis of Exposure Associated with Remedial Actions 4
1.7 Organization of the Manual 5
2 CONTAMINANT RELEASE ANALYSIS 7
2.1 Introduction 7
2.2 Contaminant Release Screening 8
2.2.1 Contaminants in Soil 8
2.2.2 Contaminants Above-Ground 10
2.3 Quantitative Analysis of Atmospheric Contamination 10
2.3.1 Fugitive Dust Emission Analysis 10
2.3.1.1 Beginning Quantitative Analysis 10
2.3.1.2 In-Depth Analysis 14
2.3.2 Volatilization Emission Analysis 14
2.3.2.1 Beginning Quantitative Analysis 14
2.3.2.2 In-Depth Analysis 21
2.3.3 Long-Term and Short-Term Release Calculation 22
2.4 Quantitative Analysis of Surface Water Contamination 22
2.4.1 Beginning Quantitative Analysis 23
2.4.1.1 Dissolved and Sorbed Contaminant Migration 23
2.4.2 In-Depth Analysis 25
2.4.3 Long-Term and Short-Term Release Calculation 27
2.5 Quantitative Analysis of Ground-Water Contamination 29
25.1 Beginning Quantitative Analysis 29
251.1 Leachate Release Rate 29
2.5.2 In-Depth Analysis 31
2.5.3 Long-Term and Short-Term Release Calculation 31
2.6 Soil Contamination 31
2.6.1 Beginning Quantitative Analysis 31
2.6.2 In-Depth Analysis 31
3 CONTAMINANT FATE ANALYSIS 35
3.1 Introduction 35
3.2 Contaminant Fate Screening 36
-------
Table of Contents (Continued)
Chapter Page
3.2.1 Atmospheric Fate 36
3.2.2 Surface Water Fate 38
3.2.3 Soil and Ground-Water Fate 40
3.2.4 Biotic Fate 40
3.3 Quantitative Analysis of Atmospheric Fate 42
3.3.1 Screening Analysis 42
3.3.2 In-Depth Analysis 46
3.3.2.1 Intermedia Transfer 46
3.3.2.2 Intramedia Transformation Processes 47
3.3.2.3 The Effects of Terrain 48
3.3.3 Computer Models 48
3.3.4 Short- and Long-Term Concentration Calculations 48
3.4 Surface Water Fate Analysis 53
3.4.1 Beginning Quantitative Analysis 53
3.4.2 In-Depth Analysis. 55
3.4.2.1 Intermedia Transformation Processes 55
3.4.2.2 Intramedia Transformation Processes 56
3.4.2.3 Computer Models 56
3.4.2.4 Short- and Long-Term Concentration Calculations 57
3.5 Quantitative Analysis of Ground-Water Fate 57
3.5.1 Discussion of Ground Water Modeling 63
3.5.1.1 The Contamination Cycle 63
351.2 Ground Water Flow Conditions 64
3.5.1.3 Multiphase Flow 65
3.5.1.4 Contaminant Flow and Hydrodynamic Dispersion 65
3.5.1.5 Transformation and Retardation 66
3.5.1.6 Higher Velocity Transport 68
3.5.2 Ground-Water Modeling Equations and Nomograph 68
3.5.2.1 Calculating Ground Water Velocity 68
3.5.2.2 Calculating the Velocity of Infiltrating Rainwater 69
3.5.2.3 Corrections for Viscosity and Density 73
3.5.2.4 Retardation Effects 73
3.5.2.5 Contaminant Velocity 75
3.5.2.6 Nomograph Technique 77
3.5.2.7 Extent of Plume 77
3.5.2.8 Use of Monitoring Data 82
3.5.2.9 VMS Model 82
3.5.3 In-Depth Methods and Models 83
3.5.4 Short- and Long-Term Concentration Calculations 93
3.6 Biotic Pathways 93
3.6.1 Estimation Procedures 93
3.6.1.1 Aquatic Animals 94
3.6.1.2 Terrestrial Animals 94
3.6.1.3 Terrestrial Plants 94
4 UNCERTAINTY IN THE ANALYSIS 95
4.1 Sources of Uncertainty 95
4.1.1 Input Variable Uncertainty 95
4.2 Modeling Uncertainty 96
4.2.1 Model Simplification 96
4.2.2 Averaging Hydraulic Conductivities 96
4.2.3 Dispersion Simulation 97
4.2.4 Numerical Models and Analytical Models 97
4.2.5 Chemical Degradation Simulation 97
4.2.6 Model Operational Parameters 97
iv
-------
Table of Contents (Continued)
Chapter Page
4.2.7 Source Shape 98
4.2.8 Steady State Modeling. 98
4.2.9 Number of Dimensions Addressed by the Model 98
4.3 Scenario Uncertainty 98
4.4 Approaches for Dealing with Uncertainty
4.4.1 Sensitivity Appraisals 98
4.4.2 Monte-Carlo Simulations 99
4.4.3 Using Monitoring Data to Calibrate the Model 99
4.5 Level of Uncertainty Appropriate for Exposure Modeling 100
4.6 Risk Communication 100
5 REFERENCES 103
APPENDIX A Analysis of Exposed Human Populations and
Exposure Calculation and Integration 113
APPENDIX 13 Possible Exposure Assessment Data Requirements for Uncontrolled
Hazardous Waste Sites and Index to Variable Terms 135
APPENDIX C Data Management Forms 145
-------
List of Tables
Number Page
1-1 Technical Resource Contacts for Superfund Exposure Assessments 5
2-1 Potential Contaminant Release Mechanisms 8
2-2 Environmental Variables and Model Parameters for
the Wind Erosion Equation 13
2-3 Diffusion Coefficients of Selected Organic Compounds 18
2-4 "C" Values for Permanent Pasture, Rangeland, and Idle Land 26
2-5 "C" Values for Woodland 26
2-6 Runoff Curve Numbers 27
2-7 Parameter Values for Permeation Equation (at 25°C) 32
2-8 Polymer Categorization for Permeation of Water 32
2-9 Permachor Values of Some Organic Liquids in Polyethylene and PVC 32
2-10 Water Permachor Value for Dry Polymers 33
3-1 Key to Stability Categories 45
3-2 Resource Requirements and Information Sources: Atmospheric Fate Models 49
3-3 Features of Atmospheric Fate Models 51
3-4 Data Requirements for Atmospheric Models 52
3-5 Resource Requirements and Information Sources: Surface Water Fate Models .... 58
3-6 Feature of Surface Water Fate Models 61
3-7 Data Requirements for Surface Water Models 62
3-8 Representative Values of Saturated Hydraulic Conductivity 70
3-9 Saturated Hydraulic Conductivity Ranges for Selected Rock and Soil Types 70
3-10 Representative Values for Saturated Moisture
Contents and Field Capacities of Various Soil Types 70
3-11 Representative Values of Hydraulic Parameters
(Standard Deviation in Parentheses) 71
3-12 Suggested Value for Cet Relating Evaporation from a US Class A Pan
to Evapotranspiration from 8 to 15-cm Tall, Well-watered Grass Turf 72
3-13 Crop Coefficients for Estimating Evapotranspiration 73
3-14 Resource Requirements and Information Sources:
Unsaturated Zone and Ground-Water Fate Models 84
3-15 Features of Unsaturated Zone and Ground-Water Fate Models 88
3-16 Data Requirements for Unsaturated Zone and Ground-water Models 91
VI
-------
List of Tables (Continued)
Number Page
A-1 Regional Census Bureau Offices 118
A-2 U.S. Home Fruit and Vegetable Garden Use, 1977 119
A-3 Summary of Human Inhalation Rates for Men, Women, and Children
by Activity Level (m3/hour) 123
A-4 Permeability Constants for Various Compounds 124
A-5 Typical Daily Soil Ingestion Rates for Children by Age Group 129
B-1 Possible Data Requirements for Estimation of
Contaminant Release and Transport and Exposed Populations 136
B-2 Index to Variable Terms 139
VII
-------
List of Figures
Number Page
1-1 Overview of the Integrated Exposure Assessment Process 2
2-1 Contaminant Release Decision Network: Contaminants in Soil 9
2-2 Contaminant Release Decision Network: Contaminants Above-Ground 11
2-3 Mean Number of Days Per Year with > 0.01 Inches of Precipitation
(i.e., "wet days") 15
2-4 Slope Effect Chart Applicable to Areas A-l in Washington,
Oregon, and Idaho, and all of A-3 24
2-5 Soil Moisture-Soil Temperature Regimes of the Western United States 24
2-6 Slope Effect Chart for Areas Where Figure 2-5 is Not Applicable 24
3-1 Environmental Fate Screening Assessment Decision Network: Atmosphere 37
3-2 Environmental Fate Screening Assessment Decision Network: Surface Water 39
3-3 Environmental Fate Screening Assessment Decision Network:
Soils and Ground-water 41
3-4 Environmental Fate Screening Assessment Decision Network: Food Chain 42
3-5 Horizontal Dispersion Coefficient as a Function of Downwind Distance
from the Source 43
3-6 Vertical Dispersion Coefficient as a Function of Downwind Distance
from the Source 44
3-7 Area Within Isopleths for a Ground-Level Source 47
3-8 Nomograph for Solutions of Time, Distance, and Concentration
for Any Point Along the Principal Direction of Ground-water Flow 78
A-1 Exposed Populations Decision Network 115
A-2 Quantitative Exposed Population Analysis 117
Viii
-------
Foreword
The Super-fund Exposure Assessment Manual presents an integrated method
to help Remedial Project Managers and their contractors define the three major
components involved in assessing human population exposure to contaminants
released from uncontrolled hazardous waste sites:
1. Analysis of toxic contaminant releases;
2. Determination of the environmental fate of such contaminants; and
3. Evaluation of the nature and magnitude of exposure to toxic
contaminants.
This report provides guidance for the development of exposure assessments
using monitoring data (which may provide the most dependable basis for
evaluating some existing exposure levels), as well as modeling techniques to
predict exposure over time.
-------
Executive Summary
The analytical process outlined in the Superfund
Exposure Assessment Manual provides a framework
for the assessment of exposure to contaminants at or
migrating from uncontrolled hazardous waste sites.
The application of both monitoring and modeling
procedures to the exposure assessment process is
outlined. This process considers all contaminant
releases and exposure routes and assures that an
adequate level of analytical detail is applied to support
the human health risk assessment process.
The analytical process is structured in five segments:
1. Analysis of contaminant releases from a
subject site into environmental media;
2. Evaluation of the transport and
environmental fate of the contaminants
released;
3. Identification
characterization
populations;
en u meration , and
of potentially exposed
4. Integrated exposure analysis; and
5. Uncertainty analysis.
The Superfund Exposure Assessment Manual
supports the development of exposure assessments
that are consistent from site to site, and provides a
means of documenting that each site receives
adequate evaluation. The procedures presented
reflect current (at the time of publication) state-of-
the-art methods for conducting an exposure
assessment. However, it is important for the analyst
to recognize that exposure assessment is a
developing science. Although the overall protocol for
conducting exposure assessments at Superfund sites
will not change significantly over time and the basic
parameters needed as input to the analysis are not
likely to change, alternative analytical methods may
be developed for many parts of the assessment. The
methods presented in this manual can serve as a
benchmark against which such new methods can be
compared.
XI
-------
Acknowledgments
This document was developed by EPA's Office of Emergency and Remedial Response (OERR).
Dr. Craig Zamuda of OERR's Toxics Integration Branch was the EPA Project Officer. Additional
guidance was provided by Peter Tong and Mary-Virginia Wandless of the Toxics Integration
Branch.
Assistance was also provided by the following people:
Bob Ambrose
Doug Ammon
Brint Bixler
Robert Carsel
Richard Daley
Carl Enfield
Tom Evans
Kevin Garrahan
Mark Garrison
Steve Golian
Karen Hammerstrom
Seong T. Hwang
Joe Keeley
Ashok Kumar
Steve Ostrodka
Zubair Saleem
Paul Schumann
James Spatarella
Richard L. Stanford
Sherry Sterling
David Tetta
Louis J. Thibodeaux
Jawed Touma
Georgia Valaoras
Paul K.M. van der Heijde
Larry Zaragoza
ORD (Office of Research and Development)
Clean Sites, Inc. (formerly USEPA)
CH2M Hill (formerly USEPA)
ORD (Office of Research and Development)
OWPE (Office of Waste Programs Enforcement)
ORD (Office of Research and Development)
ORD (Exposure Assessment Group)
ORD (Exposure Assessment Group)
USEPA Region III
OERR (Office of Emergency and Remedial Response)
OTS (Office of Toxic Substance)
ORD (Office of Research and Development)
Oregon Graduate Center
University of Toledo
EPA Region V
OSW (Office of Solid Waste)
OSW (Office of Solid Waste)
Versar, Inc. (formerly USEPA)
Roy F. Weston, Inc. (formerly USEPA)
OWPE (Office of Waste Programs Enforcement)
EPA Region X
University of Arkansas
OAQPS (Office of Air Quality Planning and Standards)
OWPE (Office of Waste Programs Enforcement)
Holcomb Research Institute
OSWER (Office of Solid Waste and Emergency Response)
Versar, Inc. assisted OERR in the development of this document in fulfillment of Contract Nos.
68-01-6271, 68-03-3149, and 68-01-7090. The Versar project team included H. Lee
Schultz, Walter A. Palmer, Mark L. Mercer, Ruth A. Dickinson, Gary Whitmyre, Alan F. Gleit, Gina
H. Dixon, and Van Kozak (currently Texas Department of Agriculture).
XII
-------
Chapter 1
Introduction
1.1 Purpose
The Superfund Exposure Assessment Manual
provides Remedial Project Managers (RPMs) with the
guidance necessary to conduct exposure
assessments that meet the needs of the Super-fund
human health risk evaluation process. Specifically,
the manual:
1. Provides an overall description of the integrated
exposure assessment process as it is applied to
uncontrolled hazardous waste sites; and
2. Serves as a source of reference concerning the
use of estimation procedures and computer
modeling techniques for the analysis of
uncontrolled sites.
This manual provides guidance for the assessment of
human population health risk only. Guidance for
ecological risk assessment will be provided
separately.
1.2 Background
The Comprehensive Environmental Response,
Compensation, and Liability Act of 1980 (CERCLA -
42 USC 9601 et. seq.), as amended by the
Superfund Amendments and Reauthorization Act of
1986 (SARA), was enacted to provide the Federal
Government with the authority to respond to releases
or threatened releases of hazardous substances,
pollutants, or contaminants into the environment. As
prescribed in the revised National Contingency Plan
(see 47 FR 137, July 16, 1982), all sites designated
for in-depth evaluation are included on the National
Priorities List. These sites are evaluated for remedial
action through the application of a Remedial
Investigation, which defines the nature and extent of
contamination, and a Feasibility Study, in which
potential remedial alternatives are developed and
analyzed. Guidance for conducting these two major
components of the remedial response process is
provided in USEPA (1985a and 1985b, respective-
ly - currently under revision). As discussed in that
guidance, a part of the Feasibility Study is the
development of a risk assessment that projects those
health impacts resulting from the uncontrolled site.
The risk assessment is based on the results of a site
exposure assessment, which evaluates:
1. The type and extent of contaminant release from
a site to environmental media;
2. The environmental transport and transformation of
contaminants following release; and
3. Implications of the resulting contact with exposed
populations.
Section 110 of SARA mandates that health
assessments be conducted by the Agency for Toxic
Substances and Disease Registry for all sites on the
National Priorities List. These health assessments can
be based on the results of site-specific exposure
assessments. The exposure assessment, therefore, is
an analytical tool that is used to comply with the
mandates of CERCLA.
1.3 Scope
This manual provides guidance for the use (but not
the acquisition) of field data in the exposure
assessment process. It does not serve as an all-
encompassing guide to the use of computer models
in the site remediation process, or direct the analysis
of health risks that result from predicted exposure.
This manual is intended to be used in conjunction
with other related guidance, such as that for the
acquisition of field data. As detailed in USEPA
(1987a), field sampling Data Quality Objectives
(DQOs) establish a phased sampling strategy
designed to guide the efficient acquisition of field data
for site-specific exposure and public health
assessments, and provide sampling plan guidance
addressing the location of sampling points. Field
operating procedures for obtaining and handling
samples have also been developed (USEPA 1987b).
Other references, (USEPA 1986a, 1986b, 1987c, and
1987d), address the utility, applications, and
limitations of computer models for predicting
contaminant concentrations and transport through
various environmental media. The process for
-------
developing a human health risk assessment for
Superfund sites has been detailed in USEPA (1985c).
When conducting a comprehensive risk assessment,
the analyst will need to refer to all of the above-cited
guidance. While none of these guidance manuals
stands alone, taken as a whole, they provide an
overall, integrated approach to analysis of site
contamination and health risk.
1.4 Use of the Manual
This manual is used to apply state-of-the-art
exposure assessment procedures to the unique
analytical needs of uncontrolled hazardous waste
sites. The ultimate goal of human exposure
assessment at Super-fund sites is the determination of
the type and magnitude of potential exposure to
contaminants present at and migrating from the site.
To achieve this goal, many sites may require a mix of
qualitative and quantitative exposure analysis. The
latter may range from simple analytical techniques
(e.g., contaminant release or dispersion estimation
equations) to more complicated computer modeling
approaches.
The general procedure for conducting an integrated
exposure analysis is illustrated in Figure l-l. This
procedure is based on EPA's published Guidelines for
Exposure Assessment (USEPA 1986c) and other
related guidance (USEPA 1985d-i) and is an
adaptation of that process to the analytical problems
posed by abandoned hazardous waste sites. As
previously mentioned, target chemicals are selected
as part of the human health risk assessment process
(see USEPA 1985c). Once these chemicals are
chosen, the exposure assessment proceeds through
the following stages:
. Contaminant Release Analysis
Each on-site release point is identified for every
target chemical, and the level of release (mass
loading) to each environmental medium is
determined. Determination of contaminant
release may be made either by direct
measurement (monitoring) of such releases or
by estimation. Although difficult to achieve for all
media, monitored release values provide a more
sound basis for projection of contaminant
migration later in the exposure assessment
process than do modeled estimates. When it is
not possible to obtain measured release rates,
estimates can be based on measurements of
contaminant concentrations in pertinent source
media (e.g., estimates of contaminant release to
ground water based on measured concentrations
in contaminated soil). The results of the
Figure 1-1.
Overview of the integrated exposure
assessment process.
Evaluation of Contaminant
Properties and Selection of
Target Chemicals1
Contaminant Release
Analysis (Multimedia)
— Monitoring Data
— Modeling Estimates
(Chapter 2)
Contaminant Transport and Fate Analysis
— Exposure Pathways
— Environmental Distribution
and Concentrations
• Monitoring Data
• Modeling Estimates
(Chapter 3)
Exposed Populations
Analysis
(Appendix A)
Integrated Exposure
Analysis
(Appendix A)
Uncertainty Analysis
(Chapter 4)
'Part of Human Health
Risk Assessment Process.
Refer to USEPA (1985c).
contaminant release analysis provide the basis
for evaluating the potential for contaminant
transport, transformation, and environmental
fate.
2. Contaminant Transport and Fate Analysis
This analysis describes the extent and
magnitude of environmental contamination (i.e.,
contaminant concentrations in specific
environmental media). When possible, direct
measurement of contaminant concentrations is
preferred, and collection of samples during site
evaluation will provide a clear basis for
determining exposure potential for some
exposure routes. However, the human health
risk assessment process also requires projection
of potential exposure over a lifetime (see Section
1-5), which can only be accomplished using
estimation procedures.
-------
3. Exposed Populations Analysis
The results of contaminant transport and fate
analysis allow the analyst to evaluate populations
contacting chemicals emanating from the site.
Analysis of exposed populations involves the
identification, enumeration, and characterization
of those population segments likely to be
exposed. The goal of this analysis is not only to
delineate those populations coming into contact
with contaminants emanating from the site, but
also to determine how and with what frequency
and duration such contact occurs.
4. Integrated Exposure Analysis
In this step, the individual chemical-specific
exposure estimates for each exposure route
(i.e., inhalation, ingestion, and dermal contact)
are developed. For each exposed population, all
exposures to each hazardous substance are
identified. In cases in which a population group
experiences more than one exposure by a given
route, exposures are summed to develop a
cumulative exposure value for the route
involved. For example, persons who reside in
the vicinity of a Superfund site may experience
dermal exposure to a given contaminant directly
on site as well as directly through basement
seepage, and exposures via both of these routes
should be summed for exposure integration
purposes.
5. Uncertainty Analysis
The exposure assessment concludes with an
analysis of uncertainty. In this analysis each step
in the assessment is reviewed to identify any
uncertainties involved and to evaluate their
separate and cumulative impact on assessment
results. Uncertainties may result from the use of
default values for analytical input parameters,
from the use of simplified estimation procedures
as opposed to more rigorous computer analysis
or monitoring-based analysis, from an inability
to define exposed populations with confidence,
etc. The uncertainty analysis provides necessary
input for remedial decisionmakers who must
evaluate the results of the exposure assessment
with regard to their implications for potential risks
associated with the uncontrolled site and
appropriate remedial alternative selection
This manual is intended to be used in conjunction
with various other guidance to conduct Superfund site
Remedial Investigations and Feasibility Studies. The
use of this manual is particularly linked to the Public
Health Evaluation Manual. The two are intended to be
used as two parts of the same process: the analysis
of health impacts resulting from uncontrolled
hazardous waste sites. In conducting a Superfund
evaluation of exposure and public health impact, the
analyst initially applies the indicator chemical selection
process" outlined in the Superfund Public Health
Evaluation Manual to select the chemicals on which
the site analyses will focus. Once the chemicals have
been selected, the analytical framework of the
Superfund Exposure Assessment Manual is applied.
Following completion of the exposure assessment,
the analyst returns to the Superfund Public Health
Evaluation Manual for guidance in determining the
degree of human health risk for each exposed
population.
The user of this manual should understand that these
analytical procedures are intended to be applied
site-specifically. No two sites will be exactly alike in
terms of the extent and complexity of contamination,
of contaminant migration, or of potentially exposed
populations. Therefore, the specific analytical
procedures to be applied in all Superfund exposure
assessments cannot be fixed in general. Instead the
approach and methods applied to conducting an
exposure assessment must be tailored to address
existing site conditions. In some situations
contaminant releases or exposure routes may be
adequately addressed by applying only screening
procedures. In other cases more complex,
quantitative evaluation will be necessary.
The Superfund Public Health Evaluation Manual
(USEPA 1985c) lists five factors affecting the degree
of analytical complexity for site analyses:
1. Number and identity of chemicals present;
2. Availability of appropriate standards and/or toxicity
data;
3. Number and complexity of exposure pathways
(including complexity of release sources and
transport media);
4. Necessity for precision of the results; and
5. Quality and quantity of available monitoring data.
Simplified analyses may be used in the following
instances: only a small number of chemicals must be
evaluated; environmental standards or criteria for
chemicals under study are available; a small number
of exposure pathways are present; release and
transport processes are relatively simple: or there is a
limited need for detail and precision in the
assessment results (e.g., screening studies).
Conversely, sites that have many contaminants for
which no environmental standards or criteria are
available, that exhibit multiple exposure pathways,
that have complex contaminant release and transport
processes in effect, or that require analytical results
in great detail and precision will require more
-.Selection of indicator chemicals will be required only at those
sites where the number of contaminants present is too large to
individually evaluate exposure to each.
-------
complex, quantitative analytical methods. Most sites
will fall somewhere between these two extremes.
obtain and review the original source
documentation cited for analytical components.
Procedures presented in this manual for conducting
quantitative analyses include both simplified "desk
top" approaches for developing order-of-magnitude
estimates and more resource-intensive, in-depth
approaches. Computer modeling and site monitoring
are included. Generally, it is appropriate to apply
simplified analysis to all pertinent exposure routes at
the beginning of quantitative evaluations so that those
causing greatest concern can be identified for
subsequent in-depth analysis.
It is important to understand that analysis of exposure
and resultant health impact is often a complex
process in which selection and application of the
most appropriate analytical tools, as well as the
insightful interpretation of their results, can be critical.
The U.S. EPA encourages ongoing communication
between site analysts and experts in various exposure
and health impact assessment fields. Thus, when
questions arise regarding the utility of a particular
model or mathematical solution, it is recommended
that the analyst review the pertinent sections
described in this manual or contact the Toxics
Integration Branch of the Hazardous Site Evaluation
Division of the Office of Emergency and Remedial
Response (FTS 475-9486). In addition, Table 1-1
lists specific EPA contacts who can provide insight
into particular site assessment problems.
In developing this manual, an attempt was made to
compile analytical methods appropriate for assessing
exposure to chemicals migrating from uncontrolled
hazardous waste facilities. There are limitations to the
application of these analytical tools and to the
interpretation of the results obtained, including:
1. While some of these tools have been developed
specifically for application to Superfund sites,
others were originally developed for different
purposes and had to be adapted or directly
applied to evaluation of conditions present at
uncontrolled hazardous waste sites. The analyst
must be careful in interpreting the results
obtained from application of these tools and must
consider their inherent uncertainties.
2. This manual assumes that the analyst has a
strong technical background in engineering or the
sciences. This background is essential to ensure
that analyses are carried out in a technically
sound fashion and that interpretations of the
results obtained are realistic.
3. It was not possible to include discussion of all
technical limitations and caveats pertaining to
each analytical tool or procedure reviewed in this
manual. It may be beneficial for the analyst to
4. Results obtained through application of these
tools must be interpreted based on conditions at
the site being analyzed. These tools are provided
to aid the analyst in making decisions, not to
make decisions for the analyst. When possible,
models used in analyzing a given site should be
verified with field monitoring data that test and
validate model predictions at that site.
5. The approach to conducting exposure
assessments outlined in this manual is
conservative as are human health risk studies.
However, the analyst needs to be sensitive to and
to compensate, at least qualitatively, for the
additive effect of multiple conservation
assumptions. The degree of conservatism should
not be so extreme that the conclusions drawn
from the analysis are unrealistic.
1.5 Timeframe of Analysis
Quantitative exposure assessments generate
estimates of the long-term (chronic daily intake) and
short-term (subchronic daily intake) exposure to
contaminants. The output of each analytical
component (contaminant release, environmental fate,
etc.) must be expressed in the same long-term and
short-term form. Long-term releases are defined as
the release rates of each contaminant migrating from
the site averaged over an assumed 70-year human
lifetime. Short-term contaminant releases are
defined (USEPA 1985c) as those that occur over a
short period (usually 10 to 90 days) during the first
year following site investigation.
1.6 Analysis of Exposure Associated with
Remedial Actions
The analytical tools presented in this Superfund
Exposure Assessment Manual are those appropriate
for analyzing exposure associated with the baseline
condition (i.e., the uncontrolled site prior to
implementation of any remedial action). It should be
noted, however, that waste treatment processes used
as part of a remediation strategy can themselves
contribute significant releases of contaminants to the
environment. Stripping volatiles from wastewaters, for
example, generally involves artificial acceleration of
the natural volatilization process, resulting in forced
transfer of the volatile contaminants from wastewater
to air. Thus, analysts must evaluate the engineering
design of each remedial alternative to determine the
level of contaminant release associated with its
implementation. The user of this manual should refer
to Farino et al. (1983) for a discussion of methods to
estimate wastewater treatment air emissions. When
incinerating toxic wastes other than those containing
PCBs, Destruction and Removal Efficiency (ORE)
-------
Table 1-1. Technical Resource Contacts for Superfund Exposure Assessments
Office
Commercial
phone number
FTS phone number
I. U.S. Environmental Protection Agency:
Office of Air Quality Planning and Standards; Research Triangle Park, N.C. (919) 541-5381 629-5381
Office of Toxic Substances; Washington, D.C. (202)382-3886 382-3886
Office of Research and Development, Exposure Assessment Group; Washington, D.C. (202) 475-8919 475-8919
Office of Research and Development, Hazardous Waste Engineering Research
Laboratory; Cincinnati, Ohio (513) 569-7418 684-7418
Environmental Research Laboratory; Ada, Okla. (405) 332-8800 743-2011
Environmental Research Laboratory; Athens, Ga. (404) 546-3134 250-3134
Center for Exposure Assessment Modeling; Athens, Ga. (404) 546-3585 250-3546
II. Centers for Disease Control:
Agency for Toxic Substances and Disease Registry; Atlanta, Ga. (404) 454-4593 236-4593
III. International Ground Water Modeling Center:
Holcomb Research Institute, Butler University; Indianapolis, Ind. (317) 283-9458
requirements can be found in 40 CFR 264.343
(Environmental Protection Agency Regulations for
Owners and Operators of Permitted Hazardous Waste
Facilities; Subpart 0 - incinerators). For incineration
of wastes contaminated with PCBs, the analyst can
refer to 40 CFR 761.70 (Polychlorinated Biphenyls
(PCBs) Manufacturing, Processing, Distribution in
Commerce, and Use Prohibitions - Incineration).
Well engineered remedial alternatives planned for
uncontrolled hazardous waste sites are not expected
in themselves to cause additional releases of toxic
contaminants to ground-water systems. Even if an
unexpected spill of toxics occurs when remedial
action is taken, contaminant release should be slow
enough to allow spilled substances to be isolated
prior to their reaching the saturated zone. Short-
term release of contaminants to air may occur while
excavating contaminated soil and loading it for
removal from the site. In such situations, the analyst
should refer to USEPA (1983a), for release equations
for material transfer.
The effectiveness of contaminant control, however,
may vary among different remediation technologies.
To evaluate post-remediation control effectiveness,
many of the analytical procedures presented in this
manual may be useful. For example, reductions in
contaminant releases can be estimated by
recalculating releases using altered (from the baseline
case) site-specific input variables based on the
remedial action under consideration. Alternatively, one
can obtain a rougher approximation by applying the
expected remedial action percent control (based on
engineering experience) to the source release
estimates calculated for the baseline case. In
addition, the analyst should refer to USEPA (1985j)
for a detailed discussion of both simplified analytical
methods and numerical modeling approaches that can
be used to estimate remedial effectiveness.
1.7 Organization of the Manual
The following chapters of this manual detail methods
for evaluating exposure to chemicals migrating from
Superfund sites. The body of the manual provides
guidance for the qualitative and quantitative evaluation
of contaminant release, migration, and fate in the
environment, along with that for evaluating uncertainty
in the analysis. Procedures for conducting exposed
populations analysis and for developing an integrated
exposure analysis are provided in appendices to this
report.
-------
Preceeding Page Blank
Chapter 2
Contaminant Release Analysis
2.1 Introduction
This chapter provides guidance for the analysis of
contaminant releases from uncontrolled hazardous
waste sites. The goal of this analysis is to determine
contaminant release rates to specific environmental
media over time. The following sections address the
release of contaminants to air, surface water, and
ground water from wastes placed both above-
ground and below-ground. In particular, guidance is
provided for the evaluation of the following categories
of contaminant releases:
1. Air releases:
a. Fugitive dust resulting from:
Wind erosion of contaminated soils
Vehicular travel over contaminated
unpaved roadways
b. Volatilization releases from:
Covered landfills (with and without internal
gas generation)
Spills, leaks, and landfarming
- Lagoons
2. Surface water releases: contaminated runoff
3. Ground-water releases:
a. Landfilled solids (lined or unlined)
b. Landfilled liquids (lined or unlined)
c. Lagoons (lined or unlined).
Contaminant release analysis is conducted in two
stages - screening of contaminant release
mechanisms and quantitative analysis. The screening,
which is a qualitative evaluation of site conditions,
identifies each potential contaminant release source,
determines the environmental media affected by each
release, and broadly defines the possible extent of
the release. The following section is designed to
establish a consistent basis for the qualitative
screening of contaminant release from site to site.
Once the potential sources of on-site contaminant
release have been screened, those requiring further
evaluation are quantitatively analyzed. This may
involve the application of a simplified "desk-top"
estimation approach, or a more in-depth method
such as computerized modeling or additional site
monitoring. The goal of this analysis is to generate
release rate estimates (mass per unit time) for each
source of contaminant release. Release rate values
are necessary as input for subsequent environmental
fate analysis (see Chapter 3). Individual on-site
releases of each contaminant are summed to
generate an overall, medium-specific release rate for
each chemical migrating from the site. Short-term
(worst-case) release rates are developed, as are
long-term rates (averaged over 70 years).
The simplified estimation procedures that follow allow
the analyst to make release approximations based on
chemical- and site-specific factors. However, these
calculations do not take into account the full range of
variables that affect on-site contaminant release.
These approaches (with one exception) assume
steady state conditions. They do not directly address
the reduction in contaminants present (due to release
losses), or the associated reduction in release loading
over time corresponding with the decreasing
contaminant reservoir.*
When possible, monitoring should be used to quantify
rates of contaminant release. In some cases,
however, this may not be feasible because methods
to directly measure releases from certain settings are
still being developed. Moreover, it may not always be
possible to monitor contaminant releases under the
conditions of concern (e.g., dust releases under high
wind conditions, surface water runoff releases during
storm events, etc.). It may often be necessary to
estimate release rates in the exposure assessment
process. All of the release rate estimation procedures
presented here, however, require some monitored
values as input. (Examples are measured contaminant
concentrations in soil, soil characteristics.) The
analyst should be aware of the need to develop a
monitored data base that is adequate to support the
needs of the contaminant release analysis portion of
the exposure assessment.
In general, the procedures to estimate the rate of
contaminant release are complete. When analyzing
Estimation of the variation in the level of release over time is
calculated separately. See Long-Term and Short-Term
Release calculation subsections in this chapter.
-------
wind erosion releases, however, the analyst should
consult other published guidance that addresses the
application of the wind erosion equation in various
regions of the country. Depending on the location of a
particular site, one of the following three manuals will
be necessary:
Craig and Turelle (1964): Great Plains
- Haynes (1966): Northeast
- Skidmore and Woodruff (1968): entire
United States.
2.2 Contaminant Release Screening
The manner of waste placement at an abandoned site
determines whether contaminant release* occurs by
any or all of the mechanisms summarized in Table
2-1. In contaminant release screening, the likelihood
of release from each source, the nature of the
contaminants involved, and the probable magnitude of
their release (relative to other on-site sources) are
considered.
Figures 2-1 and 2-2 present the decision networks
that guide contaminant release screening analysis.
Figure 2-1 deals with contaminants in or under the
soil and Figure 2-2 addresses above-ground
wastes. Any release mechanisms evident at the site
will require a further screening evaluation to
determine the likely environmental fate of the
contaminants involved (see Chapter 3).
2.2.7 Contaminants in Soil (see Figure 2-1)
The following numbered paragraphs help to interpret
and apply the steps of the contaminant release
decision network presented in Figure 2-1. Each
paragraph refers to a particular numbered box in the
figure.
1. Most uncontrolled hazardous waste sites will
exhibit some degree of surface or subsurface soil
contamination. This contamination may be the
result of intentional waste disposal underground
(landfilling) or in surface soils (surface application
or landfarming), or it may be caused by
unintentional waste releases from spills or leaks.
2. Landfilled wastes may become mobile if they are
not contained in impervious containers, or if the
containers are leaking. Release of such wastes
may contaminate subsoils, ground water (through
percolation), or air (through volatilization).
3. Landfilled wastes will be covered with soil;
however, soil cover will not necessarily isolate
wastes from the environment. If the cover can be
penetrated by rainwater or run-on, wastes can
be leached from the landfill cells and contaminate
subsoils, ultimately reaching ground water.
Similarly, the soil cover may not be deep enough
to prevent the migration of volatile contaminants
into the atmosphere. Estimations are that 60
percent of hazardous waste is in liquid (sludge)
form (USEPA 1980a). Infiltrating rainwater can
increase the migration rate of liquid or semiliquid
materials by increasing the hydraulic head
affecting them, as well as by the leaching of toxic
components. Such factors as erosion or extreme
drying (and cracking) can reduce the ability of a
soil or clay cover to maintain the isolation of
wastes. Also, contaminated soil may cover the
waste cells themselves. When evaluating the
potential for landfill releases, current conditions,
along with the long-term integrity of the landfill
and its soil cover, should be evaluated. If the
landfill soil cover does not assure long-term
For the purposes of this manual, contaminant "release" is
defined as any process that results in migration of contaminants
across the site boundary. Within this context, volatilization,
generation of surface runoff, or leachate, are considered to be
release mechanisms. Contaminant transport equates with those
processes that carry released contaminants to points distant
from the site.
Table 2-1. Potential Contaminant Release1 Mechanisms
Process
Media directly affected
(media indirectly affected)
Timeframe
Volatilization
Overland flow2
Direct discharges
Leachate generation4
Fugitive dust generation5
Generation of surface runoff
Combustion3
Air
Soils, surface water (ground water)
Soils, surface water (ground water)
Soils, ground water
Air
Soils, surface water (ground water)
Air
Chronic
Chronic, episodic
Chronic, episodic
Chronic
Chronic, episodic
Chronic, episodic
Episodic
1See Section 2.2 for a definition of contaminant "release" as used in this manual.
Impoundment overflow/failure, drum leakage, etc.
Includes on-site treatment releases (e.g., wastewater/runoff treatment, incineration).
4Buried wastes, wastes stored above ground (leaks), land application, lagoons.
Contaminated soils, particulate wastes.
-------
Figure 2-1. Contaminant release decision network: contaminants in soil.
Are Toxics Present In:
Soil?
t
Are Toxics La ndfilled?
15
*
Are Toxics Spilled, Leaked,
or Surface Applied? n
t
Is Site Accessible?
E
Does Soil Cover
Prevent Percolation
of Precipitation? f
CD
Consider Long-
Term Integrity
of Soil Cover [3]
Is Leaching
Release to Subsurface
Soils or Ground Water
Possible?
Consider Long
Term Integrity
of Soil Cover Ig
Is Leakage of
Containerized Liquid
Waste to Ground Water
Possible?
Is Volatilziation"
Release to Air
Possible?
Is Runoff Release
to Soils, Surface
Water, Ground Water or
Air (Via Volatilization)
Possible? |
Does Soil Cover
Prevent Vapor
Release to Air? R
.
Is Surface
Soil
Contaminated? R
_
E
Is Soil'
Cover
roding? fj
Is Release to
Ground Water
(Leaching) Possible?
Is Releas
or Surfa
(Runoff)
a to Soils
:e Water
Possible? [5
Is Fugitive Dust
Release to Air
Possible? |e
Is Volatilization
Release to Air
Possible?
Go on to Environmental
Fate Analysis for Contaminants
Released Via Each Existing
or potential Release Mechanism
-------
isolation of the wastes, one should evaluate the
leachability and volatilization potential of the
landfilled wastes.
4. At some hazardous wastes sites, toxic materials
may have been purposefully incorporated into
surface soils to promote their microbial
destruction. In such cases, toxic components may
still remain in the soil. At most sites, however,
surface soils have become contaminated because
of hazardous material spills or leaks during
manufacturing, processing, storage, or transfer
operations. In these situations, the potential for
release of contaminants in surface soils through
four mechanisms should be evaluated. These
mechanisms are: (1) release of volatile
components to the atmosphere (via evaporation);
(2) release of toxic particulate matter (via wind
erosion); (3) surface runoff -related releases; and
(4) percolation of contaminants or leachate to
ground water.
5. The percolation of contaminated runoff may
contaminate surface soils and underlying ground
water. Surface water systems may be similarly
degraded by contaminated runoff inflow. Runoff
may also serve as a source of volatilization
release to air, although releases from
contaminated soils would be expected to
constitute a greater threat than that from
contaminated runoff. Hydrophobic wastes may
contaminate surface waterbodies by adsorbing
onto soil material that can be eroded from the site
and enter a waterbody in surface runoff. In a
waterbody, sediment transport is much slower
than is water movement, and contaminated
sediments may remain in the vicinity of the
contamination source for a long time.
6. Under high wind conditions, wind erosion may
carry solid particulate wastes or soil particles with
sorbed hydrophobic toxic materials from the site.
7. If the site is accessible, direct contact with
contaminants may occur. Also children may
ingest some contaminated soil during play. Such
ingestion may result from "pica" behavior (i.e.,
intentional eating of soil by very young children)
or from normal hand to mouth contact.
2.2.2 Contaminants Above-Ground
The following numbered paragraphs help to interpret
and apply the steps of the contaminant release
decision network presented in Figure 2-2. Each
paragraph refers to a particular numbered box in the
figure.
1. Wastes can be stored above-ground in
lagoons/ponds, in containers (drums, tanks), or in
piles. Unless containers effectively isolate wastes
from the environment, above-ground storage
can result in the direct introduction of
contaminants into air, soils, surface water, or
ground water.
2. Lagoons may introduce hazardous materials to
the environment through a number of pathways.
Erosion or overflow resulting from heavy rainfall
can breach the lagoon and result in the outflow of
liquid wastes that contaminate surface soils,
ground water, and surface waterbodies. In
addition, unlined lagoons may introduce toxics
directly into ground water via percolation through
the lagoon bottom. Also, because lagoons are
uncovered, the release of volatile compounds to
the atmosphere is a common problem.
3. Wastes stored above-ground in containers may
not be effectively isolated from the environment.
Over time, container corrosion and leakage occur.
Leaked wastes will contaminate soils in the
storage area; may percolate to ground water; or
may contaminate surface runoff, which, in turn,
can extend the area of soil contamination or can
enter local surface waterbodies. Leaked materials
may also evaporate into the atmosphere.
4. If the site is accessible to the public, direct
contact with contaminants may occur. Also,
children may ingest contaminated soils, either
inadvertently or as a result of pica behavior.
2.3 Quantitative Analysis of Atmos-
pheric Contamination
2.3.1 Fugitive Dust Emission Analysis
Emissions of contaminated fugitive dusts (airborne
wastes or contaminated soil particles) originating at
uncontrolled hazardous waste sites can result from a
combination of such factors as (1) wind erosion of
wastes and contaminated soils, and (2) vehicles
traveling over contaminated, unpaved roads.
Methods for analyzing such contaminant releases are
presented below.
2.3.1.1 Beginning Quantitative Analysis
The following procedures are useful in estimating total
fugitive dust releases likely to result from the two
factors cited above. Once total suspendible dust
generation levels have been calculated using these
equations, one can project the amounts of hazardous
substances expected to enter the atmosphere in
fugitive dust using either of the following approaches:
*es Multiply the amount of dust generated by the
weight percent of the toxic substance in soil or
waste. This approach does not take into account
factors relating to such aspects as particle size or
adsorption potential, which can affect the amount
10
-------
Figure 2-2. Contaminant release decision network: contaminants above-ground.
Are Toxics Present in:
Lagoon, Above-Ground
Storage? ry
Yes
No
Is Lagoon Overflow
or Breaching
a Potential Threat?
No
Is Direct Overland Flow
of Toxics to Soils or
Surface Water of Indirect
Percolation to Ground Water
Possible?
Is Volatilization
Release to Air from
Lagoon Possible?
•
•
Yes
i
i
No
[2
Is Lagoon Lined
(i.e., Preclude
Leakage)?
Is Leakage a
Potential Threat?
Do Containers
Isolate Wastes
(i.e., Preclude
Leakage)? [3
Is Site
Accessible?
L J.
Yes
Is Soil,
Ground water,
or Runoff
Contamination
Possible?
Yes
Is Percolation
to Ground Water
Possible?
Yes
Yes
No
Is Gaseous
Release to Air
Possible?
No
Yes
Consider Direct
Contact with
Contaminants
No
Go on to Environmental
Fate Analysis for Contaminants
Released Via Each Existing
or Potential Release Mechanism
-------
of contaminant actually entering the atmosphere
as dust.
Multiply the estimates for total dust generation by
percentages (by weight) of the substances of
concern in actual fugitive dust samples obtained
with on-site air monitoring. This approach takes
into account those chemical-specific and site-
specific factors that affect release of
contaminated dust in the field.
(1) Wind Erosion Analysis*
Wind erosion of agricultural soils and, by
extrapolation, other disturbed soils, depends upon a
variety of factors. These include surface roughness
and cloddiness; surface soil moisture content, kind,
amount (and orientation, if applicable) of vegetative
cover: wind velocity; and the amount of soil surface
(length) exposed to the eroding wind force. The U.S.
Soil Conservation Service (SCS) has developed a
method to estimate wind erosion based on a series of
graphs relating variables presented below. The
graphical method for calculating wind erosion based
on the functional relationship of these variables is not
presented in this manual; instead, the analyst is
directed to the Skidmore and Woodruff (1968) source
document.
E = f(l', C', K', L', V)
where
(2-1)
E = potential annual average wind erosion soil
loss, (mass/area/time).
soil credibility index, (dimensionless).
climatic factor, (dimensionless).
K' = soil ridge roughness factor, (dimensionless).
L' = field length along the prevailing wind direction,
(feet).
v = vegetative cover factor, (dimensionless).
Multiplying E times the contaminated area will yield a
release rate in units of mass per time.
Table 2-2 identifies the factors that determine the
values of the five variables used in the SCS equation.
Note that the vegetative cover factor (V) specifically
applies to crop residues, and care must be taken
when extrapolating to the cover conditions present at
uncontrolled waste sites. For Remedial Investigation
and Feasibility Study estimation purposes, one can
use a "zero pounds per acre" vegetative cover value.
This assumes a worst-case situation (from a
vegetation-related wind attenuation perspective) and
Applied to nonadhering, noncompacted contaminated soil or
waste materials.
provides a conservative estimate of contaminated
fugitive dust release.
A series of publications issued by the U.S.
Department of Agriculture provides directions for
applying this equation to a site-specific situation.
Craig and Turelle (1964) present estimation
procedures for the Great Plains; Haynes (1966)
addresses the Northeast; and Skidmore and Woodruff
(1968) offer procedures for the entire nation.
Although it is strongly recommended that site-
specific soils data be obtained for each site under
evaluation, it is not necessary to do so in order to
obtain parameter values for use with the wind erosion
equation (or other fugitive dust generation equations).
Instead, when necessary, soils data can be obtained
from the local Soil Conservation Service office. SCS
has on record a range of pertinent soils data for sites
across the country where soil surveys have been
conducted. In addition, SCS maintains an extensive
computerized soil properties data base called the
Soils 5 File. This data base lists estimated soils data,
based on surveys of surrounding soils properties, for
areas where surveys have not been conducted to
date. These data are readily available from local SCS
officials. Users of this manual should consult SCS to
obtain more detailed information regarding the nature
and accessibility of information contained in the soil
surveys and the Soils 5 File.
The SCS wind erosion equation is one of a number of
approaches for estimating particulate release from
abandoned hazardous materials facilities. One such
source (Cowherd et al. 1985) is specifically designed
to guide rapid (less than 24 hours) evaluation of the
potential degree of particulate emission from
uncontrolled hazardous waste sites. This method
uses an emission factor approach to estimate release
and procedures adapted from computerized
dispersion models for approximating concentration
isopleths. Concentration estimates and Bureau of the
Census data are used to identify the exposed
population and estimate the level of exposure. This
approach includes the three key components of
exposure analysis: release rate estimation,
contaminant migration analysis, and population
exposure determination. However, Cowherd et al.
(1985) caution that their method is designed for
emergency evaluations or as a preliminary
assessment tool, which may then be used in
undertaking a more detailed investigation.
Nevertheless, the degree of accuracy attained using
this method is consistent with simplified quantitative
estimation procedures. This approach provides the
analyst with estimates of short-term (worst-case,
24-hour) release and exposure estimates, as well as
long-term (average annual) estimates.*
The SCS wind erosion equation is designed to
provide annual erosion losses only, and cannot be
12
-------
Table 2-2. Environmental Variables and Model Parameters for the Wind Erosion Equation
Equivalent SCS wind erosion equation primary wind erosion variables
Parameters
Soil erodibility index, I (function of soil particle size distribution; read
from a table)
Knoll erodibility, ls (function of knoll slope steepness; read from a
graph)
Surface crust stability, Fs
Soil ridge roughness, K,., (function of height, width, and spacing of
clods and furrows)
Annual average wind velocity, v (read from map)
Surface soil moisture, M (estimated using Thornthwalte's (1931)
precipitation-evaporation index)
Distance across field, Df (field width in direction of primary erosive
wind)
Sheltered distance, Db (calculated from barrier height upwind of field)
Quantity of vegetative cover, R' (mass of standing or fallen vegetative
residue per unit area)
Kind of vegetative cover, S (factor related to erosron-reducing
effectiveness of residues from different crops)
Orientation of vegetative cover, KQ (factor relating erosion reduction to
standing vs. fallen crop residues)
Soil and knoll erodibility, I' (equal to I x I,)
Disregarded-crust is transient
Soil ridge roughness factor, K (estimated by comparison to a set of
standard photographs Included in SCS wind erosion equation
users' manuals)
Local wind erosion climatic factor, C' (may be calculated but commonly
read from maps of C')
Field length, U (the difference between Df and Db)
Equivalent vegetative cover, V (the product of R', S, and KJ - can
often be assumed = 0 for abandoned waste sites (see text)
Source: Smith et al. 1982.
reliably altered to generate short-term estimates.** In
addition, it cannot be used with data delineating
climatic extremes for a given location, but must be
based on average annual climatic data. Instead, for
exposure assessment purposes the short-term
release, estimated using the wind erosion equation, is
assumed to equal the average release over the first
year following site investigation.
The user of this manual should review Cowherd et al.
(1985) and compare that method with the SCS wind
erosion procedure before selecting an analytical
approach for estimation of particulate contaminant
release and related exposure. The analyst can also
refer to USEPA (1983b), Farino et al. (1983) Sehmel
(1980), and Smith et al. (1982) for a review of other
possible approaches.
As noted in USEPA (1983b), the SCS wind equation
computes the total wind erosion soil loss resulting
from the combination of surface creep, saltation, and
suspension. Although appropriate for studies of
agricultural soil loss, in exposure evaluations the
analyst is generally concerned only with that fraction
of the total soil loss that consists of particles of
suspendible, wind transportable, and inhalable size.
When the wind erosion equation is used to estimate
contaminated fugitive dust exposure situations, the
Note: EPA (1985c) defines short-term concentrations to
equate with a 10- to 90-day period. Thus, the 24-hour
maximum exposure may not adequately represent subchronic
exposures.
Personal communication between Lee Schultz (Versar Inc.)
and Thomas George (US. Soil Conservation Service), July 24,
1985.
total soil loss results obtained from the wind erosion
equation must be adjusted (reduced) to reflect only
that portion of the total soil loss that is suspendible
and transportable over significant distances by wind.
Considerable discussion of the cut-off point for
suspendible soil particle size exists in the literature
(Sehmel 1980, Smith et al. 1982, and USEPA 1983a.
b). As a group, particles < 100 urn aerodynamic
equivalent diameter include those that can be
suspended by and transported in the wind and those
that can be inhaled (see Miller et al. 1979 and USEPA
1986d for a discussion of the extent to which various
particle sizes penetrate the human respiratory
system). Particles in the 30 to 100 pm diameter range
will often settle within a few hundred feet of the
source (USEPA 1983a), while those particles < 30
pm in diameter can be transported considerable
distances downwind. To estimate inhalation exposure,
only the inhalable fraction of suspended particulates
{< 10 pm in diameter) must be considered.
For particles in the 2- to 20-um size range, the
particle size distribution of the parent soil determines
the size distribution of suspended particles (Smith et
al. 1982). Therefore, that proportion which is < 10 urn
in diameter can be determined based on the soil size
distribution of the parent soil. It can be assumed that
this proportion of the total soil loss, as calculated via
the SCS wind erosion equation, is lost to suspension
and is available for inhalation.
(2) Unpaved Roads Analysis
The following equation (USEPA 1983a) can be used
to estimate fugitive dust releases associated with
vehicles traveling on contaminated unpaved roads.
13
-------
/ s WSpWW\a7/w\a5/'365-Dp\
BvT-«">(ii)(»)(7) d) (-sr)
or in metric form
0.7 , , 0.5,
(2-2)
/ s \/SpW W\ /w\ /366-Dp\
E___ = k<1.7)( — )( — )( — 1 I—) -
VT \12/\48/\2.7/ \4/ V 365 /
where
EVT =
k =
s =
Sp =
w =
w =
Dp =
emission factor for vehicular traffic, (Ib/vehicle
mile traveled: kg/vehicle kilometer traveled)
0.45 = particle size multiplier for particles
< 10 pm (i.e., particles that may remain
suspended once they become airborne and
which can be inhaled into the respiratory
system).*
silt content (of road surface material),
(percent)."
mean vehicle speed, (mph; kph).
mean vehicle weight, (tons; Mg).
mean number of wheels.
number of days with at least 0.254 mm (0.01
in) of precipitation per year (see Figure 2-3).
The emission factor (EVj) 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 assumed drought conditions at the site. Figure
2-3 reflects the range of average "Dp" values for
locations in the U.S. 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. 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
comply with the following source conditions:
x Road surface silt content = 4.3 - 20 percent;
eses Mean vehicle weight = 3-157 tons (2.7-142
Mg);
esx 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 application of emission factors to
particulate release estimation problems, the user of
See EPA (1983a) for "k" values used when release of
specific particle size groups other than < 10 pm is desired.
"soil silt content can oe 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, DC, May 1, 1984.)
this manual can refer to USEPA (1983a, b), Farino et
al. (1983) Sehmel (1980), and Smith et al. (1982).
2.3.1.2 In-Depth Analysis
For contaminated fugitive dust emissions, in-depth
analysis will consist of monitoring and modeling
activities. Generally, air sampling will be conducted
downwind and upwind of the uncontrolled hazardous
waste site. The difference in particulate loading
obtained at the two (or more) sampling locations will
quantify the particulate mass loading attributable to
the site alone (assuming that air sampling stations
can be sited to eliminate interference from other
sources). Using these data, either simple dispersion
equations or computerized air dispersion modeling*
can be used to back-calculate the emission level at
a "virtual point source." The use of dispersion
modeling to back calculate emission levels, however,
is often quite unreliable because of the difficulty in
obtaining accurate ambient monitoring and
meteorological input data.
The virtual point source is a hypothetical source
located upwind of the subject site that has a
hypothetical release rate which would result in the
contaminant concentrations observed at the
uncontrolled hazardous waste site (area source). The
virtual point source release rate can then be used in
subsequent contaminant transport analysis for the
subject site. The user of this manual should refer to
USEPA (1983c) and Seely et al. (1983) for a detailed
presentation of ambient air sampling strategies and
procedures appropriate for abandoned hazardous
waste facilities.
2.3.2 Volatilization Emission Analysis
Volatilization of contaminants at uncontrolled
hazardous waste sites can occur at the following
sources:
(1) Covered landfills - without internal gas
generation;
(2) Covered landfills - with internal gas
generation;
(3) Spills, leaks, landfarms - concentrated
wastes on the surface or adhered to soil
particles below the surface; and
(4) Lagoons - wastes dissolved in or mixed
with water.
In the baseline situation, one or more of these
sources will contribute to the overall air loading
originating at the site, and will need to be controlled
through remedial action.
2.3.2.1 Beginning Quantitative Analysis
This section presents simplified analytical procedures
for estimating releases from the above source
categories. Because the chemical properties of a
given substance largely determine the volatilization
rate, the equations presented require input of
14
-------
Figure 2-3. Mean number of days per year with >0.01 inches of precipitation (i.e., ' 'wet days") (USDC 1968).
/ '--xrJ^i
Alaska
0 200 400
100 300
-------
quantified property values. These data are available
for many chemicals that may be present at
uncontrolled hazardous waste sites, and are found in
various chemical reference texts. In cases where
chemical data are missing, the analyst must estimate
the property values. This section provides equations
for estimating certain requisite chemical properties.
Comprehensive guidance for chemical property
estimation is provided in reference materials such as
Lyman et al. (1982). Readily accessible computerized
systems are available to predict a range of pertinent
chemical properties. The computerized Graphic
Exposure Modeling System (GEMS), and its
subsystem CHEMEST, is an example. The EPA
Office of Toxic Substances in Washington, D.C. has
developed and is managing this system. Essentially a
computerized version of Lyman et al. (1982), it can
be rapidly accessed to estimate the chemical
characteristics necessary for volatilization estimation.
The user of this manual can refer to Farino et al.
(1983) for a detailed review and evaluation of existing
equations for estimating volatilization from
uncontrolled hazardous waste sites. This report
presents a survey of available air release models for
volatile substances and a critical analysis of the
applications and limitations of each.
(1) Landfills Without Internal Gas Generation
Equation 2-3 can be used to estimate volatile
releases from covered landfills containing toxic
materials alone, or toxic materials segregated from
other landfilled nonhazardous wastes. Equations 2-4
through 2-7 are used to calculate certain input
variables that are required to apply Equation 2-3.
Farmer et al. (1978) developed an equation to
estimate the effectiveness of various landfill cover
types and depths in controlling volatile releases. This
equation, based on Fick's First Law of steady state
diffusion, assumes that diffusion into the atmosphere
occurs at a plane surface where concentrations
remain constant. It ignores biodegradation, transport
in water, adsorption, and production of landfill gas.
Diffusion of the toxic vapor through the soil cover is
the controlling factor. It also assumes that there is a
sufficient mass of toxicant in the landfill so that
depletion of the contaminant will not reduce the
emission rate.
Equation 2-3, simplified by Farmer et al. (USEPA
1980b), incorporates a number of assumptions (see
Farino et al. 1983 for a complete discussion), such as
completely dry soil (worst case) and zero
Although computerized dispersion modeling can be used to
obtain contaminant release rates, it is primarily a tool for
determining contaminant atmospheric fate. Thus, refer to
Chapter 3, Environmental Fate Analysis, for detailed
discussions of air dispersion models applicable to uncontrolled
hazardous waste facilities.
concentration of volatilizing material at the soil
surface. Shen (1981) converted Farmer's simplified
equation for calculating the vapor flux rate to a form
that provides a toxic vapor emission rate by
multiplying the basic equation by the exposed
contaminated surface area. In addition, Shen modified
the equation to allow calculation of the volatilization
rate of a specific component of the overall toxic
mixture by multiplying by the weight fraction of the
component in the mixture. However, as pointed out
by Farino et al. (1983), a more accurate approach
would be to multiply by the mole fraction of the toxic
component in the buried mixture. Thus, Farmer's
equation, as modified by Shen (1981) and Farino et
al. (1983) is:
M.
. = D.C.A( ?*•>) —
1 1 SI \ t / A
asc
(2-3)
where
A
p
sc
= emission rate of component i, (g/sec).
= diffusion coefficient of component i in air,
(cm2/sec).
= saturation vapor concentration of component
i, (g/cm3).
exposed area, (cm2).
= total soil porosity, (dimensionless).
= mole fraction of toxic component i in the
waste, (gmole/gmole).
= effective depth of soil cover, (cm).
Note that total soil porosity, rather than air-filled soil
porosity, is used in this equation. The presence of
water in a soil cover will tend to decrease the flux rate
of a volatile compound by effectively decreasing the
porosity, and also by increasing the geometric
complexity of the soil pore system (because water
adheres to soil particles), thus effectively increasing
the vapor path (USEPA 1980b). Farmer et al.
suggest, however, that when using their equation to
design a landfill cover, the total porosity value be
used (USEPA 1980b), thereby designing for the worst
case (i.e., dry conditions). In most instances, it will be
appropriate to apply this same worst-case logic to
the analysis of volatilization release from landfilled
wastes, assume that landfill cover soils are dry, and
use a value for total porosity in Equation 2-3. It is
recognized, however, that there may be situations
where it can be shown that cover soils exist in a wet
condition more often than in a dry one. In these
cases, the air-filled soil porosity (Pa) may be more
appropriate, and this value can be substituted for Pt
in Equation 2-3 when analyzing volatilization release.
If not provided in existing literature, D,, a compound's
diffusion coefficient (required for the above equation),
can be calculated by Fuller's Method (Perry and
Chilton 1973):
16
-------
0.001T
1.75
MW. MW
D.=
where
T
MWi;MWa
Pa
EVi;SVa
)1/3]2
(2-4)
absolute temperature, (°K).
molecular weights of toxic
substance and air (28.8),
respectively, (g/mole).
absolute pressure, (atm).
molecular diffusion volumes of
toxic substance and air (20.1).
This is the sum of the atomic
diffusion volumes of the
compound components,
(cm3/mole).
To estimate short-term (maximum) release rates,
use a value for the temperature that reflects the
expected summer maximum temperatures. Annual
average temperatures should be used to initially
estimate long-term (average) release rates. This
initial estimated long-term release value will be
revised as described in Section 2.3.3 to develop final
long-term release estimates.
Relevant atomic diffusion volumes for use in
estimating D, are (Perry and Chilton 1973):
Aromatic ring = -20.2
Heterocyclic ring = - 20.2
C = 16.5 Cl = 19.5
H = 1.98 Br = 35.0
O = 5.48 F = 25.0*
N = 5.69 S = 17.0
Table 2-3 presents diffusion coefficients that have
been calculated for a variety of compounds, some of
which may be present at abandoned sites.
An alternative method (Shen 1981) for approximating
DJ involves the identification of a compound listed in
Table 2-3 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:
D. = D'f
VMW.
(2-5)
where
D, = 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
' This value is from Shen (1981).
MW
MW,
weight and atomic diffusion, volume of
which are close to that of the unknown.
= molecular weight of the selected
compound D'.
= molecular weight of the compound to
be estimated.
Total soil porosity, Pt, can be calculated as follows
(USEPA 1980b):
(2-6)
where
Pt
8
(dimensionless).
_3\.
= total soil porosity,
= soil bulk density,* (g/cmj): Generally
between 1.0 and 2.0 a/cm
= particle density, (g/cm ): usually 2.65
g/cm3 used for most mineral material.
For estimation, Pt can be assumed to be
approximately 0.55 for dry, non-compacted 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 (Pa) when analyzing the
volatilization release from soils with a high moisture
content (Shen 1981). Alternatively, the local Soil
Conservation Service office can be contacted to
obtain site-specific estimated air-filled soil porosity
values for specific locations.
Saturation vapor concentration, Csi can be
determined by (USEPA 1980b):
C .=
SI
pMW.
RT
(2-7)
where
R
T
= saturation vapor concentration of
component i, (g/cm3).
= vapor pressure of the chemical," (mm
Hg).
= mole weight of component i, (g/mole).
= molar gas constant, (62,361 mm Hg-
cm3/mole-°K).
= absolute temperature, (K).
Again, use maximum summer temperatures to
estimate short-term release and annual average
temperatures to initially estimate long-term release.
Values for soil bulk density for specified locations can be
obtained from the U.S. Soil Conservation Service, Soils 5 File
data base.
" If the vapor pressure of a chemical under consideration is not
available in standard reference texts, estimate it as described in
Lyman et al. (1982).
17
-------
Table 2-3. Diffusion
Compound
Acetaldehyde
Acetic acid
Acetone
Aniline
Benzene
Bromoethane
Bromoform
Carbon tetrachloride
Chlorobenzene
Chloroethane
Chloroform
Chloromethane
Cyclohexane
Dichloroethane
Dichloroethylene
Dicchloropropalene
Dimethylamrne
Ethanol
Ethyl acetate
Ethylamine
Ethylbenzene
Fluorotoluene
Heptane
Hexane
Isopropanol
Methanol
Methyl acetate
Methyl chloride
Methylethyl ketone
PCB (1 Cl)
Pentane
Phenol
Styrene
Tetrachloroethane
Tetrachloroethylene
Toluene
Tricyhloroethane
Trichloroethylene
Trichlorofluoromethane
Vinyl chloride
Xylene
Coefficients of Selected Organic Compounds
Atomic Diffusion
Molecular diffusion
Formula weight volume at 10°C
C2H40
C2H402
C3H60
C6H7N
C6H6
CH3Br
CHBr3
CCI4
C6H5CI
C2H5CI
CHC13
CH3CI
C6H12
C2H4CI2
C2H2CI2
C3H6CI2
C2H7N
C2H60
C4H802
C2H7N
C8H10
C7H7F
C7H16
C6H14
C3H80
CH40
C3H602
CH2CI2
C4H80
C12HGCI
C5H12
C6H60
C8H8
C2H2CI4
C2CI4
C7H8
C2H3CI3
C2HCI3
CCI3F
C2H3CI
C«H,ft
44
60
58
93
78
95
118
154
113
65
120
51
84
99
97
113
45
46
88
45
116
110
100
86
60
32
74
85
72
189
72
84
104
168
166
92
133
131
138
63
106
46.40
51.88
66.86
118.55
90.68
57.44
53.48
94.50
128.40
62.40
76.89
57.94
122.76
75.96
106.96
100.38
52.55
50.36
92.80
52.55
151.80
154.36
146.86
126.72
37.82
29.90
72.34
59.46
87.32
235.32
106.26
96.16
137.84
1143.96
111.00
111.14
97.44
93.48
100.00
58.44
131.60
.11758
. 10655
. 09699
.07157
.08195
.09611
.09655
.07500
. 06769
.09789
. 08345
. 10496
.07139
.08557
. 07442
.07519
.11161
.11297
.07991
.11161
.06274
.06262
. 06467
.07021
. 12004
. 14808
.09054
.09610
.08417
. 04944
.07753
.07919
.06620
.06858
.06968
.07367
.07496
.07638
.07391
. 10094
.06742
coefficients (cm2/sec)
at 30°C at 50°C
. 13249
.12007
. 10930
. 08065
. 09234
. 10830
. 10880
.08451
.07627
.11031
. 09404
.11827
. 08045
. 09643
.08386
.08473
.12577
.12730
.09005
.12577
.07070
.07056
.07287
.07912
.13526
. 16686
. 10203
. 10830
.09485
.05571
.08737
. 08924
. 07460
.07729
.07852
.08301
. 08447
.08606
.08329
.11375
.07597
.14816
.13427
.12223
.09019
.10327
.12111
.12167
.09451
.08530
. 12336
.10517
. 13226
. 08996
. 10784
.09377
.09475
. 14065
. 14236
. 10070
. 14065
. 07906
.07891
.08149
. 08848
.15126
. 18660
.11410
.12111
. 10607
. 06230
.09770
.09980
.08343
.08643
.08781
.09283
.09446
.09625
.09314
. 12720
.08495
Source: Shen 1981
18
-------
See Section 2.3.3 for directions for calculating a final
long-term release rate.
(2) 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 that, because of
their considerable organic content, generate landfill
gases (e.g., H2, CH4, C02). In these cases, the
upward movement (convective sweep) of the landfill
gas becomes the significant controlling factor, greatly
accelerating the upward migration and subsequent
release to the atmosphere of the co-disposed toxic
substances. In fact, review of Thibodeaux's work
indicates that the toxic gas migration accelerating the
effect of the landfill gas is so great that both soil and
gas phase diffusion essentially become insignificant.
The following simplified equation is recommended for
estimating the volatilization of toxic substances from
co-disposal landfills:
Ei = Ci*VyA
where
(2-8)
= emission rate, (g/sec).
Ei = concentration of compound i in the soil
pore spaces, (g/cm3)
v = mean landfill gas velocity in the soil
y pore spaces, (cm/sec). Thibodeaux
(1981) provides an average value of
1.63 x 10-3 cm/sec for this factor.
A = area, (cm2).
Recalculation of the toxic vapor release estimates
presented in Thibodeaux (1981) using this simplified
equation yields results within = 1 percent of the
values obtained using the full computation cited in the
paper. Thibodeaux (1981) notes, however, that
various site factors such as the presence of saturated
soils will tend to reduce the rate of volatile chemical
release from landfills. The degree to which this model
is able to accurately reflect contaminant release rates
for gases, especially soluble gases, generated at sites
with moist or wet soils is unknown.
(3) Spills and Leaks
Equations 2-9 and 2-11 will estimate the volatile
releases from fresh and old (respectively) chemical
spills on soil. Equations 2-10 and 2-12 through 2-
14 provide means of estimating certain input variables
required to solve Equations 2-9 and 2-11.
As discussed in Farino et al. (1983) one can apply
Equation 2-9 (adapted from Thibodeaux and Hwang
1982) to estimate volatile releases 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
contaminants 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 compounds. The original equation presented in
Thibodeaux and Hwang (1982) has been modified to
include a contaminated surface area term, thereby
resulting in the calculation of a release rate rather
than a flux rate value:
Ei = kiGCi*A
where
(2-9)
A
= emission rate of chemical i, (g/s).
= gas phase mass transfer coefficient of
chemical i, (cm/s).
= vapor concentration of chemical i,
(g/cm3).*
= area, (cm ).
Hwang (1982) has developed the following simplified
means of estimating a compound's gas phase mass
transfer coefficient.
k..
MW.
MW.
0.335
'/
V298
1.005
(2-10)
where
MWH20; MW, =
klG, H20 =
gas phase mass transfer
coefficient of chemical i, (cm/s).
molecular weight of water;
compound i, (g/mole).
temperature, (°K).
gas phase mass transfer
coefficient for water vapor at
25°C, (cm/sec).
" For conservative analyses, one can assume that the actual
contaminant vapor concentration in the soil pore spaces is the
same as the equilibrium vapor concentration. In such cases,
C can be used in place of C,. Direct measurements of C,,
h&ever, may be developed during the site investigation. When
such data are available, their use is preferred.
For conservative analyses, one can assume that the actual
contaminant vapor concentration in the soil pore spaces is the
same as the equilibrium vapor concentration. In such cases, Cj,
can be used in place of C,*. Direct measurements of C, ,
however, may be developed during the site investigation. When
such data are available, their use is preferred.
19
-------
When estimating short-term (maximum) release
rates, the highest (summer) seasonal temperature
expected at the site can be used in calculating the
gas phase mass transfer coefficient. For initial
estimation of long-term release rates, the seasonal
average temperature should be used. Final long-
term release notes are developed as discussed in
Section 2.3.3.
(4) Landfarming
In cases where past spills, leaks, or intentional
disposal directly onto or into surface soils
(landfarming) have resulted in contaminated surface
soils with liquids in the pore spaces, Equation 2-11
can be used to estimate volatilization 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. It
also assumes that there is sufficient liquid
contaminant in the pore spaces so that volatilization
will not deplete the reservoir of contaminant to the
point where it affects the rate of volatilization.
Modeling the release from soils with sorbed
contaminants and no free liquids requires another
model.
Two models for predicting the time-varying
volatilization of sorbed contaminants on soil are
presented in USEPA (1986e). The equation presented
here is adapted from Thibodeaux and Hwang (1982),
which presents a volatilization release estimation
equation designed for application to active or planned
landfarms for petroleum wastes. Farino et al. (1983)
determined it to be preferable to other approaches for
estimating volatilization release of chemicals spilled or
incorporated into soils, because it directly 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 remain constant (until all contaminant is
lost to the air), and that contaminant release occurs
by the "peeling away" of 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. The original equation has been adjusted
somewhat for use at uncontrolled waste sites, and
has also been simplified as discussed in Farino et al.
(1983), by assuming that the oil layer diffusion length
value is low (i.e., that the spilled contaminant has
become incorporated into surface soils and is not
present as a discrete film).
2DCSA
where
D
Cs
CB
A
d
average emission rate of component i
over time t, (g/sec).
phase transfer coefficient, (cm2/sec).
the liquid-phase concentration of
contaminant i in the soil, (g/cm3).
bulk contaminant concentration in soil
(g/cm3)
contaminated surface area, (cm2)
depth of dry zone at sampling time,
(cm).
time measured from sampling time,
(seconds).
D (cm2/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:
where
D
DI
Pt
Hi'
(2-12)
phase transfer coefficient, (cm /sec).
diffusion coefficient of component i in
air, (cm2/sec).
total soil porosity, (dimensionless).
Again, use of total soil porosity in this
equation results in a worst-case (dry
soil) estimate for D. As previously
discussed, however, 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 2-
3 for a discussion of the application of
and values for these two terms.
Henry's Law constant in concentration
form, (dimensionless).
Hi', 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. 1983):
where
H.
-^
RT
(2-13)
R
(2-11)
= Henry's Law constant of contaminant i,
(atm-m3/mol).
= gas constant,, (8.2 x 1 0~5 atm-
m3/mol-°K).
T = absolute temperature, (°K).
Again, use summer maximum temperatures to
estimate short-term release and annual average
20
-------
temperatures for the initial estimation of long-term
release. Final long-term release rates are developed
as discussed in Section 2.3.3.
Note that Equation 2-11 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 td (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:
'B
2D
(2-14)
where
td
CB
= the time at which all contaminant has
volatilized from the soil, (sec).
= depth from soil surface to the bottom
of the contaminated region, (cm).
= depth of dry zone at sampling time,
(cm).
D = phase transfer coefficient, (cm /sec).
= bulk contaminant concentration in soil,
(g/cm3)
Cs = contaminant liquid phase concentration
(g/cm3)
(5) Lagoons
Mackay and Leinonen (1975) have developed an
equation for estimating volatilization releases of low
solubility compounds from waterbodies such as
hazardous waste lagoons. This is presented as
Equation 2-15. Equations 2-16 and 2-17 provide
means of calculating certain input parameters
required by Equation 2-15. 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. As pointed out in Farino
et al. (1983), if it can be assumed that atmospheric
background levels of the contaminant of concern are
negligible, (as would usually be the case at
abandoned hazardous waste facilities), 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):
where
K,
(2-15)
= emission rate, (g/sec).
= overall mass transfer coefficient,
(cm/sec).
Cs = contaminant liquid phase concentration,
(g/cm3)
A = area, (cm ).
The overall mass transfer coefficient (K,) can be
calculated via the following relationship:
1 E RT
— = — (2-16)
K. k. V '
where
KI
kiL
R
T
-5
iL HjkiG
= overall mass transfer coefficient,
(cm/sec).
= liquid phase mass transfer coefficient,
(cm/sec). See Equation 2-17.
= ideal gas law constant, (8.2 x 10
atm-m3/mol-°K).
= temperature, (°K).
H | = Henry's Law constant of compound i,
(atm-m3/mol).
klG = gas phase mass transfer coefficient,
(cm/sec). See Equation 2-10.
Hwang (1982) provides a simplified method for
determining a compound's liquid phase mass transfer
coefficient for use in the above equation. To estimate
kiL, use the following equation:
k-r =
iL
where
MW.
MW.
\ 298
kL'°2
(2-17)
kiL
T
kL,02
= liquid phase mass transfer
coefficient, (cm/sec).
= molecular weight of oxygen;
compound,.
= temperature, (°K).
= liquid phase mass transfer
coefficient for oxygen at 25°C,
(cm/sec).
The value for kL,O2 can be obtained from chemical
reference texts or can be calculated (the preferred
method) as described in Farino et al. (1983).
2.3.2.2 In-Depth Analysis
In-depth analysis of volatile release can be executed
in the same manner as that described for particulates.
Subtract the monitored upwind (control) ambient toxic
vapor concentration from the monitored downwind
concentration. Use the difference between these two
values in an air dispersion model to estimate the
release rate at a "virtual point source" that would
correspond with the source of the measured
downwind concentration.
The user of this manual should again refer to USEPA
(1983c) and Seely et al. (1983) for detailed
discussions of the planning and execution of air
monitoring studies. Refer to Chapter 3 of this manual
21
-------
for a description of air contaminant dispersion
modeling tools.
2.3.3 Long-Term and Short-Term Release Cal-
culation
Long-term release values (70 years) for lagoons with
dilute solutes can be estimated as follows:
E. .=
Ai
where
EAi
Vc
e
E
V C.
C 1
~70~
1-e
V C.
C I
-<2.2xl09>
(2-18)
= average annual release of contaminant
volume of contaminated region, (cm3).
initial average concentration of
contaminant i in site soils, (g/cm3).
2.71828
initial combined 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 2-18 to
mass per second by dividing by 3.16 x
107 seconds.
Note that Vc and C, must be based on the same
value. (VcCj) is equal to the total mass of
contaminant; it can be the total mass of contaminant
in a lagoon.
For landfills and wind erosion of contaminated
particulates, the release rate is assumed constant.
The 70-year average annual release rate can be
calculated by first ascertaining if contaminant will
remain after 70 years. If so, then the release rate
itself is the 70-year average annual release rate. If
not, then the 70-year average annual release rate is
the total initial mass divided by 70 years.
To estimate long-term release from contaminated
surface soils, Equation 2-14 (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 > td), simply
determine the total amount of contaminant present at
the time of site investigation and divide by 70 years
(in seconds) to get a conservative long-term release
value (i.e., ACs(h - d)/2.21 x 109). If some
contaminant is expected to remain after 70 years (i.e.,
70 < td), use the following equation to estimate
long-term release:
E,
Ai
where
AC /
.= —(CdJ
1 70 V
+ 4.4xl
-------
toxics can be quantified directly by measuring
(sampling) the source material and determining the
volume and rate of release. Alternatively, runoff
release estimation procedures, less costly than
monitoring or modeling approaches, can also be
applied to uncontrolled sites.
In addition, surface waters may be contaminated by
inflows of ground water through bank seepage and
springs. In order to estimate the rate of such inflows,
one must conduct modeling of ground-water/surface
water linkages (see Chapter 3 for a discussion of
ground-water modeling options).
This section reviews methods for estimating toxic
releases of uncontrolled hazardous waste sites to
surface waterbodies. Note, however, that only the
surface runoff component of release to surface water
is addressed here. Other sources must be estimated
for each site based on judgment and experience.
2.4.1 Beginning Quantitative Analysis
2.4.1.1 Dissolved and Sorbed Contaminant
Migration
Many of the organic substances of concern found at
Superfund sites are relatively nonpolar, hydrophobic
substances (Delos et al., 1984). Such substances can
be expected to sorb to site soils and migrate from the
site more slowly than will polar compounds. As
discussed in Haith (1980) and Mills et al. (1982),
estimates of the amount of hydrophobic compounds
released in site runoff can be calculated using the
Modified Universal Soil Loss Equation (MUSLE) and
sorption partition coefficients derived from the
compound's octanol-water partition coefficient. The
MUSLE allows estimation of the amount of surface
soil eroded in a storm event of given intensity, while
sorption coefficients allow the projection of the
amounts of contaminant carried along with the soil,
and the amount carried in dissolved form.
(1) Soil Los Calculation
Equation 2-20 is the basic equation for estimating
soil loss. Equations 2-21 through 2-24 are used to
calculate certain input parameters required to apply
Equation 2-20. The modified universal soil loss
equation (Williams 1975), as presented in Mills et al.
(1982), is:
Y(S)E = a(Vrqp)a56 KLSCP
(2-20)
where
Y(S)E
a
Vr
= sediment yield (tons per event, metric
tons per event).
= conversion constant, (95 English, 11.8
metric).*
= volume of runoff, (acre-feet, m3).
q = peak flow rate, (cubic feet per second,
p m3/sec).
K = the soil erodibility factor, (commonly
expressed in tons per acre per
dimensionless rainfall erodibility unit). K
can be obtained from the local Soil
Conservation Service off ice.
L = the slope-length factor, (dimension-
less ratio).
S = the slope-steepness factor, (dimen-
sionless ratio).
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).
Soil erodibility factors are indicators of the erosion
potential of given soil types. As such, they are highly
site-specific. K values for sites under study can be
obtained from the local Soil Conservation Service
office. The slope length factor, L, and the slope
steepness factor, S, are generally entered into the
MUSLE as a combined factor, LS, which is obtained
from Figures 2-4 through 2-6. The cover
management factor, C, is determined by the amount
and type of vegetative cover present at the site. Its
value is "1" (one) for bare soils. Consult Tables 2-4
and 2-5 to obtain C values for sites with vegetative
covers. The factor, P, refers to any erosion control
practices used on-site. Because these generally
describe the type of agricultural plowing or planting
practices, and because it is unlikely that any erosion
control would be practiced at an abandoned
hazardous waste site, use a worst-case
(conservative) P value of 1 (one) for uncontrolled
sites.
Storm runoff volume, Vr, is calculated as follows
(Mills etal. 1982):
Vr = aAQr
where
(2-21)
a = conversion constant, (0.083 English,
IOO 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) (2-22)
Metric conversions presented in the following runoff
contamination equations are from Mills et al. (1982).
23
-------
Figure 2-4. Slope effect chart applicable to areas A-1 in
Washington, Oregon, and Idaho, and all of A-
3: MO Figure 2-6 (USDA 1974 as presented
in Mills et al. 1982).
Slope Length, Meters
20 3040 6080100150200300400600800
40.0
20.0
10.0
6.0
2 4.0
o
-------
values of uncontrolled hazardous waste sites from
Table 2-6.
The peak runoff rate, qp, is determined as follows
(Haith 1980):
aArtQr
Tr(Rt-0.2Sw)
(2-24)
where
qp
a
A
Rt
Qr
Tr
the peak runoff rate, (ft3/sec, m3/sec).
conversion constant, (I.OI English,
0.028 metric).
contaminated area, (acres, ha).
the total storm rainfall, (in, cm).
the depth of runoff from the watershed
area, (in, cm).
storm duration, (hr).
water retention factor, (in, cm).
(2)Dissolved/Sorbed Contaminant Release
As discussed in Mills et al. (1982), the analyst can
predict the degree of soil/water partitioning expected
for given compounds once the storm event soil loss
has been calculated with the following equations.
First, the amounts of adsorbed and dissolved
substances are determined, using the equations
presented below as adapted from Haith (1980):
S8 =
+
and
where
S8
ec
P
Ci
A
Ds
i) (A)
(Kdp)/ec)](Ci)(A)
(2-25)
(2-26)
sorbed substance quantity, (kg, Ib).
available water capacity of the top cm
of soil (difference between wilting point
and field capacity), (dimensionless).
sorption partition coefficient, (cm3/g).
soil bulk density, (g/cm3).
total substance concentration, (kg/ha-
cm, Ib/acre-cm).
contaminated area, (ha-cm acre-
cm). (Actually a volume; assumption is
contamination in upper 1 cm is
available for release.)
dissolved substance quantity, (kg, Ib).
This model assumes that only the contaminant in the
top 1 cm of soil is available for release via runoff.
The soil sorption partition coefficient for a given
chemical can be determined from known values of
certain other physical/chemical parameters, primarily
the chemical's octanol-water partition coefficient,
solubility in water, or bioconcentration factor. Lyman
et al. (1982) present regression equations that allow
the analyst to determine sorption coefficients for
specified groups of chemicals (e.g., herbicides,
polynuclear aromatics). If parameter values required
by the appropriate equations are not available in
chemical reference literature, they can be estimated
according to procedures described in Lyman et al.
(1982). Initially, the octanol-water partition coefficient
can be estimated based on the substance's molecular
structure. If necessary, this value can be used, in
turn, to estimate either solubility in water or
bioconcentration factor.
After calculating the amount of sorbed and dissolved
contaminant, the total loading to the receiving
waterbody is calculated as follows (adapted from
Haith 1980):
PXi=[Y(S)E/100p]Ss
plus
(2-27)
(2-28)
where
PXi = sorbed substance loss per event, (kg,
Ib).
Y(S>E = sediment yield, (tons per event, metric
tons).
P = soil bulk density, (g/cm3).
Ss = sorbed substance quantity, (kg, Ib).
PQi - dissolved substance loss per event,
(kg, Ib).
Qr = total storm runoff depth, (in, cm).
Rt = total storm rainfall, (in, cm).
Ds = dissolved substance quantity, (kg, Ib).
PXi and PQi can be converted to mass per volume
terms for use in estimating contaminant concentration
in the receiving waterbody by dividing by the site
storm runoff volume (Vr, see Equation 2-21).
2.4.2 In-Depth Analysis
Releases to surface waterbodies at uncontrolled
hazardous waste sites can be quantified most
accurately by direct measurement (sampling and
analysis) of the contaminant flow. Alternatively,
upcurrent and downcurrent sampling can be
conducted to determine the release level at the site
that would be used to estimate the ambient
concentration (i.e., the difference between the
upcurrent and downcurrent concentrations). Either
simple dispersion equations or sophisticated
computer modeling approaches (see Chapter 3) can
be used to "back calculate" the measured ambient
concentration to the "virtual point source."
25
-------
Table 2-4. "C" Values for Permanent Pasture, Rangeland, and Idle Land
Cover that contacts the surface
Vegetal canopy
Type and height
of raised canopyb
No appreciable canopy
Canopy of tall weeds or short
brush (0.5 m fall height)
Appreciable brush or brushes
(2m fall height)
Trees but no appreciable low
brush (4 m fall height)
Canopy
cover0
(%)
25
50
75
25
50
75
25
50
75
Percent ground cover
Typed
G
w
G
w
G
W
G
W
G
W
G
W
G
W
G
W
G
W
G
W
0
0.45
0.45
0.036
0.036
0.026
0.026
0.17
0.17
0.40
0.40
0.34
0.34
0.28
0.28
0.42
0.42
0.39
0.39
0.36
0.36
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
20
.20
.24
.17
.20
.13
.16
.10
.12
.18
.22
.16
.19
.14
.17
.19
.23
.18
.21
.17
.20
40
0
0
0
0
0
0.
0
0
0
0.
0
0.
0.
o
0.
0.
0.
0.
0.
0.
.10
.15
.09
.13
.07
11
.06
.09
.09
14
.085
13
08
.1 2
10
14
09
14
09
13
60
0.042
0.090
0.038
0.082
0.035
0.075
0.031
0.067
0.040
0.085
0.038
0.081
0.036
0.077
0.041
0.087
0.040
0.085
0.039
0.083
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0.
80
.013
.043
.012
.041
.012
.039
.011
.038
.013
.042
.012
.041
.012
.040
.013
.042
.013
.042
.012
,041
95-100
0.003
0.011
0.003
0.011
0.003
0.011
0.003
0.011
0.003
0.011
0.003
0.011
0.003
0.011
0.003
0.011
0.003
0.011
0.003
0.011
Source: Wischmeier 1972
aAII values shown assume: (1) random distnbution of mulch or vegetation, and (2) mulch of appreciable depth where it exists.
bAverage fall height of waterdrops from canopy to soil surface: m = meters.
CPortron of total-area surface that would be hidden from view by canopy in a vertical projection (a bird's-eye view).
dG: Cover at surface is grass, grasslike plants, decaying compacted duff, or litter at least 5 cm (2 in) deep.
W: Cover at surface is mostly broadleaf herbaceous plants (as weeds) with little lateral-root network near the surface and/or
undecayed residue.
Stand condition
Well stocked
Medium stocked
Poorly stocked
Tree canopy per-
cent of area3
100-75
70-40
35-20
Forest litter per-
cent of areab
100-90
85-75
70-49
Undergrowth0
Managedd
Unmanagedd
Managed
Unmanaged
Managed
Unmanaged
"C" factor
0.001
0.003-0.01 1
0.002-0.004
0.01-0.04
0.003-0.009
0.02-0.09"
Source: Wischmeier 1972
awhen tree canopy is less than 20 percent, the area will be considered as grassland or cropland for estimating soil loss.
bForest litter is assumed to be at least 2 in deep over the percent ground surface area covered.
°undergrowth is defined as shrubs, weeds, grasses, vines, etc., on the surface area not protected by forest litter. Usually
found under canopy openings.
dManaged - grazing and fires are controlled.
Unmanaged - stands that are overgrazed or subjected to repeated burning
eFor 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 3-4. The factor of 0.7 adjusts for much higher soil organic matter on permanent woodland.
26
-------
Table 2-6. Runoff Curve Numbers
Soil
group Description
Site type
Overall site" Road/right of way Meadow
Woods
Lowest runoff potential: Includes
deep sands with very little silt and
clay, also deep, rapidly perme-
able loess (infiltration rate =
8-12 mm/h).
Moderately low runoff potential:
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
infiltratipn after thorough wetting
(infiltration rate = 4-8 mm/h).
Moderately high runoff potential:
Comprises shallow soils and soils
containing considerable clay and
colloids, though less than those of
group D. The group has below-
average infiltration after
presafuration (infiltration rate =
1-4 mm/h).
Highest runoff potential: Includes
mostly clays of high swelling
percent, but the group also in-
cludes some shallow soils with
nearly impermeable subhorizons
near the surface (infiltration
rate = 0-1 mm/h).
59
74
74
84
30
58
45
66
82
90
71
77
92
78
83
Source: Adapted from Schwab et al. 1966.
Values taken from farmstead category, which is a composite including buildings, farmyard, road, etc.
2.4.3 Long-Term and Short- Term Release
Calculation
For surface runoff releases, the long-term release
value can be calculated as follows:
I Characterize an average storm event for the area
in terms of duration. This can best be
accomplished by consulting local or regional
climatological experts, or the National
Climatological Data Center in Asheville, North
Carolina. Then, using USDC (1961) determine
the amount of rainfall corresponding to the
selected duration rainfall event on a one year-
return frequency basis. Divide this amount into
the mean annual rainfall for the area to obtain the
average number of average rainfall events per
year.
I Use these data and the equations presented in
this section to calculate runoff contaminant
release associated with each yearly average
storm.
I Estimate the potential total long-term release for
both dissolved and sorbed runoff loss* as follows:
* This approach is overly conservative as it assumes that the
contaminant concentration in surface soil remains essentially
the same during the entire 70-year period.
EAi = BN
where
EAi =
B
N
(2-29)
long-term release of contaminant i in
runoff, (mass/70 years).
dissolved or sorbed loss per storm
event, (i.e., PX, or PQ,; see Equations
2-27 and 2-28).
number of "average" storm events in
70 years.
Determine the total amount of soil that will erode
from the site over 70 years. This can be
accomplished by applying the Universal Soil Loss
Equation (USLE-Wischmeier and Smith 1978).
This equation, from which the MUSLE (see
Equation 2-20) was developed, estimates annual
soil losses in runoff. The USLE takes the same
form as the MUSLE, except that the storm
event-specific volume and flow rate variables
are replaced by a factor, R, the rainfall runoff
factor. Therefore, the USLE is:
Y(S)A = R,KLSCPASd
where
(2-30)
Y(S)A = annual soil loss in runoff, (tons/yr,
tonnes/yr).
27
-------
Rr = the rainfall and runoff factor,
expressing the erosion potential of
average annual rainfall in the locality
(can be obtained from the local Soil
Conservation Service office), (dimen-
sionless).
K = the soil-erodibility factor, commonly
expressed in tons per acre per Rr unit
(can be obtained from the local Soil
Conservation Service office) (in metric
tons/ha/Rr unit), (English K'1.292 =
metric K).
L = the slope length factor, (dimen-
sionless).
S = the slope steepness factor, (dimen-
sionless).
C = the cover factor, (dimensionless ratio:
1.0 for soil, see the following
discussion for vegetated site "C"
values).
P = the erosion control practice factor,
(dimensionless: 1.0 for uncontrolled
hazardous waste sites).
A = acreage of area, (acres, ha).
Sj = the sediment delivery ratio, (dimen-
sionless).
The sediment delivery ratio (Sj) can be estimated
using the following equation:
Sd = Dd-0.22
where
(2-31)
Da = the overland distance between the site
and the receiving waterbody (ft).
Mills et al. (1982) note that this equation was
empirically derived from data for D values from 0 to
800 feet, and caution that it may require further
testing, particularly in sites located in the Midwest and
Central U.S.
Note that in certain areas of the Pacific Northwest
and central western states, thaw and snowmelt may
contribute most of the runoff erosive force on an
annual basis. In such cases, an additional erosion
factor, Rs, must be added to the rainfall and runoff
factor, R, to calculate the total R value for use in the
USLE. Limited field data have indicated that an
approximate estimate of Rs may be obtained by
multiplying 1.5 times the local average total rainfall (in
inches) for the period December 1 through March 31
(Wischmeier and Smith 1978). However, the local Soil
Conservation Service office can provide the overall R
value (Rr plus Rs).
• Based on site monitoring data, estimate the
average contaminant concentration in the layer of
soil that must be eroded to equal the total
estimated amount of soil lost over 70 years
(based on site soil sampling data and the
calculated vertical depth of soil that will erode
over that time period).
• Multiply the average contaminant concentration
on site by the site area to calculate the mass of
contaminant present in that amount of soil
estimated to be eroded over 70 years. This
represents the maximum amount of contaminant
available for erosion losses over the 70-year
period.
• Compare the estimated potential contaminant
runoff losses over 70 years with the mass of
contaminant present (in 70-year erodible soils at
the site). If the estimated total loss to runoff is
less than the amount available, divide the
estimated total 70-year losses by the total
volume of stormwater runoff estimated over 70
years. This will approximate the contaminant
concentration in runoff (both dissolved and
sorbed).
• If the total estimated contaminant runoff losses
exceed the amount of contaminant present in
70-year erodible site soils, divide the total mass
of contaminant present in such soils by the
volume of runoff estimated to leave the site over
70 years to develop adsorbed and dissolved
contaminant loss estimates in concentration form.
In either case, the runoff value needed to
estimate contaminant transport and dispersion in
surface waterbodies can be estimated by dividing
the total volume of runoff estimated to leave the
site over 70 years by the number of seconds,
minutes, etc. in 70 years to estimate runoff
volume per unit time.
Many factors can influence the actual degree of
contaminant loss in given storm events. Because
such factors vary from locale to locale, no single
method will guarantee accurate estimates of short-
term contaminant losses in runoff from all sites.
However, the following approach should yield
reasonable approximations of the magnitude of such
short-term loss. While short duration, high intensity
storm events (thunderstorms) clearly cause significant
erosion, the water quality effects of such storms are
too ephemeral to adequately reflect short-term
releases as defined (i.e., 10 to 90 days). Therefore, a
storm event is needed that will generate contaminant
releases adequate to affect water quality over a time
period approaching the 10-day lower bound of the
short-term time frame. For this analysis, a 1-year,
24-hour storm event has been selected. Data
quantifying the amount of rainfall that corresponds to
the 1-year, 24-hour storm event (as well as similar
data for other storm return periods and durations) are
provided in USDC (1961). To estimate short-term
runoff release, the average site contaminant
concentration should be estimated based on sampling
28
-------
data for the top cm of soil only. This value is then
used in Equations 2-27 and 2-28 to estimate runoff
losses on a single storm event basis.
Research based on the work of Haith et al. (1980) is
currently underway at Cornell University* to develop
runoff loading factors for organic chemicals in soils.
After these factors are devised, the analyst will be
able to obtain average loading values based solely on
a chemical's octanol/water partition coefficient and
the geographic location under study. This will greatly
simplify the generation of long-term average release
estimates.
Note that in order to estimate long-term and short-
term contaminant concentrations in surface water, the
long-term and short-term release values are used,
along with average and minimum streamflow data as
described in Chapter 3, Environmental Fate Analysis.
2.5 Quantitative Analysis of Ground-
Water Contamination
Surface soils at uncontrolled hazardous waste sites
may become contaminated with toxic materials as a
result of (1) the intentional placement of wastes on
the ground (dumping, landfarming), (2) spills, (3)
lagoon failure (overland flow), or (4) contaminated site
runoff. Leaching of toxics from a contaminated soil
surface can carry contaminants into subsurface
layers.
2.51 Beginning Quantitative Analysis
2.5.1.1 Leachate Release Rate
This section presents simplified approaches for
estimating contaminant release rates to ground water.
Such estimation can be determined for dry landfills,
lagoons, or wet landfills, whether unlined or lined with
clay or flexible membrane liners.
(1) Estimating Release Rate from Facilities Lined with
Clay or Natural Soil
Release rate estimation involves the determination of
both the contaminant concentration in the leachate
and the volumetric flux of leachate. The determination
of contaminant concentration is made using
equilibrium conditions (steady state), whereas the
volumetric flux can be ascertained with instantaneous
time-varying models or with steady state equations.
Modeling the release rate of toxic constituents can
thus be done in terms of either the instantaneous
time-varying releases or the annual average release
(i.e., steady state release rate based on an annual
average). This section discusses the determination of
the steady state release rate (annual average); the
* Contact Douglas A. Haith, Cornell University, Ithaca, N.Y.,
(607)256-2280.
equations are simpler than the computer models
necessary for instantaneous time-varying releases.
Analysts interested in performing instantaneous
time-varying release rate determinations are referred
to Chapter 3, where the HELP and SESOIL models
are discussed. HELP and SESOIL are appropriate for
modeling dry solid waste in a landfill or landfarm
situation; they are not appropriate for modeling the
release rate of liquids from lagoons, landfills, or
landfarms. Rainstorms come in discrete intervals
separated by dry periods. Using steady state
equations to model rainfall-induced leaching,
however, assumes that 1/365th of the annual
recharge occurs each day. Although this is an
assumption, it is felt to be a useful one for most
cases. Most abandoned hazardous waste sites have
received liquids in the past; very few have received
only dry solids. Hence, the question of the
assumption of steady state conditions is relatively
moot. For the bulk of the modeling situations (liquid
wastes), the steady state and the instantaneous rates
are the same, and since the steady state equations
are simpler, they are the method of choice.
For lagoons, the analyst should use the concentration
of contaminant in the lagoon as the concentration of
the contaminant leaving the lagoon, since the
"leachate" is the waste itself. The waste leaves the
lagoon by percolating through the clay liner or the
native soil, or it permeates the flexible membrane
liner (FML).
For landfills, the analyst should use the equilibrium
solubility of the solid waste, assuming that the
contaminant will have fully equilibrated with the
percolating rainwater. The use of the equilibrium
solubility concentration as the leachate concentration
is an assumption, it is based on a typical residence
time of 21 years for rain percolating through a
covered (1 09-7 cm/sec) secure landfill. The
assumption is that the time used for determining the
equilibrium solubility of the chemical is much shorter
than the residence time in the fill. If the fill is
uncovered (or covered with a permeable cover), the
travel time through the landfill may be too short for
the above assumptions to be valid. In these cases,
the analyst should calculate the travel time and
compare it to the time used in the solubility test. If the
travel time is not longer than the test time, the analyst
should estimate the leachate concentration as a
fraction of the equilibrium solubility concentration.
Additionally, the above assumptions assume a landfill
of only one waste stream, if the fill has only a small
quantity of the subject waste in it, the contact time is
the time for travel through the isolated material. In
these conditions, the leachate concentration will
typically be a fraction of the equilibrium solubility. The
analyst may wish, in some instances, to model the
solubility of the contaminant within a complex
leachate. In this case, the solubility of a hydrophobic
29
-------
contaminant can be increased by the organic fraction
of the complex leachate.
For landfarms, the assumption that adequate
residence time is available for contaminants to reach
equilibrium solubility may not be viable, and the
analyst should estimate the degree of solubilization.
This can be done by dynamic modeling of the kinetics
of dissolution, or it can be approximated based on
experience and engineering judgment. Because of the
complexities of dynamic modeling, this approach
usually is not worth the slightly increased accuracy
gained, especially since other parameters may affect
the accuracy of the final answer. Concentration is
typically estimated as a fraction of the equilibrium
solubility.
The volumetric flux of contaminated water can be
calculated in two ways, one for solid wastes and one
for liquid wastes.
(a) For landfilled solids, the only liquid present is
water percolating into the fill. For uncovered landfills,
this can range from the infiltration fraction of the
rainfall, to the full precipitation (if no rain runs off of
the fill before infiltrating), to larger flows of water if the
site is exposed to stormwater run-on from an
adjacent area. For covered landfills, the infiltration
fraction may be limited by the permeability of the
cover. Typically in wet climates the cover permeability
is limiting, while in dry climates the permeability does
not limit percolation, and normal soil percolation ratios
can be used.
The loading rate to ground water can be calculated
with the following equation:
where
(2-32)
contaminant loading rate, (mass/time).
percolation rate, see Equation 3-14
for calculation of q, (length/time).
area of landfill, (length squared).
solubility of solid chemical,
(mass/volume).
(b) For lagooned or landfilled liquids, precipitation
has a minimal influence on leachate generation, as
liquid waste will percolate to the watertable under the
influence of gravity. The rate-determining step is the
permeability of the liner or underlying soil (if there is
no liner). For liquids, the following form of Darcy's law
should be used to estimate the volumetric flux leaving
the site.
= K*i*A
(2-33)
Q1 = volume loading rate, (volume/time).
Ks = Darcy's coefficient; for unlined lagoons
use native soil hydraulic conductivity;
conductivity (length/time) (see Chapter
3 for sources of hydraulic conductivity).
i = hydraulic gradient, (length/length).
Equations 2-33 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 watertable (when
the Ks is for native soil). In these
cases the term "i" can be taken as 1.
A = area of lagoon, (length squared).
This QJ is then used to estimate mass loadings
with the following equation:
Lc=C,*Qi
where
LC
Cs
(2-34)
Q
= contaminant loading rate, (mass/time).
= contaminant concentration in lagoon
fluid, (mass/volume).
= volume loading rate, (volume/time).
Equations 2-33 and 2-34 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 watertable. 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 2-33 and 2-34 are appropriate
when analyzing lagoon releases, but should not be
used for spills or other conditions where the
chemicals on the surface do not pond for a long time.
In these conditions, the assumption of saturated flow
(through the liner or soil) may be violated.
Equations 2-33 and 2-34 apply to liquids that are
mostly water. For lagoons that contain organic fluids,
however, the equations may need to be corrected.
For liquids with a density or viscosity that differs from
water, correct Ks for this different viscosity and
density by calculating the term Kc, using the
following:
K, = Kw * DC/DW * Uw/uc
where
(2-35)
s term = hydraulic
of contaminant,
where
Kc = corrected K
conductivity
(length/time).
Kw = hydraulic conductivity of ground water,
(length/time).
D = density of liquids: c=contaminant,
w = water, (mass/volume).
30
-------
U = dynamic viscosity of liquids: c = contam-
inant, w = water, (mass/length * time).
and then substituting Kc for Ks in Equation 2-33.
(2) Estimating Release Rate from Facilities Lined with
Flexible Membranes
The release rate from an intact lined landfill or lagoon
can be calculated for a small group of contaminants.
Failed liners can be modeled as a function of the
extent of the failure using the modeling equations for
clay or natural soil-lined facilities. Although a flexible
membrane (FML) liner appears to allow no migration
through the barrier, it may indeed be penetrated by
organic compounds and contaminated water, although
the rate of permeation is understandably small. The
rate at which a contaminant permeates through a
polymeric material has been shown to be dependent
upon various properties of the permeant, such as
size, shape, polarity, and other factors (Steingiser et
al. 1978).
Salame and others proposed the use of a
permeability equation to predict the rate of permeation
of liquids and gases through various polymers
(Salame 1961, 1973, 1985; Steingiser et al. 1978):
Ps = Ap0e-sH
where
(2-36)
Ps
AP
SH
0
permeation rate, (g-mil/100
in2*day*cmHg).
constant solely dependent on the type
of polymers used, (g-mil/100
inn*day*cmHg).
constant solely dependent on the type
of polymers used, (cc/cal).
the polymer "permachor" calculated for
each polymer-permeant pair, (cal/cc).
Salame lists values for these parameters obtained
from his extensive experimental work. These values
are shown in Tables 2-7, 2-8, 2-9, and 2-10.
For permeation of water through FMLs, polymers are
categorized into five groups based on the values of
the solubility parameter as shown in Table 2-8. 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 amide
(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 2-9 and
2-10, respectively. The water "permachor" values
for various polymers given in Table 2-10 apply under
dry conditions. For water permeation under wet
conditions, permachor values may be reduced by
about 20 percent.
The term P can be used to calculate the release rate
in grams/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:
LC = P,*A*p/d!
where
(2-37)
LC
PS
A
P
de
contaminant loading rate, (mass/time).
permeation rate, (g-mil/IOO
in2*day*cmHg).
area of liner, (in units of 100 in2).
vapor pressure, (cmHg).
thickness of the liner, (mils).
2.5.2 In-Depth Analysis
In-depth analytical approaches for quantification of
baseline contaminant release to ground water involve
the use of computerized models. Refer to Chapter 3
of this manual for a detailed discussion of the nature
and applications of such modeling tools.
2.5.3 Long-Term and Short-Term Release
Calculation
For toxic substance release to ground-water
systems, directly calculate the short-term
(maximum) release values from the measured surface
and subsoil contaminant concentrations using the
tools discussed in this section. Obtain long-term
(average) values by applying the procedure previously
outlined for particulate releases to air (see Section
2.3.3).
2.6 Soil Contamination
2.6.7 Beginning Quantitative Analysis
No estimation methods are presented for analysis of
surface soil contamination. Site soils will be sampled
directly and the degree and extent of their
contamination delineated during the Remedial
Investigation. Sampling and analysis may also have
been conducted for subsurface soils. In certain cases,
however, it may be desirable to project subsurface
contamination without conducting unsaturated zone
sampling. USEPA (1987a) covers soil sampling
strategies.
2.6.2 In-Depth Analysis
Surface soil monitoring, usually conducted during the
Remedial Investigation, constitutes in-depth
quantitative analysis. Subsurface (unsaturated zone)
in-depth analysis will usually involve application of
sampling and modeling approaches. Sampling and
31
-------
Table 2-7. Parameter Values for Permeation Equation (at 25 °C)
Liquid organics in3
Water in polymer categoryb
Parameter
A a-mil
100 in2 day cmHg
s (cc/cal)
0(cal/cc)
PE
1 x 104
0.506
Table 2-gc
NVC
1 x 104
0.23
1 2
11.5 10.2
0.16 0.135
Table 2-K
3
5.4 x 102
0.115
3C
4
25
0.035
5
(d)
0.099
aSource: Salame n.d.; Salame 1961.
bSee Table 2-8 regarding polymer category. Source: Salame 1985.
cSee the table indicated for these values.
dA = 0.33 exp (0.064 xO2), where o is the solubility parameter, (cal/cc)1'2.
Table 2-8. Polymer Categorization for Permeation of
Water
Polymer group
Categorization
1 Any polymer with 8 < 8.9a
2 Any polymer with 8.9 < 8 < 10
3 Any polymer with 10 < 8 < 11
4 Polymer containing nitrile (CN) group with 10.5 < 8
5 Polymer with H bonding and 11 < 8
a8 = Solubility parameter (cal/cc)1/2
Source: Salame 1985.
analysis can provide a direct quantification of the
degree of contamination in subsurface soils.
Alternatively, computer models (e.g., SESOIL;
Bonazountas and Wagner 1981) are used to project
the level of unsaturated zone contamination over time
from surface placement of toxics. Refer to Chapter 3
of this manual for a detailed discussion of computer
models that can be applied to the unsaturated zone
contamination estimation.
Table 2-9. Permachor Values of Some Organic Liquids in
Polyethylene and PVCa
In nonpolar polymer In polar polymer
Liquid
Acetic acid
Benzaldehyde
Benzene
2-Butoxy ethanol
Butyl acetate
Butyl alcohol
Butyl ether
Butyraldehyde
Capryllc acid
Carbon tetrachloride
p-chlorotoluene
Cyclohexane
Dibutylphthalate
Diethylamine
Ethanol
Heptane
Hexane
Methyl ethyl ketone
Methanol
Nitroethane
i-Pentyl propionate
i-Propyl amine
Trichloroethylene
o-Xylene
p-Xvlene
13.0
15.9
5.4
24.4
13.0
18.0
10.4
13.5
19.0
5.8
7,6
7.0
31.4
10.0
16.0
7.0
6.0
12.5
15.0
15.4
15.0
11.0
5.4
9.4
7,4
44.0
4.0
7.0
75.0
5.0
50.0
46.0
0.0
50.0
22.0
7,5
45.0
17.0
5.7
48.0
44.0
43.0
1.0
47.0
7,0
7,0
6.7
3.0
11.0
9.0
aPolyethylene and PVC are nonpolar and polar polymers,
respectively.
Sources: Salame n.d.; Steingiser et al. 1978.
32
-------
Table 2-10. Water Permachor Value for Dry
Polymers
Permachor
Polymer value (0)
Polyvinyl alcohol TCU
Polyacrylonitrile 109
Cellulose (dry) 97
Polyvinylidene chloride 87
Polycaprolactam (dry) 80
Polyacrylonitrile styrene (70/30) (Lopac) 76
Polyacrylonitrile styrene/butadiene 75
(70/23/7) (Cycopac930)
Polychlorotrifluoroethylene 71
Polyethylene terephthalate 68
Polyvinylidene fluoride (Kynar) 67
Polyacrylonitrile styrene/ = tibutadiene 65
(56/27/4/13) (Cycopa\c 920)
Polyvinyl chloride 62
Polyoxymethylene (Delrin) 57
Polymethyl methacrylate 55
Polyvinyl acetate (dry) 45
Polystrene/acrylonitrile (74/26) 45
Polyethylene (HD) 40
Polysulfone 34
Polypropylene 33
Polycarbonate (Lexan3) 33
Polystyrene 28
Polyethylene (LD) 26
Polyisobutylene 17
Polyethylene/vinyl acetate (85/15) 15
Polybutadiene 8
Polymethyle pentene (IPX) 8
Polydimethyl siloxane (dry) -4
Sources: Salame 1961; Salame n.d.; Steingiser et al.
1978.
33
-------
Preceeding Page Blank
Chapter 3
Contaminant Fate Analysis
3.1 Introduction
This chapter provides guidance for evaluating the
transport, transformation, and fate of contaminants in
the environment following their release from an
uncontrolled hazardous waste site. The contaminant
release rate estimates described in the previous
chapter provide the basis for contaminant fate
analysis. The results form the basis for subsequent
analysis of exposed populations and estimation of the
levels of exposure incurred (see Appendix A). The
goal of contaminant fate analysis is to identify off-
site areas affected by contaminant migration and to
determine contaminant concentrations in these areas.
The following sections address analysis of
atmospheric fate, surface water fate, ground-water
fate, and biotic fate. Within each of those sections,
contaminant transport is addressed (except for biotic
fate analysis, which does not involve contaminant
transport). A screening analysis is conducted to
provide an initial qualitative assessment of
contaminant transport in the environment. It is
designed to (1) identify each transport process
governing the movement of various contaminants
within and among environmental media, (2) determine
the direction and roughly gauge the rate of
contaminant movement from the site, and (3) identify
areas to which contaminants have been or may be
transported. Screening analysis is designed both to
provide initial organization and direction for
subsequent in-depth analysis of contaminant
environmental transport, and to provide a consistent
basis for analysis from site to site.
When likely pathways of contaminant migration have
been identified by screening analysis, those pathways
requiring further evaluation are quantitatively
addressed. Like analysis of contaminant release, this
analysis can involve either the use of "desktop"
analytical solutions or numeric methodology.
Simplified environmental fate estimation procedures
are based on the predominant mechanisms of
transport within each medium, and they generally
disregard intermedia transfer or transformation
processes. In general, they produce conservative
estimates (i.e., reasonable upper bounds) for final
ambient concentrations and the extent of hazardous
substance migration. However, caution should be
taken to avoid using inappropriate analytical methods
that underestimate or overlook significant pathways
that impact human health.
When more in-depth analysis of environmental fate
is required, the analyst must select the modeling
procedure that is most appropriate to the
circumstances. In general, the more sophisticated
models are more data-, time-, and resource-
intensive.
The following criteria should be considered when
selecting an in-depth environmental fate model or
method:
Capability of the model to account for important
transport, transformation, and transfer
mechanisms;
* "Fit" of the model to site-specific and
substance-specific parameters;
x* Data requirements of the model, compared to
availability and reliability of site information; and
x Form and content of model output. This refers to
the model's ability to address important questions
regarding human exposure or environmental
effects and to provide all data required as input to
further analysis.
Information regarding the major environmental
processes that may affect the fate of hazardous
substances in each medium is provided. These
processes include transformation and intermedia
transfer mechanisms, as well as the more complex
transport mechanisms that are not incorporated into
estimation procedures. By comparing the list of
important processes identified for the site with the
summary of model features presented at the end of
each section, the analyst can select the model best
suited to the requirements of the site.
The Graphical Exposure Modeling System (GEMS),
developed by the EPA's Exposure Evaluation Division
(EED), Office of Toxic Substances (OTS) is a set of
35
-------
computer models that is easily accessible and has
the ability to produce sophisticated analyses of
environmental fate. GEMS consists of models
capable of assessing contaminant fate in air, surface
water, ground water, and soil. These fate models
contain pertinent data files (including nationwide soil,
land use, and meteorological data, and data on many
major river systems, lakes, and reservoirs); user-
input data manipulation and storage capabilities;
statistical processing programs; and such graphics
capabilities as presentation of results in map form.
GEMS is designed to be user-friendly. Although
environmental fate modeling experience is highly
desirable, personnel with no computer programming
background can also use the system because of its
progressive menu and user prompting formats. At
each decision point, the user is presented with a list
of possible selections. When specific data are
required to activate a program, the system requests
each type of data needed and the units required. At
any point in the procedure, the user can request help
from the system, and a clear explanation of the
choices or steps facing the user is provided.
The GEMS host computer is a Vax-11/780, which is
located at the EPA National Computer System at
Research Triangle Park, North Carolina. The system
can be accessed and used with the following terminal
types: DEC UT-100 series, Tektronix 4014 series,
and ASCII.
Terminals must be capable of transmitting or
receiving ASCII data in full duplex mode, using even
parity and seven-bit data word length, with
communication rates of 300 or 1200 bits per second.
Most common acoustic modems are compatible
(GSC 1982).*
Monitoring data can also be useful in analyzing
contaminant transport and fate. Monitoring results can
provide, however, only a measurement of the existing
extent of contamination. In addition, monitoring data
alone may not allow the analyst to discriminate the
contributions of specific sources to measured
contaminant loadings. In all assessments, some
degree of modeling contaminant movement within and
among environmental media will be necessary to
predict the associated exposure over a 70-year
lifetime. Thus, a combination of monitoring and
modeling techniques will be necessary to conduct an
analysis of contaminant fate for exposure assessment
purposes.
For in-depth guidance in selecting and running a
computer model to use in analyzing contaminant
migration from a particular site, the analyst should
review the following guides:
-USEPA(1977a):
USEPA (I986b):
-USEPA(1987d):
-USEPA (1986a):
Guidelines for Air Quality
Maintenance Planning and
Analysis, Volume 10
(Revised): Procedures for
Evaluating Air Quality Impact
of New Stationary Sources
Guideline on Air Quality
Models (Revised) 1986 and
Supplement A (1987)
Surface Water Model Se-
lection Criteria
Ground Water Model Se-
lection Criteria
* Contact personnel within the EED are Ms. Patricia Harrigan,
Mr. Loren Hall, or Mr. Russell Kinnerson. They can be reached
at EPA, Washington, D.C., (202) 382-3931.
- USEPA (1985J): Modeling Remedial Actions
In addition, it is recommended that the analyst obtain
the user's manual for any model selected before
attempting its application.
For contaminant fate in estuaries and reservoirs, the
analyst should review Mills et al. (1982).
To evaluate the retardation of contaminant plumes
composed of mixed wastes in ground-water systems
the analyst is referred to the following references for
detailed guidance: Nkedi-Kizza et al. (1985), Rao et
al. (1985), Woodburn et al. (1986).
3.2 Contaminant Fate Screening
Figures 3-I through 3-4 present the decision
networks for screening contaminant fate in air,
surface water, ground water, and biota. Any migration
pathways (identified in the qualitative evaluation) that
will require additional analysis are described in
Sections 3.3 through 3.6. These pathways will be
further evaluated to determine the likelihood of
population exposure as described in Appendix A.
In Sections 3.2.1 through 3.2.4, brief guidance is
provided for the qualitative evaluation of contaminant
migration pathways. The paragraphs presented below
are keyed to the accompanying decision networks
and are intended to provide further elaboration of
those boxes in the decision networks.
3.2.7 Atmospheric Fate
The following numbered paragraphs each refer to
particular numbered boxes in the Figure 3-1.
1. The atmospheric fate of contaminants must be
assessed whenever it is determined that significant
36
-------
Figure 3-1. Environmental fate screening assessment decision network: atmosphere.
Contaminant Release
Screening Assessment
Potential Volatilization of
Contaminants from Site
_E
Potential Release of Fugitive Dust/
Contaminated Particulates from Site
Consider Direction and Rate of Contaminant
Migration within Air Medium.
Major Mechanisms: Wind Currents, Dispersion
s.
Consider Direction and Distance of
Paniculate Movement with Wind Currents.
Major Mechanisms: Wind Speed, Particle Size.
Gravitational Settling, Precipitation.
El
CO
Will Settleout and Rainout
Potentially Result in Sufficient
Soil Contamination to Bring
About Leaching to Ground Water?
Will Contaminants Potentially
Reach Agricultural, Hunting,
or Fishing Areas?
Yes
J3
Determine Probable Boundaries
of Elevated Concentrations.
No
Yes
Consider Contaminant Transfer
to Ground Water. Assess Fate
Associated with
This Medium (See Figure 3-3)
Consider Transfer of Contaminants
to Biota Used by Humans. Assess Fate
with This Medium (See Figure 3-4)
J
Will
Contaminants Potentially Reach
Surface Waterbodies? .-
Yes
Identify
Populations Exposed Directly
to Atmospheric Contaminants
(See Appendix A)
Consider Transfer of Contaminants
to Surface Water. Assess Fate
Associated with
This Medium (See Figure 3-2)
-------
gaseous or airborne particulate contaminants are
released from the site. The atmospheric fate of
contaminants released originally to other media, but
eventually partitioning to the atmosphere beyond site
boundaries, must also be assessed whenever this
intermedia transfer is likely to be significant.
2. The predominant directions of contaminant
movement will be determined by relative directional
frequencies of wind over the site (as reflected in
area-specific wind rose data). Atmospheric stability
and wind speeds determine off-site areas affected
by ambient concentrations of gaseous contaminants.
Usually, high stability and low wind speed conditions
result in higher atmospheric concentrations of
gaseous contaminants close to the site. High stability
and moderate wind speeds result in moderate
concentrations over a larger downwind area. Low
stability or high wind speed conditions cause greater
dispersion and dilution of contaminants, resulting in
lower concentrations over larger areas.
For particulate contaminants (including those
adsorbed to dust or soil particles), ambient
concentrations in the atmosphere and areas affected
by airborne contaminants are determined by
windspeed and stability and also by particle size
distribution. High winds result in greater dispersion
and cause particulates to remain airborne longer
(which may also increase release rates). Low winds
and high stability will result in rapid settleout of
particulates and in a more concentrated contaminant
plume closer to the site. Larger particles will settle
rapidly, decreasing the atmospheric concentrations
with distance from the site. Finer particles will remain
airborne longer, and their behavior will more closely
approximate that of gaseous contaminants, as
described above.
3. Settleout and rainout are important mechanisms
of contaminant transfer from the atmospheric media
to both surface soils and surface waters. Rates of
contaminant transfer caused by these mechanisms
are difficult to assess qualitatively; however, they
increase with increasing soil adsorption coefficients,
solubility (for particulate contaminants or those
adsorbed to particulates), particle size, and
precipitation frequency.
Areas affected by significant atmospheric
concentrations of contaminants exhibiting the above
physical/chemical properties should also be
considered as potentially affected by contaminant
rainout and settleout to surface media. Contaminants
dissolved in rainwater may percolate to ground water,
run off or fall directly into surface waters, and adsorb
to unsaturated soils. Contaminants settling to the
surface through dry deposition may dissolve in or
become suspended in surface waters, or may be
leached into unsaturated soils and ground water by
subsequent rainfall. Dry deposition may also result in
formation of a layer of relatively high contamination at
the soil surface. When such intermedia transfers are
likely, one should assess the fate of contaminants in
the receiving media.
4. If areas identified as likely to receive significant
atmospheric contaminant concentrations include
areas supporting edible biota, the biouptake of
contaminants must be considered as a possible
environmental fate pathway. Direct biouptake from
atmosphere is a potential fate mechanism for
lipophilic contaminants. Biouptake from soil or water
following transfer of contaminants to these media
must also be considered as part of the screening
assessments of these media; for example,
hexachlorobenzene was found to accumulate in
plants (Russell et al. 1971, Gillet 1980, Trabelka and
Garten 1982).
3.2.2 Surface Water Fate
The following numbered paragraphs each refer to
particular numbered boxes in the Figure 3-2.
1. The aquatic fate of contaminants released from
the CERCLA site as well as those transferred to
surface water from other media beyond site
boundaries must be considered.
2. Direction of contaminant movement will usually
only be clear for contaminants introduced to rivers
and streams. Currents, thermal stratification or
eddies, tidal pumping, and flushing in impoundments
and estuaries render qualitative screening
assessment of contaminant directional transport
highly conjectural for these types of waterbodies. In
most cases, entire waterbodies receiving
contaminants must be considered potentially
significant human exposure points. More in-depth
analyses or survey data may subsequently identify
contaminated and unaffected regions of these
waterbodies.
3. Similarly, contaminant concentrations in rivers or
streams can be roughly assessed based on rate of
contaminant introduction and dilution volumes.
Estuary or impoundment concentration regimes are
highly dependent on the transport mechanisms
enumerated above. Contaminants may be localized
and remain concentrated, or disperse rapidly and
become diluted to insignificant levels. The
conservative approach is to conduct a more in-depth
assessment and use model results or survey data as
a basis for determining contaminant concentration
levels.
4. Important intermedia transfer mechanisms that
must be considered where significant surface water
contamination is expected include transfers to ground
water where hydrogeology of the area indicates
significant surface-water/ground-water exchange;
transfers to biota where waters contaminated with
38
-------
Figure 3-2. Environmental fate screening assessment decision network: surface water.
Contaminant Release
Screening Assessment
Potential Release of Hazardous
Substance to Surface Water-body
IT
Consider Direction and Rate of Contaminant
Migration Within Waterbody.
Assess Distance Downstream, or Areas of Lakes and Estuaries
Major Mechanisms: Currents in Affected Rivers or Streams;
Dispersion in Impoundments; Tidal Currents and .»_
Flushing in Estuaries I 2
Co
CO
Estimate Surface Water Contaminant
Concentrations
Major Factors: Source Release Strength, Dilution Volume
Is Exchange of Water
Between Surface Waterbodies
and Ground Water Significant?
Li
JL
Is Water Used for Irrigation or Watering
Livestock, or Does Waterbody Support
Commercial or Sport Fish Population?.-
Consider Transfer of Contaminants
to Ground Water. Assess Fate in This
Medium (See Figure 3-3)
Consider Transfer of Contaminants
to Biota Used by Humans. Assess
Fate Associated with
This Medium
(See Figure 3-4)
Is Hazardous Substance Volatile?
Identify Human
Populations Exposed Directly
to Surface Waters
(see Appendix A)
Consider Transfer of Contaminants
to Air Medium. Assess Fate Associated
with This Medium
(See Figure 3-1)
-------
lipophilic substances support edible biotic species;
and transfer to the atmosphere where surface water
is contaminated by volatile substances. High
temperatures, high surface-area-to-volume ratios,
high wind conditions, or turbulent stream flow also
enhance volatilization rates.
Contaminant transfer to bed sediments represents
another significant transfer mechanism, especially in
cases where contaminants are in the form of
suspended solids, or are dissolved, hydrophobic
substances that can become adsorbed by organic
matter in bed sediments. For the purposes of this
manual, sediments and water are considered part of a
single system because of their complex
interassociation. Surface water/bed sediment transfer
is reversible; bed sediments often act as temporary
repositories for contaminants and gradually re-
release contaminants to surface waters. Sorbed or
settled contaminants are frequently transported with
bed sediment migration or flow. Transfer of sorbed
contaminants to bottom-dwelling, edible biota
represents a fate pathway potentially resulting in
human exposure. Where this transfer mechanism
appears likely, the biotic fate of contaminants should
be assessed.
3.2.3 Soil and Ground-water Fate
The following numbered paragraphs each refer to
particular numbered boxes in Figure 3-3.
1. The fate of contaminants in the soil medium is
assessed whenever the contaminant release
atmospheric or fate screening assessments results
show that significant contamination of soils is likely.
2. The most significant contaminant movement in
soils is a function of liquid movement. Dry, soluble
contaminants dissolved in precipitation, run-on, or
human-applied water will migrate through percolation
into the soil. Migration rates are a function of net
water recharge rates and contaminant solubility.
Liquid contaminants may percolate directly into soils.
Organic liquids may alter soil permeabilities or may be
of lower viscosity and/or higher density than water,
resulting in percolation rates many times greater than
that of water. Contaminants with high soil adsorption
coefficients may bind to soils and become relatively
immobile.
3. Important intermedia transfer mechanisms
affecting soil contaminants include volatilization or
resuspension to the atmosphere and biouptake by
plants and soil organisms. These, in turn, introduce
contaminants to the food chain.
4. The fate of contaminants in ground water is
assessed whenever site contaminant release
screening analysis indicates direct introduction of
contaminants to ground water (e.g., through disposal
wells or fluid releases to an aquifer near the ground
surface), or whenever the screening assessments of
atmospheric, surface water, or soil contaminant fates
(as outlined above) indicate potential contaminant
transfer to ground water.
5. The qualitative assessment of ground-water flow
is often based on the assumption that subsurface
hydrologic gradients (which determine flow directions
and rates) approximate surface topography. This
approach is unreliable and should be used only in the
absence of hydrogeologic data. Ground-water flow is
influenced by many factors including hydraulic
conductivity of soils, hydraulic gradient, presence of
subsurface impermeable barriers, presence of
discharge areas (e.g., streams intercepting ground-
water flow) and presence of fissures, cavities, or
macropores. Hydrogeologic survey data (where
available) provide a more reliable basis for
contaminant transport assessment than do surface
topographs.
6. Site and surrounding community survey data
describing the location of wells are compared with the
expected subsurface contaminant plume boundaries
to identify locations of potential exposure points.
7. Important mechanisms of contaminant transfer
from ground water to other environmental media
include contaminated water exchange between
surface waters and ground water and uptake of
contaminants by edible biota. The former mechanism
must be considered whenever surface waters are
downgradient from the CERCLA site; it increases in
likelihood with closer proximity of these surface
waters to the site. Available hydrogeologic information
for the site and surroundings should be reviewed for
any indication that the aquifer underlying the site is
connected to surface waters.
The second major intermedia transfer mechanism,
biouptake, may occur through two pathways: (1)
direct exposure of plants and lower trophic level
animals to contaminated ground water in regions
where the ground-water level is close to or at the
soil surface (e.g., marshy areas, areas adjacent to
aquifer discharge points), and (2) biotic exposure to
ground water resulting from human activities such as
irrigation or watering of livestock with well water.
3.2.4 Biotic Fate
The following numbered paragraphs each refer to
particular numbered boxes in Figure 3-4.
1. A screening environmental fate assessment for
the biotic medium is performed after the fate of
contaminants in the atmosphere, surface waters, or
ground water has been assessed. Starting with the
expected distribution of contaminants in each of these
media, potential points of biotic contact with
40
-------
Figure 3-3. Environmental fate screening assessment decision network: soils and groundwater
Contaminant Release
Screening Assessment
Release to Soils at
or Surrounding Site
Consider Rate of Contaminant
Percolation Through Unsaturated
Soils, Based on Soil Permeabilities,
Water, or Liquid Recharge Rates.rj
Release to Ground Water
Beneath Site _
Will Contaminants
Potentially
Reach Ground Water?
Does Contaminated
Soil Support
Edible Species?
Consider Direction and Rate of
Ground Water Flow, Using Available
Hydrogeologic Data, or by Assuming These
Will Approximate Surface Topography,
No
Are Contaminants Volatile?
Are Contaminants in
Fine Paniculate Form on—
Sorbed to Particulates? I3
t
Could Contaminants Reach
A Surface Waterbody?
[7
\
t
Could Contaminants Reach
Any Wells Located
Downgradient? rr-
i '
F
1
Is Plume Sufficie
Surface to Allow Dir
of Contaminated
i
Consider Transfer of Contaminants
to Surface Water Medium, Assess Fate
in This Medium (See Figure 3-2)
M
!Ez
Is Well Water Used for Irrigation
or Watering of Livestock?
nl
Identify
Human Populations
Directly Exposed to
Well Water (Appendix A)
Yes
Consider Transfer of Contaminants
To Biota Used by Humans. Assess Fate
Associated with This Medium
(See Figure 3-4)
Consider Transfer of Contaminants
to Atmosphere. Assess Fate Associated
with This Medium (See Figure 3-1)
Identify
Human Populations
Exposed Directly to Soils
(Appendix A) ;
-------
contaminated media and important affected biotic
species are identified.
2. Important species are those used directly by man
(game animals, sport or commercial fish, crustaceans
and mollusks, agricultural crops and livestock;
naturally-occurring fruits, herbs, other edible
vegetation), and those that introduce contaminants to
species used by man through the food chain (e.g.,
livestock feed crops; or plants and lower trophic-
level animals consumed by any of the animal groups
listed above).
3. Assessed mechanisms of transport in the biotic
medium include the food chain, natural animal
migration, or human commercial activity. Food chain
transport can result in high concentrations of
contaminants in the tissue of edible species not in
direct contact with contaminated air or water. Human
commercial transport and natural migratory behavior
of contaminated species can result in a wide
distribution of edible species or tissue-containing
contaminants.
4. Edible tissue concentrations are a function of the
level and type of biotic exposure to contaminants, the
partitioning of contaminants between organic tissue
and substrate media, the biodegradability of
contaminants, organism-specific metabolic charac-
teristics, and ecosystem characteristics.
3.3 Quantitative Analysis of Atmospheric
Fate
3.3.1 Screening Analysis
The atmospheric fate of substances released from
uncontrolled hazardous waste sites can be estimated
by using the following equation to estimate ground-
level atmospheric concentrations of pollutants at
selected points on a centerline of a plume directly
downwind from a ground-level source (Turner 1970):
Figure 3-4.
C(X) = -
Q
no a u
y *
(3-1)
Environmental fate screening assessment
decision network: food chain.
Ambient Contaminant Concentration
and Distribution Estimates from
Air, Surface Water, Ground Water
Screening Fate Analyses
Potential Biotic
Exposure to
ContamlmBTte
Consider Biotic Species Within Areas of
Elevated Ambient Hazardous Substance
Concentrations as Potential Vectors
of Hazardous Substances r
Consider Transport of Hazardous Material
Within Biologic Medium
Major Mechanisms: Human Commercial Activity,
Organism Migration, Movement of Hazardous
Material Through Food Chain.
Identify Edible Biotic Species
Affected Indirectly Through
Food Chain I
Assess Potential Edible Tissue
Concentrations, Distribution of
Contaminated Organisms
Identify
Exposed Human Populations
(Appendix A)
where
C(X)
Q
= concentration of substance at distance
x from site, (mass/volume).
= release rate of substance from site,
(mass/time).
= dispersion coefficient in the lateral
(crosswind) direction, (distance).
= dispersion coefficient in the vertical
direction, (distance).
= mean wind speed, (distance/time).
= the value pi = 3.14.
The appropriate dispersion coefficients can be
obtained from Figures 3-5 and 3-6. These figures
provide values for 6y and 6Z, respectively, as
functions of downwind distance, x, and stability
classes A though F. These stability classes are based
on the Pasquill stability classification system, where
Class A is very unstable and Class F is moderately
stable (Pasquill 1961). Table 3-1 presents a brief
illustration of how stability classes are defined.
42
-------
Figure 3-5. Horizontal dispersion coefficient as a function of downwind distance from the source (from Turner 1970).
10,000-
1,000-
P
e
-------
Figure 3-6. Vertical dispersion coefficient as a function of downwind distance from the source (from Turner 1970).
1,000
10-
B-^
E --
Fj_.
1.0'
0.1
1 10
Distance Downwind, km
100
"Curves designated A through F represent dispersion coefficient functions for atmospheric stability classes A through F. See text
for sources of atmospheric stability data.
44
-------
Table 3-1. Key to Stability Categories
Night
Thinly
Surface wind overcast or
speed at a Day incoming Solar radiation > 4/8 Low < 3/8
Height of 10 (insolation) cloud Cloud
m (m/sec)
<2
2-3
3-5
5-6
>6
Strong Moderate Slight
A A-B B
A-B B C
B B-C C
C C-D D
C D D
Cover
E
D
D
D
Cover
F
E
D
D
The neutral class (D) should be assumed for all overcast
conditions during day or night.
"Appropriate insolation categories may be determined through the
use of sky cover and solar elevation information as follows:
Sky cover
4/8 or Less or
Any Amount of
High Thin Clouds
5/8 to 7/8 Middle
Clouds (7000
feet to 16,000
foot base)
5/8 to 7/8 LOW
Clouds (less than
7000 foot base)
Solar
elevation
angle > 60°
Strong
Moderate
Slight
Solar
elevation
angle < 60°
but > 35°
Moderate
Slight
Slight
Solar
elevation
angle < 35°
but > 15°
Slight
Slight
Slight
Source: USEPA 1977b
To obtain the maximum hourly concentration, select
the calculational methodology for coning and fanning
plumes in USEPA (1977b). To obtain the estimated
maximum concentration for a 3-, 8-, or 24-hour
averaging time, multiply the l-hour maximum by the
factors given in USEPA (1977b).
To estimate long-term mean atmospheric
concentrations, obtain STAR (Stability Array) data
specific to the site. These data provide seasonal or
annual joint frequencies for each stability class, wind
direction, and wind speed category. Assume an
annual average wind speed of 3 meters/second, and
calculate the long-term mean atmospheric
concentration for each exposed population by
applying a weighted average, based on the relative
frequency of each stability class and of wind flow
toward selected exposure points. Equation 3-2
provides a rough weighted average estimate (Turner
1970):
C (x) = W(x) [CA(x)fA + CB(x)fB + Cc(x)fc
+ CD(x)fD + CE(x)fE + CF(x)fF] (3 - 2)
where
C(x)
= average concentration at point x over
long term.
CA(x) = concentration at point x during stability
class A (from Equation 3-1).
fA = relative annual frequency of stability
class A for the specified wind direction.
and subscripts B through F represent the various
stability classes.
Note that this estimate is a rough approximation
because it is simplified by the assumption that the
mean wind speed is 3 meters/second for all stability
classes. A more sophisticated estimate can be made
by incorporating site-specific wind speed frequency
data, and performing similar weighted average
calculation of ambient concentrations. This is a time-
consuming procedure, however, and the use of
computer-based estimation procedures may be
more cost-effective if sophisticated estimates are
required. STAR data are available from the National
Climatic Center (NCC), Asheville, North Carolina
(phone: (704) 259-0205) for all National Weather
Service (NWS) locations in the U.S. The NWS Station
that is most representative of the site should be used.
The area within which the ground-level
concentration of a hazardous substance is above a
predetermined critical concentration (i.e., the plume
isopleth) can be described using the following
procedures. Calculate the crosswind distance from
any point along the plume centerline (i.e.,
perpendicular to the plume centerline) to the isopleth
boundary by Equation 3-3 (Turner 1970):
where
C(CL)
V(x)
C(x)
(3-3)
predetermined critical concentration
level, (mass/volume).
perpendicular distance from point on
plume centerline to the C(CL) isopleth
boundary, (length units).
concentration at plume centerline, x
distance from source, (mass/volume,
as calculated by Equation 3-1).
lateral dispersion coefficient, (length
units).
Vary the value for x (downwind distance from the
source) input into Equations 3-1 and 3-3, starting
at a point near the site* and increasing this value until
the value for C(x) (obtained from Equation 3-1)
equals the predetermined critical concentration
C(CL). Values calculated for y describe the isopleth
boundary on either side of the plume centerline.
Equations are generally considered applicable to downwind
distances of at least 200 m.
W(x) = relative annual frequency of wind flow
towards point x.
45
-------
Estimate the area within a plume isopleth using Figure
3-7 which plots the value CfCLJ/j (relative
concentration times wind speed versus isopleth area,
for each stability class A through F).
All of the preceding simplified equations provide
atmospheric fate estimates based on several simple
assumptions, one of which requires special mention.
This is the assumption that the hazardous substance
released from a site is in a form that can remain
airborne indefinitely (i.e., either gaseous or consisting
of particles less than 20 microns in diameter) (Turner
1970).
In cases where fugitive dust blown from the site
includes solid hazardous substances (or soil
particulates carrying adsorbed hazardous substance)
of greater diameter than 20 microns, relatively rapid
gravitational settling of the larger particles occurs.
Consequently, much of the hazardous material
reaches the ground before advection and dispersion
can transport and dilute the plume as described by
the above equations. Thus, areas close to the
uncontrolled hazardous site may experience
significant soil contamination, and human exposure
points farther from the site may experience lower
atmospheric concentrations than estimated by these
equations. Hanna and Hosker (1980) present a
procedure for estimating the gravitational settling rate,
distance of travel from the source, and deposition rate
of airborne particulates.
All of the above simplified procedures incorporate the
following additional assumptions:
• Steady-state condition, i.e., windspeed is steady
at rate u, and the hazardous substance release is
continuous, at average rate Q. Wind direction is
also assumed to be steady; short-term
fluctuations are disregarded.
• Longitudinal dispersion is negligible (substance
travels at wind speed in the downwind direction).
• The substance is refractory (all removal and
decay processes are disregarded).
• The substance is distributed normally, or
according to a Gaussian distribution, both
vertically and in the crosswind direction.
• The air environment is homogeneous; wind
speeds and stability are equal at all heights above
the ground, and no obstructions to wind flow or
dispersion exist other than at the ground.
Complete reflection occurs at the ground/air
interface.
3.3.2 In-Depth Analysis
Where estimates of ambient atmospheric
concentrations of hazardous substances developed
by the preceding simplified procedures indicate that
these concentrations pose potential health hazards,
more accurate, in-depth analysis of atmospheric fate
may be required. Numerous computer models are
available for this purpose and are listed in USEPA
(1986b). These models vary in sophistication and
capability, and in their ability to incorporate
expressions describing the effect of various
processes on the atmospheric fate of hazardous
substances. The most important of these processes
are briefly described below. Consider the importance
of each of these processes to the atmospheric fate of
the substances under analysis before selecting a
computer model.
3.3.2.1 Intermedia Transfer
The following are the most important processes that
affect the removal of hazardous substances from the
air medium and their transfer to other sectors of the
environment.
(1) Dissolution
This is the process whereby hazardous substances in
the gaseous state are dissolved into water droplets
present in the atmosphere. This process, followed by
precipitation, distributes the substance over the
surface media, and percolation to ground water may
follow. Direct dissolution may also occur between
gaseous substances in the atmosphere and surface
waters at the air/water interface. Dissolution is a
constant, reversible process, the amount of haz-
ardous substance in the aqueous phase is de-
termined by the partition coefficient of the substance
between the gas and aqueous phases. This partition
coefficient is in turn a function of the vapor pressure
and water solubility of the substance, its
concentration in the air, and temperature. See Lyman
et al. (1982) or Hanna and Hosker (1980) for methods
of estimating this partition coefficient and atmospheric
half-lives resulting from dissolution/ rainout.
(2) Adsorption
Through the process of adsorption, hazardous
substances in the vapor phase become attached to
particulate matter suspended in the air (aerosols), or
onto soil particles at the air/soil media interface.
Suspended aerosols settle to surface media, thereby
removing adsorbed substances from the air
environment. The adsorption rate of a particular
substance is principally a function of the number and
surface area of aerosols per volume of air, the
molecular weight of the substance in question, its
concentration in the air, and its saturation vapor
pressure. Cupitt (1980) provides a method for
estimating atmospheric contaminant removal rates
due to adsorption to particulates and settleout.
(3) Gravitational Settling
This mechanism is most important for particulate
hazardous substances, or hazardous substances
46
-------
Figure 3-7. Area within isopleths for a ground-level source (Hilsmeir and G if ford 1962. as presented by Turner, 1970)
10"
10"
107
10"
105
104
103
10*
10'
10"
10'
10'
*Curves designated A through F represent functions for atmospheric stability classes A through F. See text for sources of atmospheric
stability data.
adsorbed onto suspended particulates, if the
particulate matter is more than 20 |jm in diameter.
These particles settle to the surface media at a rate
that is a function of their density, shape, and
diameter, and of wind speed (Hanna and Hosker
1980).
(4) Precipitation
Precipitation itself is a major mechanism for removal
of particulate and aerosol matter. Raindrops require
particulates or aerosols to serve as nuclei for their
condensation from the vapor state of water.
Moreover, raindrops generally remove particulates
and aerosols > 1.0 urn in diameter as they fall below
the cloud level.
3.3.2.2 Intramedia Transformation Processes
Many hazardous substances are subject to decay or
transformation to other substances with new
properties while entrained in the air environment. The
two most important of these processes are described
below. While the product of such transformation
processes will usually have different properties from
those of the original hazardous substance, the new
substance produced may also have hazardous
properties. Cupitt (1980) provides estimates of
47
-------
constants that determine the rate of each
transformation process below, as well as of the
importance and likely products of these processes,
for 46 hazardous materials. Hendry and Kenley (1979)
provide rate constants and estimation procedures for
these processes.
(1) Photolysis
This is the breakdown of substances because of
photochemical reaction brought about by solar
energy. Photolysis can be direct, when the hazardous
substance is itself affected by solar radiation, or
indirect when the hazardous substance reacts with
other substances that have been raised to a reactive
state by solar radiation. Photolysis rates depend on
solar radiation availability, the light absorption
coefficient of the hazardous substances, and a
reaction yield constant (which describes the efficiency
of transformation of the hazardous substance with the
available sun energy).
(2) Oxidation
The reaction of substances with oxidants in the
atmosphere can result in their transformation. The
two most important atmospheric oxidants are ozone
and the hydroxyl radical. Reaction rate constants for
oxidation are chemical specific; the overall rate of
transformation of a hazardous substance by oxidation
depends on the concentration of the oxidant and the
reaction rate constant.
3.3.2.3 The Effects of Terrain
Features such as vegetation, large buildings, urban
areas, rough topography, hills, or mountains can all
profoundly affect the atmospheric fate of airborne
substances, principally by altering the laminar flow of
transporting wind currents. The effects of terrain on
wind currents may include increased turbulence,
downwash in the lee of large obstacles, or localized
alterations in the direction of flow. Because the
release of substances from hazardous waste sites
usually occurs at ground level, the fate of these
substances is especially susceptible to the effects of
terrain. Select a model capable of accounting for
these effects in any case where these listed terrain
features exist between the site and points of human
exposure.
3.3.3 Computer Models
Tables 3-2, 3-3, and 3-4 provide general
information about computer-based models that could
be appropriate to in-depth analysis of the
atmospheric fate of substances released from
CERCLA sites. Table 3-2 contains resource
requirements, references, and sources for each
model; Table 3-3 summarizes their features and
capabilities; and Table 3-4 discusses the data
requirements of each. By comparing the information
in these tables with identified site features, site data
availability, final output requirements, and resource
availability, one can select the most applicable and
cost-effective model.
The Industrial Source Complex (ISC) long-term
model and the TOXBOX area source model are
presently integrated into the GEMS system. These
models are accessed under a subsystem of GEMS
referred to as the GEMS Atmospheric Modeling
System (GAMS). A brief description of ISC is
provided below.
The ISC (Bowers et al. 1979) is a Gaussian
dispersion model, capable of estimating the
concentration and deposition rates of gaseous and
particulate pollutants around a point, area, or line
source. Because it is integrated into the GEMS
system, it is especially useful for the analysis of the
atmospheric fate of hazardous substances. Based on
a user-input release location (in the form of
latitude/longitude coordinates or zip code), stored
climatological data from the nearest meteorological
monitoring stations are retrieved (GSC 1982).
The integration of ISC with a population distribution
model called SECPOP gives it the capability of
expressing atmospheric fate of pollutants in terms of
numbers of people affected at various concentration
levels (this capability is discussed in more detail in
Appendix A, Exposed Populations).
The ISC model can estimate the concentration of
pollutants released from point, area, or line sources.
Area sources are simulated by use of a virtual point,
and line sources by a series of points. Short-term
(hourly) or long-term (seasonal, annual average)
concentration estimates can be developed, and
gravitational settling can be simulated based on
user-input half-life data (GSC 1982).
ISC can be used with IBM, CDC, or VAX computers.
The model is implemented within GEMS on EPA's
VAX 11/780 and can be accessed with a variety of
user terminal types. (See Section 3.1 for access
instructions.)
3.3.4 Short- and Long-Term Concentration
Calculations
Long-term average ambient air concentrations of
hazardous substances at human exposure points are
estimated using the long-term average release rate
over the time period of interest, and the weighted
averaging algorithm presented as Equations 3-1 and
3-2. Annual average climatological data, or STAR
data including long-term frequencies of all
climatological parameters, should be used as input to
these equations.
Where site-specific data are unavailable, short-
term concentration levels are estimated using the
maximum short-term release rate and climatological
assumptions presented in Table 3-1. When using
48
-------
Table 3-2. Resource Requirements and Information Sources: Atmospheric Fate Models
Model
Description
Resource Requirements, comments
References, sources of documentation,
software
Box Model
Climatological Dispersion
Model (COM)
Industrial Source Complex
co
Ram
CRSTER
xx Area Source.
xx Vertical dispersion or no vertical
dispersion option.
xx Basic box model.
• Long-term seasonal or annual.
• Point or area sources.
xx Gaussian plume model.
• Simulates nonconservative pollutants.
xx Can simulate turbulence over urban
areas.
• Outputs long-term average
concentrations at user-specified
receptors.
xx Operates in both long-term and short-
term modes.
x Accounts for settling and dry deposition
of particles; downwash, area, line, and
volume sources; plume rise as a function
of downwind distance: separation of point
sources; and limited terrain adjustments.
xx Appropriate for industrial source
complexes, rural or urban areas, flat or
rolling terrain, transport distances less
than 50 kilometers, and one hour to
annual averaging times.
-, Steady-state Gaussian plume model.
Appropriate for point and area sources,
urban areas, flat terrain transport
distances 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.
xx Steady-state Gaussian dispersion
model.
xx Designed to calculate concentrations
from point sources at a single location.
xx Highest and high-second high
concentrations are calculated at each
receptor.
xx Appropriate for single point sources,
rural or urban areas, transport distances
less than 50 kilometers, and flat or rolling
terrain.
xx Available through GEMS (see Section
3.1).
xx Requires stability array data.
xx FORTRAN V program language; has
been Implemented on the UNIVAC 1110.
j«22 K bytes storage required.
x Software available as part of UNAMAP
package for $420.
r Integrated into GEMS (see Section 3.1).
< Source data: location, emission rate,
physical stack height, stack gas exit
velocity, stack inside diameter, and stack
gas temperature. Optional inputs include
source elevation, building dimensions,
particle size, distribution-with
corresponding setting velocities, and
surface reflection.
< Meteorological data: includes stability
wind rose (STAR deck), average
afternoon mixing height, average morning
mixing height, and average air
temperature.
< 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
temperature. Area sources require
location, size, emission rate, and height
of emission.
< Meteorological data: hourly surface
weather data from the preprocessor
RAMMET. Actual anemometer height is
also required.
? Available on UNIMAP (Version 6).
? Source data: emission rate, physical
stack height, stack exit velocity, stack
inside diameter and stack gas
temperature.
? Meteorological data: hourly surface
weather data from the preprocessor
RAMMET. Actual anemometer height is
also required.
Documentation: Busse and Zimmerman
1976
Software: Computer Products, NTIS,
Springfield, VA. 22161
Documentation: Bowers et al. 1979
Software: Computer Products, NTIS,
Springfield, VA. 22161
Reference: Turner and Novak, 1978.
Reference: USEPA 1977b.
(Continued)
-------
Table 3-2. (Continued)
Model
Description
Resource Requirements, comments
References, sources of documentation,
software
Texas Climatological Model Control (TCM)*
Texas Episodic Model (TEM)*
CJ!
o
Model MPTER
VALLEY"
x* Long-term (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 2,500 receptor locations on
downwind user-specific grid.
• Outputs average concentration data.
• Steady-state model.
• Point or area sources.
• Short-term - 10 minutes to 24 hours.
• Produces maximum and average
concentrations over time periods selected
by user.
• User can select up to 2,500 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.
* Multiple point source algorithm useful for
estimating air quality concentration of
relatively non-reactive pollutants.
• Appropriate for point sources, rural or
urban areas, flat or rolling terrain,
transport distances less than 50
kilometers, and one hour to one year
averaging times.
• Short- or long-term.
• Simulates plume impact in complex
terrain.
x Provides screening estimates of worst-
case short-term concentrations.
• Provides annual average concentrations.
• 12-receptor grid.
xx Requires stability array data.
* FORTRAN program language; has been
Implemented on Burroughs 6810/11.
• Batch mode.
x 17 K bytes memory required.
• Technical background in meteorology, air
pollution useful.
FORTRAN program applicable to a wide
range of computer types; has been
Implemented on Burroughs 6810/11.
Requires approximately 26 K bytes
memory.
Engineering, meteorology, atmospheric
transport background useful.
Source data: location, emission rate,
physical stack height, stack gas exit
velocity, stack inside diameter, stack gas
temperature, and optional ground level
elevation.
Meteorological data: hourly surface
weather data from the preprocessor
RAMMET. Actual anemometer height is
also required.
May require careful analysis of output by
experienced air quality modeler.
Documentation: Texas Air Control Board
1980.
Reference: Christiansen 1976.
Documentation: Pierce and Turner 1980.
Chico and Catalano 1986.
Reference: Burt 1977.
Software: Computer Products, NTIS,
FORTRAN V program, applicable to wide Springfield, VA 22161.
range of computers.
Approximately 13 K bytes memory
required.
Sources: Bonazountas et al. 1982; USEPA 1979; USEPA 1982a.
'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.
**Thus model is recommended for screening applications only.
-------
Table 3-3. Features of Atmospheric Fate Models
en
TEH
TON'
COM
VALLEY*
ISC
BOXMOO
RAM
GRSTER
MfTER
Source: Bonazountes et al. 1982; USEPA 1979; USEPA 1982a.
**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 concentrations equivalent to those provided by the preferred
models, e.g., CPM, RAM, ISC, MPTER, CRSTER. for a given application.
-------
Table 3-4. Data Requirements for Atmospheric Models
en
TEH'
TCNT
COM
VALLEY'
ISC
BOXMOD
RAM
CRSTER
MPTER
Source: Bonazountas el al. 1982; USEPA1979; USEPA 1982a.
• This model to recommended tor screening applications only.
" These models are not EPA preferred models. These models can be used H they can be
demonstrated to estimate concentrations equivalent to those provided by the preferred
models, e.g., CPM, RAM, ISC, MPTER, CRSTER, for a given application.
-------
site-specific data, the most stable atmospheric
conditions, lowest wind speed, and greatest percent
of wind flow toward the exposed population should be
used as input to Equation 3-1, along with maximum
release rate estimates for the duration of interest.
Usually, the population nearest the point or area of a
ground-level release experiences the highest short-
term exposure.
As indicated in Table 3-2, several atmospheric fate
models have the capability of producing short-term
maximum and long-term average ambient
concentration estimates where in-depth analysis is
desirable.
3.4 Surface Water Fate Analysis
The environmental fate of hazardous materials
entering surface waterbodies is highly dependent on
the type of waterbody. The three major classifications
are rivers and streams, impoundments, and estuaries.
Methods for estimating contaminant concentrations in
the first category are provided below.
As mentioned in the introduction to this chapter,
contamination of flowing waterbodies will probably be
a more common occurrence with regard to
uncontrolled hazardous waste facilities than will
contamination of impoundments or estuaries. Thus, in
this section guidance for estimating contaminant fate
in flowing waterbodies is presented. In those cases
where contaminant fate in an impoundment or estuary
is necessary, the analyst is referred to Mills et al.
(1982) for guidance.
The Probabilistic Dilution Model is an analytical tool
that can be used to extend the qualitative screening
analysis presented in the previous section and that in
some cases may make application of the quantitative
analyses discussed in following sections
unnecessary. This model has been adapted by the
U.S. Environmental Protection Agency, Office of
Toxic Substances, to support the exposure
assessment process for contaminants in surface
water. The model is based on the fact that, in
general, the most important process affecting a
contaminant's concentration in a surface waterbody is
the degree of its dilution. Thus, the model uses
streamflow data for a given subbasin and contaminant
loading data (from the contaminant release analysis
discussed in Chapter 2) to predict the number of
times per year a given contaminant concentration will
be exceeded. For contaminants that have health-
based concentration standards (or for which health-
based concentration cut-off values can be
calculated), the model can be used to predict the
annual number of occurrences (days) that
unacceptable health risks may result for persons
using the affected waterbody. This model can be
applied to the Superfund exposure assessment
process as an extended screening tool to highlight
contaminant releases to surface water that actually
require detailed environmental fate (and subsequent
exposed populations) analysis. Contact the USEPA
Office of Toxic Substances, Exposure Evaluation
Division (Pat Kennedy, (202) 382-3916) for more
detailed information on accessing the Probabilistic
Dilution Model.
3.4.1 Beginning Quantitative Analysis
The following equation (adapted from Delos et al.
1984) provides a rough estimate of the concentration
of a substance downstream from a point source
release into a flowing waterbody, after dilution of the
substance by the receiving waterbody:
c=-
where
(3-4)
Qe
Qt
= concentration of substance in stream,
(mass/volume).
= concentration of substance in effluent,
(mass/volume).
= effluent flow rate, (volume/time).
= combined effluent and stream flow
rate, (volume/time).
This equation predicts the concentration of substance
in the waterbody resulting from contaminant releases
from the subject site alone; it does not take into
account additional sources of contamination
("background" concentrations) that may also
contribute to the total level of contamination in the
waterbody.
In cases where hazardous waste is introduced into a
stream through intermedia transfer from air, soil,
ground water, or nonpoint source, or where the
release rate is known in terms of mass per unit time
rather than per unit effluent volumes, in-stream
concentrations can be estimated by use of the
following equation:
where
Tr
Qt
(3-5)
intermedia transfer rate,
(mass/time)
stream flow rate after intermedia
transfer has occurred,
(volume/time).
Assumptions implicit in these equations are the
following:
* Mixing of the hazardous substance in the water is
instantaneous and complete.
53
-------
&& The hazardous material is refractive (i.e., all
decay or removal processes are disregarded).
^ Stream flow and rate of contaminant release to
the stream are constant (i.e., steady-state
conditions).
The assumption of complete mixing of a hazardous
substance in a flowing water body is not valid within a
mixing zone downstream from the point or reach
where the substance is introduced. Under certain
conditions, this mixing zone can extend downstream
for a considerable distance, and concentrations can
be considerably higher within the mixing zone than
those estimated by the foregoing dilution equations.
The length of the mixing zone is estimated by the
following equation (adapted from Fischer et al. 1979,
Liu 1977, Neely 1982):
MZ =
where
MZ
w
u
si
g
0.4
0.6
(3-6)
mixing zone length, (length units).
width of waterbody, (length units).
stream velocity, (length/time).
stream depth, (length units).
slope of the stream channel,
(length/length).
acceleration due to gravity, (32 ft/sec2).
These equations provide in-stream contaminant
concentrations resulting from site releases only. If
total in-stream contaminant concentrations are
desired, these should be estimated by adding
background (i.e., upstream from the site) in-stream
contaminant concentrations to those estimated by
Equations 3-4 and 3-5.
If the hazardous substance is introduced into a
flowing waterbody over a length of that body, rather
than from a point source, assume that the mixing
zone begins at the downstream end of the reach over
which introduction takes place. Neely (1982) presents
an estimation procedure for hazardous substance
concentration at exposure points within a mixing zone
that incorporates an expression for dispersion.
The dilution equations (3-4, 3-5) and the procedure
presented by Neely (1982) assume that the introduced
hazardous substance is conservative. Therefore, they
predict an estimated stream/river concentration that
remains constant from the downstream end of the
mixing zone throughout the remaining length of the
stream, or decreases only with further dilution
resulting from additional stream flow from tributaries.
This is useful as a basic model for the fate of
conservative hazardous substances; for
nonconservative substances, it provides a useful
worst-case estimate. If the released substance is
found through this estimation procedure to be diluted
to concentrations below a predetermined level of
concern, and no important exposure points exist
within the mixing zone, the fate of the substance in
this medium may need no further analysis. However,
where the concentration after dilution of a
nonconservative substance is still above a
predetermined critical level, it may be useful to
estimate the distance downstream where the
concentration will remain above this level, as well as
the concentration of the substance at selected
exposure points downstream.
This type of estimation can be performed through use
of an overall decay coefficient, which represents a
combination of all decay and loss rates affecting the
removal of a substance from a waterbody. The
concentration of a nonconservative substance at a
selected point downstream from the release point and
below the mixing zone (complete mixing is assumed)
can be estimated by the following equation (from
Delos et al. 1984), which employs the concept of an
overall decay coefficient:
W(x) = W(O)e
(3-7)
where
W(x)
W(0)
e
K
x
concentration at downstream distance x,
(mass/volume).
concentration immediately below point of
introduction, (from Equations 3-4, 3-
5).
2.71828.
overall decay coefficient, (time)'1.
distance downstream from point of
introduction, (length).
stream velocity, (length/time).
The overall decay coefficient can also be used to
estimate the distance downstream over which a
nonconservative substance remains above a
predetermined critical concentration level W(CL). This
is estimated by substituting W(CL) for W(x) in
Equation 3-7, and solving this equation for x, as
follows:
u
x= — —
K
where
u
K
W(CL)
/W(CL)\
V W(O) /
(3-8)
distance downstream from point of
introduction, (length).
stream velocity, (length/time).
overall decay coefficient, (time)-1.
predetermined critical concentration
level, (massVolume).
54
-------
W(0) = concentration immediately below point of
introduction, (from dilution Equations 3-
4, 3-5).
This equation incorporates the following assumptions:
eses Mixing is complete.
* Conditions are steady state.
esx Longitudinal dispersion is negligible; the
substance transports downstream at stream
velocity.
esx All decay and transfer processes can be
described as first-order coefficients (i.e., decay
rates are a direct function of hazardous
substance concentration).
Values for K can be derived empirically where
monitoring data are available, or can be estimated
based on decay rate constants available for many
hazardous substances in the technical literature.
Concentration data from immediately below the point
of substance release into a stream (after complete
mixing of waste stream into the waterbody), and from
at least one point downstream of the mixing zone are
required for the empirical estimation of K. Note that
overall decay coefficients are substance- and site-
specific and can vary with climatic and hydrologic
conditions. Care must be taken in calibrating the
coefficient empirically. Data covering seasonal
fluctuations must be used, and seasonal values for K
corresponding to the various observed conditions, or
a worst-case K value (i.e., lowest reasonable value)
for the purpose of conservative estimation, should be
developed.
For estimation of K through the summation of
published decay rate constants, the most important
removal process affecting the compound of concern
in the receiving waterbody must be known. For this
information, see the discussion below (Section 3.4.2),
or see Callahan et al. (1979), or Mabey et al. (1992).
Additional references that provide decay rate constant
values for a wide variety of compounds include:
Verschueren (1984) Dawson et al. (1980) USCG
(1974), and Schnoor et al. (1987).
Reliable values for K, which have been developed for
a given waterbody and hazardous substance under
no-action conditions (i.e., during remedial
investigation), can be used to estimate the fate of this
same substance resulting from the release rates
projected after implementation of various remedial
action alternatives.
3.4.2 In-Depth Analysis
When aquatic concentration estimates developed by
the above simplified methods (or methods covering
estuaries or impoundments provided by Mills et al.
1982) indicate that these concentrations pose a
potential human health hazard at one or more
exposure points, more accurate estimates of short-
term and long-term concentrations of the hazardous
substance may be required. A large number of in-
depth methods and computer models exist to assess
the fate of substances in the aquatic environment.
Each of these models differs in the number and types
of aquatic fate processes that it incorporates. The
most important of these aquatic processes are
described below, and information is provided to allow
identification of those processes most likely to be
significant at the site, and for the hazardous
substances under analysis.
3.4.2.1 Intermedia Transfers
The major processes by which hazardous substances
can be transferred from surface water to other
environmental media are as follows:
(1) Volatilization
Volatilization of a substance from water depends on
the physicochemical properties of the substance and
characteristics of the waterbody and body of air
involved. Volatilization increases in importance for
substances with higher vapor pressure, and for
waterbodies with higher surface area-to-volume
ratios and higher turbulence (Delos et al. 1984).
Callahan et al. (1979) stress the importance of
volatilization as a route of intermedia transfer for 129
priority pollutants. If volatilization is considered an
important process for the substance being studied, or
if the importance of volatilization is unknown, the rate
of volatilization can be estimated by the method
provided by Mills et al. (1982) for quiescent
waterbodies or by Delos et al. (1984) for turbulent
bodies. Lyman et al. (1982) provide methods for
estimating volatilization rates from water.
(2) Sedimentation
Hazardous substances released to a surface
waterbody in the solid, particulate form will settle out
over time and become mixed into the bottom
sediment. In addition, liquid hazardous substances
with high affinities for adsorption to suspended
particulates will settle out of surface waters with these
particulates. The rate of sedimentation is governed by
the difference between settling velocity and
resuspension velocity. The former increases with
mean particle size and density and with water
temperature, and can be estimated by the procedure
presented by Delos et al. (1984). Resuspension
velocity is a function of bottom shear stress. Delos et
al. (1984) provide a procedure to estimate this rate.
Where sedimentation is considered to be an important
process, use a surface water fate model that has the
capability of accounting for bed-water exchange and
sediment load transport.
55
-------
(3) Sorption
Substances dissolved in surface waters can sorb onto
solids suspended in the water or onto bed sediments.
This process, in effect, transfers the substances from
the water to the sediment medium, and proceeds until
an equilibrium point is reached. This equilibrium point
(and the resulting water and sediment concentrations
of the substance) is determined by the soil-water
partition coefficient (a parameter that is a function of
sediment type, water pH, cation exchange capacity,
and organic content of sediment) and the
physicochemical properties of the hazardous
substance. In general, metals and hydrophobic,
nonpolar organic compounds have a high tendency to
sorb onto entrained or bottom sediment. See Lyman
et al. (1982) for methods of estimating sediment
adsorption of waterborne contaminants.
3.4.2.2 Intramedia Transformation Processes
The following is a brief description of the important
intramedia transformation processes that may be
significant for the surface water fate of hazardous
substances. Rate-controlling factors are stated for
each. Callahan et al. (1979), Mabey et al. (1982),
Verschueren (1984), and Sax (1984) provide rate
constants for these processes for numerous
compounds.
(1) Photolysis
Chemical transformation due to photolysis utilizes
energy from sunlight, and for some chemicals, can
occur by several processes. Direct photolysis rates
are a function of photon availability, light absorption
coefficients for the chemical in question, and a
reaction yield constant (i.e., the efficiency of
substance transformation with the available solar
energy). Indirect photolysis occurs through the action
of intermediate substances naturally occurring in the
medium. These intermediates absorb light energy by
various processes and in this energized state, react
with the hazardous substance. Indirect photolysis is a
function of photon availability, concentration and light
absorption coefficient of the intermediate, and a rate
constant for the reaction between the energized
intermediate and the hazardous material.
(2) Oxidation
Oxidation is the reaction of substances with oxidant
species. Oxidation rates are a function of the
concentrations of the substance in question,
concentration of the oxidant, and a rate constant for
reaction between them.
(3) Hydrolysis
Hydrolysis is the nucleophilic displacement of an
electronegative substituent on a carbon atom by an
hydroxyl group. The nucleophilic reactant can be
either a water molecule or an hydroxyl ion. Hydrolysis
of most compounds is highly dependent on the pH of
the waterbody medium and can be promoted by both
acid and base conditions. The rate of hydrolysis is a
function of the concentration of the hazardous
substance and the rate constants for the acid- and
base-promoted processes at each pH value.
(4) Biodegradation
Biodegradation is the breakdown of substances
through the enzymatic action of biota present in the
water. Most biodegradation is carried out by microbial
biota. It depends on the metabolic rates and
characteristics and the population density of the biotic
agents, which are in part functions of the availability
of other nutrients, pH and temperature of the
medium,, and sunlight availability, among other
factors.
3.4.2.3 Computer Models
Tables 3-5, 3-6, and 3-7 summarize the features,
data requirements, resource requirements, and
references or contacts for selected computer-based
models appropriate to the in-depth analysis of the
aquatic fate of hazardous releases from Superfund
sites. Additional details for certain of the models
addressed in the tables are provided below:
Exposure Analysis Modeling System (EXAMS-II)
(Burns et al., 1982) is a steady-state and dynamic
model designed for rapid evaluation of the behavior or
synthetic organic chemicals in lakes, rivers, and
estuaries. EXAMS-II is an interactive program that
allows the user to specify and store the properties of
chemicals and ecosystems, modify the characteristics
of either via simple English-like commands, and
conduct rapid, efficient evaluations of the probable
fate of chemicals. EXAMS-II simulates a toxic
chemical and its transformation products using
second-order kinetics for all significant organic
chemical reactions. EXAMS-II, however, does not
simulate the solids with which the chemical interacts.
The concentration of solids must be specified for
each compartment; the model accounts for sorbed
chemical transport based on solids concentrations
and specified transport fields. Benthic exchange
includes pore-water advection, pore-water
diffusion, and solids mixing. The latter describes a net
steady-state exchange associated with solids that is
proportional to pore water diffusion.
A data set of average or typical values for
waterbody-specific data is presently being
developed by Battelle Northwest Laboratories, under
contract to EPA. This data file will contain parameter
values for a number of major U.S. river systems,
lakes, and reservoirs, and will be integrated with the
EXAMS program. These values will be accessible for
fate modeling of the waterbodies included (GSC
1982).
MINTEQA1 (Felmy et al., 1984; Brown and Allison,
1987) is a geochemical model that is capable of
56
-------
calculating equilibrium aqueous speciation,
adsorption, gas phase partitioning, solid phase
saturation states, and precipitation-dissolution of 11
metals (arsenic, cadmium, chromium, copper, lead,
mercury, nickel, selenium, silver, thallium, and zinc).
MINTEQA1 contains an extensive thermodynamic
data base and contains six different algorithms for
calculating adsorption. Proper application of
MINTEQA1 requires applicable expertise, because
kinetic limitations at particular sites may prevent
certain reactions even though they might be
thermodynamically possible.
Hydrological Simulation Program - FORTRAN
(HSPF) (Johanson et al., 1984; Donigian et al., 1984)
is a comprehensive package for simulation of
watershed hydrology and water quality for both
conventional and toxic organic pollutants. HSPF
incorporates the watershed-scale ARM (Agricultural
Runoff Model) and NPS (Non-Point Source) models
into a basin-scale analysis framework that includes
pollutant transport and transformation in stream
channels.
The model uses information such as the time history
of rainfall, temperature, and solar radiation; land
surface characteristics such as land use patterns and
soil properties; and land management practices to
simulate the processes that occur in a watershed.
The result of this simulation is a time history of the
quantity and quality of runoff from an urban or
agricultural watershed. Flow rate, sediment load, and
nutrient and pesticide concentrations are predicted.
The program takes these results, along with
information about the stream network and point
source discharges, and simulates instream processes
to produce a time history of water quantity and quality
at any point in a watershed - the inflow to a lake,
for example. HSPF includes an internal data base
management system to process the large amounts of
simulation input and output.
Water Analysis Simulation Program (WASP4)
(Ambrose et al., 1986, 1987) is a generalized
modeling framework for contaminant fate and
transport in surface water. Based on the flexible
compartment modeling approach, WASP can be
applied in one, two, or three dimensions. WASP is
designed to permit easy substitution of user-written
routines into the program structure. Problems that
have been studied using WASP include biochemical
oxygen demand, dissolved oxygen dynamics,
nutrients and eutrophication, bacterial contamination,
and toxic chemical movement.
A variety of water quality problems can be addressed
with the selection of appropriate kinetic subroutines
that may be either selected from a library or written
by the user. Toxics WASP (TOX14; Ambrose et al.,
1987) combines a kinetic structure adapted from
EXAMS with the WASP transport structure and simple
sediment balance algorithms to predict dissolved and
sorbed chemical concentrations in the bed and
overlying waters.
Eutrophication WASP (EUTR04; Ambrose et al.,
1987) combines a kinetic structure adapted from the
Potomac Eutrophication Model with the WASP
transport structure. EUTR04 predicts dissolved
oxygen, carbonaceous biochemical oxygen demand,
phytoplankton, carbon, and chlorophyll a, ammonia,
nitrate, organic nitrogen, and orthophosphate in the
bed and overlying waters.
SARAH (Ambrose and Vandergrift, 1986) is a
steady-state mixing zone model for back-
calculating acceptable concentrations of hazardous
wastes discharged to land disposal or waste water
treatment facilities. For steady or batch waste
streams, SARAH considers the following
concentration reductions: dilution and loss during
treatment, initial Gaussian mixing at the edge of a
stream, lateral and longitudinal diffusion in the mixing
zone, sorption, volatilization, hydrolysis, and
bioaccumulation in fish. The user must specify,
appropriate in-stream criteria for protection of the
aquatic community, and humans through consumption
of fish and water. The benthic community is not
presently considered. Treatment loss is handled
empirically. The human exposure pathways
considered include ingestion of treated drinking water
and consumption of contaminated fish.
3.4.2.4 Short- and Long-Term Concentration
Calculations
Long-term average ambient water concentrations
should be calculated using (1) the average release
rate (from Chapter 3) projected for the time period of
interest, and (2) the annual average stream flow rate
as input to the above estimation procedures.
Short-term concentration levels are obtained through
use of the short-term release rate developed during
contaminant release analysis and the lowest
reasonable 24-hour flow rate, or the 7-day, 10-
year (7-Q-10) low flow rate for the period of
record, as presented in the above estimation
procedures.
Table 3-6 indicates several aquatic fate models
capable of estimating both short- and long-term
ambient water concentrations that are appropriate to
in-depth analysis of the aquatic fate of contaminants
released from Superfund sites.
3.5 Quantitative Analysis of Ground-
Water Fate
To model the migration of contaminants in ground
water the following factors should be estimated:
57
-------
Table 3-5. Resource Requirements and Information Sources: Surface Water Fate Models
Model
Description
Resource Requirements, comments
References, sources of documentation,
software
Water Quality Assessment Methodology
(WQAM)
Simplified Lake/Stream Analysis (SLSA)
Michigan River Model (MICHRIV)
Chemical Transport and Analysis Program
(CTAP)
01
09
Exposure Analysis Modeling System
(EXAMS-II)
• Steady-state, 1 -dimensional model
* Requires only desk top calculations
xx Provides canonical information
xx Models lakes, rivers, and estuaries
xx Steady-state, 1 -dimensional model
x Solution either by desk top calculations or
simple FORTRAN program
• Suitable for simplified lake and river
systems
xx Steady-state, 1 -dimensional model
xx Computer program written in FORTRAN
x Similar to SLSA, but can model more
than one reach
xx Intended for metals
xx Models rivers and streams
xx Steady-state, d-dimensional
compartmental model
xx FORTRAN IV program suitable for
numerous computers
x Similar to SLSA except more
sophisticated; each CTAP compartment
is equivalent to one SLSA "lake"
xx Models streams, stratified rivers, lakes,
estuaries, and coastal embayments
xx Steady-state, 3-dimensional
compartmental model
• Complex computer program
xx Contains comprehensive second-order
decay kinetics for organics; most models
only have first-order kinetics
xx Models organic chemicals
xx Suitable for freshwater, non-tidal aquatic
systems
x Easy to set up and use
xx No computer programming needed;
requires only hand calculator
xx Recommended if time, costs, or
information are restrictive
xx Easy to set up and use
• Computer programming not necessary; if
used, only 280 bytes are required;
suitable for microcomputers
*«rWell documented and suggested for use
before use of a more sophisticated model
xx May be used with hand calculator
xx Easy to set up and use
xx Requires minimal computer programming
xx Requires extensive data input
xx FORTRAN program - suitable for IBM
360/370, UNIVAC 108, CDC 6600
mainframe computers
xx Microcomputer version available requiring
32 K bytes storage
xx One of the better documented models,
which may make it more desirable than
other complex models
xx Requires extensive data input
xx Has been incorporated into EPA-OTS
GEMS system (see Section 4.1)
*«rWell documented and recommended for
use over most other models
• Available on magnetic tape for installation
on mainframe or small computers (e.g.,
PDP-11 or HP 3000); batch version
requires 64 K bytes memory at a
minimum, more for complex modeling
ji^Also available in interactive version,
requiring 164 K bytes memory plus 2 K
bytes for each chemical and 2.5 K bytes
for each environment
An estimated 350 hours required for
installation and setup, assuming all data
are readily available
Reference: Mills et al. 1982
Documentation:
ORD Publications
USEPA, Cincinnati, Ohio 45268
(513) 684-7562
Reference: HydroQual 1982
Documentation:
William Gulledge
2581 M Street, N.W.
Washington, DC. 20037
(202) 887-1183
Reference: Delos et al. 1984
Technical Assistance Available from:
Bill Richardson
USEPA
Environmental Research Laboratory -
Duluth
Large Lakes Research Station
Reference: HydroQual 1982
Documentation:
William Gulledge
Chemical Manufacturers Association
2581 M Street, N.W.
Washington, D.C. 20037
(202) 887-1183
Reference: Burns et al. 1982
Documentation:
ORD Publications, Center for
Environmental Research Information
USEPA
Cincinnati, Ohio 45268
(513) 684-7562
Center for Water Quality Modeling
Environmental Research Laboratory
USEPA
Athens, Ga. 30613
(404) 546-3585
(Continued)
-------
Table 3-5.
(Continued)
Model
Description
Resource Requirements, comments
References, sources of documentation,
software
Metals Exposure Analysis Modeling System
(MINTEQA1)
Hydrological Simulation Program
FORTRAN (HSPF)
Transient One-Dimensional Degradation
and Migration Model (TODAM)
01
CD
Channel Transport Model (CHNTRN)
Finite Element Transport Model (FETRA)
Steady-state, 3-dimensional
compartmental model
Complex computer program
Designed for modeling of metal loadings
Suitable for freshwater, non-tidal aquatic
systems
Time-varying, 1 -dimensional model
Designed for year-round simulation
Models organic pollutants
Second-order decay mechanisms
Models non-tidal rivers, streams, and
mixed lakes
Time-varying, 1 -dimensional model
Second-order decay mechanisms
Models river and estuarine systems
Requires exterior hydrodynamic model
(e.g., EXPLORE) to provide channel and
flow velocities to TODAM
•f Time-varying, 1 -dimensional model
* Models organic pollutants
* Second-order decay mechanisms
•s Models rivers, lakes, estuaries, and
coastal waters
• Can be coupled with a hydrodynamic
model, CHNHYD, to estimate flow
dynamics where such data are not
available
* Time-varying, P-dimensional model
(longitudinal and lateral)
* Second-order decay mechanisms for
organic pollutants
*s Models rivers, estuaries, coastal systems,
and completely mixed lakes
* Can be coupled with EXPLORE
hydrodynamic model to generate flow
velocities where these are unknown
Complex metal dynamics requiring
extensive data input
Can be used with mainframe or small
(e.g., PCP 11/70 or HP 3000) computers
ineractive format
Contains data base with thermodynamic
properties of 7 metals
Requires extensive data input
Most suitable to minicomputers (e.g., HP
3000, PRIME. HARRIS) as model
utilizes direct access input-output, which
can be costly on mainframe computers
Requires 250 K bytes of overlay-type
storage
Has been used on IBM 370 series
computers
Requires extensive data input
Complex FORTRAN program, written in
the preprocessor language FLECS or in
FORTRAN IV
Applicable to VAX or POP 11/70
computers (batch mode)
TODAM has been applied; however,
documentation is currently under review;
release date unknown
Requires extensive data input, and
extensive setup time
Has not been field-tested, and
documentation is currently under review
FORTRAN IV program language
Applicable to IBM 3933 computer, and
others
Input data requirements are extensive
Computer program written in FORTRAN
IV
Can be used on IBM. VAX, or CDC-
7600 computers
Has been field-validated
Setup and execution time requirements
are extensive
Further information:
Yasuo Onishi
Battelle Pacific Northwest Laboratories
Richland, WA 99352
(509) 376-8302
Reference: Johanson et al. 1984
Software:
Center for Water Quality Modeling
Environmental Research Laboratory
USEPA
Athens, GA 30613
(404) 546-3585
Reference: Onishi et al. 1982
Further information:
Yasuo Onishi
Battelle Pacific Northwest Laboratories
Richland, WA 99352
(509) 376-8302
Reference: Yeh 1982
Documentation:
Dr. G. T. Yeh
Environmental Sciences Division
Oak Ridge National Laboratory
P.O. Box X
Oak Ridge, TN 37830
(615) 574-7285
Reference: Onishi 1981
Further information:
Yasuo Onishi
Battelle Pacific Northwest Laboratories
Richland, WA 99352
(509) 376-8302
(Continued)
-------
Table 3-5. (Continued)
Model
Description
Resource Requirements, comments
References, sources of documentation,
software
Sediment-Contaminant Transport
(SERATRA)
xx Time-varying, P-dimensional model
(longitudinal-and vertical)
x* Complex sediment transport mechanisms
xx Second-order decay mechanisms for
organic pollutants
• Models rivers and lakes
Estuary and Stream Quality Model (WASP4) xx Time-varying, 3-dimensional model
xx Sophisticated second-order organic
decay kinetics
xx Models rivers, lakes, and estuaries
Surface Water Back Calculation Procedure
(SARAH)
xx Steady-state, 1 -dimensional analytical
solution
xx FORTRAN Code
xx Models contaminated leachate plume
feeding the downgradient surface
waterbody (stream or river)
xx Monte Carlo simulated generic
environment
xx Degradation, dilution, sorption, and
volatilization
xx Broaccumulation in fish
xx Requires extensive data input
xx Computer program written in FORTRAN
preprocessor language FLECS < in
batch mode
xx Has been field-tested and is available
for use
xx Requires an estimated 750 man-hours
for setup, assuming all required data are
readily available
xx Very data-intensive model
xx User must provide hydrodynamic flows
between model compartments
x Applicable to IBM 370 or POP 11/70
systems
xx FORTRAN IV program requires 64 K
bytes memory
xx Requires 150-300 man-hours for
setup
xx Generic environment, minimal data input
• FORTRAN model
Reference: Onishi and Wise 1982a, Gnishi
and Wise 1982b
Documentation:
ORD Publications
Center for Environmental Research
Information
USEPA
Cincinnati, OH 45268
(513) 684-7562
Technical Assistance:
Robert Ambrose
EPA Athens Environmental Research Lab
Center for Water Quality Modeling
Athens, GA 30613
(404) 546-3546
Documentation and Software:
Dr. John Connolly
Environmental Engineering and Science
Manhattan College
Bronx, N.Y. 10471
(212) 920-0276 or:
Dr. Parmely H. Prichard
Environmental Research Laboratory
Gulf Breeze, FL 32561
(904) 932-5311
Robert Ambrose
Center for Water Quality Modeling
USEPA
Athens, GA 30613
(404) 546-3546
Documentation: Jan. 14, 1986
Federal Register, Hazardous Waste
Management System, Land Disposal
Restrictions, Proposed Rule
Software.
David Disney, Environmental Research
Laboratory, Environmental Protection
Agency, College Station Road, Athens,
GA 30613, (404) 546-5432, or (404)
546-3123
Source: Versar 1983a.
-------
Table 3-6. Features of Surface Water Fate Models
MCHMV
CT»f
EXA
MMTEQA1
TODAM
CMNTM
FfTKA
KMTRA
WASP 4
Sourcti: USEPA 1985h; Dttot »l (I. 1984
-------
Table 3-7. Data Requirements for Surface Water Models
en
IVJ
CTAP
EXAM I
MMTEOA1
TOOAM
CNNTRN
FITHA
MRATMA
WASP 4
Sourcti: USEPA 1985H; D*k>i •! al. 1984
-------
Direction - The direction of contaminant migration is
important in predicting the potentially exposed
population.
Velocity - The migrating contaminant's velocity is
important in assessing when contamination will reach
the exposed population and how long the
contamination will be affecting that population.
Concentration - Concentration of the contaminant in
ground water at the exposure locations is used to
calculate dose to the population. This factor is used
to convert the amount of water consumed each day
to the mass of contaminant received each day. The
mass information is then used to predict health
effects associated with exposure to the contaminant
(USEPA1985d).
Volume - The contaminated region's volume is
important in evaluating the extent of the
contamination, which is essential to estimating costs
of remedial measures and viability of specific
alternative remedial measures for the particular site. It
is also useful for determining how long a remedial
measure will have to be taken.
The following ground-water discussions are divided
into three sections:
1. The minimum technical foundation that is needed
in order for the analyst to apply and interpret the
equations and models for ground water. This
discussion is meant to support the hydrologist
familiar with water supply calculations, providing
an introduction to contaminant hydrology.
Readers needing a more complete introduction to
hydrology may wish to read EPA's Handbook
titled "Groundwater" (EPA/625/6-87/016).
2. Equations that can predict average contaminant
velocity and mass flux for dilute solute and
concentrated contaminant plumes. Knowing the
travel time and the degradation half-life, one can
predict contaminant attenuation. A nomograph is
provided for predicting dilution and contaminated
front velocity of dilute solute plumes, as are
equations that are useful in assessing the extent
of contamination. The narrative contains guidance
for interpreting available monitoring data from
existing wells and from monitoring wells. All of the
equations apply to homogeneous and isotropic
media; fractured rock flow and karstic terrain flow
are not addressed.
3. Computer models that predict dilution,
attenuation, and contaminated front velocity of
dilute solute plumes only. All of the computer
models assume homogeneous and isotropic
media. Computer models that predict organic fluid
migration are not discussed, nor are models that
describe karstic terrain flow. The state of the art
for these models is not well-developed, and thus
they are considered beyond the scope of this
report. The analyst wishing to model organic fluid
migration in porous media should use the
equations in Section 3.5.2.
3.5.7 Discussion of Ground- Water Modeling
3.5.1.1 The Contamination Cycle
The two primary types of ground-water
contamination at uncontrolled hazardous waste sites
involve leaching of solid contaminants and percolation
of liquid contaminations to the underlying aquifer.
Solid material itself does not generally contaminate
ground water directly, because it does not move
through the porous soil. Thus, it will not migrate until
precipitation (or ground water) leaches (dissolves)
some of it and carries it down to the water table.
Ground-water contamination by this route depends
on the precipitation rate and the solubility of the solid
contaminant. A variation of this route involves
dissolution of the solid contaminant by a complex
leachate that contains organic constituents as well as
water. The existence of dissolved organic
constituents in the leaching fluid causes organic
contaminants to have a higher solubility. The
importance of this phenomenon is greatest for
contaminants with a high octanol/water partition
coefficient (Enfield 1984, Jaw-Kwei n.d.).
Liquids do not need infiltrating precipitation to carry
them down to the water table; they move on their own
with help from gravity. Thus, ground-water
contamination by liquids is not dependent on the
precipitation rate or the solubility of the contaminant.
The viscosity and density of a liquid affect its rate of
migration. After the liquid has percolated through the
soil, some will remain in the interstitial pore spaces;
this material will dissolve into the percolating
precipitation and migrate downward as a function of
its water solubility and the rainfall rate. Another
source of contamination by liquid material arises from
intentional injection into the aquifer itself (deep-well
injection) or "injection" into the vadose (unsaturated)
zone (unlined lagoons).
Hazardous waste is often assumed to be primarily
solid waste; however, studies showing the relative
proportion of solid to pourable hazardous RCRA
wastes indicate that pourable hazardous waste
constitutes 60 to 95 percent of the total (Skinner
1984). The equations for modeling liquid waste
migration pertain to a larger percentage of the waste
migration situations than the dilute solute transport
models (computer models/nomograph).
Two other types of ground-water contamination may
also occur. These are contamination by gaseous
contaminants and contamination by intermedia
transfers. Gases constitute a relatively small source
of ground-water contamination, since they are more
63
-------
likely to contaminate air than ground water. The main
mechanism for gases contaminating ground water is
equilibration of gases leaking from buried containers
or injected into the ground, with percolating rainwater
causing subsequent downward migration and mixing
of this contaminated water with ground water.
Intermedia contamination of ground water can come
from either air or surface water. Contamination from
air can result from two mechanisms: rain-out and
wash-out. Rain-out occurs when airborne con-
taminated particulates form condensation nuclei for
the formation of rain drops. Wash-out occurs when
falling rain captures gaseous or particulate con-
taminants as it falls to earth. The concentrations of
contaminants entering ground water as a result of
gaseous contamination or intermedia transfers are
generally very small, and these are not considered to
be significiant sources of ground-water contam-
ination in most cases.
A third source of contamination that may be
significant at some sites is through ground-
water/surface-water system interconnections. That
is, contaminated surface water may recharge a
ground-water system. This occurs only in reaches
where the surface-waterbody is a "losing stream"
(i.e., one that supplies water to the ground-water
system). Frequently, ground water feeds surface
water (gaining reaches). For gaining reaches, the
ground water, if contaminated, contaminates the
surface-waterbody into which it discharges.
One aspect of the contamination cycle that should be
considered is the ratio of contaminant to con-
taminated ground water. A very small quantity of
concentrated contaminant can contaminate a large
volume of ground water to the ppm or ppb level.
3.5.1.2 Ground-Water Flow Conditions
After precipitation infiltrates the surface of the ground,
it travels vertically down through the vadose zone
(unsaturated zone) where it meets the water table,
and it then flows approximately horizontally. The
horizontal flow within the aquifer is saturated.
(1) Saturated Zone
A simplified flow equation is used to describe the
volumetric 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. This equation is called
Darcy's law. The properties of the liquid (water or
contaminant) and the permeability of the porous
medium determine the hydraulic conductivity. The soil
has an intrinsic property of permeability, which is
determined by the size, orientation, and con-
nectedness of the pore spaces.
Soil permeability is a function of soil pore space,
which is determined by soil particle size. Small
diameter clay soil particles cause clay soil to have low
permeability, while larger diameter sandy soil particles
result in the high permeability of sandy soils. The
permeability, and therefore the hydraulic conductivity,
of a homogeneous soil is constant under conditions of
saturated flow.
In cases where the vadose zone is saturated and the
flow direction is vertical, the change in height of the
water per unit of vertical travel distance is always
one. Thus, the hydraulic gradient for vertical saturated
flow is unity, and the volumetric flows are proportional
to the permeability alone.
(2) Unsaturated Zone
Darcy's law governs flow anywhere in the porous
medium, including the vadose, or unsaturated, zone.
In the vadose zone, however, the pore spaces are not
saturated with water or any other liquid. The hydraulic
conductivity of any liquid through a porous medium is
partly dependent on the amount of liquid in the pore
spaces, and hydraulic conductivity for unsaturated soil
can be expressed as a fraction of the hydraulic
conductivity at saturation.
When the pore spaces are entirely filled with liquid
(i.e., saturated), the hydraulic conductivity for that
medium is at its maximum value. This is called the
saturated hydraulic conductivity (or simply hydraulic
conductivity), and it is essentially constant for a
specific liquid saturating a specific soil medium.
The unsaturated hydraulic conductivity at residual
moisture content is very small. When the soil is very
dry, most of the moisture is tightly bound by capillary
forces in the void spaces, and the water will not flow
easily. Unsaturated hydraulic conductivity increases,
gradually at first and then more rapidly, as the degree
of saturation increases from the residual moisture
content to the saturated moisture content. Since the
hydraulic conductivity is dependent upon the moisture
content, the specific discharge through the vadose
zone varies with the degree of saturation at any
depth.
The rate of infiltration at the ground surface may be
limited by the capacity of the soil to accept water or
by the delivery rate of water at the ground surface
(e.g., the precipitation rate). The infiltration rate into
soil cannot exceed the value for that soil's saturated
hydraulic conductivity. When the hydraulic loading to
the surface of the ground is low, such as light rainfall
alone, the flow of water through the vadose zone is
64
-------
unsaturated; however, when the hydraulic loading is
large, such as beneath a lagoon, the flow of water
through the vadose zone can be saturated. When the
hydraulic loading is small, it is limiting and the vertical
flow through the vadose zone is unsaturated. When
the hydraulic loading is larger than the flow that can
move through the soil with saturated flow, the
permeability of the soil is limiting the flow, and the
vertical flow through the vadose zone is saturated.
351.3 Multiphase Flow
The water solubility of any particular chemical will
determine whether it will be transported as a solute,
as a colloid, or as a separate, concentrated phase.
Many chemicals that have been identified as
contaminants in ground water are sparingly soluble in
water. When introduced to the ground-water system
as liquids, such chemicals can flow as an
independent species through the porous medium.
When the immiscible contaminant comes into contact
with the water in the pore spaces of the vadose zone
or at the water table (phreatic surface), the liquids do
not mix but essentially remain as two separate
phases. Some of the chemical will go into solution
with the water, but since the solubility of the chemical
is very low, the bulk of the contaminant will remain as
a separate layer that could saturate the pore spaces it
is flowing through. Thus, the migration of two
immiscible liquids in porous media is called two-
phase flow.
Complete descriptions of two-phase flow require an
additional equation for each separate phase present
in the flow system. Several general rules that can be
applied in analyzing ground-water contamination
problems involving immiscible chemicals, are as
follows:
(1) Floaters
The specific gravity of an immiscible liquid
contaminant will determine whether water will displace
it or it will be displaced by water. In downward flow,
water can displace the lighter, immiscible liquid so the
water is found below the immiscible liquid. In
horizontal flow, the less dense, immiscible liquid will
tend to float upward until the separate immiscible
phase floats on top of the water table. Thus, the
immiscible liquids that have a specific gravity of less
than one are sometimes referred to as "floaters." As
a general rule, immiscible hydrocarbons that are
nonchlorinated are floaters (less dense than water).
(2) Sinkers
Immiscible contaminants more dense than water,
whose specific gravity values are greater than one,
can displace water when flowing through the porous
medium. Gravity will cause dense immiscible liquids
to sink as they flow horizontally through the porous
medium. Thus, the immiscible liquids more dense
than water are often referred to as "sinkers."
Generally, chlorinated hydrocarbons that contaminate
water are more dense than water.
Density and specific gravity are intrinsic properties of
a chemical, and values for natural or manufactured
chemicals are usually published (Verschueren 1984;
Callahan et al. 1979).
(3) Hydraulic Gradient for Immiscible Fluids
The hydraulic gradient, the difference in the hydraulic
heads at two points divided by the distance (along the
flow path) between the points, is the driving force for
ground-water movement in a porous medium. With
regard to an immiscible separate phase, however, the
gradient that causes the immiscible liquid to flow is
not necessarily the same as that which influences the
ground water. If contaminants in an immiscible phase
that is more dense than water reach the bottom of the
aquifer, that separate phase may alter its flow
direction to conform to the shape and slope of the
aquitard surface. In some cases, the base of the
aquifer may be sloped in a different direction from the
direction of flow determined by the hydraulic gradient.
This possibility should be considered when the
analyst tries to identify the direction of the
contaminant plume's migration.
The assumption that the hydraulic gradient of the
separate, immiscible phase approximates that of
ground water is quite reasonable for the less dense
immiscible liquids. Since these contaminants float on
the water table, the hydraulic gradient of the phreatic
surface is probably also the gradient of the immiscible
phase.
(4) Hydraulic Conductivity of Immiscible Fluids
If the saturated hydraulic conductivity of water
through a porous medium is known, it is very easy to
modify that value to calculate the hydraulic
conductivity of that same porous medium saturated
with a different liquid, such as a separate layer of an
immiscible phase.
3.5.1.4 Contaminant Flow and Hydrodynamic
Dispersion
In contaminant transport, contaminants can be
thought of as a mass flowing through a cross-
sectional area of the porous medium that is
perpendicular to the flow direction. The discussion
presented here is for solute transport (mass that is
transferred with the flowing ground water), but basic
concepts also apply to the flow of immiscible,
separate phases.
The movement of contaminants in ground water can
be described by two principal mechanisms: gross
fluid movement (advective flow) and dispersion.
Gross fluid movement can be either ground-water
movement or organic fluid movement (the waste itself
moving as a concentrated liquid). Dispersion also can
be described by two principal mechanisms: fluid
65
-------
mixing (mechanical dispersion) and diffusion. The
next section addresses the underlying mechanisms
for fluid mixing.
Fluid mixing is important for two reasons: (1) precise
modeling of contaminant movement and (2) modeling
of dilution of the contaminant concentration between
source and exposed population,
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. In
ground water, the magnitude of dilution is much
smaller. Flow in both air and surface water can be
turbulent. Turbulent flow means that all the flow paths
are not essentially parallel to the gross direction of
motion; some flow paths are at right angles to the
bulk fluid motion. The flow components that are
perpendicular to the bulk fluid motion cause the
plume to spread laterally. This reduces the
concentration in the plume, while making the plume
contaminate a larger volume of air or surface water.
In ground water, 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. Since the interconnecting pore spaces do
not make a continuous flow channel, some lateral
mixing will occur in real soil.
Dispersion in air and surface water is caused by the
eddy currents (and diffusion). If the flow is broken up
into two components, longitudinal flow and eddy flow,
the gross motion is due to the longitudinal flow, and
the eddy flow is responsible for mixing. The
magnitude of the eddy currents is the same in all
directions (longitudinal, transverse, and vertical).
Since the concentration gradients are weaker in the
longitudinal direction than they are in the transverse
and vertical directions (for continuous steady state
sources), the net effect of mixing in the longitudinal
direction is small compared to the effect of mixing in
the directions perpendicular to the flow direction.
When air and surface water are modeled, the
longitudinal mixing is often neglected: lateral mixing is
modeled as the principal mixing phenomenon.
Dispersion in ground water is not caused 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 dispersion; the last phenomenon causes
lateral dispersion. In ground water, the magnitude of
the mixing is much greater for longitudinal mixing than
for lateral mixing. Researchers have reported
longitudinal dispersivity values ranging from 2 to 25
times higher than transverse dispersivity values
(Gelharetal. 1985).
In ground water, dilution occurs at a much slower rate
than it does in air or surface water. The overall
magnitude of mixing is smaller, and the component of
mixing that is most important to dilution (lateral) is the
smaller component of ground-water mixing. For
short-term releases (spills), longitudinal mixing is
useful in diluting plume concentrations. This is
because the plume can effectively mix with the
uncontaminated water in front of and behind the slug
of contamination, whereas continuous sources make
the length of the plume so long that its middle section
cannot effectively mix with clean water in front or
behind it.
3.5.1.5 Transformation and Retardation
Movement of contaminants can be modeled by fluid
movement, fluid mixing, and diffusion; however, for
more accurate modeling, chemical transformation and
retardation should also be considered. Some
contaminants are subject to transformation and
retardation while others are not; the relative
significance of transformation and retardation for
specific contaminants determines the need to model
these mechanisms. Transformation is the term used
to describe loss of the contaminant from the plume.
The mass of the contaminant is not lost; rather, the
molecular structure is changed so that the toxicity
associated with the initial molecular structure is no
longer present. When the molecular structure of
degradation products is more toxic than the original
contaminant, degradation is not considered
attenuation. Attenuation is used to describe chemical
structure changes that reduce or eliminate the toxicity
of the contaminant, and to describe phenomena that
function as sinks for toxic contaminants. Phenomena
that are reversible are not sinks for toxic
contaminants.
Chemical interactions between contaminants and the
soil matrix that are reversible delay the migration of
contaminants but do not act as a sink. The effect of
these chemical interactions is modeled as retardation.
Retardation is modeled using a coefficient to scale
down the velocity of ground water to the slower
effective velocity of the contaminant mass.
Attenuation reduces population risk; retardation delays
population risk.
Many reversible interactions can cause retardation
phenomena; however, only two retardation
mechanisms apply to wide classes of contaminants
and are well enough understood to be modeled on a
regular basis. Organic retardation and cationic
retardation are the most frequently modeled
phenomena. Organic retardation refers to hydrophobic
contaminants sorbing onto organic material in the soil
66
-------
matrix. Cationic retardation refers to positive charged
ions associating with the soil matrix. This association
can be due to polar species in the ground water being
attracted to the ionic double layer surrounding clay
particles in the soil, or it can be due to ionic bonding
with the soil matrix.
(I) Retardation of Organics
Organic retardation, which refers to hydrophobic
contaminants sorbing onto organic material in the soil
matrix, is estimated in ground water by the use of a
retardation coefficient. The velocity of each
compound in ground water is a function of the
characteristics of the soil media and the compound's
octanol-water partition coefficient. The octanol-
water partition coefficient measures the compound's
degree of hydrophobicity. The parameter of the soil
media that determines the presence of sorption sites
is the percent of organic carbon in the soil.
When the contaminant concentration in the water is
high and the quantity of contaminant on the surface
of the soil organic carbon is low, the net transfer is
from the water to the soil. Since the transfer is an
equilibrium process, it reverses when the
concentration in the water is low and the quantity of
contaminant on the surface of the soil organic carbon
is high.
If the leachate contains sufficient quantities of organic
material to affect the solubility of the contaminant, the
modeling of retardation is more difficult. The toxic
constituent flow will still be retarded, but not as much.
Instead of partitioning between the water and soil
organic carbon, the contaminant will partition between
the polar-organic fluid and the soil organic carbon.
The toxic contaminant will spend a smaller fraction of
time on the solid soil particles and a larger fraction of
time in the fluid; this will increase its migration
velocity. Modeling this phenomenon, however, is
complex and has already been well documented
elsewhere. The analyst interested in modeling
retardation in complex leachates is referred to
Nkedi-Kizza et al. (1985), Rao et al. (1985), and
Woodburn et al. (1986).
Once a contamination source stops contaminating the
ground water (either a one-time slug or the end of a
long-term loading), the saturated sorption sites start
to lose contaminants to the clean ground water that
flows after it. This phenomenon causes the
development of a plume shape that has a long tail of
decreasing contamination. Since the rate of
desorption is high when the degree of saturation is
high, and is lower as the quantity of contaminant on
the sorption sites diminishes, the desorption
phenomenon can provide a degrading influence on
the ground water for a long time.
(2) Retardation of Cations
In cationic retardation, positively charged ions'
associate with the soil matrix (clay particles). There is
a smaller effect for anion exchange. Anion exchange
is due to positive charges associated with hydrous
oxides. Since soils typically have more negatively
charged clay particles than positively charged
hydrous oxides, cations flow with a more retarded
velocity than do anions. Contaminants that are not
charged are not subject to ionic retardation.
Contaminants that are compounds or complexed ions
also are not retarded by ionic retardation.
Cationic retardation is reversible, as is organic
retardation, and it forms a trail of low-level
contamination after the source of contamination
stops. Once a source stops contaminating the ground
water, the saturated ion exchange sites start to lose
contaminants to the clean ground water. This
phenomenon causes the development of a plume
shape that has a long tail of decreasing
contamination. Since the rate of release is high when
the degree of saturation is high, and lower as the
quantity of contaminant on the ion exchange sites
diminishes, the reversible ion exchange phenomenon
can provide a degrading influence on the ground
water for a long time.
(3) Transformation/Attenuation
Transformation/attenuation is the term used to model
sinks for contaminants. The particular type of
chemical fate modeled depends on the contaminant
and the soil characteristics. The following is a list of
different fate mechanisms:
x* Hydrolysis
xx Complexation-chelation
xx Acid/base reactions
xx Oxidation/reduction reactions
xx Biodegradation
xx Radioactive decay
xx Chemical precipitation
xx Coagulation
• Peptization reactions.
Attenuation is modeled with the use of a "half-life"
parameter. Whether the degradation is due to
hydrolysis or biodegradation, the time necessary for
the concentration to drop by half is the measure of
degradability.
Appropriate individual decay rates or overall decay
coefficients have been developed for some
substances and are available in the technical
literature. Sources for such data include: Callahan et
al. (1979); Dawson et al. (1980); Mabey et al. (1982);
Sax (1984); USCG (1974); and Verschueren (1984).
Methods of estimating decay coefficients are
presented by Lyman et al. (1982).
67
-------
3.5.1.6 Higher Velocity Transport
Some situations can cause the migration velocity of
contaminants to be faster than the ground-water
velocity. Macromolecules can themselves move faster
than the ground water, and any hydrophobic
contaminants that are sorbed onto them will also
move faster. Until recently, hydrophobic contaminants
were thought to flow with a retarded velocity only
because of preferential sorption onto stationary
organic soil particles. In such cases, the time the
contaminant spends on the stationary soil particles
lowers its average velocity. Conversely, the time a
contaminant spends on a "high speed"
macromolecule raises the average velocity of the
contaminant. Since hydrophobic contaminants sorb
onto both stationary and higher velocity
macromolecules, both must be considered in order
for the modeling of transport for hydrophobic
contaminants to be complete.
Macromolecules may be found in ground water in
concentrations ranging from 1 mg/l to 10 mg/l and are
large enough that only the large pore spaces are
available for migration. This means that their average
velocity is the average velocity of the large pore
spaces, and not the average velocity of all the pore
spaces. The velocity of flow through each pore space
is a function of the size of the pore space, and the
larger pore spaces allow faster flow than do the small
spaces. The velocity difference between the average
large pore space and the average pore space is
approximately one order of magnitude.
Macromolecules with large hydrophobic surface area
and small polar surface area will flow with a retarded
velocity because of reversible sorbtion onto soil
carbon. These macromolecules will not cause higher
speed transport. Macromolecules with large polar
surface areas and small hydrophobic surface areas
will travel faster than the ground water. These
molecules can speed up the migration velocity of
hydrophobic contaminants.
Macromolecular transport is not frequently modeled;
however, when such modeling is necessary, the
analyst can refer to Enfield and Bengtsson (n.d.) for
detailed guidance.
3.52 Ground- Water Modeling Equations and
Nomograph
This section provides a number of hydrologic
modeling equations and a nomograph. In no cases
will all equations be necessary; depending on the
observed chemical contaminant, a discrete subset of
the equations will be useful in assessing the ground-
water contamination problem at a specific
uncontrolled hazardous waste site.
Five discrete classes of contaminant are discussed.
Each class is based on a different technique for
calculating contaminant migration. The five classes of
contaminant can have dramatically different calculated
velocities and concentrations; use of the appropriate
analytical techniques for each class is thus necessary
for accuracy.
Estimating contaminant velocity is based on
estimating water velocity. For those contaminants that
flow as water flows, contaminant velocity equals water
velocity (vertical or horizontal). For those that flow at
rates different from water, the estimated water
velocity must be adjusted to approximate that of the
contaminant.
3.5.2.1 Calculating Ground-Water Velocity
Ground-water velocity can be determined for both
the saturated zone and the vadose (unsaturated)
zone. Vadose zone velocity is discussed in the next
section; saturated zone velocity is discussed in this
section.
Ground-water velocity in the saturated zone is
calculated using Darcy's Law (Bouwer 1978):
v = Ksi
where
Ks
i
(3-9)
Darcy velocity of water, also termed
superficial velocity, or specific
discharge, (length/time).
hydraulic conductivity of soil or aquifer
material, (length/time).
hydraulic gradient, (length/length).
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 more precise modeling, by
effective porosity (thus taking into account the fact
that the entire cross-section of the pore is not
flowing (i.e., due to boundary layer effects). For clay
soils, the effective porosity also corrects for the effect
of electro-osmotic 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
where
(3-10)
VPW
V
= ground water (pore water) velocity,
(length/time).
= Darcy velocity (superficial velocity,
specific discharge), (length/time).
-------
p = 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 the few
parameters that are not available. Literature values for
saturated hydraulic conductivity are presented in
Table 3-8 (Rawls et al. 1982) and Table 3-9
(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 developed during site
investigation. Water levels in existing nearby wells
can also provide an indication of hydraulic gradient.
Table 3-10 provides values for saturated moisture
content, which is roughly equal to the effective
porosity, or Pe, for several soil types.
It must be emphasized that site-specific data are
highly preferable to regional data, or data obtained
from any of the above-referenced tables. If site-
specific information on effective porosity is available,
it should be used; however, literature values for soils
with the same hydraulic conductivity provide sufficient
accuracy. Effective porosity (P,) can be approximated
by the difference between the moisture content at
saturation and at the wilting point (-15 bar)*. The
equation is as follows (Rawls 1986):
Pe = e8-9(-15)
where
(3-11)
This estimation procedure addresses the fraction of
the pore spaces that is contributing to flow, but does
not address the effect of electro-osmotic
counterflow and the development of electrokinetic
streaming potentials. For clays, this can be a
significant difference. Literature values listed in Table
3-10 should be used for clay solids (these values
incorporate the effects of the clays 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 approximation
and further refinement of this approach would
improve accuracy. Corrections for the path length
difference between the straight line distance versus
the tortuous path through which ground water flows
can improve the precision (Freeze and Cherry 1979),
although the literature does not provide a consistent
correction factor to apply. To provide a feel for the
magnitude of this correction the analyst can review
Das (1983) which suggests a correction of 1.41. This
value can be used to correct the velocity or the
distance (not both) by dividing the number by 1.4.
However, the analyst must interpret the results
obtained through such correction with care, as the
degree to which the factor cited in Das applies to any
given site is uncertain.
3.5.2.2 Calculating the Velocity of Infiltrating
Rainwater
This section discusses the calculation of the velocity
of percolating rainwater flowing through the vadose
zone. Darcy's law can be used to calculate the
unsaturated flow velocity; however, the hydraulic
conductivity 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 et al. 1982):
'pw
= q/0
(3-12)
Pe = effective porosity, (fraction, where
dimensionless).
9S = water content when the pores are fully
saturated, (fraction, dimensionless).
9(-15) = wilting point moisture content, (fraction,
dimensionless).
pw
q
8
'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. Bar is a measure of pressure (dynes/cm2).
= interstitial ground water (pore water)
velocity, (length per unit time).
= average percolation or recharge rate,
(depth per unit time).
= volumetric moisture 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. For
unsteady hydraulic loading, the "q" and "0" will vary
with time and depth. Additionally, the distribution of
"q" and "0" will vary as the moisture migrates down.
This makes determination of the average transport
velocity burdensome. For situations where steady-
state conditions cannot be assumed, the analyst
should use a computer model; for example, SESOIL
(one of EPA's GEMS computer system) calculates
the time of travel for seasonally varying rainfall rates.
The volumetric water content (0) in the unsaturated
zone can be estimated using the following equation
(Clapp and Hornberger 1978*:
69
-------
Table 3-8. Representative Values of Saturated
Hydraulic Conductivity
Hydraulic
conductivity
Soil texture Number of soils3 («,.; cm/sec)b
Sand
Loamy sand
Sandy loam
Loam
Silt loam
Sandy clay loam
Silt clay loam
Clay loam
Sandy clay
Salt clay
Clay
762
338
666
383
1,206
498
366
689
45
127
291
5.8 x
1.7 x
7.2 x
3.7 x
1.9 x
1.2 x
4.2 X
6.4 x
3.3 x
2.5 x
1.7x
10-3
10-3
10-4
10-4
10-4
10-4
10-5
10-5
10-5
10-5
10-5
Table 3-9. Saturated Hydraulic Conductivity Ranges
for Selected Rock and Soil Types
Saturated Hydraulic Conductivity (cm/sec)
aNumber of Individual soil samples included in data
compiled by Rawls et al. 1982.
bpredicted values based on compiled soil properties.
Source: Adapted from Rawls et al. 1982.
Soils
Unweathered marine
clay
Glacial till
Silt, loess
Silty sand
Clean sand
Gravel
Rocks
Unfractured
metamorphic and
Igneous rock
Shale
Sandstone
Limestone and
dolomite
Fractured Igneous and
metamorphic rock
Permeable basalt
Karst limestone
5xlO-"
10-10
IO-J
10-5
10-4
10-1
10-2
5 x ID-12
10-8
5 x 10-8
10-6
10-5
ID-4
- io-J
ID-4
I0-3
- 10-1
i
- 102
- 10-8
IO-J
5 x IO-4
5 x 10-4
_ 10-2
1
1
Source: Adapted from Freeze and Cherry 1979.
Table 3-10. Representative Values for Saturated Moisture Contents and Field Capacities of Various Soil Types
Saturated moisture content (9s)a
Field capacity (cm3/cm3)b
Number of soils
Sand
Loamy sand
Sandy loam
Loam
Silt loam
Sandy clay
loam
Clay loam
Silty clay loam
Sandy clay
Silty clay
Clay
762
338
666
383
1,206
498
366
689
45
127
291
Mean
0.437
0.437
0.453
0.463
0.501
0.398
0.464
0.471
0.430
0.479
0.475
± 1 Standard deviation
0.347
0.368
0.351
0.375
0.420
0.332
0.409
0.418
0.370
0.425
0.427
- 0.500
- 0.506
- 0.555
0.551
0.582
- 0.464
_ 0.519
- 0.524
- 0.490
- 0.533
- 0.523
Mean
0.091
0.125
0.207
0.270
0.330
0.255
0.318
0.366
0.339
0.387
0.396
± 1 Standard deviation
0.018
0.060
0.126
0.195
0.258
0.186
0.250
0.304
0.245
0.332
0.326
- 0.164
-0.190
0.288
- 0.345
- 0.402
- 0.324
- 0.386
- 0.428
- 0.433
- 0.442
- 0.466
aFrom total soil porosrty measurements compiled by Rawls et al. (1982) from numerous sources.
bwater retained at -0.33 bar tension; values predicted based on compiled soil property measurements
Source: Rawls et al. 1982.
70
-------
0 = (0s)*(q/Ks)l/<2b + 3)>
where
(3-13)
0
0s
q
= volumetric water content in the
unsaturated zone, (volume/volume or
unitless).
= volumetric water content of soil under
saturated conditions, (volume/volume
or unitless).
= percolation rate (assumed to be equal
to the unsaturated hydraulic
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 "b" and the term
" 1/(2b+ 3)" are listed in Table 3-11.
Table 3-11. Representative Values of Hydraulic Para-
meters (Standard Deviation in Parentheses)
Soil texture No- of bb — 1— esC
soils3 2b + 3
Sand
Loamy sand
Sandy loam
Silt loam
Loam
Sandy clay
loam
Silt clay loam
Clay loam
Sandy clay
Silt clay
Clav
13
30
204
384
125
80
147
262
19
441
140
4.
4
4.
5
5
7
7
8
.05
.38
.90
.30
.39
.12
.75
.52
10.40
10.40
11
.40
(1.78)
(1.47)
(1.75)
(1.87)
(1.87))
(2.43)
(2.77)
(3.44)
(1.64)
(4.45)
(3.70)
0
0
0
0
0
0
0
0
.090
.085
.080
.074
.073
.058
.054
.050
0.042
0
0
.042
.039
0.395
0.410
0.435
0.485
0.451
0.420
0.477
0.476
0.426
0.492
0.482
(0.056)
(0.068)
(0.086)
(0.059)
(0.078)
(0.059)
(0.057)
(0.053)
(0.057)
(0.064)
(0.050)
aNumber of individual soil samples included in data compiled by
Clapp and Hornberger (1978).
bEmpirical parameter relating soil matric potential and moisture
content; shown to be strongly dependent on soil texture.
cWolumetric soil moisture content (volume of water per volume of
soil).
Source: Adapted from Clapp and Hornberger 1978.
The saturated volumetric water content (@s)>
saturated hydraulic conductivity (Ks), and the
exponential function (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, values in Tables 3-10
through 3-11 provide guides for the rough estimation
of 0S, Ks, and the term 1/2b + 3 . Representative
values from two different sources are presented for
Ks (Tables 3-8 and 3-9) and 0S (Tables 3-10 and
3-11), in order to demonstrate the variability in
estimates for these values.
Note that the value 6 cannot exceed 0S, the
saturated soil moisture content. When 0 calculated
by Equation 3-13 equals or exceeds 0s, it must be
assumed that saturated conditions exist. In such
cases, use Equations 3-9 and 3-10.
Similarly, the minimum value for 0 that is applicable
to Equation 3-13 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 due to 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 3-10.
Wherever Equation 3-13 results in a value for 0 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 Vpw is equal to
zero.
Note that the percolation rate (q) cannot exceed the
saturated hydraulic conductivity (Ks) for the site soil.
Whenever q > Ks (and therefore 0 as calculated by
Equation 3-13 > 0S) for the duration of the study
period, it must be assumed that saturated conditions
exist and that saturated flow prevails. Equations 3-9
and 3-10 in the preceding subsection provide a
means of estimating saturated flow velocities.
The following equation provides an estimate of the
term q (Enfield et al. 1982):
Pr-ET-
(3-14)
where
HL
Qr =
= hydraulic loading from manmade
sources, (depth per unit time)
= precipitation, (depth per unit time)
= evapotranspiration, (depth per unit
time)
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 two sources,
71
-------
Table 3-12. Suggested Value for Cet Relating Evaporation from a US Class A Pan to Evapotranspiration 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 fetch of
crop (m from
Wind pan)
Light < 170 km/day
Moderate 170-425 km/day
Strong 425-700 km/day
Very strong > 700 km/day
0
10
100
rtioo
0
10
100
1000
0
10
100
1000
0
10
100
1000
r-vvciayc i cv
20-40
0.55
0.65
0.7
0.7
0.5
0.6
0.65
0.7
0.45
0.55
0.6
0.65
0.4
0.45
0.5
0.55
jiui iai iciau v<
%*
40-70
0.65
0.75
0.8
0.85
0.6
0.7
0.75
0.8
0.5
0.6
0.65
0.7
0.45
0.55
0.6
0.6
y' Upwind fetch of
dry fallow (m
>70 from pan)
0.75
0.85
0.85
0.85
0.65
0.75
0.8
0.8
0.6
0.65
0.7
0.75
0.5
0.6
0.65
0.65
0
10
100
1000
0
10
100
1000
0
10
100
IOOO
0
10
100
1000
rvvciayc icyiuii
relative humidity,
20-40 40-70
0.7
0.6
0.55
0.5
0.65
0.55
0.5
0.45
0.6
0.5
0.45
0.4
0.5
0.45
0.4
0.3
0.8
0.7
0.65
0.6
0.75
0.65
0.6
0.55
0.65
0.55
0.5
0.45
0.6
0.5
0.45
0.4
ai
%*
>70
0.85
0.8
0.75
0.7
0.8
0.7
0.65
0.6
0.7
0.65
0.6
0.55
0.65
0.55
0.5
0.45
'Mean of maximum and minimum relative humidities.
Source: Jensen 1973, as presented by Enfield et al. 1982.
the National Weather Service, Forest Service offices,
National Oceanic and Atmospheric Administration
(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 (P,) for
the study period can be obtained from various local
weather authorities such as those listed above.
ET is estimated by using measured Class A pan
evaporation rates (a measure of local evaporation
rates under standardized conditions, available from
the nearest NOAA gauging station) in the equation:
C
veg
correction factor for converting
evapotranspiration from turf grass to
evapotranspiration from other
vegetative cover types, (unitless).
ET = EVAP x Cet x Cve
where
(3-15)
EVAP = region-specific or site-specific
measured evaporation rates, (depth per
unit time).
Get = correction factor for converting
measured pan evaporation rates to
evapotranspiration rates from turf
grass, (unitless).
Values for Cet are taken from Table 3-12, which
requires climatological and pan descriptive in-
formation.
The term Cveg is available mainly for agricultural
crops (Table 3-13), 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 conservative default
value for this term is the smallest reasonable value,
or 0.6.
Qr, or the average runoff over the study period, is
estimated by the method presented in Section 2.4 of
this manual. A more reliable value for this term can
be obtained from local USGS gauging stations. For
relatively level sites, a reasonable conservative
default value for the purposes of this estimation
procedure is that Qr = 0, where site-specific data
are unavailable or cannot be estimated.
72
-------
Table 3-13.
crop
Crop Coefficients for Estimating
Evapotranspiration
Period
Coefficient
(cveg)
Alfalfa April 1 - October 10 0.87
Potatoes May 10 - September 15 o.e5
Small grains April 1 - July 20 o.e
Sugar beets April 10 - October 15 o.e
Source: Jensen 1973, as presented by Enfield et al.
1982.
The above method for predicting the velocity of water
migrating through the vadose zone is the best
approximation available; however, real world non-
homogeneities, 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.
3.5.2.3 Corrections for Viscosity and Density
When the movement of liquids other than water is
calculated, the saturated and the unsaturated
hydraulic conductivity must be corrected for the
density and viscosity of the non-water liquid. The
equation for this correction is as follows:
Kc = Kc*(density of chemical/density of water) (3-16)
'(viscosity of water/viscosity of chemical)
where
Kn
= hydraulic conductivity of water (Darcy's
coefficient), (saturated or unsaturated)
= hydraulic conductivity of chemical,
(saturated or unsaturated).
When the migration velocity through the vadose zone
is calculated, density and viscosity should be
corrected with the above equation. For horizontal flow
below the water table, density and viscosity should be
factored in when the hydraulic gradient is the slope of
the chemical plume. In many cases, one can assume
that the thickness of the concentrated chemical
plume is relatively constant. For such situations, the
slope of the concentrated chemical is zero and the
analyst should not correct for the density. The slope
(hydraulic gradient) is that of water, and the Darcy
coefficient reflects the density of water. However, the
viscosity of the chemical is the viscosity of the
flowing fluid of concern, and the analyst should
correct for the viscosity.
3.5.2.4 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 complex leachates containing organic
constituents will show some retardation, although not
as much as in pure ground water.
When a hydrophobic contaminant flowing in a dilute
plume flows past a soil particle that contains organic
carbon, the contaminant partitions between the polar
solvent (water) and the solid organic carbon. When
the concentration in the water is high and the
concentration on the soil particle low, the net
migration is from the water to the soil. When the
reverse occurs and the concentration in the water is
low and the concentration on the soil particle is high,
the net migration is from the soil particle to the water.
When the water and soil concentrations are in
equilibrium, there is no net migration. However, the
flux from the soil to the water and the flux from the
water to the soil are not zero; rather, they are positive
fluxes that are equal and are in opposite directions.
When the partitioning is between concentrated
chemical and soil particles, the contaminant does not
prefer the solid "solvent" effects of the organic
carbon in the soil to the organic liquid solvent effects
of the concentrated chemical plume. Hence,
hydrophobic contaminants partition out of polar
solvents (water) but not out of hydrophobic solvents,
and thus, retardation effects are modeled for dilute
plumes only.
Retardation can be modeled for complex leachates,
but the methods are not presented in this report. The
reader is referred to Nkedi-Kizza et al. 1985, Rao et
al. 1985, and Woodburn et al. 1986, for guidance on
performing these calculations.
The retardation protocol is based on the assumption
that adsorption of hydrophobic contaminants is due to
sorption 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 (function of particle size). Numerous
studies of the correlation of Kd 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 protocol estimates hydrophobic retardation
based on soil organic carbon, but it should not be
taken to 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 (no appreciable sorption to
nonexistent organic carbon). Unfortunately, methods
73
-------
for estimating adsorption coefficients under these
conditions are not currently available (Lyman et al.
1982). The protocol discussed in this report relies on
the percent of organic carbon content of the soil.
To simplify modeling, equilibrium conditions are
modeled as the contaminant velocity being a fraction
of the ground-water velocity. If the analyst thinks of
the time an individual portion of the contaminant mass
is in the water as the time it has ground-water
velocity, and the time the contaminant is on the soil
particles as the time the contaminant does not have a
velocity, the contaminant velocity is related to the
ground-water velocity by the ratio of time on soil
particles to time in the water. The ratio of time in the
water to time on the soil particles is the same ratio as
the concentration ratio at equilibrium.
In complex leachates containing organics, the time a
hydrophobic contaminant spends on the solid carbon
is reduced because the ratio of the contaminant's
solubility in the fluid to its solubility on soil carbon is
increased. The hydrophobic contaminant partitions
between the organics in the flowing fluid and the
organics that are solid.
The same logic applies to cation retardation, and the
contaminant velocity for cations is also modeled as
fraction of ground-water velocity.
The equation used to calculate the retardation is as
follows (Kent et al. 1985):
Rd - 1
where
(B* Kd)/pt
(3-17)
R=
pt=
Kd=
retardation factor, (unitless).
bulk density, (g/ml).
total porosity, (unitless).
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):
where
Rd =
V =
vpw
vd =
(3-18
retardation factor, (unitless).
velocity of ground water, (same
units as Vc' length/time).
velocity of contaminant, (same
units as V^' length/time).
The term Kd is based on sorption isotherm column
studies. While this is the more precise approach, the
analyst will typically have to work with estimated
parameters. For hydrophobic contaminants, the term
Kd can be estimated from the term Koc (Lyman et al.
1982):
Knn = Kr|/f,
d"oc
where
Kd =
partition coefficient for organic
carbon, (ml/g).
distribution factor for soil, (ml/g).
fraction of organic carbon in the
soil.
The term "fraction of organic carbon" (foc) is precise
when taken from empirical measurements of the soil
in the study area. For cases where this is not
possible, estimates can be made. For the vadose
zone velocity, a value of foc from Rawls (1986)
provides a good estimate. Rawls' work focused on
soils near the surface, the area of interest to
agriculture. For saturated zone velocity, the analyst
has two choices. If the subsoil came from igneous or
metamorphic rock, the foc 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 percent
for the foc. If the subsoil came from sedimentary rock,
the foc distribution may be similar to the distribution for
agricultural soils done by Rawls. The variation of foc
with depth may be relatively constant. The carbon
was at the surface at one time, and has been buried
over geological time. Hence, the analyst should use a
value of foc from the Rawls (1986) distribution for the
saturated zone velocity determination (Trask and
Patnode 1942). Soil/water partition coefficients have
been developed for many contaminants of importance
(Callahan et al. 1979 and Mabey et al. 1982).
If Koc is not known, it can be estimated from
regression equations that relate Koc to Kow
(octanol/water partition coefficient). There are six
regression equations that relate Koc to Kow. The
equation that was based on a chemical class closest
to the subject contaminant should be used. If the
contaminant does not fit into a specific class, the first
regression equation should be used because it was
based on the largest sample. The regression
equations are as follows (Lyman et al. 1982):
Log KOC = 0.544 log K™ + 1.377
(3-20)
based on a wide variety of contaminants, mostly
pesticides
or
74
-------
log Koc = 0.937 log Kow - 0.
006
(3-21)
based on aromatics, polynuclear aromatics, triazines
and dinitroaniline herbicides
or
log Koc = 1.00 log K,,w - 0.21 (3-22)
based on mostly aromatic or polynuclear aromatics
or
log Koc = 0.94 log Kow + 0.02 (3-23)
based on s-triatines and dinitroaniline herbicides
or
logKoc= 1.029 logKow -0.18
(3-24)
based on a variety of insecticides, herbicides, and
fungicides
or
log Koc = 0.524 log Kow + 0.855
(3-25)
based on substituted phenylureas and alkyl-N-
phenylcarbamates.
The retardation effects are computed from the
octanol/water partition coefficient (Kow), which relates
the concentration in polar solvent (water) to the
concentration in hydrophobic solvent (octanol
simulating the soil organic carbon). If the
contaminated plume has a large concentration 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 Rd,
(i.e., the contaminant will migrate at a speed closer to
that of ground water). Much of the solubility of
extremely hydrophobic contaminants in the water of
an octanol/water partition coefficient test is due to
dissolution in the octanol that is dissolved in the water
rather than dissolution into water. This effect depends
on the degree to which the water is not pure water;
for most low-level contamination situations, this
effect can be ignored. This manual does not present
equations for calculating a numerical correction for
this effect. The analyst should be aware of the
general influence of this effect, but not model the
precise numerical difference. For dilute plumes, the
analyst should model full retardation; for concentrated
plumes, the analyst should model no retardation.
3.5.2.5 Contaminant Velocity
The velocity of concern is the actual contaminant
velocity. The determination of ground-water velocity
discussed earlier is done to provide a foundation for
calculating the contaminant velocity. The particular
method used for determination of the contaminant
velocity is dependent on the type of ground-water
transport the chemical undergoes. Thus, the first step
in calculating the velocity is classifying the subject
contaminants as to migration class.
Once the molecular identity of the contaminant is
known, three determining parameters can be taken
from literature:
1. Physical state at room temperature (i.e., is it a
solid or a liquid?)
2. Hydrophobicity (i.e., is it hydrophilic or
hydrophobic?)
3. Density (i.e., is it less dense than water?, Is its
density near that of water?, Or is it more dense
than water?)
The five migration classes are as follows:
Migration
class #
A)
B)
C)
D)
E)
Vadose zone transport
Solid/carried by
precipitation
Hydrophilic liquid/
waste percolation
Hydrophobic liquid/
waste percolation
Hydrophobic liquid/
waste percolation
Hydrophobic liquid/
waste percolation
Saturated zone
transport
Solute transport
Solute transport
Low density/
floater transport
Medium density/
buoyant transport
High density/
sinker transport
Although the specific chemical will migrate according
to the above classes, it is important to note that the
concentrated plumes will also have a dilute plume
near them. For mass flux considerations, the
concentrated plume will dominate.
(1) Migration Class #A: Solid Material
Solid material will dissolve into percolating
precipitation and migrate as a solute. Precipitation
provides the hydraulic loading that drives the rate of
release. The plume exists as a single plume (for
single chemical contaminant) that has a single
average velocity. Unretarded contaminants move with
the ground water, and hence, the ground-water
velocity is the contaminant velocity. Retarded
contaminants move with a velocity that is slower than
ground-water velocity, and therefore the
contaminant velocity is based on the ground-water
velocity adjusted for retardation. Typically, the velocity
is a fraction of the ground-water velocity.
(2) Migration Class 49: Hydrophilic Liquids
Liquids will directly percolate into the soil (i.e., without
waiting for precipitation to cause leaching). The
75
-------
hydraulic loading is due to the combination of
chemicals' hydraulic loading and that due to
precipitation. The velocity of transport through the
vadose zone must be calculated with corrections for
the density and viscosity of the contaminant. The
plume exists as a single plume (for a single chemical
contaminant) that can be considered to have a single
average velocity. Unretarded contaminants move with
the ground water, and hence, the ground-water
velocity is the contaminant velocity. This is only exact
after the plume has mixed with the ground-water to
the point that its density and viscosity are similar to
those of water. When the plume first reaches the
water table, it has not mixed with very much water,
and its density and viscosity differences suggest
calculating a contaminant velocity that is different
from the ground-water velocity. Since the velocity
difference varies gradually from the source to the
point downgradient where it is well mixed, this
calculation is complex. Therefore, the analyst should
calculate as if the ground-water velocity represented
the contaminant velocity for the length of the plume.
The analyst should be aware of the limitations of this
method. Retarded (cationic) contaminants move with
a velocity that is slower than ground-water velocity.
In this case, contaminant velocity is based on the
ground-water velocity adjusted for retardation, and is
a fraction of the ground-water velocity.
(3) Migration Class #C: Hydrophobia Liquids Low
Density
Once hydrophobic liquids reach the water table, they
form two distinct plumes (for a single chemical
contaminant), with each having its own average
velocity. The concentrated plume will float on the
surface of the water table and move in the same
direction as the ground-water flow. Its velocity is a
function of the contaminant's viscosity. If mounding is
significant, the density must also be considered. The
dilute plume is formed by small amounts of the
chemical dissolving in water as limited by the
hydrophobic chemical's solubility. This plume will be
found below the concentrated plume, with the highest
concentration near the concentrated plume. From the
point where the contamination leaves the
concentrated plume to form the dilute plume, the
dilute plume will move with the ground-water flow (at
a retarded velocity). The concentrated plume will
have a single average velocity, and it will start at the
location of the source. The dilute plume will have a
single average velocity, but its starting point can be
from the location of the source, or it can form from
the concentrated plume anywhere along the length of
the concentrated plume.
Retarded contaminants in the dilute plume move with
a velocity that is slower than ground-water velocity.
Thus, contaminant velocity, based on the ground-
water velocity adjusted for retardation, is typically a
fraction of the ground-water velocity. Contaminants
in the concentrated plume do not move with the
ground-water velocity; their velocity must be
determined by considering the effect of the
hydrophobic contaminant's viscosity. The
concentrated plume does not exhibit retardation
effects. If mounding is significant, the analyst also
must factor in the density.
(4) Migration Class #D: Hydrophobic Liquids/Medium
Density
This class of compounds migrates similarly to Class
#3, except that the concentrated plume will not float
or sink, but will have more or less neutral buoyancy. It
will move in the direction of ground-water flow, but
its migration velocity will be a function of its viscosity.
Again, the dilute plume will surround the concentrated
plume, forming a transition zone between the
uncontaminated water and the concentrated plume
body. From the point where the contaminant leaves
the concentrated plume to form the dilute plume, the
dilute plume will move with the ground-water flow (at
a retarded velocity). The concentrated plume will
have a single average velocity, and it will start at the
location of the source, or it can form from the
concentrated plume anywhere along the length of the
concentrated plume.
Retarded contaminants in the dilute plume move with
a velocity that is slower than ground-water velocity.
Thus, the contaminant velocity, based on the
ground-water velocity adjusted for retardation, is a
fraction of the ground-water velocity. Contaminants
in the concentrated plume do not move with the
ground-water velocity; their velocity must be
determined by considering the effect of the
hydrophobic contaminant's viscosity. The
concentrated plume does not flow with retardation
effect.
(5) Migration Class #E: Hydrophobic Liquids/High
Density
As with low and medium density hydrophobic% once
a high density plume reaches the water table, it forms
two distinct plumes (for a single chemical
contaminant) with each having its own average
velocity. The concentrated plume will sink to the
bottom of the aquifer. Its velocity is a function of the
contaminant's viscosity. If mounding on the aquitard
is significant, the density must also be considered.
The dilute plume will be above the concentrated
plume, with the highest concentration near the
concentrated plume and the lowest concentration at
the farthest distances from the concentrated plume.
The concentrated plume will have a single average
velocity and will start at the location of the source.
The dilute plume will have a single average velocity,
but its starting point can be from the location of the
source, or it can form from the concentrated plume
anywhere along the length of the concentrated plume.
76
-------
Retarded contaminants in the dilute plume move with
a velocity that is slower than ground-water velocity.
Thus, the contaminant velocity, based on the
ground-water velocity adjusted for retardation, is a
fraction of the ground-water velocity. Contaminants
in the concentrated plume do not move with the
ground-water velocity; their velocity must be
determined by considering the effect of the
hydrophobic contaminant's viscosity. If the sinker
mounds above the aquitard significantly, the density
should be taken into consideration. The concentrated
plume does not flow with retardation effects.
3.5.2.6 Nomograph Technique
The following nomograph is based on a solution to
solute transport in an aquifer from a point source that
extends throughout the thickness of the aquifer.
Contaminant transport from the source includes
advective flow with the ground water and longitudinal
and transverse dispersion (see Wilson and Miller
1978). The nomograph is taken from Kent et al.
(1985); the analyst is referred to this source
document for further discussion of the use of the
nomograph and its limitations.
The nomograph, which is a one-dimension model
(results restricted to a line, dispersion is two-
dimensional), is intended as a rapid means to obtain
an approximate solution. Scale factors are used to
translate Wilson and Miller (1978) to nomograph form.
Dilution/dispersive mixing and retardation parameters
are included in the solution.
Three scale factors that must be calculated before
using the nomograph are:
X =— (3-26)
RdDx
Dy
(3-27)
(3-28)
Two of the three ratios are computed directly, and the
third is found using the nomograph (Figure 3-8). The
procedure for calculating the scaling factors and using
the nomograph is presented as follows:
(1) Scale Factor Development
This nomograph models the same variety of
conditions that the Wilson and Miller model (from
which it was derived) does, yet it does it with only
one graph. This was achieved by scaling the
parameters to make them dimensionless. Distance X
is made dimensionless by dividing by the distance
scaling factor (XD, the characteristic dispersion
length). The mass flux (Q * C,) is made
dimensionless by dividing by the mass flux scaling
factor (QD). And time (T) is made dimensionless by
dividing by the time scaling factor (TD). Obtain XD
using the following:
Dx
(3-29)
where variables are defined as in Figure 3-8,
Definition of Terms.
Calculate TD using
(3-30)
V*
where
V
Rd
Td,Dx =
and where
(Ks*i)/Pe.
1 + p* Kd/Pt.
defined in Figure 3-8, Definition of
Terms,
= foe * Koc.
Calculate QD using:
QD = Pe * m * (Dx * Dy)1/2
(3-31)
where variables are defined as in Figure 3-8,
Definition of Terms.
(2) Application of Scale Factors
Use the three scale factors and the nomograph
(Figure 3-8) to calculate the concentration at time T
and distance X.
(a) Find T/Td curve desired.
(b) Find X/Xd on the x-axis.
(c) Plot the point of intersection of the T/Td curve
and X/Xd.
(d) Use this point and the point on the Q * C0/Qd line
to draw a straight line. Where this line intersects
the concentration line, the concentration at
distance X and time T is indicated.
3.5.2.7 Extent of Plume
As discussed earlier, a large volume of contaminated
ground water can result from a small volume of
chemical release. For example, a lO-gallon spill of
solvent can contaminate a billion gallons of ground
water to 10 ppb. Similarly, a 5000-gallon tanker
truck can contaminate 500 billion gallons of ground
water to 10 ppb. The analyst must be aware of the
relationship between volume of contaminant released
and volume of contaminated ground water. The
77
-------
Figure 3-8. Nomograph for solutions of time, distance, and concentration for any point along the principal direction of ground-
water flow.
Nomograph for
Plume Center-Line
Concentration
io-6H
1 fl-
IC-3-1
QC0
H
(Ib/ft3)
J«H
103-
104-
105
•10-"
•io-1
mg/l
102
104
10s
10'
-107
•10'
•'
100
1,000 10,000
100,000
X
XD
MO"
C
(mg/l)
" i
-10
F-102
HO3
=-104
Source: Kent et al. (1985)
-------
Figure 3-8. (Continued)
Definition of Terms
Primary Variables
C — concentration of leachate at a specific time and distance.
X = distance from source where concentration of leachate is computed.
distance is measured in direction of ground-water flow (perpendicular to gradient).
Y = transverse distance measured from the centerline of ground-water flow
(assumed to be zero in the nomograph).
t = sample time from beginning of leachate source flow.
Units
(M/L3)
(L)
(L)
(T)
Aquifer Parameters :
m = effective aquifer thickness or zone of mixing.
Pe = effective porosity of aquifer or zone of mixing.
v = velocity of ground-water flow within voids, estimated directly from:
(L)
(Dimensionless)
Ki
where
coefficient of permeability or hydraulic
conductivity of aquifer or zone of mixing.
gradient of ground-water flow.
(Dimensionless)
Transport Parameters:
= longitudinal dispersion coefficient (mixing rate) with respect to
distance in x direction and time, estimated directly or from:
(L2/T)
where
longitudinal dispersivity
molecular diffusion coefficient, which is assumed
to be negligible for velocities typical of permeable
aquifers. D* may be the dominant process in
aquitards where ax V would be negible.
2
(L2/T)
79
-------
Figure 3-8. (Continued)
Definition of Terms
Dy = Transverse dispersion coefficient (mixing rate) with respect to
distance in the y direction and time, estimated directly or from:
Dy=ayv+D*
where
Units
transverse dispersivity,
(I)
or estimated as:
Dx divided by a ratio, which commonly ranges between
5 and 10 for medium to coarse sand aquifers.
Retardation factor estimated directly or from:
(Dimensionless)
where
0b(Kd) v
R,= l+ (or)R,= —
d p d vd
Pb = bulk density of aquifer medium. (M/L3)
Pt = total porosity. (Dimensionless)
Kd = distribution factor for sorption on aquifer
medium (from sorption isotherm column studies) (L3/M)
v = velocity of ground water. (UT)
vd = observed velocity of leachate for a given concen-
tration and chemical species. (UT)
Y = coefficient for radioactive or biological decay. For
no decay, the value of y is one. (Assumed to be
one in the nomograph.) Calculated from: (Dimensionless)
4Dx 4Dxlog(2)
~ + 2 2t
vtl/2
where
X= decay constant =
log (2)
(I/I)
t1/2 = Half-life: time when half of the original mass remains.
(T)
80
-------
Figure 3-8. (Continued)
Definition of Terms
Units
Source Rate of Leachate:
QC0 = Mass flow fate: (M/T)
where
Q = Volume flow rate estimated directly or from: (L3/T)
Q=Aq
where
A = area of source. (L2)
q = recharge rate. (L/T)
C0 = Initial concentration (M/L3)
Intermediate Variables (used for nomograph only):
Xd = A characteristic dispersion length or scale factor given by: (L)
TD= A characteristic dispersion time or scale factor given by: (T)
R,D
T = —
Q = A characteristic dilution-dispersion flow given by: (L/T)
Q =P m*/D D
0> e > x y
81
-------
equation is a simple mass balance equation and is
expressed as follows:
For liquid contaminants:
= Vgw*Cgw
(3-32)
where
V., = volume of liquid chemical released.
Vgw = volume of contaminated ground water.
C1 = average concentration of chemical
contaminant in the released liquid.
Cgw = average concentration of contaminant
in ground water.
Both volumes and concentrations should be in the
same units.
For solid contaminants:
MC*CC = Cgw*Vgw
where
(3-33)
Mc = mass of solid waste, (in milligrams).
Cc = concentration expressed as mass
fraction, fraction of contaminant in
waste, (dimensionless).
Cgw = concentration of contaminant in ground
water, (mg/liter).
Vgw = volume of contaminated ground water,
(liters).
To convert the quantity of contaminated ground water
to a volume of contaminated soil, the following
equation is used:
(Vgw*0.13368)/Pt = Vc
where
(3-34)
Vgw = volume of contaminated ground water,
(in gallons).
pt = total porosity, (dimensionless fraction).
Vc = volume of contaminated soil, (in cubic
feet).
Or alternatively:
VgW/Pt = Vc (3-35)
where both volumes are in the same units.
3.5.2.8 Use of Monitoring Data
The analyst should take care when using monitoring
data to assess the depth of contamination in order to
calculate volume or mass in the plume. The
difference between monitoring and pumping wells will
affect the interpretation of the concentrations found in
the wells. Monitoring wells are the more desirable, but
since most existing wells will be pumping wells,
monitoring wells will typically have to be installed. The
cost associated with drilling monitoring wells most
likely will 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. They do not induce a large
vertical component in the ground-water flow, and
thus 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 (a pumping well provides water). 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 the length of the well screen. Water will
be drawn from above and below the well screen. The
well water does not reflect the concentration of a
particular depth, but rather reflects an average
concentration from a range of depths. This makes an
assessment of the depth of contamination difficult.
However, it makes assessing the mass in the plume
easier since the well draws a sample that represents
the concentrations at a wide range of depths near the
well screen depth.
3.5.2.9 VMS Model
In addition to the nomographic technique, the Office
of Solid Waste (OSW) has developed a simplified
model for its delisting program that relates leachate
concentration to receptor well concentration 500 feet
downgradient from the edge of a landfill. The
approach is called the VMS model (Vertical and
Horizontal Spread model). The only reduction in
concentration provided by the VMS model is that due
to vertical and horizontal dispersion (OSW plans to
add hydrolysis and biodegration for organics). The
approach involves back calculating from a health-
based ground-water concentration at the exposed
population location to an acceptable leachate
concentration at the site. Wastes with leachate above
this concentration must be managed as hazardous
wastes. Those with leachate below this concentration
can be managed in a municipal landfill or
82
-------
nonhazardous industrial landfill (i.e., outside the
hazardous waste system).
The only data the VMS model requires are the
leachate concentration and the annual volume of
wastes disposed of (constituent concentrations of
toxicants are also required in order to 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. For Superfund purposes, the total
volume of waste at the site would be used as the
"Annual Volume of Waste" term. A small volume of
waste can rely on greater dilution, while a large
volume of waste is assigned a smaller dilution
potential. All other input parameters are fixed at
reasonable worst-case values. By fixing the
environmental parameters, the model assumes a
generic environment that is consistent with OSW's
requirements. For CERCLA purposes, the model is
considered to be useful as a simplified analytical
procedure, and use of the VMS model for site-
specific, in-depth analysis is not recommended.
When using this model, one should keep its
limitations in mind. The VMS model simulates soluble
toxic constituents dissolving into percolating
precipitation and moving with the ground water. It
does not address solvent transport of organics (two-
phase flow) or the percolation of organic fluids into
the ground.
Critics of the VMS model have pointed out two
weaknesses of the approach. The first point is that
the model upon which the VMS model was based (the
Domenico and Palciaukas model) does not relate
leachate concentration to exposed population well
concentration. This model relates the concentration in
ground water immediately below the hazardous waste
sites to the exposed population well concentration.
When leachate enters ground water, it will be mixed
with ground water. This contaminates ground water
and at the same time dilutes the concentration of
leachate. It is wrong to use the C0 term in the model
as leachate concentration, because it represents the
concentration in ground water at the vertical point
where leachate enters. This concentration must be
measured on a site-specific basis to make the use
of the model consistent with the boundary conditions
used in the derivation of the model. The model is
derived from the following assumptions:
1. Steady-state concentrations are achieved under
the conditions that the concentration C0 in
ground water is maintained on a vertical plane of
finite size.
2. No longitudinal dispersion occurs; dispersion only
in the y and z directions is assumed.
3. Recharge or dilution mechanisms, other than
ground-water flow and the above-mentioned
dispersion, are ignored.
4. The contaminant velocity in ground water is
known.
The second weakness is the method used to
determine the cross-sectional area of the plume at
the edge of the landfill. The depth of the plume is
determined by the horizontal velocity of ground water
and the vertical velocity of the contaminant. The
model presumes that the vertical velocity of the
contaminant in the vadose zone is also the vertical
velocity of the contaminant in the saturated zone. In
the vadose zone, the contaminants are under the
influence of gravity; in the saturated zone, the vertical
velocity is much smaller because the effect of gravity
is canceled by the buoyancy forces. The VMS model
assumes that the velocities are the same.
These two weaknesses were present in the VMS
model at the time this document was written:
subsequent revisions may address these problems.
3.53 In-Depth Methods and Models
Several references are available that provide detailed
derivations and outline the application of more
sophisticated equations for the analysis of
contaminant migration in the saturated and
unsaturated zones. The analyst is referred to the
following documents: USEPA 1985J; Van Genuchten
and Alves (1982) Walton (1984), and Javendel et al.
(1984), USEPA (1986a), Geotrans (1986), and van
derHeijde (1985 and 1987).
Tables 3-14, 3-15, and 3-16 provide information
regarding several modeling procedures for the in-
depth assessment of the ground-water fate of
hazardous substances. Note that in order to provide
the analyst with an indication of the large number of
computer models that could be applied to analysis of
contaminant fate in ground water, Table 3-15
(Features of Unsaturated Zone and Ground-Water
Fate Models) provides data for 24 models in addition
to the 11 for which more detailed information is
provided in Tables 3-14 and 3-16. Two of the
models addressed in these tables are part of GEMS:
SESOIL and AT123D. The latter is described in
greater detail below, because it is more versatile and
is applicable to a wide range of fate analysis
situations. Additionally, following that discussion
further detail for certain of the models addressed in
Tables 3-14, 3-15, and 3-16 is also provided.
AT123D (Analytical Transient 1-, 2-, or 3-
Dimensional Simulation Model) is capable of
simulating the transport and fate of hazardous
material under 300 different user-selected situations
(Yeh 1981). One of eight source configurations can
be selected: a point source; line sources aligned in
83
-------
Table 3-14. Resource Requirements and Information Sources : Unsaturated Zone and Ground-Water Fate Models
Model
Description
Resource requirements, comments
References, sources of documentation,
software
Unsaturated zone
Seasonal Soil Compartment Model
(SESOIL)
CD
.b.
• Long-term fate simulations
x*s Accounts for numerous hydrologic,
meteorologic characteristics of site
xx Accounts for numerous transfer,
transformation processes: adsorption
volatilization, degradation, brodegradation
• Models organics, inorganics
xx Produces contaminant concentration
distribution in Unsaturated zone, quality of
ground-water runoff
& Handles up to three layers of soil types,
permeabilities
xx integrated into GEMS (see Section 3.1)
• Versatile, easy to use
xx FORTRAN program language; has been
Implemented on IBM 370, VAX 11/780
Documentation: Bomazountas and Wagner
1981
Contact for access to GEMS system:
Mr. Loren Hall
US. EPA, Exposure Evaluation Division
Washington, D.C.
(202) 382-3931
PRZM (Pestcide Root Zone Model)
• One-dimensional
xx Organic substances
xx Degradation is simulated
xx Provides pollutant velocity, distnbution,
and concentration data
xx Accommodates various release rates,
schedules
xx PC Based Model
xx Requires 256 K RAM minimum, 640K
preferred,
x Intel 8089 or 80287 math coprocessor
xx Has been field-verified with pesticides
xx FORTRAN program language
Reference: Carsel et al. 1984
Information:
David Disney
USEPA Environmental Research Laboratory
Athens, Ga. 30613
(909) 546-3132
PESTAN
xx One-dimensional
xx Organic substances
xx Degradation is simulated
xx Provides pollutant velocity, distribution,
and concentration data
xx Accommodates various release rates,
schedules
xx Considered a screening model
xx Rapid evaluations
xx Inexpensive, easy to use; requires only
hand-held calculator
xx Has been field-verified with pesticides
Reference: Enfield et al. 1982
(Continued)
-------
Table 3-14. (Continued)
Model
Description
Resource requirements, comments
References, sources of documentation,
software
Hydrologic evaluation of landfill
performance (HELP) (as modified by
Anderson-Nichols)
One-dimensional
Models leaching from landfills to
unsaturated soil beneath landfill
Has four options to handle modeling the
solubilization of toxic constituents
Models organics/inorganics
Uses rainfall and waste solubility to
model leachate concentrations leaving
landfill
Four options allow modeling with available
data
Information:
Brian Bicknell
Anderson-Nichols
Palo Alto, Calif. 94303
(415) 493-1864
oo
Ol
Saturated zone
Random Walk Solute Transport Model
(RWSTM)
(a.k.a. TRANS)
(requires PLASM for flow modeling)
xx One- or two-dimensional
xx Time-vanant release rates
xx Accommodates well-injected release
xx Incorporates dispersion, retardation
xx Handles nonconservative pollutants
xx Accounts for well pumping
xx Provides contaminant concentration at
user-selected points
xx Requires mathematical programming,
hydrogeological knowledge on part of
user
xx Has been field-validated
Documentation: Prickett et al. 1981
Coupled Fluid, Energy and Solute
Transport (CFEST) Combined with
UNSAT-ID
xx Three-dimensional
xx Accommodates heterogeneous,
anisotropic, multilayered soil
configurations
x Handles saline aquifers as well as fresh
water
xx Transport mechanisms of dispersion,
advection simulated
xx Sorption, degradation mechanisms not
incorporated
xx Time-variant release and flow rates
xx Combination covers unsaturated and
saturated zones
< Has been applied for arsenic and
organic wastes
Documentation: Gupta et al. 1987
(Continued)
-------
Table 3-14. (Continued)
Model
Description
Resource requirements, comments
References, sources of documentation,
software
Sandra Waste Isolation Flow and
Transport Model (SWIFT and SWIFT
CD
Leachate Plume Migration
Model (LPMM)
Three-dimensional
Transport processes of advection,
dispersion simulated
Sorption, degradation processes
accounted for
Appropriate for waste-infection,
waste-isolation modeling
Code was based on SWIP Model
• Has been field-verified
• Has associated user's guide in self-
teaching format
• FORTRAN program; has been
implemented on various CDC systems
including CDC 7600
• 1986 version has been released
XMS Continuous source model
xx Dispersion is simulated
xx Degradation processes accounted for
xx Has been field verified
x A simplistic model; results may not be as
sophisticated as necessary for Level III
work
Can be used in nomographic, hand-held
calculator, or computer form
Relatively easy to use
Documentation: Reeves and Cranwell
1981; Finley and Reeves 1968
Software:
National Energy Software Center
Argonne National Laboratories
Argonne, III. 60439
Information:
Intera Environmental Consultants, Inc.
11999 Katy Freeway, Suite 610
Houston, Tex. 77079
References: Kent et al. 1982
Analytical Transient One-, Two-, and
Three-Dimensional Simulation Model
(AT123D)
See Section 4.4.2 of text
• FORTRAN program applicable to wide
range of computers
• May require extensive setup time
• Available through GEMS (see Section
4.1)
Documentation: Yeh 1981
(Continued)
-------
Table 3-14. (Continued)
Model
Description
Resource requirements, comments
References, sources of documentation,
software
Unsaturated and Saturated Zones
Finite-Element Model of Waste
(FEMWASTE) and
Finite Element Model of Water Now
(FEMWATER)
Two-dimensional
Aeflnterzone transfer is modeled
^Incorporates convectron, dispersion
£gSimulates degradation of nonconservative
substances
&& Absorption is accounted for
«KCapable of modeling layered,
heterogeneous soil zones
.e^f FEMWATER is a model for ground-
water flow, while FEMWASTE simulates
the transport/fate of contaminants
.efts'Has been Implemented on IBM 360
,«•,«• May require background in hydrogeology,
differential equations, programming
g& Field-verified
Documentation: Yeh and Ward 1981
Information;
Dr. George T. Yeh
Oak Ridge National Laboratory
Environmental Science Division
P.O. Boxx
Oak Ridge, Tenn. 37830
(615) 574-7285
Solute Transport and Dispersion Model && Two-dimensional e& Field Verified
.efts' Conservative substances (no decay .efts' Relatively inexpensive, easy to use
simulation)
.efts' Heterogeneous soil conditions accounted
for
.efts' Pumping or recharging well effects
modeled
• Thickness of saturated zone may vary
Documentation:
Kowikow and Bredehaeft 1974
Sources: USEPA 1982b; Brown et al. 1983; Kufs et al. 1983; Versar 1983.
-------
Table 3-15. Features of Unsaturated Zone and Ground-Water Fate Models
\
SOIL/ROCK
CHARACTERISTICS
\
FLUID
CONDITIONS
\ FLOW \
\CONDITIONS \
AT1IJO
LPMM
OME-D
ptwna (to)
R«**«.
•LM
VHS
MW-Analyl.
•ta*tr Belt
Klnztlkach
WlHon M
Belult
FtiMrait
F*mw«l«r/F«m
•
itcrenenve CIIAIIUK ^
Ttk 1M1 1
K*M tl M. 1MB
«WI tttniMMtll MM Ahw* IBM
C*«*H tl •. tMI
JntmM tl H. IM4
ItaycMm tl H. 1MT
OMIKMC* mt PikjIiMkjt tMI
urn «tf HtlkJt m* Bikilittai tMt
•Mr mn4 VtmUfl tMT
KMutlkM* 1MB
WHIM tMt
B*yn tMI
MMIHwi tMI
T»* tMT. Tth, »M Wtrt tMI
NwwM M M. 1MB
Onto unt Btfrt IBM
Caratl at «. IBM
BWMMUMOT mt WtfMf IBM
mi atBMMtn tBTB
V*M IBM
•wtflhlM IBM
Ol»U tl Bl tMT
Ttft n« MM tMB
KlBP tMT
KwilWw unt BtOMtn 1BTB
Bswlwil Mitf RwiNiMV IMfl
Traqr IBM
PrkkMI tl M. tMt
MMtra tMt
Rttvw tl M. tMB
Trmto IBM
HwBhswn M •! 1W7
Mill tl tl. tMT
EnlltM tl al. tMI
••• taklt S-14
00
do
•NON-AQUEOUS PHASE LIQUIDS 1) FOR UNSATURATED ZONE ONLY.
-------
Table 3-15. (Continued)
OO
CD
\
GEOMETRY
\
TRANSPORT AND
FATE PROCESSES
\
REFERENCE CITATION
ATI MO
LPMM
ONE-D
Plwiw (SO)
R««q
SLN
VH»
RW-AMlyt.
BVAVOT J>tl
KlKMltlMlt
wftlMfi 3S
••tut*
F*«tr*n
F<«w«t«f>FMN»Mt«
Peril*
oat/osi
PRZM
• MOIL
•UMATRA-1
•UTRA
TRIPM
CF68T
FEWA'FEMA
HSTID
MOC(U»QS S-0)
MOC OCNSC
HOC NRC
RW8TM
•WENT
•WIFT •
TRACR M>
TRAFRAP/WT
• IOPLUME-H
PE3TAN
HELP
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
• ^
•
•
•
•
•
•
•
•
•
•
^
•
•
•
•
•
•
•
4
•
•
•
\
•
•
•
•
•
^
•
•
•
•
•
" \
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
^ \
•
' \
•
•
•
•" i
•
•
•
•
•
•
•
•
•
•
' \
•
*
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
* \
•
' \
^
•
•
•
^
•
•
•
•
•
•
•
ft* 1M1 1
K*M M *. IMS
•Ml OWWOMM *Ht AtVM 1MI
C»4tH M M. IMS
JcvMltfcl M •!. 1M4
Hur«K*rn M d. 1M7
Dcmmlo* tut PctotaMlIM 1M1
VMI 4>f tMMto m4 •ftalvMWi 1MC
•MC *n4 VwraV 1M7
KNn»»*Mi 1M«
W««m IMf
•*l|Ni 1MI
MMftnMS • ••*
Ted 1MT. Vik, mt W«N 1M1
Rim»lMt •! tf. 1M»
Dwte ant •** 1M»
C«M4 M M. 1M4
••IMMHHM0 M|4 WMMf 1M4
•M owNwMwi irr»
V«M 1M4
QufMhtan IMS
Oupu M ri. 1MT
T««l (Ml NMI IMf
KI|V 1MT
KMliMr «M >l«l»lltl im
•wrt*r« *M Kwilnw IMS
Tragy IMS
PiMwn •( «L 1M1
Inln* 1MI
RWVM M •!. ISM
Travta 1M4
Hufikom M d. 1MT
RNd M d. 1MT
EnlMd M «. IMS
8M trill* 1-14
PHASE LIOMM 1, FOR UNSATUMnGO ZONE ONLK.
-------
Table 3-15. (Continued)
K BOUNDARY/SOURCE \
CHARACTERISTICS \
CAPAilLITIES
A
Modal
REFERENCE CITATION
AT123D
LPMM
ONE-D
Plum. (3D)
R»tq
SLM
VMS
RW-Analyt.
•aavar Sett
Klntalbaeh
Walton 3$
Soluta
Famtran
Fam watar/Famwaata
Porllo
GS2/OS3
PRZM
SESOIL
SUMATRA-1
SUTRA
TRIPM
CFEST
FEWA/FEMA
HST1D
MOC(USGS 2-D)
MOC DENSE
HOC NRC
RWSTM
SWENT
SWIFT II
TRACR 3D
TRAFRAP/WT
•IOPLUME-II
PEST AN
HELP
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
Y«a 1M1 1
Kant at al. IMS
van QamioMan and Alvaa 1M2
Catfafl M •!. tSS2
JavanM at al. ISM
HuyaKorn at •!. 1M7
DcMMnfco mtt PrietankM 1SS2
nn «w H^fd* •>* SrlnhriMn 1Mt
SMT mnt V«nul|i 1W7
Kfei»lbMh ISM
WiHon IMS
toll* IMS
Mcrthwi IMS
Y«* 1M7, V«k, and Ward 1M1
Runekal at al. IMS
Davfc an* 8*g«l IMS
Caracl M al. 1M4
SwuMuntu ami Wagnw 1M4
van QanueMan 1*T»
VOM 1M4
Ouragklan 1M3
Oupta al al. 1M7
Va* tnt Hutf IMS
KlM> »M7 •
Konfcow and Badahecti 107»
Sanfetd and Konlkow IMS
Tracy 1M2
Prtekan « al. 1M1
Mara 1M3
Rwvaa •( al. 1MC
Trarb 1M4
Huyakem a( al. 1M7
Mai at al. 1M7
EnftaM M al. 1M2
Saa tabto 3-14
•NON-AQUEOUS PHASE LIQUIDS 1) FOR UNSATURATED ZONE ONLY.
90
-------
Table 3-16. Data Requirements for Unsaturated Zone and Ground-Water Fate Models
CD
HWSTM PL»SM
MMMSTC • FCMMTtft
SOURCES: USEPA 1982b; Brown It ll. 1913; Kufe el il. 19*3; Versir 1983.
-------
one of three different ways with respect to ground-
water flow; area sources, also aligned in one of three
different configurations; or a volume source (existing
plume). Release types can be instantaneous,
longer-term but finite, or constant. Aquitard locations
can be specified below or on both sides of the aquifer
in any configuration; or the aquifer can be treated as
infinite in all directions. Advection and dispersion
transports are simulated. Losses resulting from
volatilization, degradation, and adsorption are
modeled. The model predicts contaminant movement
in one, two, or three dimensions (Yeh 1981).
Use of AT123D requires the following information:
dispersion coefficients in horizontal, vertical, and
longitudinal direction: geometry of the aquifer,
especially regarding configuration of aquitards; soil
properties, including bulk density, effective porosity,
hydraulic conductivity (permeability); source type; and
release duration and strength, soil-waste stream
partition coefficient, hydraulic gradients, and an
overall decay constant (or soil half-life figures) for
the substance studied (Yeh 1981).
The model determines contaminant concentration at
any point, at a downstream and lateral distance and
depth specified by the user, as a function of time
from the beginning of source release.
AT123D can be accessed through the GEMS system
(see Section 3.1). It is written in FORTRAN and can
be installed on a wide range of computer types.
In addition, the Office of Solid Waste (OSW) has
developed a national model that uses the Monte Carlo
simulation for relisting hazardous wastes on a generic
basis. This FORTRAN computer model is a three-
dimensional advective-dispersive transport model.
The model currently considers the mechanisms of
hydrolysis, dispersion, and rainfall recharge into the
ground-water plume. OSW is using the model to
back-calculate from a health-based standard at the
exposed population well to an acceptable on-site
leachate concentration. If a treated waste produces
leachate with a contaminant concentration below the
acceptable concentration, then it is considered
protective of the public health.
The model currently uses the HELP model to provide
leachate release rates. Leachate strength
(concentration) is provided by the Toxicity
Characteristic Leaching Procedure (TCLP). OSW
plans to add the geochemical model MINTEQ to
handle metal speciation. Biodegradation processes
are being evaluated for incorporation into the model.
Since EPA's model is a national model that uses a
generic environment, the data requirements are
minimal. The model approximates an average
environment by making multiple runs (typically several
thousand runs for each chemical constituent) with
varying environmental data. By applying this
approach, called a Monte Carlo simulation, one can
model the dilution potential of all possible sites as a
cumulative frequency distribution versus expected
concentration at an exposed population well. The
extent to which a particular CERCLA site matches the
OSW model depends on the closeness of site
characteristics and the model assumptions. If a
particular CERCLA site has adequate hydrogeologic
data and satisfies the model assumptions, the model
can be used for site-specific analyses. Before final
assessment of the desired level of cleanup, however,
application of the model on a site-specific basis will
typically be required. Generic modeling is appropriate
for OSW's purposes, but may suggest cleanup levels
beyond those necessary at a particular site.
Preliminary work or screening-level efforts at
CERCLA sites where adequate, good quality
hydrogeologic data do not exist can benefit from the
model's minimal data requirements for site-specific
environmental parameters.
The model is being updated to incorporate flow
through fractured media and the unsaturated zone.
The data base for MINTEQ is being enlarged to
handle additional metals, and more data are being
collected to validate the model results.
Since OSW's model uses a Monte Carlo simulated
environment, it should be applied with this limitation in
mind. Other limitations in the use of this model derive
from two sources: (1) limitations in the scope of the
model, and (2) specific modeling choices made so
that the model would support OSW's requirements.
The model's scope is limited by the leachate release
algorithm HELP, which models soluble toxic
constituents dissolving into percolating rainwater and
moving with that water. It does not address
percolation of organic fluids into the ground or
associated leaching by concentrated organics.
Additionally, the TCLP does not fully predict leachate
concentrations due to leaching with water containing
dissolved solvents. It does assume the presence of
acetic acid in leach water, thereby providing some
measure of hydrophobic solubility. Although HELP
can model a variety of landfill cover situations, OSW's
requirements were such that it modeled a landfill with
a failed liner but an intact (aged) cover. The
permeability of the hypothetical cover was chosen at
1 x 10-6 cm/sec to represent an aged (deteriorated)
cover with an initial permeability of 1 x 10-7 cm/sec.
OSW states that it found the range of permeabilities
for aged clay actually to be between 1.4 x 10-6 and
43 x 10-6 (USEPA 1986). For CERCLA sites,
selection of a permeability within that range may be
more appropriate. Also, many CERCLA sites do not
have a cover, or the cover may be breached. In either
case, the mass flux leaving the site will be
considerably larger. Even if the site has an intact
cover, one may wish to predict long-term potential
92
-------
releases and also to consider the eventual
subsidence and breaching that may occur in the
future.
Pesticide Root Zone Model (PRZM) (Carsel et al.,
1984) simulates the vertical movement of pesticides
in unsaturated soil, both within and below the plant
root zone, and extending to the water table using
generally available input data that are reasonable in
spatial and temporal requirements. The model
consists of hydrology and chemical transport
components that simulate runoff, erosion, plant
uptake, leaching, decay, foliar wash off, and
volatilization (implicitly) of a pesticide. Predictions can
be made daily, monthly, or annually.
3.5.4 Short- and Long-Term Concentration
Calculations
Long-term average ground-water concentrations of
contaminants at exposure points are a function of the
concentration profile over the time period of study,
which are, in turn, a function of hydrologic
fluctuations, release rate fluctuations, and the
effectiveness of remedial actions. Average
concentration values are obtained from steady-state
methods. Several of the in-depth analysis models
tabulated in Section 3.5.3 accept time-weighted
input data, and provide long-term average
concentrations, as well as the concentration profile as
a function of time.
Short-term concentrations at exposure points are
obtained by examining the ground-water
concentration profile at the selected exposure point
over time, and identifiying of the period of maximum
concentration.
3.6 Biotic Pathways
3.6. J Estimation Procedures
After the fate of a contaminant in air, water, and
ground water has been estimated, one can assess its
fate in biotic populations. Using the ambient
concentration data developed for each of these
media, a determination is made whether any biotic
populations that can potentially serve as pathways for
human exposure to hazardous materials (i.e., vector
organisms) are within zones of elevated hazardous
material concentrations. Such vector populations may
include agricultural crops; agricultural livestock; fish,
shellfish, or crustaceans that are important
commercial or sport species; and game populations in
hunting areas.
In assessing the biological fate of hazardous
materials, the following processes, which determine
the rate of introduction of hazardous material to and
the final concentration of hazardous material within
vector organisms, should be considered:
^ The concentration of hazardous material in
environmental media containing or supporting
vector organisms.
A* The metabolic rate of the vector organisms.
Metabolic rates are functions of several
environmental parameters including temperature
and the availability of sunlight, oxygen, nutrients,
and water or other factors.
MS* Substance bioavailability: the affinity of each
hazardous substance for partitioning into the
organic phase or its availability for other forms of
uptake. The bioavailability of each substance
differs, as does that of various chemical species
of an individual substance: the octanol/water
partition coefficient is an indication of this
parameter. Bioavailability of a given substance
can vary with environmental conditions. Factors
that influence the physiochemical speciation of
substances, and thus their bioavailability, include
salinity, pH, Eh, organic carbon concentration,
and temperature.
** Characteristics of species metabolic processes.
These characteristics differ among species and
include feeding habits and ability to metabolically
degrade, store, and eliminate the substance.
Bioconcentration factors (or BCFs, the ratios of
organism tissue concentration to ambient
environmental concentration) for many species
and hazardous substances have been empirically
determined and are discussed below.
Consider the following transport mechanisms in
assessing the distribution of hazardous substances
within the biologic medium and identifying the
potential points of human exposure:
* Transport and distribution of vector organisms as
a result of human commercial or sport activity.
MS* Migration of organisms, or movement of these
organisms with advective flow of environmental
substrate media.
x* Movement of contaminants through the food
chain. This mechanism often results in very high
concentrations of hazardous materials in the
tissue of higher trophic level organisms within and
without contaminated areas.
General theoretical relationships between the above
factors and concentrations of hazardous substances
at human exposure points are not available. This is
because such relationships are highly specific to
individual ecologies, biotic species, hazardous
substances, and human activities associated with
involved biotic species.
93
-------
For this reason, the assessment of biotic
concentrations of hazardous substances at human
exposure points is limited to the qualitative
identification of major pathways, and the rough
quantification of exposure levels wherever some
means of relating ambient soil, water, or air
concentrations to edible tissue concentrations are
available.
The available methods of estimating tissue
concentrations in aquatic animals, terrestrial animals,
and terrestrial plants are discussed in the following
sections.
3.6.1.1 Aquatic Animals
Because aquatic animals are immersed in the
contaminated water medium to which they are
exposed, it is commonly assumed that tissue
contaminant concentrations are a function of
contaminant equilibrium partitioning between water
and organic tissue, and are therefore directly related
to contaminant ambient water concentrations. The
bioconcentration factor (BCF) represents the ratio of
aquatic animal tissue concentration to water
concentration. This ratio is highly contaminant-
specific and is also dependent on the aquatic species
and on site parameters.
The most reliable source of aquatic animal BCF
values is monitoring data for the site. Wherever water
concentrations and biotic tissue concentrations have
been surveyed simultaneously, a site-specific BCF
can be calculated for the species and substance
involved (assuming water column concentration
values represent relatively steady concentrations over
at least the previous several weeks, and not short-
term high or low concentrations). This BCF can be
used to project changes in tissue concentrations
resulting from projected changes in ambient water
concentrations of the involved hazardous substance.
In cases where site monitoring data are insufficient for
development of a BCF, one can use the BCF values
reported in technical literature. A substantial amount
of research is available regarding the bioconcentration
of hazardous substances, especially in aquatic
organisms (see USEPA Office of Water Regulations
and Standards: Ambient Water Quality Criteria
documents, for a review of research current to 1980;
or Verschueren 1984; Dawson, English, and Petty
1980; Mabey at al. 1982; and Callahan et al. 1979 for
BCF factors). Exercise care to match contaminants,
species, and site conditions (e.g., temperature, pH,
water salinity) for which reported BCF values were
measured with conditions at the site. BCF values for
different species or contaminants or those measured
under dissimilar conditions may not be applicable.
A third alternative for derivation of BCF values is to
calculate these values based on the structure or
physiochemical properties of the hazardous
substance. See Lyman et al. (1982), Kenaga and
Goring (1978), and Veith et al. (1980) for instructions
on BCF estimation procedures.
3.6.1.2 Terrestrial Animals
Little data are available allowing the quantification of
contaminant concentrations in edible terrestrial animal
tissue based on ambient environmental
concentrations. Kenaga (1980) compiled and studied
data comparing dietary concentrations of several
organic compounds with the concentration of these
compounds in the fat of beef cattle. He found that the
fat/diet BCFs for these compounds correlate
reasonably well with the water solubility (negative
correlation) and octanol-water partition coefficient
(positive correlation) of these compounds. BCFs
could only be predicted within three to four orders of
magnitude, however. This method of tissue
concentration estimation must be considered
semiquantitative at best.
Human exposure to contaminants through the
terrestrial animal pathway can be reliably determined
only through identification of potential vector
organisms and exposure points, and through a
sampling and analysis program for determining tissue
concentrations at these exposure points.
3.6.1.3 Terrestrial Plants
Plant adsorption of environmental contaminants has
been studied by various researchers, and some data
are available regarding the uptake of pesticides and
other contaminants by edible crops. These data cover
specific crop uptake of specific contaminants (see
CDHS 1985 for a review of pesticide research),
however, and no relationships allowing reliable
extrapolation of soil/plant tissue concentration ratios
are presently identified. Where plant/soil BCF data are
available in the technical literature for the specific
plant species, contaminant, soil type, and tissue type
of concern in a Superfund exposure assessment,
these BCF data can be used for a semiquantitative
estimation of edible tissue concentrations.
As is the case with terrestrial animals, the most
reliable technique for assessing contaminant
concentrations at points of human exposure to plant
tissue is the identification of potential vector
organisms and exposure points, and the surveying of
tissue contaminant concentration in these organisms.
94
-------
Chapter 4
Uncertainty in the Analysis
This chapter provides a brief introduction to the
evaluation of uncertainties inherent in the exposure
assessment process. When applying the exposure
assessment tools outlined in the preceding sections,
uncertainty may be a factor at each step. Such
uncertainty can involve variations in the values of
variables used as input to a given model, the
accuracy with which the model itself represents actual
environmental processes, and the manner in which
the exposure scenario is developed. Each of these
categories of potential uncertainty is discussed below.
Once the exposure assessment is completed, its
results must be reviewed and evaluated to identify the
degree of uncertainty involved. This factor should
then be considered when using the assessment
results for remedial decisionmaking.
The following discussions focus on the uncertainties
of assessing the average daily exposures to toxic
chemicals; uncertainties related to the' human health
response to these exposures are not discussed. The
information provided here does not constitute a
comprehensive treatment of uncertainties in the
exposure assessment process. It is intended to make
the analyst aware of the categories of uncertainties
that may be involved in exposure assessments. In-
depth guidance for the execution of uncertainty
analyses is provided in various references in the
literature. Specifically, the analyst may wish to review
the following sources of information concerning
various aspects of uncertainty analysis pertinent to
the exposure assessment process:
- Cohen (1950)
- Eisenhart (1968)
- Henrion and Morgan (1984)
- Hoffman et al. (1984)
- Kleijnen (1974)
- Morgan et al. (1984)
- Rubinstein (1981)
- USEPA (1987e)
4.1 Sources of Uncertainty
4.1.1 Input Variable Uncertainty
Most of the analytical procedures presented in this
manual are quantitative in nature, and their results
may be highly dependent upon the accuracy of the
input variables used. For example, hydraulic
conductivity and other parameters that determine the
velocity of ground water and the contaminants that it
may carry can vary significantly over relatively short
distances, thereby affecting one's ability to estimate
average contaminant velocities with confidence.
Similarly, the presence of hydrogeologic hetero-
geneities can affect the speed with which
contaminants arrive at a given well from their point of
release and also their direction of travel. Often, the
presence of such heterogeneities may be unknown.
Thus, the accuracy with which values for such
parameters can be quantified is critical to the degree
of confidence that the decisionmaker has in the
assessment results.
Most scientific computation involves a limited number
of input variables, which are tied together by a model
to provide the desired output. The input parameters
can be broadly classified into the following categories:
constants, state variables, and natural variables.
A constant has a single value irrespective of the
nature of other variables. In some cases, the
variability of a parameter may be so small that it can
be considered constant. In other cases, even if the
value varies, its effect on the final answer may be
minimal. The results are not sensitive to variation in
that parameter's value.
A state variable is one that has a fixed value, but that
value is not known accurately. The errors in such
variables are due to limitations in experimental
techniques. A relevant example is the octanol/water
partition coefficient. While this has a single value for a
given system, some degree of uncertainty is
introduced through experimental errors. In some
instances the values of state variables are estimated
rather than measured; therefore, the uncertainties for
such values are even higher.
A natural variable is one that can exhibit different
values. An example is soil porosity, which can exhibit
different values within a range because the soil matrix
varies with location, and because a given area may
include many soil types.
95
-------
If the actual values for such variables are not
accurately known for the location in question, the
estimated exposure may be significantly in error. This
problem is illustrated by a study where the values of
parameters needed to calculate the velocity of a
solute in ground water were varied randomly, using
Monte Carlo simulation techniques (Mercer, Silka,
and Faust 1985). This analysis determined that the
velocity estimates may vary over four orders of
magnitude.
The selection of accurate input parameters is
essential to estimate the contaminant velocity and
other components of the exposure assessment.
Often, however, the analyst will not be able to
determine the value of such parameters with absolute
certainty. It is important that one be aware of the type
and degree of uncertainties involved at each stage of
the analysis, and interpret the results obtained
accordingly.
The different values of input parameters that are
measured many times can be expressed as a
parameter distribution. A parameter distribution
typically appears as a bell-shaped curve. The mode,
or the most likely value, is represented by the peak of
the bell-shaped curve. The tails to either side
represent the relative frequency of times when the
measured values are greater or less than the mode.
For a parameter that varies considerably, the bell-
shaped curve will be wide (standard deviation is
large). For those that do not vary appreciably, it will
be narrow (standard deviation is small).
Input parameter distributions can be used to generate
the output parameter distribution. The shape of the
parameter distribution conveys the degree of
uncertainty of the parameter (input or output). This is
the most rigorous way to define the uncertainty of the
predicted output parameter; however, it is used
infrequently in the environmental field due to the lack
of input parameter distributions upon which to base
the predicted output parameter distribution. This
subject will be discussed further in the section on the
Monte Carlo technique.
In the environmental field, the methods used for
discussing the degree of uncertainty are often
qualitative rather than quantitative. Qualitative
methods involve discussing whether the data are
thought to be representative or not. Some exposure
modeling is done based on literature values rather
than measured values. In such cases the degree of
certainty may be expressed as whether the estimate
was based on literature values or measured values,
not on how well defined the distribution of the
parameter is. Some exposure estimates are based on
estimated parameters; the qualitative statement that
the exposure was based on estimated parameters
defines the certainty in a qualitative manner.
4.2 Modeling Uncertainty
4.2.1 Model Simplification
The degree to which a specific contaminant transport
and fate model accurately represents the actual
conditions that are present in the environment
constitutes a large source of potential uncertainty.
The analyst must choose the model that addresses
the appropriate aspects of interest.
Models are typically simplifications of the complexities
of reality. There is some accuracy lost when making
these simplifications. While such loss may be small in
some cases, in others it may be unacceptably large.
Two assumptions that illustrate this idea are the
assumptions of homogeneous soils and isotropic soils
for ground-water models. In most cases, these
assumptions do not materially change the answer. If
the soil under the site has layer cake stratigraphy, the
assumption of homogeneity is invalid. Typically, most
cases will be in-between the two extremes of
homogeneous soils and completely non-
homogeneous soils. The analyst will have to decide if
the assumptions are valid for each case.
In some cases the simplification of the real world into
an actual model is acceptable and, although
producing uncertainty, it is a necessary evil. There is
a point at which the level of the discrepancy between
the model and the real world constitutes an error in
the use of the model and not an acceptable
simplification that is necessary to model a
complicated real world. At this point, the deviation is
an error and not an uncertain prediction.
4.2.2 Averaging Hydraulic Conductivities
An example of this would be the modeling of ground
water flow by averaging the hydraulic conductivities
across all aquifer materials. For contaminant transport
modeling, this would constitute an error; however, for
modeling well production, this is an accepted
practice. Ground water modeling with numbers has
been occurring for the last 100 years. For the first 90
years of this period, most of the modeling was for
water supply; contaminant migration was not
modeled. The practice of averaging the hydraulic
conductivities across the cross-sectional area of the
aquifer produced answers that had high certainties
when predicting the volume of water that could be
produced by a well during a period of time. Some
modelers applied this technique to the problem of
modeling contaminant migration and produced
erroneous results. Although they were accustomed to
this practice, it was not acceptable in this case.
Modeling contaminant migration requires that areas of
different hydraulic conductivity be treated separately
(sometimes it is not possible to differentiate the areas
and the model results must be viewed as less
certain). For example, if the site overlies a sand layer
and a clay layer, the analyst should model the two
96
-------
layers separately. The result of the separate modeling
will show that the time of arrival in the sand is much
sooner than in the clay layer. Effectively the majority
of the contaminant mass would migrate through the
sand layer and hardly any would use the clay layer for
migration. Assuming an average hydraulic
conductivity would predict a time delay between
release and arrival that is 100 to 1000 times too long.
Such uncertainties, however, constitute an error of
approach, and are not unresolvable uncertainties.
4.2.3 Dispersion Simulation
Different ground-water models simulate dispersion in
different ways. The degree to which a particular
model accurately models the dispersion at a given
site affects the accuracy of using that model for that
site. Ground water dispersion modeling is a young
field and the state of the art is rapidly advancing. The
analyst should become familiar with the dispersion
simulation technique for each model he/she uses.
Also, some ground water models presume an aquifer
of infinite depth, while some model a finite aquifer
depth. Contaminants dispersing in an aquifer of finite
depth will effectively reflect off the lower aquitard and
cause the resulting downstream concentrations to be
larger. Use of a model appropriate to the constraints
of the site is necessary for accurate modeling of the
drop-off in contaminant concentration with travel
distance. Additionally, some models will simulate
lateral constraints of the aquifer to model this
limitation on the reduction in downstream
concentrations.
Dispersion modeling in air and surface water has
been performed for a much longer time, and as such,
the methods for modeling dispersion have coalesced
into a consistent approach. However, limitations on
the extent of dispersion for air modeling can vary. For
example, a valley model will simulate the constraint of
lateral dispersion by the valley walls. A model that
handles inversions will simulate the build-up of
contaminant concentration due to limited vertical
mixing. Surface water models may vary on the
approach they take to modeling initial mixing. Some
surface water models use compartments to manage
the modeling task. If the modeler uses a small
number of large compartments, small scale effects
may not be accurately modeled and the results will be
less certain.
4.2.4 Numerical Models and Analytical Models
Different types of models provide varying accuracy in
different situations. Two types of models are
numerical (finite-element) and analytical models.
Neither is best in all cases, but one is usually better
in a given situation. The numerical models are
typically more difficult to use, and thus ease of use
may enter into the decision of model selection.
Analytical models often involve mathematical
simplifications. These simplifications are made in
order to find a closed-form solution. In most cases
the accuracy lost is negligible; however, in extreme
cases the inaccuracy will be large.
Typically, analytical models require less computer
time than do numerical models. If the grid is large, a
numerical model requires a substantial amount of
computer time for each run. Numerical models
typically require more input data. Different program
needs cause different questions to be raised. A
preliminary scoping problem will rarely require a
numerical model; conversely, a problem that requires
maximum defensibility will suggest that the additional
data and operational burdens of a numerical model
are justified in light of the greater certainty of the
output.
In cases where the question involves simulating what
will happen in typical generic situations across the
country, an analytical model will give a better picture
than a numerical model. Numerical models address
site-specific conditions better than do analytical
models: they do not necessarily model a typical
situation with any increased accuracy.
4.2.5 Chemical Degradation Simulation
Some models do not describe all of the processes
that may potentially occur. For example, degradation
is not accounted for in some models. If the con-
taminant is extremely refractory (i.e., does not
degrade), this limitation will not materially affect the
answer. If the contaminant degrades quickly,
however, this limitation will cause the model results to
be in substantial error. Some models simulate the
effect on the reaction rate kinetics of two con-
centrations while some use only one concentration.
The simpler approach of 1st order reaction kinetics is
acceptable if the other concentration does not vary
appreciably, and is less accurate if both the con-
centrations vary substantially. The analyst must rely
on his/her judgment to ensure that the uncertainty is
minimized.
4.2.6 Model Operational Parameters
Certain modeling parameters specified by the analyst
can have a profound effect on the accuracy and
viability of the output. An example is the parameter
"time step." Time step is used on iterative models.
Models may either calculate an answer explicitly or
they may determine their solution with a successive
iteration approach. For iterative models, the analyst
will have to make many model runs, and not stop until
he/she has a good run. The challenge of choosing an
appropriate time step is that both too large and too
small time steps cause inaccuracies. The analyst
must find the optimum size for the time step. A time
step that is too small causes numerical error
97
-------
propagation (see below), while one that is too large
causes a less accurate calculation of each step.
Numerical error propagation in iterative models can
cause inaccurate answers. If the analyst uses too
many iterations, the truncation error of digital
representation of numbers can build upon the
successive iterations and produce output that is
totally erroneous, The degree of error that can be
present can make the output totally meaningless. For
example, the estimated output concentrations can
include concentrations that are greater than a million
PPM or concentrations that are negative. Clearly,
concentrations in these ranges signify a bad run. The
analyst must also watch out for iteration errors that
produce errors that are less obvious and, hence,
there is the possibility that the analyst will not be
aware of their occurrence. Conversely, the analyst
may choose too few iterations and the resulting time
step between each iteration then becomes too large.
In this case, the model will inaccurately calculate
each step. The analyst must become familiar with the
models he/she is using so as to stay in the safe area
between the two extremes. Knowing the precise limits
is difficult, but staying between them is important.
4.2.7 Source Shape
The degree to which the shape of the source is
modeled can effect uncertainty. For example, if the
analyst uses a point source model to model an area
source, the nearby concentrations will be less
accurate than they would be if the analyst used an
area source model. Line sources and volume sources
can provide the same problem. At large distances
from the source, the effect of the shape of the source
is less important, and may often be neglected. Some
sources are best modeled as a vertical line source
and some are best modeled as a horizontal line
source; hence, orientation is a factor as well as
shape. It is a matter of fit between the model and the
actual site rather than choosing the best source
shape for all cases.
4.2.8 Steady State Modeling
Use of a steady-state model to model a true
steady-state scenario provides accurate results. Use
of a steady-state model to model a truly dynamic
scenario can produce inaccurate answers. In most
cases, the analyst will have to make a judgment as to
whether the actual scenario is close enough to steady
state to justify using a steady-state model. The
analyst must match the model to the question being
asked, and to the details of the specific site, in order
to minimize the uncertainty of the output.
4.2.9 Number of Dimensions Addressed by the
Model
Choice of a one-, two- or three-dimensional
model can affect the uncertainty of the results.
Neither is best in all cases and, typically, one is
preferred in a given site-specific scenario. The
three-dimensional model generally has less
uncertainty than the one- -or two-dimensional
models, but, this is not always the case. For example,
when modeling the migration of contaminants in
ground water through a lO-foot thick aquifer, a
two-dimensional model will produce more certain
results than the blind application of a three-
dimensional model. It is not just a trade-off between
difficulty of the model and quality of the output, but a
matching situation as well.
4.3 Scenario Uncertainty
The analyst needs to be aware of uncertainties that
result from using conservative assumptions when
data are lacking. While it is traditional in exposure
assessment to make conservative assumptions in the
absence of data, such assumptions must be
reasonable and the assessment results must be
interpreted with caution. Use of reasonably
conservative assumptions at each step may produce
cumulative assessment results that are overly
conservative and thus unreasonable.
In addition, conceptual errors may result in the use of
assumptions that affect the selection of the modeling
technique applied to the exposure assessment. For
example, using a three-dimensional model in
situations where the aquifer thickness is not "large"
in relation to the areal extent of contamination would
not be appropriate. Thus, the concepts upon which
the exposure scenario is based must be carefully
considered to make sure that they adequately reflect
the situation under evaluation.
Quantitative descriptions of scenario uncertainty are
often impractical, and qualitative descriptions of the
level of uncertainty are more common for the young,
and developing, field of exposure assessment. Any
exposure prediction has cases of overstatement and
understatement of risk. Where possible, the
understatements and overstatements of risk are
minimized. Where this is not possible, the analyst
attempts to balance them so as to produce a
prediction that is most realistic.
4.4 Approaches for Dealing with
Uncertainty
4.4. J Sensitivity Appraisals
Variation in the values of input parameters causes
variation in the values of the output parameters. The
ratio of the input parameter variation to the output
parameter variation will be different for parameters in
different parts of the equation. Sensitivity appraisals
involve assessing which parameters have the highest
ratios and which have the lowest. The accuracy of
parameters that have the largest effect on the
accuracy of the output parameters should be high,
while parameters that have only a small effect on the
accuracy of the output parameters can be estimated
98
-------
or determined by less accurate and less costly
methods.
Sensitivity appraisals can be quantitative or
qualitative. A quantitative sensitivity appraisal involves
plotting the output parameter as a function of variation
of a single input parameter, while holding all of the
other input parameters constant. As one can imagine,
there may be a different functional relationship
between the output parameter and the varying input
parameter for each combination of fixed input
parameters. For complex models the approach can
become overwhelming. Typically, the analyst will be
able to interpret the equation and set up the fixed
input variables so as to minimize the number of
functional relationships produced. However, it may
still be burdensome, and it may produce results that
are more precise than necessary.
In the environmental modeling field, the qualitative
approach has strong advantages over a quantitative
approach. The qualitative approach involves
inspecting the model's equations, and ascertaining
which input variables are the most sensitive. This is
usually done by visual inspection, with an
understanding of the mathematical relationships in the
equation. For example, if one input parameter
multiplies all the other terms, the analyst can expect
the input parameter to have a sensitivity ratio of one.
If the input parameter is the exponent of the other
terms, the analyst can expect this parameter to have
a very high sensitivity ratio. If the input parameter is
part of a separate term that is added to the rest of the
equation, and it is multiplied by a constant of low
value, the input parameter can be assumed to have a
low sensitivity ratio. A qualitative appraisal is usually
the most efficient technique for determining the input
parameter accuracy needs.
4.4.2 Monte-Carlo Simulations
The Monte-Carlo technique involves running a
model a large number of times with varying input
parameters. The values for the input parameters are
chosen from the parameter distributions, with its
relative frequency of a particular value being used
being equal to the relative frequency in the parameter
distribution. This is based on the assumption that the
input variables vary independently from each other.
This technique generates an output parameter
distribution, which provides a mode and a statement
of the uncertainty associated with the prediction.
One difficulty with this technique is the assumption of
independent variation. The input variables are chosen
as if there were no relationship among them. If the
variables are truly independent, the results are
accurate. Typically, however, the variables are related
to each other and are, thus, dependent variables. For
example, if the two input variables are hydraulic
conductivity and hydraulic gradient, the analyst could
assume that they are either independent or
dependent. The analyst could assume independency
because the two variables represent different factors
that do not have a direct functional relationship
between them. But, if the analyst looked at enough
sets of data, the sites with high conductivity would
have more gradients that are flat; conversely, the
sites with low conductivity would have more gradients
that are steep. Thus, the two variables would exhibit
covariation and cannot be considered strictly
independent. This weakens the validity of using the
Monte-Carlo approach.
While it is possible to use input parameter
distributions to generate model output distributions
using Monte-Carlo simulations, it is usually not
possible to get the input parameter distributions. The
input parameter distribution shows the variation of
parameter values. It must be based on a large
number of observations (actual measurements). The
environmental field is young and growing. As such,
most sampling (to date) falls short of providing the
mass of data necessary to generate an input
parameter distribution. Faced with this dilemma, some
analysts have fallen back on assuming such
distribution. Since they do not have a way to gauge
the distribution, a uniform distribution from the lowest
to highest possible value is assumed. This distribution
states that there is an even probability that the value
could be any value between the lowest and the
highest value of the range.
Assuming an input parameter distribution does not
help to reduce uncertainty, however, as the certainty
of the output is then a function of the assumed
certainty of the input parameter. For example, if you
assume that the input parameters are very precise,
then the certainty of the output is high. Conversely, if
you assume the parameters may have an equal
probability to be any value across the range of
possible values, the certainty of the output will be low.
Using a Monte-Carlo approach with assumed input
parameter distributions that are uniform only indicates
how accurate the model is at predicting the output
parameter when you have no idea what the input
parameters are, since models predict output based on
the relationship to the input parameters. Thus, using
the Monte-Carlo technique to assess the certainty of
a model's predictions cannot be done with assumed
input parameter distributions.
4.4.3 Using Monitoring Data to Calibrate fhe
Model
One of the best ways to reduce the uncertainty of the
predicted parameter is to use monitoring data to
calibrate the model. If you have measured
contaminant concentrations that are comparable to
modeled contaminant concentrations, the analyst can
correct for over- or under- predictions. If for
example, the measured values are always 90% of the
predicted values, the analyst can multiply all of the
output values by 90%.
99
-------
The difficulty of this technique is that the values must
be comparable. In many cases the model is being
used to predict future events. Current contaminant
concentrations can be determined more accurately by
monitoring, thus the need for modeling is reduced.
In air and surface water modeling the difference
between current and future events is much smaller
than for ground water modeling. Air and surface water
move more quickly than does ground water. Hence,
calibration is a more useful technique in air and
surface water modeling than in ground water
modeling.
If the ground water model predicts a certain
contaminant concentration 1 mile from the source
after 20 years, and monitoring shows no
contamination at 1 mile from the source, this cannot
be used to calibrate the model. The plume may not
have reached the point 1 mile away, as of yet. In 20
years, monitoring may very well show the same
contaminant concentration that was predicted by the
model. Care should be taken to ensure that the
monitoring data used to calibrate the model are
comparable in time and space.
4.5 Level of Uncertainty Appropriate for
Exposure Modeling
There is no one level of certainty that is appropriate
for all situations. Each program has different needs,
and various parts of a program have diverse needs. A
screening level study has less need for accuracy than
a court case that will require a substantial sum of
money from a PRP. The level of defensibility required
will vary from one situation to another
EPA program offices have developed a multi-tiered
approach. A desk top model may be sufficient for a
first-tier analysis, an analytical model may be
sufficient for a second-tier analysis, and a numerical
model may be required for a third-tier analysis. For
example, the method of screening sites for inclusion
on the National Priorities List should be less rigorous
than the method of supporting a decision on various
site clean-up options. Data requirements will also
vary.
Although it would be nice to have maximum accuracy
in all cases, it would also imply maximum difficulty in
all cases. Clearly, a balance must be found between
difficulty and accuracy of the prediction.
4.6 Risk Communication
Once the analyst has completed the modeling task,
the results of the task must be communicated to the
analyst's supervisor. This information should include
the predictions of exposure over time, and it should
include some communication regarding the level of
uncertainty associated with the prediction. The level
of uncertainty can be expressed in a quantitative or
qualitative form. Further guidance on risk
communication can be found in USEPA (1987e).
A quantitative appraisal of the uncertainty is the most
preferable way to express the uncertainty. A
quantitative presentation may be an output parameter
distribution which tells the most probable value
(mode) and the relative probability that the value is
larger or smaller than the mode. Or, the presentation
may consist of the predicted value and a standard
deviation. The standard deviation provides the level of
precision or uncertainty. Another approach involves
providing the predicted value and the 95% confidence
limits. The 95% confidence limits express that 95%
of the possible values of the parameter will be
between the upper and lower confidence limits, The
main catch to precise numerical expression of the
uncertainty is the lack of sufficient data upon which to
base the quantitative expression of the uncertainty. In
the future, it may be possible to use this precise
approach.
A qualitative appraisal of the uncertainty is the most
viable way to express the level of uncertainty. A
qualitative presentation will describe the significant
factors that determine the level of uncertainty. The
quality of the prediction is a function of the quality of
the inputs to the prediction. Major inputs that affect
quality are: data precision, model sophistication, and
defensibility of the scenario.
Expressing the quality of the data would entail
describing the sources of the data. For example: Did
the data come from literature values or were the data
taken from actual site measurements? Were the data
measured by the best available techniques or were
they sampled by another technique? Were replicate
samples taken? Was the sampling protocol sufficient
to obtain representative samples? Are the costs of
the sampling program appropriate for the use of the
results, or could more expensive data gathering
techniques be used?
Expressing the quality of the model used would entail
a description of the type of model. For example: Is
the model a desk-top calculation, an analytical
model, or a numerical model? Has the model been in
use for some time or is it new? Is the model a
standard model used by the agency or is it new to the
agency? Have other people used the model? Does
the model address all of the important facets of the
situation, or does it neglect some potentially important
factors? Has the model been used in court cases
before? How good is the model relative to other
possible models? Is it the best available model at this
point in time? Is the model the most defensible model
available? Were monitoring data used to calibrate the
model predictions? How comparable were the
monitoring data to the model predictions?
100
-------
Expressing the quality of the scenario is more difficult.
Reasonableness of the scenario is important. Use of
similar scenarios by the agency in the past is useful
information. Questions to ask would include: Was the
scenario used in court cases, for rulemaking activity
that has been published in the Federal Register,
and/or did it receive public comment? Was the public
comment favorable or did it bring out potential
difficulties? Does the scenario neglect certain
exposure routes that have been neglected by the
agency in the past?
The important aspect to consider is how good the
prediction is, not how imperfect the model is.
Modeling is a young field that is rapidly growing.
Uncertainties are minimized but never eliminated.
Modeling produces state-of-the-art estimates, and
nothing more.
101
-------
Preceeding Page Blank
Chapter 5
References
Ambrose, R.B. and Vandergrift, S.B. 1986. SARAH, a
surface water assessment model for back
calculating reductions in abiotic hazardous waste.
Athens, GA: U.S. Environmental Protection
Agency. EPA-600/3-86/058.
Ambrose, R.B., Vandergrift, S.B. and Wool, T.A.
1986. WASPS, A hydrodynamic and water quality
model-model theory, a user's manual, a pro-
grammer's guide. Athens, GA: U.S. Environmental
Protection Agency. EPA-600/3-86-034.
Ambrose, R.B., Wool, T.A., Connolly, J.P. and
Schantz, R. 1987. WASP4, a hydrodynamic and
water quality model-model theory, a user's
manual, a programmer's guide. Athens, GA: U.S.
Environmental Protection Agency.
Bear, J. and Verruijt, A. 1987. Modeling groundwater
flow and pollution. Dordrecht, The Netherlands: D.
Reidel Publishing Company.
Beljin, M.S. 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.
Bonazountas, M. and 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.
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. Contract No. 68-
01-5146.
Bouwer, H. 1978. Groundwater hydrology. McGraw-
Hill Pub. Co., New York, NY.
Bowers, J.F., et al. 1979. Industrial source complex
(ISC) dispersion model user's guide, volumes I and
II. Washington, DC: U.S. Environmental Protection
Agency. EPA 45014-79-030. As reviewed in:
Bonazountas M, Fiksel J., et al. 1982. En-
vironmental mathematical pollutant fate modeling
handbook/catalogue (Draft). Washington, DC: U.S.
Environmental Protection Agency, Office of Policy
and Resource Management. Contract No. 68-
01-5146.
Brown, D.S. and Allison, J.D. 1987. MINTEQA1,
equilibrium metals speciation model: a user's
manual. Athens, GA: U.S. Environmental
Protection Agency.
Brown, S.M., Boutwell, S.H., Roberts, B.R. 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. Work
Assignment No. 5.
Burns, L.A., Cline, D.M., Lassiter, R.R. 1982.
Exposure analysis modeling system (EXAMS) user
manual and system documentation. Athens, GA:
Environmental Research Laboratory, Office of
Research and Development, U.S. Environmental
Protection Agency. EPA-600/3-82-023. As
reviewed in: Versar Inc. 1983. Methodology for
assessing exposures to chemical substances via
the ingestion of drinking water. Washington, DC:
U.S. Environmental Protection Agency. Contract
No. 68-01-6438.
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. Information Clearing House.
PM-211A.
Busse, A.D., Zimmerman, J.R. 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. Contract No. 68-
01-5146.
103
-------
Callahan, M., Slimak, M., Gabel, N., et al. 1979.
Water-related environmental fate of 129 priority
pollutants. Washington, DC: U.S. Environmental
Protection Agency. EPA-440/4-79-029a,b.
Carsel, RF., Smith, C.N., Mulkey, L.A., Dean, J.D.
and Jowise, P. 1984. User's manual for the
pesticide root zone model (PRZM): release 1.
Athens, Georgia: Environmental Research
Laboratory, U.S. Environmental Protection Agency,
EPA-600/3-84- 109.
Chico, T. and Catalano, J.A. 1986. Addendum to the
user's guide for MPTER. Research Triangle Park,
NC: U.S. Environmental Protection Agency.
Christiansen, J.H. 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: U.S. EPA. 1982. Environmental
Modeling Catalogue. Washington, DC: U.S.
Environmental Protection Agency. Information
Clearing House. PM-211 A.
Clapp, R.B. and Hornberger, G.M. 1978. Empirical
equations for some soil hydraulic properties. Water
Resources Research. 14:601-604.
Codell, R.B., Key, K.T. and Whelan, G. 1982. A
collection of mathematical models for dispersion in
surface water and ground water. Washington DC:
U.S. Nuclear Regulatory Commission.
NUREGICR-0868.
Cohen, A.C. 1950. Censored samples from normal
distributions. Ann. Math. Stat. Volume 21.
Cowherd, C., Muleski, G.E., Englehart, P.J. and
Gillette, D.A. 1985. Rapid assessment of exposure
to particulate emissions from surface contamination
sites. Kansas City, MO. Midwest Research
Institute. PB85-192219.
Craig, D.G., Turelle, J.W. 1964. Guide for wind
erosion control on cropland in the Great Plains
states. U.S. Department of Agriculture. Soil
Conservation Service. Washington, DC.
Cupitt, L.T. 1980. Fate of toxic and hazardous
materials in the air environment. Research Triangle
Park, NC: U.S. Environmental Protection Agency.
EPA-600/3-80-084.
Das, B.M. 1983. Advanced Soil Mechanics. New
York, McGraw-Hill Book Company.
Davis, L.A. and 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.
Dawson, G.W., English, C.J. and Petty, S.E. 1980.
Physical chemical properties of hazardous waste
constituents. Athens, GA: U.S. Environmental
Protection Agency, Environmental Research
Laboratory.
Delos, C.G., Richardson, W.L., DePinto, J.V., et al.
1984. Technical guidance manual for performing
wasteload allocations, book II: streams and rivers.
U.S. Environmental Protection Agency. Office of
Water Regulations and Standards. Water Quality
Analysis Branch. Washington, DC. (Draft Final.)
Domenico, P.A. and Palciauskas, V.V. 1982.
Alternative boundaries in solid waste management.
Ground water. Vol. 20, No. 3.
Donigian, A.S., Imhoff, J.C., Bicknell, B.R. and Kittle,
J.L. 1984. Application guide for hydrological
simulation program - Fortran (HSPF). U.S.
Environmental Protection Agency. EPA-600/3-
84-065.
Eisenhart, C. 1968. Expression of the uncertainties of
final results. Science. Volume 160. June 1968.
Enfield, C.G. and Bengtsson, G. Not dated. Macro
molecular transport of hydrophobic contaminants in
aqueous environments. EPA/ORD, Ada, OK.
Enfield, C.G., Carsel, R.F., Cohen, S.Z., Phan, T. and
Walters, D.M. 1982. Approximating pollutant
transport to ground water. Ground Water 20(6)
711-722.
Enfield, C.G. 1984. Chemical transport facilitated by
multiphase flow systems. Robert S. Kerr
Environmental Research Laboratory. U.S.
Environmental Protection Agency, Ada, OK.
Presented at seminar on Degradation, Retention
and Dispersion of Pollutants in Ground Water,
Copenhagen, Denmark, September 12-14.
Farino, W., Spawn, P., Jasinski, M. and Murphy, B.
1983. Evaluation and selection of models for
estimating air emissions from hazardous waste
treatment, storage and disposal facilities. Revised
Draft Final Report. GCA Corporation.
GCA/Tech nology Division. Bedford,
Massachusetts, prepared for U.S. Environmental
Protection Agency, Office of Solid Waste. Land
Disposal Branch. Contract No. 68-02-3168.
Farmer, W.J., Yang, M.S., Letey, J., Spencer, W.F.
and Roulier, M.H. 1978. Land disposal of
hexachlorobenzene waste: controlling vapor
movement in soils. San Antonio, TX. 4th Annual
Symposium on Land Disposal.
104
-------
Felmy, A.R., Girvin, D.C., Jenne, E.A. 1984.
MINTEQ — a computer program for calculating
aqueous geochemical equilibria. Athens, GA: U.S.
Environmental Protection Agency. EPA-600/3-
84-032.
Finley, N.C. and Reeves, M. 1968. SWIFT self-
teaching curriculum. Washington, DC: Nuclear
Regulatory Commission. NUREG/CR-1968,
SAND 81-0410. As reviewed in: Lo TYR, Scott
BH, Benjamin RR. 1983. Remedial action
assessment models for hazardous waste sites.
Review draft. Athens, GA: U.S. Environmental
Protection Agency. Contract No. 68-03-3116.
Fisher, H.B., List, E.F., Koh, R.C.Y., Imberger, J.,
Brooks, N.H. 1979. Mixing in inland and coastal
waters. New York, NY: Academic Press.
Freeze, R.A., Cherry, J.A. 1979. Groundwater.
Englewood Cliffs, NJ: Prentice-Hall, Inc.
GSC. 1982. Graphical exposure modeling system
(GEMS) user's guide. General Software
Corporation. Washington, DC: Office of Pesticides
and Toxic Substances, U.S. Environmental
Protection Agency. Contract No. 68-01-6618.
Gelhar, L.W., Mantoglou, A., Welty, C. and Rehfeldt,
K.R. 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.
Geotrans. 1986. Analytical models for evaluating
leachate migration in groundwater systems.
Washington, DC: U.S. Environmental Protection
Agency, Office of Solid Waste.
Gillett, J.W. 1980. Terrestrial microcosm technology
in assessing fate, transport and effects of toxic
chemicals. In Dynamics, Exposure and Hazard
Assessment of Toxic Chemicals, R. Hauge, editor.
Ann Arbor Science, Ann Arbor, Ml. As quoted in
Trabalka and Garten (1982).
Gupta, S.K., Cole, C.R., Kincaid, C.T., and Monti,
A.M. 1987. Coupled fluid, energy, and solute
transport (CFEST) model: formulation and user's
manual. Columbus, Ohio: Office of Nuclear Waste
Isolation, Battelle Memorial Institute.
Gureghian A.B. 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.
Haith, D.A. 1980. A mathematical model for
estimating pesticide losses in runoff. Journal of
Environmental Quality. 9(3):428-433.
Hanna, S.R., Hosker, R.P., Jr. 1980. Atmospheric
removal processes for toxic chemicals. Silver
Spring, MD: National Oceanic and Atmospheric
Administration. Technical Memorandum ERL
ARL-102.
Haynes, W.A. 1966. Guide for wind erosion control in
the northeastern United States. U.S. Department of
Agriculture. Soil Conservation Service. Washington,
DC.
Hendry, D.G., Kenley, R.A. 1979. Atmospheric
reaction products of organic compounds.
Washington, DC: U.S. Environmental Protection
Agency. EPA-56011 2-79-001.
Henrion, M. and Morgan, M.G. 1984. A computer aid
for risk and other policy analysis. Pittsburgh, PA:
Carnegie-Mellon University, Department of
Engineering and Public Policy.
Hilsmeier, W.F., Gifford, F.A. 1962. Graphs for
estimating atmospheric diffusion. Oak Ridge, TN:
Atomic Energy Commission. ORO-545. As
reviewed in: Turner DB. 1970. Workbook of
atmospheric dispersion estimates. Research
Triangle Park, NC: U.S. Environmental Protection
Agency, Office of Air Programs. AP-26.
Hoffman, F.O., Miller, C.W., Little, C.A. 1984.
Uncertainties associated with predictions derived
from models and parameters. In: Models and
parameters for environmental radiological
assessments. C.W. Miller, Ed. Oak Ridge National
Laboratory, Health and Safety Research Division.
Hwang, S.T. 1982. Toxic emissions from land
disposal facilities. In Environmental Progress. Vol.
1, No. 1.
Huyakorn, P.S., Ungs, M. J., Mulkey, L.A. and
Sudicky, S.A. 1987. A three-dimensional
analytical method for predicting leachate migration.
Ground water, Vol. 25, No. 5.
HydroQual, Inc. 1982. Application guide for CMA -
HydroQual chemical fate models. Prepared for:
Chemical Manufacturers Association, Washington,
DC. As reviewed in: Versar Inc. 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.
Intera. 1983. SWENT: A three-dimensional finite-
difference code for the simulation of fluid, energy,
and "solute radionuclide transport. Inter-a, Inc.,
105
-------
Columbus, OH: Office of Nuclear Waste Isolation,
Battelle Memorial Institute. BMI/ONWI-457.
Javandel, I., Doughty, C. and Tsang, C.F. 1984.
Groundwater transport: handbook of mathematical
models. AGU Water Resources Monograph.
10:240.
Jaw-Kwei, A.M. and Luthy, R.G. not dated. Aromatic
compound solubility in solvent/water mixtures.
Pittsburgh, PA: Carnegie-Mellon University.
Jenson, M.E., 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.
Johnanson, R.C., Imhoff, G.C., Davis, H.H. 1984.
Hydrocomp Inc. Users manual for hydrological
simulation program - FORTRAN (HSPF). Athens,
GA: Office of Research and Development, U.S.
Environmental Protection Agency. EPA-600,9-
80-015. As reviewed in: Versar Inc. 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.
Kenaga, E.E. and Goring, C.A.I. 1978. Relationship
between water solubility, soil sorption, octanol-
water partitioning, and concentration of chemicals
in biota. In: Eaton JG, Parrish PR, Hendricks AC,
eds. Aquatic toxicology. Proc. of third annual
symposium on aquatic toxicology, New Orleans,
LA, 17-18 Oct. ASTM special technical publication
707.
Kenaga, E.E. 1980. Correlation of bioconcentration
factors of chemicals in aquatic and terrestrial
organisms with their physical and chemical
properties. Environ. Sci. Technol. 14(5): 553-556.
Kent, D.C., Pettyjohn, W.A. and Prickett, T.A. 1985.
Analytical methods for the prediction of leachate
plume migration. Ground Water Monitoring Review.
Spring 1985. Vol. 5 #2 pp. 46-59.
Kent, D.C., Pettyjohn, W.A., Witz, F. and 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.
Kinzelbach, W. 1986. Groundwater modeling: an
introduction with sample programs in BASIC.
Amsterdam, The Netherlands: Elsevier Science
Publishers.
Kipp, K.L. 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.
Kleijnen, J.P.C. 1974. Statistical techniques in
simulation. New York. NY: Marcel Dekker, Inc.
Konikow, L.F., and Bredehoeft, J.D. 1978. Computer
model of two-dimensional transport and
dispersion in ground water. Washington, DC: U.S.
Geological Survey. Techniques of Water Resource
Investigation, Book 7, Chapter 2. 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.
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.
Liu, H. 1977. Predicting dispersion coefficient of
streams. J. Environmental Engineering Division.
Proceedings of the American Society of Civil
Engineers. Vol. 103.
Lyman, W.J., Reehl, W.F., Rosenblatt, D.H. 1982.
Handbook of chemical property estimation
methods. New York. McGraw-Hill.
Mabey, W.R., Smith, J.H., Podoll, R.T., et al. 1982.
Aquatic fate process data for organic priority
pollutants. Final draft report. Washington, DC:
Office of Water Regulations and Standards, U.S.
Environmental Protection Agency.
Mackay, D., Leinonen, P.J. 1975. Rate of evaporation
of low-solubility contaminants from water bodies
to atmosphere. In Environ. Sci. Technol. Vol.
9(13).
Martinez, M.J. 1985. A finite element computer
program for simulating radionuclide transport
through porous media. Albuquerque, NM: Sandia
National Laboratory. SAND 84-0747.
Mercer, J.W., Silka, L.R. and Faust, C.R. 1985.
Modeling ground-water flow at Love Canal.
Journal of Environmental Engineering, August.
Miller, F.J., Gardner, D.E., Graham, J.A., Lee, R.E.
Jr., Wilson, W.E., Bachmann, J.D. 1979. Size
consideration for establishing a standard for
106
-------
inhalable particles. In Journal of the Air Pollution
Control Association.
Mills, W.B., Dean, J.D., Porcella, D.B., 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. Environmental Research Laboratory.
Office of Research and Development. EPA-
600/6-821004 a,b,c.
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.
Morgan, M.G., Henrion, M., Nair, I., Mullin, T. and
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 Enginering and Public Policy.
Neely, W.B. 1982. The definition and use of mixing
zones. Environ. Sci. Technol. 16(9):520A-521A.
Nkedi-Kizza, P., Rao, P.S.C. and Hornsby, A.G.
1985. Influence of Organic cosolvents on sorption
of hydrophobic organic chemicals by soils.
Environmental Science and Technology Vol. 19,
No. 10, p. 975-979
Onishi, Y., Whelan, G. and Skaggs, R.L. 1982.
Development of a multimedia radionuclide
exposure assessment methodology for low-level
waste management. Athens, GA: Office of
Research and Development, U.S. Environmental
Protection Agency. As reviewed in: Versar Inc.
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
Onishi, Y. and Wise, S.E. 1982a. Mathematical model,
SERATRA, for sediment-contaminant transport in
rivers and its application to pesticide transport in
Four Mile and Wolf Creeks in Iowa. Athens, GA:
Office of Research and Development, U.S.
Environmental Protection Agency. EPA-600-
3182-045. As reviewed in: Versar Inc. 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.
Onishi, Y. and Wise, S.E. 1982b. User's manual for
the instream sediment-contaminant transport
model SERATRA. Athens, GA: Office of Research
and Development, U.S. Environmental Protection
Agency. EPA-60013-82-055. As reviewed in:
Versar Inc. 1983. Methodology for assessing
exposures to chemical substances via the ingestion
of drinking water. Washington, DC: U.S.
Environmental Protection Agency. Contract No.
68-01-6438.
Onishi, Y. 1981. Sediment-contaminant transport
model. Journal of the Hydraulics Division, ASCE.
107(HY9):1089 - 1107. Proc. Paper 16505. As
reviewed in: Versar Inc. 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.
Pasquill, F. 1961. The estimation of the dispersion of
windborne materials. Meteorol. Mag. 90, 1063,
33-49. As reviewed in: Turner DB. 1970.
Workbook of atmospheric dispersion estimates.
Research Triangle Park, NC: U.S. Environmental
Protection Agency, Office of Air Programs. AP-
26.
Perry, R. and Chilton, C. 1973. Chemical engineers
handbook. 5th edition. New York: McGraw-Hill.
Pierce, D. and Turner, D.B. 1980. User's guide for
MPTER. Research Triangle Park, NC: U.S.
Environmental Protection Agency. EPA-600:8-
80-016. NTIS No. PB-80-176361,
Prickett, T.A., Naymik, T.G. and Lonnquist. C.G.
1981. A "random-walk" solute transport model for
selected groundwater quality evaluations.
Champaign, IL: Illinois Department of Energy and
Natural Resources. ISWS/BUL-65181. 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.
Rao, P.S.C., Hornsby, A.G., Kilcrease, P.P., Nkedi-
Kizza, P. 1985. Sorption and transport of
hydrophobic organic chemicals in aqueous and
mixed solvent systems: model development and
preliminary evaluation. Journal of Environmental
Quality, Vol. 14, No. 3, July-September. Rawls
WJ, Brakensiek DL, Saxton KE. 1982. Estimation
of Soil Water Properties. Transactions of the
American Society of Agricultural Engineers.
25(5):1316-1320 and 1328.
Rawls, W.J. 1986. Computer printouts from the soils
data base, dated August 28, 1986, from W.J.
Rawls, Beltsville Agricultural Research Center,
Beltsville, Maryland.
Reeves, M. and Cranwell, R.M. 1981. User's manual
for the Sandia Waste-Isolation Flow Transport
model (SWIFT). Washington, DC: Nuclear
107
-------
Regulatory Commission. NUREG/CR-2324,
SAND81-2516. As reviewed in: Lo TYR, Scott
BH, Benjamin RR. 1983. Remedial action
assessment models for hazardous waste sites.
Review draft. Athens, GA: U.S. Environmental
Protection Agency. Contract No. 68-03-3116.
Reeves, M., Ward, D.S., Johns, N.J. and Cranwell,
R.M. 1986. The Sandia waste-isolation flow and
transport model for fractured media; release 4.84:
theory and implementation. Washington, D.C: U.S.
Nuclear Regulatory Commission. NUREGICR-
3328.
Rifai, H.S., Bedient, P.B., Borden, R.C. and
Haasbeek, J.F. 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.
Rubinstein, R.Y. 1981. Simulation and the Monte
Carlo method. New York, NY: Wiley Science
Publishers.
Runchal, A., Sagar, B., Baca. R.G. and Kline, N.W.
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.
Russell, R.S., Bartlett, B.O. and Bruce, R.S. 1971.
The significance of long-lived nuclides after a
nuclear war. In D.W. Bensen and A.M. Sparrow,
eds. Survival of food crops and livestock in the
event of nuclear war. Oak Ridge, TN: US Atomic
Energy Commission. As quoted in Trabalka and
Garten (1982).
Salame, M. (no date) Permeability-structure
relationships of high polymers. Obtained by private
communication. Monsanto Co., Bloomfield, CT.
Salame, M. 1961. The prediction of liquid permeation
in polyethylene and related polymers. SPE Trans.
1(4): 153.
Salame, M. 1973. Transport properties of nitrile
polymers. J. Polymer Sci. 41:1-15.
Salame, M. 1985. Private communication. Monsanto
Co., Bloomfield, CT.
Sanford, W.E. and Konikow, L.F. 1985. A two-
constituent solute transport model for groundwater
having variable density. Reston, VA: U.S.
Geological Survey. WRI 85-4279.
Sax, N.I. 1984. Dangerous properties of industrial
materials, 6th edition. New York, NY: Van
Nostrand Reinhold.
Schnoor, et al. 1987. Processes, coefficients, and
models for simulating toxic organics and heavy
metals in surface waters. Athens, GA: U.S.
Environmental Protection Agency. EPA-600/3-
87/015.
Schwab, G.O., Frevert, R.K., Edminster, T.W. and
Barnes, K.K. 1966. Soil and water conservation
engineering. 2nd edition. New York: John Wiley
and Sons.
Seely, D., Turina, P., Pangaro, N., et al. 1983.
Development of protocols for ambient air sampling
and monitoring at hazardous waste facilities:
methods summary report. Draft Report. GCA
Corporation. GCA/Technology Division. New
Bedford, MA. Prepared for U.S. Environmental
Protection Agency, Office of Solid Waste, Land
Disposal Branch. Contract No. 68-02-3168.
Sehmel, G.A. 1980. Particle resuspension: a review.
In Environment International. Vol. 4. Pergamon
Press, Ltd.
Shen, T. 1981. Estimating hazardous air emissions
from disposal sites. Pollution Engineering 13(8):
31-34.
Skidmore, E.L., Woodruff, N.P. 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.
Skinner. 1984. Banning wastes from land disposal.
Paper presented by EPA Office of Solid Waste and
Emergency Response at the First Public Briefing
on the 1984 Amendments to the Resource
Conservation and Recovery Act. December 11,
1984, Washington, DC.
Smith, W.J., Whicker, F.W., Meyer, H.R. 1982.
Review and categorization of saltation, suspension
and resuspension models. Nuclear Safety. 23(6).
Steingiser, S., Nemphos, S.P., Salame, M. 1978.
Barrier polymers. In: Kirk-Othmer encyclopedia
of chemical technology, 3rd edition. John Wiley
and Sons, New York, NY.
Texas Air Control Board. 1980. User's guide to the
Texas climatological Model. Austin, TX: Texas Air
Control Board. As reviewed in: U.S. EPA. 1982.
Environmental Modeling Catalogue. Washington,
DC: U.S. Environmental Protection Agency.
Information Clearing House. PM-211A.
108
-------
Thibodeaux, L.J., Hwang, ST. 1982. Landfarming of
petroleum wastes - modeling the air emission
problem. In Environmental Progress. Vol. 1, No. 1.
Thibodeaux, L.J. 1981. Estimating the air emissions
of chemicals from hazardous waste landfills. In
Journal of Hazardous Materials. Vol. 4.
Trabelka, J.R., Garten, C.T., Jr. 1982. Development
of predictive models for xenobiotic bioaccumulation
in terrestrial ecosystems. Oak Ridge, TN: Oak
Ridge National Laboratory, Environmental Sciences
Division, (NTIS DE83 003171). ORNL-5869.
Turner, D.B. 1970. Workbook of atmospheric
dispersion estimates. Research Triangle Park, NC:
U.S. Environmental Protection Agency, Office of
Air Programs. AP-26.
Tracy, J.V. 1982. User's guide and documentation for
adsorption and decay modifications to the USGS
solute transport model. Washington, D.C.: U.S.
Nuclear Regulatory Commission. NUREGiCR-
2502.
Trask, P.O., Patnode, H.W. 1942. Source beds of
petroleum. Tulsa, OK: The American Association of
Petroelum Geologists.
Travis, B. 1984. TRACR3D: A model of flow and
transport in porous/ fractured media. Los Alamos,
NM: Los Alamos National Laboratory. LA-0667-
MS.
Turner, D.B., and Novak, J.H. 1978. User's guide for
RAM. Publication No. EPA-60018-78-016 Vols
a and b. Research Triangle Park, NC: U.S.
Environmental Protection Agency.
USCG. 1974. Hazardous chemical data. Washington
DC: United States Coast Guard CG-446-2.
USDA. 1974. Department of Agriculture. Universal
soil loss equation. Agronomy technical note no.
32. Portland, Oregon. U.S. Soil Conservation
Service. West Technical Service Center.
USDC. 1961. Rainfall frequency atlas of the United
States. Washington, DC: U.S. Department of
Commerce. Weather Bureau. Technical Paper
Number 40.
USDC. 1968. Climatic atlas of the United States.
Washington, DC: U.S. Department of Commerce.
Environmental Sciences Services Administration.
Environmental Data Service.
USEPA. 1977a. Guidelines for air quality maintenance
planning and analyses, Vol. 10 (Revised).
Procedures for evaluating air quality impact of new
stationary sources. Research Triangle Park, NC:
Office of Air Quality Planning and Standards, U.S.
Environmental Protection Agency.
USEPA. 1977b. User's manual for single source
(CRSTER) model. Research Triangle Park, NC:
Office of Air Quality Planning and Standards, U.S.
Environmental Protection Agency. EPA Publication
No. EPA-45012-77-013.
USEPA. 1979. Environmental modeling catalogue.
Washington, DC: U.S. Environmental Protection
Agency. Contract No. 68-01-4723.
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.
USEPA. 1980b. Land disposal of hexachlorobenzene
wastes: controlling vapor movement in soil. U.S.
Environmental Protection Agency. Municipal
Environmental Research Laboratory. Office of
Research and Development. Cincinnati, OH.
EPA-60012-80-119.
USEPA. 1982a. Environmental modeling catalogue.
Washington, DC: U.S. Environmental Protection
Agency. Information Clearing House. PM-211 A.
USEPA. 1982b. Establishment of guidelines for
modeling groundwater contamination from
hazardous waste facilities. Preliminary groundwater
modeling profile (discussion draft). Washington,
D.C.: 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.
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-60014-83-
007.
USEPA. 1983c. Technical assistance document for
sampling and analysis of toxic organic compounds
in ambient air. Research Triangle Park, NC: U.S.
Environmental Protection Agency, Environmental
Monitoring Systems Laboratory. EPA-60014-83-
027.
109
-------
USEPA. 1985a. Guidance on remedial investigations
under CERCLA. Washington, D.C.: U.S.
Environmental Protection Agency, Office of
Emergency and Remedial Response.
USEPA. 1985b. Guidance on feasibility studies under
CERCLA. Washington, D.C.: U.S. Environmental
Protection Agency, Office of Emergency and
Remedial Response.
USEPA. 1985c. Superfund public health evaluation
manual. Draft. Prepared by ICF, Inc. for the Policy
Analysis Staff, Office of Emergency and Remedial
Response. Washington, D.C.: U.S. Environmental
Protection Agency. October 1, 1985.
USEPA. 1985d. Methods for assessing exposure to
chemical substances. Versar Inc. Washington,
D.C.: U.S. Environmental Protection Agency, Office
of Toxic Substances. EPA 56015-85-001.
USEPA. 1985e. Methods for assessing exposure to
chemical substances in the ambient environment.
Versar Inc. Washington, D.C.: U.S. Environmental
Protection Agency, Office of Toxic Substances.
EPA 56015-85-002.
USEPA. 1985f. Methods for assessing exposure from
disposal of chemical substances. Versar Inc.
Washington, D.C.: U.S. Environmental Protection
Agency, Office of Toxic Substances. EPA 560/5-
85-003.
USEPA. 1985g. Methods for enumerating and
characterizing populations exposed to chemical
substances. Versar Inc. Washington, D.C.: U.S.
Environmental Protection Agency, Off ice of Toxic
Substances. EPA 560/5-85-004.
USEPA. 1985h. Methodology for assessing exposures
to chemical substances via the ingestion of
drinking water. Versar Inc. Washington, D.C.: U.S.
Environmental Protection Agency, Off ice of Toxic
Substances. EPA 560/5-85-005.
USEPA. 19851. Methods for assessing environmental
pathways of food contamination. Versar Inc.
Washington, D.C.: U.S. Environmental Protection
Agency, Office of Toxic Substances. EPA 560/5-
85-008.
USEPA. 1985J. Modeling remedial actions at
uncontrolled hazardous waste sites. Cincinnati OH:
Hazardous Waste Engineering Research
Laboratory, Gffice of Research and Development,
U.S. Environmental Protection Agency.
EPA/540/2-85/001.
USEPA. 1986a. Mathematical model selection criteria
for performing exposure assessments: groundwater
contaminants from hazardous waste facilities.
Washington, D.C.: Office of Health and
Environmental Assessment, U.S. Environmental
Protection Agency.
USEPA. 1986b. Guideline on air quality models.
Research Triangle Park, N.C.: Office of Air Quality
Planning and Standards. U.S. Environmental
Protection Agency. EPA-450/2-78-027R.
USEPA. 1986c. Guidelines for exposure assessment.
Washington, D.C.: Office of Research and
Development, U.S. Environmental Protection
Agency. 51 FR 34042.
USEPA. 1986d. Office of Solid Waste. Hazardous
waste management system; land disposed
restrictions; proposed rule. Federal Register Vol.
51, No. 9. January 14, 1986.
USEPA. 1986e. Development of advisory levels for
polychlorinated biphenyls (PCB) cleanup.
Washington, DC: U.S. Environmental Protection
Agency, Exposure Assessment Group, Office of
Research and Development. (NTIS #PB86-
232774) May 1986. EPA/600/6-86-002.
USEPA. 1987a. Data quality objectives for remedial
response activities. Washington, D.C.: Office of
Emergency and Remedial Response and Office of
Waste Programs Enforcement, Office of Solid
Waste and Remedial Response, U.S.
Environmental Protection Agency.
USEPA. 1987b. Compendium of superfund field
operating methods. Washington, D.C.: Office of
Solid Waste and Emergency Response, U.S.
Environmental Protection Agency. EPA-540/P-
87/001. OSWER Directive 9355.0-14.
USEPA. 1987c. Proposed criteria for selection of
groundwater exposure assessment models.
Washington, D.C.: Office of Health and
Environmental Assessment. U.S. Environmental
Protection Agency. OHEA-E-219.
USEPA. 1987d. Mathematical model selection criteria
for performing exposure assessments: surface
water models. Washington, D.C.: Office of Health
and Environmental Assessment. U.S.
Environmental Protection Agency. OHEA-E-245.
USEPA. 1987e. Risk Communication, Risk
Assessment, Management Communication, Guide
to Selected Sources. USEPA Office of Information
Resource Management, Office of Toxic
Substances, March 1987. EPA-IMSD/87/002.
van der Heijde, Paul K.M., Bachmat, Y., Bredehoeft,
J., Andrews, B., Holtz, D., Sebastian, S. 1985.
Groundwater management: The use of numerical
models. Washington, DC: American Geophysical
110
-------
Union, Water Resources Monograph #5, 2nd
Edition.
van der Heijde, P.K.M. and Srinivasan, P. 1986. A
"random-walk" model for solute transport in
groundwater 'RANDOM WAL/RWH'. Indianapolis,
IN: International Ground Water Modeling Center,
Holcomb Research Institute, Butler University.
van der Heijde, Paul K.M., Beljin, M.S. 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.
van Genuchten, M.T. 1978. Mass transport in
saturated-unsaturated media: one-dimen-
sional solutions. Princeton, NJ: Department of Civil
Engineering. Princeton University. 78/WR-11.
van Genuchten, M.T. and Alves, W.J. 1982.
Analytical solutions of the one-dimentional
convective-dispersive solute transport equation.
Riverside, CA: U.S. Salinity Laboratory, Agricultural
Research Service, U.S. Department of Agriculture.
Technical Bulletin 1661.
Veith, G.D., DeFoe, D.L., Bergstedt, B.V. 1980.
Measuring and estimating the bioconcentration
factor of chemicals in fish. J. Fish Res Board Can.
36: 1040-1048.
Versar. 1983. Theoretical evaluation of sites located
in the zone of saturation. Draft final report. Versar,
Inc. Chicago IL: U.S. Environmental Protection
Agency. Contract No. 68-01-6438.
Verschueren, K. 1984. Handbook of environmental
data on organic chemicals. New York: Van
Nostrand/Reinhold Press.
Voss, C.I. 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.
Walton, W.C. 1984. Handbook of analytical ground
water models. International Ground Water
Modeling Center, Holcomb Research Institute,
Butler University, Indianapolis, Indiana.
Walton, W.C. 1985. Thirty-five BASIC groundwater
programs for desktop microcomputers
'WALTON84-35BASIC'. Indianapolis, IN:
International Ground Water Modeling Center,
Holcomb Research Institute, Butler University.
Williams, J.R. 1975. Sediment-yield prediction with
the universal equation using runoff energy factor.
In Present and prospective technology for
predicting sediment yields and sources. U.S.
Department of Agriculture. ARS-S-40.
Wilson, J.L., Miller, P. J. 1978. Two-dimensional
plume in uniform ground-water flow. Journal of
the Hydraulics Division, ASCE 104(4): 503-514.
Wischmeier, W.H., Smith, D.D. 1978. Predicting
rainfall erosion losses - a guide to conservation
planning. Washington, DC: U.S. Department of
Agriculture. Agriculture Handbook No. 537.
Wischmeier, W.H. 1972. Estimating the cover and
management factor on undisturbed areas. U.S.
Department of Agriculture. Oxford, MS:
Proceedings of the USDA Sediment Yield
Workshop.
Woodburn, K.B., Rao, P.S.C., Fukui, M., Nkedi-
Kizza, P. 1986. Solvophobic approach for
predicting sorption of hydrophobic organic
chemicals on synthetic sorbents and soils. J. of
Contam. Hydrology. 1: 227-241.
Yeh, G.T. and Huff, D.D. 1985. FEMA: A finite
element model of material transport through
aquifers. Oak Ridge, TN: Oak Ridge National
Laboratory, ORNL-6063.
Yeh, G.T. 1981. AT123D. Analytical transient one-,
two-, and three-dimensional simulation of waste
transport in the aquifer system. Oak Ridge, TN:
Oak Ridge National Laboratory, Environmental
Sciences Division Publication No. 1439. ORNL-
Yeh, G.T. 1982. CHNTRN: a chemical transport
model for simulating sediment and chemical
distribution in a stream/river network. Washington,
DC: Office of Pesticides and Toxic Substances,
U.S. Environmental Protection Agency. 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.
Yeh, G.T. 1987. FEMWATER: A finite element model
of water flow through saturated-unsaturated
porous media - first revision. Oak Ridge, TN: Oak
Ridge Natational Laboratory, ORNL-5567/RI.
Yeh, G.T., Ward, D.S. 1981. FEMWASTE: A finite-
element model of waste transport through
saturated-unsaturated porous media. Oak Ridge
National Laboratory, Environmental Services
111
-------
Division: Publication No. 1462, ORNL-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.
112
-------
Appendix A
Analysis of Exposed Human Populations and Exposure Calculation and Integration
Table of Contents
Chapter Page
1 QUANTITATIVE ANALYSIS OF EXPOSED POPULATIONS 114
1.1 Introduction 114
1.2 Exposed Populations Screening 114
1.3 Quantitative Exposed Populations Analysis 116
1.4 Identification and Enumeration of Exposed Human Populations 118
1.4.1 Populations Exposed Through Air 118
1.4.2 Populations Exposed Through Surface Water or Ground Water 119
1.4.3 Populations Exposed Through Food 119
1.4.4 Populations Exposed Through Soil 120
1.5 Population Characterization 120
1.6 Activity Analysis 120
2 EXPOSURE CALCULATION AND INTEGRATION 121
2.1 Inhalation Exposure 122
2.2 Dermal Exposure 123
2.3 Ingestion Exposure 128
2.3.1 Food/Soil 128
2.3.2 Water 128
2.4 Exposure Integration 129
3 APPENDIX A REFERENCES 131
113
-------
Chapter 1
Quantitative Analysis of Exposed Populations
1.1 Introduction
The results of contaminant release and fate
analyses provide the basis for assessing exposed
populations. This assessment compares environ-
mental contamination data with populations data to
determine the likelihood of human contact with
contaminants of concern. This chapter details
methods useful in evaluating the following
components of exposed populations analysis:
1. Identification and enumeration of exposed
populations;
2. Characterization of exposed populations; and
3. Analysis of activities that bring populations
into contact with contaminants.
Each of these components is detailed in the following
subsections.
As with other evaluations, exposed populations
analysis begins with a screening assessment, which
identifies exposure pathways that are incomplete, i.e.,
those situations where contaminants are released and
migrate from a site, but do not contact human
populations and are not likely to do so in the future.
Such situations require no further analysis. At the
same time, exposed populations screening also points
out those exposure pathways that are complete and
that will require quantitative analysis to estimate the
extent of human exposure.
Data needed to quantify potentially exposed
populations are readily available. In essence, all
quantitative exposed populations evaluations can be
considered in-depth analyses. For each population
segment identified in this portion of the exposure
assessment process, exposures are quantified and
integrated as described in Chapter 2 of this Appendix.
1.2 Exposed Populations Screening
Exposed populations screening is primarily qualitative.
This evaluation draws on the results of contaminant
fate analysis (presented in Chapter 3) to determine
the likelihood and extent of human population contact
with contaminants.
Exposed populations screening is guided by the
decision network provided in Figure A-l. The
following numbered paragraphs each refer to
particular numbered boxes in the figure.
1. Human exposure through inhalation should be
evaluated for contaminants that have migrated or may
migrate from the site into air. The assessment should
consider both contaminated dust and volatile
compounds. For screening purposes, comparing
contaminant concentration isopleths with maps of the
local area will identify the potential for such human
population inhalation exposure. The user should
realize, however, that exposure can occur in
recreational areas as well as in residential,
commercial, or industrial areas, and should interpret
local area maps accordingly.
2. In cases where surface waterbodies have been
contaminated by toxics migrating from a site, the
water's potential commercial use as a fish or shellfish
source should be evaluated. If the waters are
commercially fished, fishermen may be exposed
through dermal contact with contaminated water,
although such exposure will generally be
overshadowed by other exposure mechanisms.
3. In cases where recreationally or commercially
caught fish/shellfish are taken from contaminated
waters, persons consuming the catch may be
exposed. For chemicals that tend to bioaccumulate,
consumers may be exposed to contaminant
concentrations in fish/shellfish tissue that are many
times greater than those present in the water column
or sediments. When performing exposed populations
screening, the analyst need only determine whether
waters identified in the environmental fate analysis as
having received contaminants from the hazardous
waste site are used commercially or recreationally.
4. Individuals who swim in contaminated waters can
experience dermal exposure to toxics over their entire
body. In addition, significant quantities of
contaminated water may be ingested inadvertently
while swimming, and swimmers will be exposed to
114
-------
Figure A-1. Exposed populations decision network
Environmental Fate Analysis
+
4, + 4r + +
Have Toxics Migrated
into Air?
Have Toxics Migrated
into Surface
Water?
Have Toxics Migrated
into Ground
Water?
I Have Toxics Migrated 1 1 Is Site Accessible 1
into Ott-site Soils? i— 1 1 to Public? i— 1
fH 1 L3
Are Persons
Potentially Exposed
Via Inhalation?
Is Contaminated
Surface Water
Fished
Commercially?
Is Contaminated
Surface Water
Used
Recreationally?
Is Contaminated
Surface Water
A Drinking
Water Source? Hf
Is Contaminated
Ground-water
A Drinking
Water Source?
Are Persons
Potentially
Exposed Via
Dermal Contact?
Are Workers
Potentially
Exposed Via
Dermal
Contact? r—
[2
^V
Are Persons
Potentially Exposed
Via Consumption of
Contaminated
Fish/Shellfish?
[T
W
Are Persons
Potentially Exposed
Via Dermal Contact,
Ingestion, or
Inhalation
While Swimming?
[T
Are Persons
Potentially
Exposed Via
Ingestion?
A
Are Persons Potentially
Exposed Via
Inhalation During
Showering/Bathing?
r-ft-
Are Persons
Potentially
Exposed Via
Ingestion of
Contaminated Soil
or Home Grown Food?
Are Persons
Potentially
Exposed Via Derma
Contact or Soil
Ingestion On-site?
A.
Are Persons
Potentially
Exposed Via
Inhalation
On-site?
A
Go on to Integrated Exposure
Analysis for Each Potentially
Exposed Population
-------
volatile contaminants in the water through inhalation.
Other screening should evaluate the existing or
potential degree to which the local population uses
contaminated water-bodies (fresh or marine) for
swimming.
5. If contaminated ground water or surface water is
a source of potable water, the population served may
experience considerable ingestion exposure. Similarly,
the population may also be exposed to toxics through
both dermal absorption and inhalation (of volatiles)
while showering or bathing. When undertaking a
screening analysis, it is only necessary to determine
which residences or commercial/institutional estab-
lishments are likely to obtain their potable water from
contaminated water sources.
6. If contaminants migrate to off-site soils, persons
contacting such soil may be exposed. Individuals who
grow their own fruit or vegetables at home may
experience additional exposure from ingesting food
grown in contaminated soils, as do those consuming
contaminated commercially-grown foods. Similarly,
livestock that have grazed on contaminated
vegetation may constitute a source of ingestion
exposure for consumers. Screening analysis should
strive to correlate areas of human habitation with
areas of contaminated soil, as defined in the
environmental fate analysis.
7. Similarly, if direct access to the site is possible,
children may be attracted to the location and may
directly contact hazardous materials or contaminated
soil. Such activity may result in inhalation or dermal
exposure, as well as intentional or inadvertent
ingestion of contaminated soil. For screening
purposes, the proximity of residential areas to the site
should indicate the potential for direct access by
children.
1.3 Quantitative Exposed Populations
Analysis
Quantitative analyses of potentially exposed human
populations comprises three distinct steps, which are
illustrated in Figure A-2. First, the results of
environmental fate analysis are compared with data
identifying and enumerating nearby human
populations to provide boundaries and quantify the
population(s) potentially or actually coming into
contact with contaminated air, water, and soil.
Populations consuming contaminated food (home
grown vegetables, fish) can similarly be identified
once the areal extent of contamination is known.
Population characterization, the second step, involves
identifying those groups within the exposed population
that, because of the specific health effects of some
pollutants or factors related to the population itself,
would experience a higher risk than would the
average population as a result of a given level of
exposure. Indeed, the health effects of the
contaminants under evaluation will often dictate the
need for population characterization. For example, if
mutagenic or teratogenic substances are involved,
women of childbearing age should be considered a
high-risk group. In addition, factors relating to the
exposed population may cause certain groups to
constitute high-risk subpopulations. These include:
• Persons with a genetic predisposition to certain
health effects;
• Persons whose health or resistance to disease is
impaired by behavioral factors such as smoking,
use of alcohol or drugs, etc;
• Infants, children, and the elderly, who are more
susceptible to health impacts from a given
exposure than are persons of other ages;
• Persons who are already suffering from disease
and may be more susceptible to further
impairment as a result of a given level of
exposure than are healthy persons;
• Persons who are exposed to naturally high
background levels of contaminants (e.g., selenium
or arsenic) and may be at greater risk to small
incremental increases of hazardous substances
than are persons who are not exposed to such
background levels; and
• Nutritionally deficient populations who may be
less resistant to exposure than those with
adequate diet.
While most studies will consider only the exposed
population as a whole and not as separate discrete
subpopulations, in certain cases, such detailed
population analysis may be warranted for in-depth
studies.
Age and sex influence the average inhalation rate, the
rate of food and water intake, the body area subject
to dermal exposure, and the types of food consumed,
all of which can affect the level of exposure actually
experienced. Some quantitative assessments may
require further characterization of populations to
determine age- and sex-specific exposure factors.
The third step is activity analysis. Once population
identification and characterization have answered the
question "Who may be exposed?", the question
"How and to what level are component portions of
this population exposed?" may next be asked in order
to refine the evaluation. This refinement involves
determining the exposed population's activities.
Comprehensive analysis can encompass the range of
indoor, outdoor, and in-car activities. For Superfund
Feasibility Studies, however, average values for
activity-related considerations usually suffice.
116
-------
Figure A-2. Quantitative exposed population analysis
Site Sampling
Data
1
Identification
and
Enumeration
of ^
Exposed ^N
Population
Characterization
of
Exposed '
Population
Activity
Analysis
r
\
+ ^
JL^^"""1^
r ~W
Inhalation of Inhalation o
Contaminants Contaminant
On-Site Off-Site
^
r ^ r
Identify Populations Comb
with Possible Access to Site with
(Workers, Children Fat
L ^
V.
^
*
9
»
Environmental
Fate Analysis
4-
I
1 Ingestion
-i- T ^
1 Water 1 PICA 1 Food 1
ne Census Combine Population Combine Population Data Use Recre
Air Data Data w'tn Ground-or wjtn Environmental to Identify
e Results Surface Water pate Results and Food of Exposed S
Fate Results and Production Statistics
•« » ^ r ^ r ^
*
Determine Determine
Site-Specific Age/Sex National Age/Sex
Distribution from Distribution from
Census of Population Census of Population
± ±
Identify/Quantify
Exposure-Related
Activities
Site Sampling
Data
+ +
4
4, 4,
Swimming 1 I Bathing 1 Direct Contact 1
ation Data Combine Population Identify Populations
Population Data with Ground-or Njtn Possible Access to Site
wimmers, etc Surface Water (Workers, Children)
Fate Results and
-------
The activity analysis can also help to identify high-
risk groups. For example, those groups that may
experience a significantly higher frequency or duration
of exposure as compared with the general population
can also be considered high-risk groups.
1.4 Identification and Enumeration of
Exposed Human Populations
The major population data base that can be accessed
to determine the size, distribution, and demographic
characteristics of a geographically defined population
is the Census of Population.
The data collected in the Census are organized
according to geographic areas. Within these areas,
data are further broken down into Census-defined
statistical areas and government units. Population
data are available within Standard Metropolitan
Statistical Areas (SMSAs) down to the level of the
"block" and in non-SMSAs to the level of the
Enumeration District (ED).
These data are especially useful in quantifying and
characterizing populations exposed as a result of their
presence in a specific locale (e.g., those exposed to
toxics in ambient air or soil). An isopleth map of
varying concentrations around a source can be
overlaid with Census maps. Such maps are available
for areas within SMSAs and can be purchased from
the Bureau of the Census. Also, Census Tracts
(Series PHC80-2) contains detailed characteristics
of the population (e.g., age, sex, race, education)
within each tract, a division of an SMSA containing
4,000 residents each. Census Tracts is currently
available on microfiche by SMSA and on computer
tape.
Many Super-fund sites are not within SMSAs. Census
data for non-SMSA areas are not available on maps,
but can be transcribed from Census publications.
The most useful Census publications for this type of
data are Number of inhabitants (Series PC80-1-A)
and General Population Characteristics (Series
PC80-1-B). Each series is currently available and
consists of a separate volume for each state, together
with a national summary volume. Number of
inhabitants provides only population counts, with no
demographic data. It provides data down to the level
of county subdivision and incorporated town. General
Population Characteristics provides population counts
by age, sex, and other demographic data, and
contains data down to the level of small towns (1,000
or more inhabitants).
All printed Census information is available for
purchase through the Government Printing Office
(GPO); all series issued on microfiche, maps,
computer tapes, and technical documentation are
available directly from the Customer Services Branch
at the Bureau of the Census, Department of
Commerce, Washington, D.C., and can be ordered by
calling (202) 763-4100. Alternatively, it may be more
convenient to contact one of the Census Bureau
regional offices. Cities where such offices are located
and phone numbers for the public information service
within each regional office are listed in Table A-l.
Table A-l.
Regional Census Bureau
Offices
Atlanta. Ga.
Boston,Mass.
Charlotte, N.C.
Chicago, IL.
Dallas,Tex.
Denver.Colo.
Detroit, Mich.
Kansas City, Kans.
Los Angeles, Calif.
New York, N.Y.
Philadelphia, Pa.
Seattle, Wash.
(404) 881-2274
(617)223-0226
(704)371-6144
(312)353-0980
(214)767-0625
(303) 234-5825
(313)226-4675
(913)236-3731
(213)209-6612
(212)264-4730
(215)597-8313
(206) 442-7080
7.4.7 Populations Exposed through Air
A convenient means of accessing quantitative
population data for a specific area impacted by air
contaminants is to directly link environmental fate and
exposed populations analysis through use of an
integrated computer-based fate model, and
population data retrieval program called ATM-
SECPOP. Developed by the EPA Office of Toxic
Substances, Exposure Evaluation Division (OTS-
EED), this model primarily analyzes point source
emissions, but can also be adapted to area or line
source analyses. ATM-SECPOP integrates the
output of a concentration prediction model (ATM)
(Patterson et al. 1982); a population distribution data
base (the proprietary 1980 Census Master Area
Reference File (MARF)), which is accessed via a
population distribution model called SECPOP; and
graphic and mapping information displays. This
integration affords a rapid and efficient means of
generating and presenting exposure data relating to
the airborne release of chemical substances. The
graphic display functions can be used to illustrate the
relationship of variables such as the distribution of
exposure or concentration versus distance for any or
all directions around a facility. Graphic displays may
be in the form of bar charts, scatter plots, rose
diagrams, or maps. Because of the proprietary nature
of the data contained in MARF, ATM-SECPOP's use
is restricted to personnel and contractors of EPA,
Office of Toxic Substances (EPA-OTS). Special
arrangements can be made for others to use the
data. Inquiries should be directed to the Modeling
118
-------
Section of the Exposure Assessment Branch of
EPA-OTS in Washington, D.C. A detailed discussion
of ATM is presented in Chapter 3 of this manual.
Where sites are accessible, the possibility that
children may enter and explore or play on the site
should be evaluated. On-site, children may
experience inhalation exposure to contaminated dust,
volatiles, or both. In some cases, the site boundary
may adjoin residential properties, and the area of
contamination may actually include such residences.
Accurate estimation of the potentially exposed
population in such a case is difficult; it can be
assumed that each household with children in the
immediate vicinity of the site has one child who may
find the site inviting. This should provide an upper
bound estimate on the actual number of children who
may enter the site. The Bureau of the Census (1986)
reports that in 1984, 50.1 percent of all U.S.
households included children. This percentage can be
applied to the total number of local households to
enumerate those in the area with children. The
analyst must decide which households are close
enough to the site to be considered.
Similarly, workers conducting activities at the site may
also experience inhalation exposure. Local authorities
(e.g., Zoning Board) may be able to supply
information on the likelihood of on-site work-related
activities that can be used to estimate the number of
workers who may become exposed. Remediation
workers are not included in this estimated exposed
population.
7.4.2 Populations Exposed through Surface Water
or Ground Water
Environmental fate analysis results can be used to
identify geographically-defined sources of
recreational (aquatic) dermal exposure, such as river
reaches downstream of an uncontrolled hazardous
site. The exposed population comprises swimmers in
those specific contaminated waters. The local
government agency concerned with recreation should
be able to provide estimates of the populations
swimming in local waters; this will usually be the
state, city, or county Department of Parks or
Recreation. Alternatively, one can use the following
national average value from the Bureau of Outdoor
Recreation (USDOI 1973): 34 percent of the total
population swims outdoors in natural surface
waterbodies (including oceans, lakes, creeks, and
rivers).
All persons served by a water supply system that
draws water from a contaminated water source must
be considered as potentially exposed through
ingestion and dermal exposure while bathing.
Information concerning local surface drinking water
sources and populations served can be obtained from
the local Department of Public Works, Planning
Department, or Health Department. These sources
should be able to provide information on public
departments or private drinking water treatment
companies that use ground water as their raw water
supply, and also on the number of households
drawing water from private wells.
1.4.3 Populations Exposed through Food
Exposure to contaminated food will usually be
associated with fruit and vegetables grown in home
gardens or with game residing in or using
contaminated areas. In order to identify the number of
persons consuming contaminated home grown fruit
and vegetables, first consult General Population
Characteristics, Series PC80-1-B to learn the total
number of households in a given geographic area.
Then the data presented in Table A-2, which provide
estimates of the percent of households in urban and
rural areas that have fruit and vegetable gardens and
the average number of persons per household, can
be applied to the local population data to estimate the
number of persons likely to consume contaminated
home grown produce.
The USDA Food Consumption of Households report
series can be consulted to estimate the local
population using a given food item for urban, rural
non-farm, and rural farm locales. These reports
present seasonal food use survey data on the
following bases: Northeast (USDA 1983a), North
Central (USDA 1983b), South (USDA 1983c), and
West (USDA 19834). More aggregated data are also
provided for the entire United States in a companion
report (USDA 1983e). The percent of households
using a given food item can be obtained from these
reports. The product of this value and the total
resident population of an area is an estimate of the
local exposed population. Similar national level data
are also provided on the basis of age and sex in Food
and Nutrient Intakes of Individuals in 1 Day in the
United States (USDA 1980). In addition, the U.S.
Food and Drug Administration (FDA) can be
contacted for data concerning daily intakes of various
food items. Such data have been compiled for the
FDA Total Diet Study (Pennington 1983).
Table A-2. U.S. Home Fruit and Vegetable Garden
Use, 1977
Percent of
households Household Percent of
with size (no. of total U.S.
Urbanization
Urban
Rural non-farm
Rural farm
gardens
43
41
84
persons)
3.17
3.44
3.86
population
32
9
3
Source: USEPA 1980.
119
-------
Monitoring data may indicate whether fish and game
are contaminated in the subject area. One can
estimate the fishing population by contacting the local
agency responsible for issuing fishing licenses; this
may be the state fish and game commission or the
state department of natural resources. Since there are
2.69 persons in the average household (Bureau of the
Census 1986) one can estimate the actual exposed
population by multiplying 2.69 by the number of
licensed hunters or fishermen in the area.
7.4.4 Populations Exposed through Soil
Exposure to contaminated soil constitutes a potential
exposure route for workers or children playing
outdoors. Neighborhood children playing at the site
can be exposed to high levels of contaminants. Soil-
related exposure in such cases would be through
direct dermal contact with the contaminated soil.
Another potentially significant, but infrequently
encountered, exposure mechanism involves children
who eat dirt; this eating behavior, known as pica, may
lead to their actually ingesting contaminated soil.
Hand-to-mouth contact during normal play is a
more common means of ingesting soil, however. For
any site located near residential areas, the degree of
accessibility to children should be considered. Bureau
of the Census data can be used as described in
Section A-1.4.1 to estimate the number of local
children who may have access to the site.
In addition, workers conducting activities at the site
(other than remediation) may have direct dermal
contact with contaminated soils. Section A-1.4.1
provides general guidance to identify and enumerate
exposed worker populations.
1.5 Population Characterization
After exposed populations have been identified and
enumerated, they can be characterized by age and
sex factors. The physiological parameters that
determine the dose received per a given level of
exposure (e.g., breathing rate, skin surface area, and
ingestion rate) are often age- or sex-specific. Also,
from a toxicity standpoint, subpopulations defined by
age or sex, such as the elderly or women of
childbearing age, may be especially susceptible to a
chemical substance. Superfund studies will generally
use average values, but by characterizing exposed
populations, one can determine exposure distributions
within the population at large and delineate specific
high-risk subpopulations.
The Census Publication series General Population
Characteristics (PC80-1-B) cites figures for the
age and sex structure of the population residing in a
specific area. Separate volumes for each state
contain age and sex breakdowns at the level of
county subdivisions and small towns. If more detail is
required, the Census Bureau microfiches containing
this information at the Census tract level (only
available by SMSAs).
In the case of exposure resulting from ingestion of
food, the food consumption surveys of the USDA
(1983a-e) record age and sex data for the sampled
population. These data are contained in five separate
regional reports; the appropriate one should be
consulted.
In lieu of obtaining site-specific data, one can use
the population characteristics of the U.S. as a whole,
provided in the yearly Statistical Abstract of the
United States (for example, see Bureau of the
Census 1986), to approximate the population
distribution in the area of concern.
1.6 Activity Analysis
Activities engaged in by members of a given
population or subpopulation can dramatically affect
the level of human exposure to environmental
contaminants. For example, persons whose lifestyle
or employment involves frequent strenuous activity
will inhale larger volumes of air per unit time than will
those living a less strenuous life, and will experience
a higher level of exposure to airborne contaminants.
Activity analysis allows refinement of certain
parameters used in the calculation of exposure,
including:
x* Inhalation rate;
x Frequency of exposure; and
xx Duration of exposure.
The procedure for integrating activity-related
inhalation, frequency, and duration data into the
exposure assessment process is detailed in the
following chapter.
120
-------
Chapter 2
Exposure Calculation and Integration
This chapter provides guidance for calculating and
integrating exposures to all populations affected by
the various exposure routes associated with a given
uncontrolled hazardous waste site. Specifically,
guidance is provided to estimate exposure from:
1. Inhalation
a. Ambient air
b. Indoor air (contaminants released during
showering)
2. Dermal contact
a. Water (swimming)
b. Soil
3. Ingestion
a. Food
b. Water
c. Soil
This analysis is based on the results of all previous
analyses, and is the final stage of the exposure
assessment. This guidance is complete; no additional
documentation is required to finish the analysis.
Integrated exposure analysis is conducted for only
those contaminants having complete exposure
pathways (i.e., those contaminants that are released
and migrate from the site and that do contact human
populations). Therefore, no screening evaluation is
included in the exposure integration process. While
calculating the exposure incurred is traditionally the
final step in the quantitative exposure assessment
process, it can also be viewed as a component of the
human health risk assessment. Therefore, the
material detailed in this chapter is also discussed in
the Superfund Public Health Evaluation Manual
(USEPA1985).
Exposure is defined as the amount of pollutant
contacting body boundaries (skin, lungs, or
gastrointestinal tract). Exposure calculation considers
how often populations come into contact with
contaminants in specific environmental media, the
mode of such contact, and the amount of
contaminated medium that contacts the internal or
external body surface during each exposure event.
The goal of this analysis is to quantify the amount of
contaminant contacted within a given time interval.
Short-term and long-term exposures are calculated
in the same manner. First, for each exposure scenario
under consideration, an exposure per event is
estimated. This exposure value quantifies the amount
of contaminant contacted during each exposure
event, with "event" being defined differently
depending on the nature of the scenario under
consideration (e.g., each day spent swimming in a
contaminated river is a single swimming exposure
event, each day's inhalation of contaminated air is an
inhalation exposure event). Event-based exposure
estimates take into account the concentration of
contaminant in the medium through which exposure
occurs, the rate of contact with such media
(inhalation rate, ingestion rate, etc.), and the duration
of each event.
The analyst can convert event-based exposure
values to final exposure values by multiplying the
exposure per event by the frequency of exposure
events over the timeframe being considered. Short-
term exposure is based on the number of exposure
events that occur during the short-term timeframe
(10 to 90 days), while long-term exposures are
based on the number of events that occur within an
assumed 70-year lifetime. The 70-year assumed
average lifetime is traditionally used in exposure
assessments, and it provides a conservative upper
bound of lifetime exposure. Certain exposure
scenarios, however, may only apply to short-term
exposure. Whenever practical, the analyst should
strive to determine the timeframe over which a given
exposure pathway would be expected to affect the
exposed population. Once determined, the timeframe
will indicate whether that pathway should be
evaluated on a short- or a long-term basis.
Exposure estimates are expressed in terms of mass
of contaminant/unit of body mass/day by dividing daily
exposure by the value for total body mass of an
average individual in the exposed population. For
Superfund studies, an average adult body mass of 70
kg will usually be adequate for this conversion. In
cases where exposure to specified subpopulations
must be evaluated, values for other than average
121
-------
adults may be required. Consult Anderson et al.
(1984) to obtain alternative body mass values.
Similarly, average values for activity-related
parameters (e.g., inhalation rate) generally will be
adequate for Superfund site evaluations. For special
situations and detailed exposure analysis, analysts
can refer to the discussion of activity data in Freed et
al. (1985). An exposure factors handbook is currently
under development (USEPA 1987), and the analyst
performing exposure assessments after publication of
this manual should consult that document for the
most up-to-date exposure factors.
The following sections address the exposure
calculation process specific to each exposure
mechanism. Data management forms designed to
organize and tabulate the data in the exposure
calculation process are presented in Appendix C.
2.1 Inhalation Exposure
Inhalation exposure per event is estimated based on
the hours per event, the inhalation rate of the
exposed individual during the event, and the
concentration of contaminant in the air breathed. The
formula for calculating event-based exposure is the
following:
IEX = t xIxC(X)xF-BW-2.56xlO\.i, .
e lifetime
where
IEX = estimated inhalation exposure,
(mg/kg/day).
te = duration of an exposure event,
(hours/event).
I = average inhalation rate of exposed
persons, (m3/hr).
C(X) = contaminant air concentration
throughout the exposure period,
(milligrams/m3).
BW = average adult body weight, (70 kg).
F = frequency of exposure event,
(number/lifetime).
Short-term exposure is calculated using the short-
term contaminant air concentration, and long-term
exposure is based on the long-term concentration.
Inhalation exposures are keyed to geographic
locations delineated during the Environmental Fate
Analysis. Ambient concentration is generally assumed
to be homogeneous throughout a limited area or
sector (within an isopleth); however, this assumption
is not always well-founded. Numerous studies have
shown that there can be marked differences in indoor
and outdoor concentrations of pollutants (Budiansky
1980, Moschandreas et al. 1978) or among
microenvironments in the same area (Ott 1981). To
account for these differences when calculating
exposure, several investigations have coined the term
"microenvironment," which refers to a type of
physical setting where concentrations of pollutants
can be expected to be similar. For Superfund studies,
it is usually unnecessary to disaggregate analysis on
a microenvironment basis. Instead, it can generally be
assumed that contaminants have been present long
enough for indoor-to-outdoor concentrations to
have reached equilibrium.
To calculate exposure duration, the analyst considers
the amount of time exposed persons actually spend in
the contaminated area. For example, if a site is in a
residential area, one can conservatively estimate
exposure by assuming that all residents spend the
entire 24-hour day within the contaminated zone. If a
site is located in an industrialized area, it may be
more appropriate to base duration on an 8-hour
workday, if it can be reasonably assumed that
workers do not also live in the immediate
industrialized area. Such factors must be evaluated
on a case-by-case basis. For inhalation exposure,
frequency is assumed to be daily.
For a general application, use an average adult value
for inhalation rate. An example of an adult average
derived from experimental results (USEPA 1981) is an
inhalation rate of 1 m3/hour. This value can be used
to conservatively estimate exposure regardless of
microenvironments or activity.
Generating time-weighted average inhalation rates
provides a more precise estimate of inhalation rate.
This calculation is based on microenvironment-
related data and activity stress levels/ventilation rates
associated with the individual microenvironment. If
this level of detail is warranted, the inhalation rates
presented in Table A-3 can be used. Freed et al.
(1985) cite directions for developing time-weighted
average inhalation rates.
To calculate ambient inhalation exposure, one should
obtain contaminant air concentration values from the
results of the environmental fate analysis. In one
case, however, concentration values will have to be
calculated in the exposure integration stage of the
exposure assessment. Persons showering or bathing
in potable water contaminated with toxics may be
exposed through inhalation if the contaminants are
volatile. This is especially true of showering, since the
high turbulence, combined with the elevated
temperature of the shower water, can produce a
significant release of volatile components.
Various approaches are available to estimate
contaminant concentrations indoors. These
approaches depend on a number of factors, including
the room air volume, air exchange and mixing factors,
contaminant concentration in the water, the amount of
water used, and the manner in which a contaminant
122
-------
Table A-3. Summary of Human inhalation Rates for
Men, Women and Children by Activity Level
(m3/hour)a
Resting" Light0 Moderated Heavy6
Adult male
Adult female
Average adult'
Child, age 6
Child, age 10
0.6
0.6
0.6
0.4
0.4
1.3
1.3
1.3
1.4
1.7
2.6
2.4
2.6
2.1
3.3
7.1
4.9
6.0
2.4
4.2
aValues of inhalation rates for males, females, and children
presented in this table represent the midpoint of ranges of
values reported for each activity level in Anderson et al. (1984)
"includes watching television, reading, and sleeping.
Includes most domestic work, attending to personal needs and
care, hobbies, and conducting minor indoor repairs and home
improvements.
dlncludes heavy indoor cleanup and performance of major
indoor repairs and alterations and climbing stairs.
Includes vigorous physical exercise and climbing stairs carrying
a load.
'Derived by taking the mean of the adult male and adult female
values for each activity level. A representative 24-hour
breathing rate for an average adult is 1.1 m3/hour. This value
is based on the assumption that the average adult spends 93.2
percent of the time at the light/resting level of activity, 5.8
percent at a moderate level of activity, and 0.9 percent at a
heavy level of activity. Values for the percent of time spent at
each activity level are from Freed et al. (1985).
is released into room air (instantaneously, con-
tinuously, time-dependent). If showering/bathing
exposure estimation is required for a Superfund
exposure assessment, the analyst is referred to
Versar (1984) for a detailed discussion of techniques
to estimate indoor air contaminant concentration. For
both showers and baths, the analyst should assume a
continuous contaminant release during the bathing/
showering period. Values for the other variable factors
mentioned above can be obtained from Versar
(1985).
To evaluate inhalation exposure to contaminants
volatilizing from potable water while showering, the
analyst should again assume frequency to be daily.
Each shower is assumed to last 15 minutes.
Inhalation exposure to swimmers can be based on
monitored or estimated ambient air concentrations
above a contaminated water body. To estimate
concentrations, calculate the rate of volatilization of
the contaminant from the water body and use this
value as the input to a "box model" air migration
model. The dynamic release rate can be calculated
using Equations 2-10, 2-15, 2-16, and 2-17. The
recommended air model is BOXMOD (in EPA's
GEMS system, see Chapter 3).
2.2 Dermal Exposure
Dermal exposure is determined by the concentration
of hazardous substance in a contaminated medium
that is contacted, the extent of contact (i.e., the body
surface area contacted), and the duration of such
contact. For exposure to contaminated water, dermal
exposure per event is calculated as follows:
DEX=tPxAVxCxPCxFxl liter/1000 cm£
4_days_
lifetime
(A-2)
where
DEX
te
AV
C
PC
F
BW
= estimated dermal exposure,
(mg/kg/day).
= duration of exposure, (hours/event).
= skin surface area available for contact,
(cm2).
= contaminant concentration in water,
(mg/liter).
= dermal permeability constant for the
subject contaminant, (cm/hr).
= frequency of exposure events per
lifetime.
= average adult body weight, (70 kg).
The term 1 liter/1,000 cm3 is a volumetric conversion
constant for water.
When possible, it is important to consider the degree
to which a given contaminant is actually able to enter
the body. Some compounds will not readily penetrate
the skin, while others may do so at a rapid rate. The
above equation can only be used in cases where
dermal permeability constants for the contaminant(s)
of concern are known. Table A-4 lists dermal
permeability constants for selected compounds. For
many compounds, however, dermal permeability
constants will not be available. In such cases, the
analyst must assume that contaminants are carried
through the skin as a solute in water which is
absorbed (rather than being preferentially absorbed
independently of the water), and that the contaminant
concentration in the water being absorbed is equal to
the ambient concentration. Thus, the permeation rate
of water across the skin boundary is assumed to be
the factor controlling the contaminant absorption rate.
Short-term dermal exposure per event is calculated
using the short-term contaminant concentrations in
water or soil, and long-term exposure is based on
the long-term contaminant concentrations.
The local recreation department may have detailed
data quantifying the duration and frequency of water
use for swimming. When such locale-specific data
are not available, the following national average
figures, based on data from the Bureau of Outdoor
Recreation (USDOI 1973), can be applied:
* Frequency of exposure = 7 days/year.
123
-------
Table A-4. Permeability Constants for Various Compounds
Permeability
Compound
SURFACTANTS
Decanoic acid
Dodecanoic acid
Tetradecanoic acid
Hexadecanoic acid
Octadecanoic acid
Sodium dodecyl sulfate
Sodium dodecyl isothionate
Sodium p-1 -dodecyl
benzenesulphonate
Sodium laurate
IONS
Aluminum
Potassium
Bromide
Palmitate
Laurate
DRUGS
Methotrexate
Benzoyl peroxide
Estradiol
Amphetamine
Ouabain
Burimamide
Metramide
Cimetidine
PHENOLS
Resorcinol
p-Nitrophenol
n-Nitrophenol
Phenol
Methylhydroxybenzoate
n-Cresol
o-Cresol
p-Cresol
beta-Naphthol
o-Chlorophenol
p-Ethylphenol
3,4-Xylenol
p-Bromophenol
p-Chlorophenol
Thymol
Chlorocresol
constant3 (cm//hr)
1 .OOE-03
2.00E-03
6.00E-04
1 .20E-05
6.00E-06
2.00E-03
5.40E-05
6.00E-06
1 .OOE-03
7.20E-06
6.70E-05
1 .80E-05
4.20E-05
3.00E-03
6.00E-10
5.10E-07
3.90E-03
1 .40E-05
3.90E-06
1 .70E-07
1.10E-07
3.30E-07
2.40E-03
5.58E-02
5.58E-02
8.22E-02
9.12E-02
1.52E-01
1.57E-01
1.75E-01
2.79E-01
3.31 E-01
3.49E-01
3.60E-01
3.60E-01
3.60E-01
5.28E-01
5.50E-01
Reference
Howes 1975
Howes 1975
Howes 1975
Howes 1975
Howes 1975
Howes 1975
Howes 1975
Howes 1975
Tregear 1966
Tregear 1966
Tregear 1966
Tregear 1966
Tregear 1966
Tregear 1966
McCullough et al. 1976
Nachtet at. 1981
Galey et al. 1976
Galey et al. 1976
Sutton 1973
Sutton 1973
Sutton 1973
Sutton 1973
Roberts et al. 1977
Roberts et al. 1977
Roberts et al. 1977
Roberts et al. 1977
Roberts et al. 1977
Roberts et al. 1977
Roberts et al. 1977
Roberts et at. 1977
Roberts et at. 1977
Roberts et al. 1977
Roberts et at. 1977
Roberts et at. 1 977
Roberts et at. 1977
Roberts et al. 1977
Roberts et al. 1977
Roberts et al. 1977
(Continued)
124
-------
Table A4. (Continued)
Compound
PHENOLS (Continued)
Chloroxylenol
1 ,4,6-Trichlorophenol
2,4-Dichlorophenol
STEROIDS
Progesterone
Pregnenolone
Hydroxypregnenolone
Hydroxyprogesterone
Cortexone
Testosterone
Cortexolone
Corticosterone
Cortisone
Hydrocortisone
Aldosterone
Estrone
Estradiol
Estriol
Dihydroepiandrosteroneb
Dihydrotestosteroneb
ALCOHOLS
Methanol
Ethanol
Propanol
Butanol
Pentanol
Hexanol
Heptanol
Octanol
Nonanol
Decanol
GLYCOL ETHERS
2-Methoxyethanol
2-Ethoxyethanol
2-Ethoxyethanol acetate
2-n-Butoxyethanol
1 -Methoxypropan-2-OI
2-(2-Methoxyethoxy)ethanol
2-(2-Ethoxyethoxy)ethanol
2-(2-n-Butoxyethoxy)ethanol
Permeability
constant3 (cm//hr)
5.90E-01
5.94E-01
6.01 E-01
1 .50E-03
1 .50E-03
6.00E-04
6.00E-04
4.50E-04
4.00E-04
7.50E-05
6.00E-05
1 .OOE-05
3.00E-06
3.00E-06
3.60E-03
3.00E-04
4.00E-05
1 .70E-04
3.90E-04
5.00E-04
8.00E-04
1 .20E-03
2.50E-03
6.00E-03
1 .30E-02
3.20E-02
520E-02
6.00E-02
8.00E-02
2.89E-03
8.42E-04
8.07E-04
2.14E-04
1 .25E-03
2.06E-04
1 .32E-04
3.60E-05
Reference
Roberts et al. 1977
Roberts et al. 1977
Roberts et al. 1977
Scheuplein et al. 1969
Scheuplein et al. 1969
Scheuplein et al. 1969
Scheuplein et al. 1969
Scheupliln et al. 1969
Scheuplein et al. 1969
Scheuplein et al. 1969
Scheuplein et al. 1969
Scheuplein et al. 1969
Scheuplein et al. 1969
Scheuplein et al. 1969
Scheuplein et al. 1969
Scheuplein et al. 1969
Scheuplein et al. 1969
Schaefer et al. 1 982
Schaefer et al. 1982
Scheuplein and Blank 1971
Scheuplein and Blank 1971
Scheuplein and Blank 1971
Scheuplein and Blank 1971
Scheuplein and Blank 1971
Scheuplein and Blank 1971
Scheuplein and Blank 1971
Scheuplein and Blank 1971
Scheuplein and Blank 1971
Scheuplein and Blank 1971
Dugard et al. 1984
Dugard et al. 1984
Dugard et al. 1984
Dugard et al. 1984
Dugard et al. 1984
Dugard et al. 1984
Dugard et al. 1984
Dugard et al. 1984
(Continued)
125
-------
Table A4. (Continued)
Compound
PESTICIDES0
Azodrin
Ethion
Guthion
Malathion
Parathion
Baygon
Carbaryl
Aldrin
Dreldrin
Lindane
24-D
Diquat
OTHER
Water
Ethylbenzene
Styrene
Toluene
Anilrne"
N-nitrosodiethanolamine
Ethyl ether
2-Butanone
1-Butanol
2-Ethoxyethanol
2,3-Butanediol
Benzeneb
Desoximetasoneb'd
Linoleic acidb'd
Dithranolb
Theophyllineb
Caffeineb
8-Methoxypsoralenb.d
p-Butoxyphenylacethydroxamic
acidbd
Triacetoxyanthraceneb
Heparinb'd
Carbon disulfideb
Permeability
constanta (cm//hr)
9.80E-04
2.20E-04
1 .06E-03
5.50E-04
650E-04
1 .31 E-03
4.90E-03
5.20E-04
5.10E-04
6.20E-04
3.90E-04
2.00E-05
8.00E-04
1 .OOE-03
6.00E-04
9.00E-04
2.00E-02
5.50E-05
1.70E + 01
5.00E + 00
4.00E + 00
3.00E-01
5.00E-02
4.10E-01
3.40E-05
1 .60E-05
2.10E-04
2.50E-05
3.30E-04
9.90E-04
9.80E-05
5.80E-05
8.20E-05
5.50E-02
Reference
Feldman and Maibach 1974
Feldman and Maibach 1974
Feldman and Maibach 1974
Feldman and Maibach 1974
Feldman and Maibach 1974
Feldman and Maibach 1974
Feldman and Maibach 1974
Feldman and Maibach 1974
Feldman and Maibach 1974
Feldman and Maibach 1974
Feldman and Maibach 1974
Feldman and Maibach 1974
Blank et al. 1984
Dutiewicz and Tyras 1967
Dutiewicz and Tyras 1968
Dutiewicz and Tyras 1968
Baranowska-Dutkiewicz
1982
Bronaugh et al. 1981
Scheuplein and Blank 1971
Scheuplein and Blank 1971
Scheuplein and Blank 1971
Scheuplein and Blank 1971
Scheuplein and Blank 1971
Baranowska-Dutkiewicz
1982
Schaefer et al. 1982
Schaefer et al. 1982
Schaefer et al. 1982
Schaefer et al. 1982
Schaefer et al. 1982
Schaefer et al. 1982
Schaefer et al. 1982
Schaefer et al. 1982
Schaefer et al. 1982
Baranowska-Dutkrewicz
1982
a Permeability constants are for contaminants as a dilute solution in water, except
as noted.
b Calculated permeability constant, subject to error.
c Permeability constants are for contaminants in acetone. These values should not
be used for dermal exposure due to contact with contaminated water. These
values should be used for dermal exposure to pure wastes.
d Permeability constants are for contaminants in gel. These values should not be
used for dermal exposure due to contact with contaminated water. These values
may be used for dermal exposure to pure wastes.
126
-------
ss Duration of exposure = 2.6 hours/day.
Dermal absorption of waterborne contaminants may
be a significant exposure route. The factors that
influence dermal absorption of chemicals are: (1) the
nature of the compound (molecular weight,
lipophilicity), (2) the presence of other compounds
that might facilitate passage of a chemical though the
skin (e.g., chelating or complexing agents), and (3)
the permeability of the skin. Generally only lipid-
soluble, non-ionized compounds are absorbed
significantly through the skin. Also, the skin is
normally permeable only to compounds whose
molecular weights are less than 500 Daltons. The
permeability of the skin to larger molecular weight
compounds and to less lipophilic compounds can be
increased when corrosive agents such as acids are
present or when there are skin abrasions. For
waterborne chemicals, exposure through the skin is
almost directly proportional to concentration.
Brown, Bishop, and Rowan (1984) recently reported
that when compared with ingestion, dermal absorption
of volatile organic contaminants in drinking water
accounted for approximately 29 to 91 percent of the
total dose incurred, with the average being about 64
percent. The dermal exposure route becomes
especially pertinent when organic contaminants are
present in very dilute aqueous solution, as may often
be the case at Super-fund sites. In certain cases, then,
dermal exposure to contaminants contained in ground
or surface water may actually overshadow ingestion
exposure.
When persons become exposed to contaminants in
drinking water, the dermal exposure associated with
bathing or showering should also be considered. One
can use the same approach to assess
bathing/showering as was used for swimming.
Generally, an average frequency of one bath or
shower per day can be assumed, and each event can
be estimated to last 15 minutes.
For swimming or bathing exposure, the surface area
available for dermal exposure is assumed to equal the
total amount of human skin surface area. Average
availability values are given below for adults and
children. If the exposed population is not separated by
age groups, both availability values should be used to
represent a general range of exposure for the total
swimming or bathing population. Both availability
figures cited below are taken from Anderson et al.
(1984):
• Average adult (male and female, 20-30 yrs) =
18,150 cm2.
• Average child (male and female, 3-12 yrs) =
9,400 cm2.
Direct dermal contact with contaminants present in
soil is calculated as follows:
DEX=ClXAVxDAxF
t days
lifetime
(A-3)
where
DEX
AV
DA
F
BW
dermal exposure, (mg/kg/day).
weight fraction of chemical substance
in soil, (unitless).
skin surface area available for contact,
(cm*).
dust adherence, (mg/cm ).
frequency of exposure events per
lifetime.
average adult body weight, (70 kg).
Values for contaminant weight fraction in the
contaminated soil will be available from the site
survey. Skin surface availability depends on the
nature of activity being conducted, and can vary for a
given activity depending on the season of the year.
Anderson et al. (1984) provide data on skin surface
areas of different parts of the body for adults and
children. Based on a projection of the type of activity
at the site and the age of the exposed population
(i.e., workers or children), the data in Anderson et al.
can be used to develop skin surface estimates for
use in estimating direct dermal exposure.
Data on dust adherence to skin (DA) are limited,
although the following experimental values for (soil-
related) dust adherence were reported by the Toxic
Substance Control Commission of the State of
Michigan (Harger 1979):
es Commercial potting soil adheres to hands at 1.45
mg/cm2.
* Dust of the clay mineral kaolin adheres to hands
at 2.77 mg/cm2.
The degree to which these values represent dust
adherence at any given site is uncertain, as such
adherence will depend on a variety of site-specific
factors. Therefore, instead of selecting one of the
above values to estimate direct dermal exposure, it is
suggested that the analyst use both values and
generate an exposure range. The lifetime frequency
of direct dermal exposure will also vary considerably
and will depend on the nature of the site, its ease of
access, and a variety of other factors. Thus, contact
frequency should be estimated on a case-by-case
basis, based on knowledge of the site and its
environs.
127
-------
Note that this approach is conservative in that it
assumes that all of the contaminant adsorbed to the
soil (dust) particles is available for absorption through
the skin. In fact, only a percentage of the total
adsorbed contaminant mass may actually be available
for such absorption, as some percentage may remain
bound to the soil particle.
The site survey will provide values for the
contaminant weight fraction in the contaminated soil.
Skin surface availability depends on the nature of the
activity being conducted, and can vary for a given
activity depending on the season of the year.
Anderson et al. (1984) provide data on skin surface
areas of different parts of the body for adults and
children. Based on a projection of the type of activity
at the site and the age of the exposed population
(i.e., workers or children), one can use the data in
Anderson et al. to develop skin surface estimates for
use in estimating direct dermal exposure.
The lifetime frequency of direct dermal exposure will
vary considerably and will depend on the nature of
the site, its ease of access, and other factors.
Contact frequency should be estimated on a case-
by-case basis, based on knowledge of the site and
its environs.
2.3 Ingestion Exposure
2.3. J Food/Soil
Food ingestion exposure is estimated as the product
of contaminant concentration in the food consumed
and the amount of food consumed per day.
Frequency is daily for foods that are a regular part of
the diet. For recreationally caught fish, frequency can
be estimated based on the seasonal nature of fishing
involved.
USDA source materials listed in Section A-1.4.3 are
also useful in quantifying the amount of contaminated
food ingested. The Food Consumption of Households
report series provides data quantifying the amount of
various food categories consumed by households on
a seasonal basis. Similar data are presented in food
and Nutrient Intakes of Individuals in 1 Day in the
United States. The first source can be used to derive
estimates of the amount of various foods consumed
by the overall exposed population by applying
seasonal percentage use values to local population
census data. The second source is used in
subpopulation analyses by applying sex- and age-
specific consumption values to census data for the
exposed population.
Consumption of fish caught in contaminated waters
may be an important ingestion route, since certain
contaminants of concern tend to biomagnify in the
food chain. This phenomenon results in tissue
concentrations of contaminants in predator fish
exhibiting levels that greatly exceed the ambient
concentration in the waterbody. An average daily fish
ingestion rate for the U.S. population has been
estimated as 6.5 grams per day (USEPA 1980b).
Persons for whom fish constitutes a major portion of
the overall diet may consume up to 124 grams per
day (USDA 1980). A West Coast study of
consumption of fish caught in contaminated waters by
sport fishermen (Puffer et al. 1981) reports a median
fish ingestion rate of 37 grams/day. This report also
lists a maximum rate of 225 grams/day.
Ingestion exposure estimates are calculated in the
same manner, regardless of the type of food
ingested. Multiplication of the contaminant
concentration in the ingested food by the amount of
contaminated food ingested per day yields exposure
per day.
Children may ingest soil during play both inadvertently
and intentionally (pica behavior). In those
assessments where the exposed populations analysis
has found that children may have access to areas of
contaminated soil, this exposure route should be
evaluated. Data quantifying the amount of soil
ingested by children are conflicting and vary
considerably. For example, Calabrese et al. (1987)
report that estimates range from a low of 10 mg/day
(for 2-year-old children) to a high of 10,000 mg/day
(for 1.5- to 3.5-year-old children). Within this
range, reasonable typical values can be identified and
associated with various age groups, if desired. For
studies warranting such detail, the daily soil ingestion
rate values presented in Table A-5 can be used. For
studies that do not require such detail, one can use
an overall average soil ingestion value of 100 mg/day.
Table A-5. Typical Daily Soil Ingestion
Rates for Children by Age
GrouP Soil ingestion range
Age (mg/day)
0-9 months
9-18 months
1.5 - 3.5 years
3.5 - 5 years
5-18 years
0
50
200
50
10
Source: Calabrese et al. 1987.
2.3.2 Water
Event-based water Ingestion exposure equals the
daily total amount of contaminant ingested from either
surface or ground waters affected by the Superfund
site. This exposure is determined by the contaminant
concentration in the water and the amount of water
ingested per day. On average, an adult ingestion
coefficient of 2.0 liters per day (USEPA 1980b) can
be used for Superfund site analyses. Frequency of
drinking water exposure is daily.
128
-------
When contaminated surface waters are used
recreationally, it may be appropriate to estimate
exposure that results from inadvertently ingesting
contaminated water while swimming. For this analysis,
the same values for event frequency and duration
previously presented in Section A-2.2 should be
used. In addition, to estimate the amount of
contaminated water ingested per event, an assumed
value of 50 ml per hour can be used.
2.4 Exposure Integration
The final step in the exposure assessment process
for uncontrolled hazardous waste sites is the
integration of all exposures experienced by individual
exposed populations. This simply involves organizing
the results of the previous analyses to total all
exposures to a given hazardous substance
experienced by each population segment. Because
different chemicals exhibit different toxicological
properties, exposures to each contaminant of concern
are considered separately. Note that in some cases,
individual populations may be exposed to a given
chemical in a particular medium through more than
one exposure scenario. For example, persons who
swim in contaminated waters may obtain their
drinking water from the same contaminated
waterbody. In such cases, the dermal exposure
experienced while swimming can be added to that
experienced during bathing or showering to generate
an overall dermal exposure value for that population
segment. The data management forms supplied in
Appendix C are designed to help organize the results
of exposure calculation and integration.
129
-------
-------
Chapter 3
Appendix A References
Anderson E, Browne N, Duletsky S, et al. 1984.
Development of statistical distribution or ranges of
standard factors used in exposure assessments.
Revised draft final report. Washington, DC: Office
of Health and Environmental Assessment, U.S.
Environmental Protection Agency. Contract No.
68-02-3510.
Baranowska-Dutkiewicz B. 1982. Skin absorption of
aniline from aqueous solutions in man. Toxicology
Letters 10:367-72.
Blank IH, Moloney J, Ernslie AG, et al. 1984. The
diffusion of water across the stratum corneum as a
function of its water content. The Journal of
Investigative Dermatology. 82:188-194.
Bronaugh R, Condon E, Scheuplein RJ. 1981. The
effect of cosmetic vehicles on the penetration of
n-nitrosodiethanolamine through excised human
skin. Journal of Investigative Dermatology 76:94-
6.
Brown HS, Bishop DR, Rowan CA. 1984. The role of
skin absorption as a route of exposure for volatile
organic compounds (VOCs) in drinking water.
Amer. J. of Public Health. 74 (5).
Budiansky S. 1980. Indoor air pollution. Environ. Sci.
Technol. 14(g): 1023-1027.
Bureau of the Census. 1986. Statistical abstract of the
United States: 1986 (106th edition). Washington,
DC: U.S. Department of Commerce, U.S.
Government Printing Office.
Calabrese EJ, Kostecki PT, Gilbert CE. 1987. How
much soil do children eat? An emerging
consideration for environmental health risk
assessment. Paper submitted to Comments in
Toxicology.
Dugard PH, Walker M, Mawdsley SJ, Scott RC. 1984.
Absorption of some glycol ethers through human
skin in vitro. Environment Health Perspectives
57: 193-97.
Dutkiewicz T, Tyras H. 1967. A study of skin
absorption of ethylbenzene in man. British Journal
Ind. Med. 24:330-32.
Dutkiewicz T, Tyras H. 1968. Skin absorption of
toluene, styrene, and xylene by man. Department
of Toxicological Chemistry and Industrial
Toxicology, Medical Academy, Lodz, Poland.
Feldman RJ, Maibach HI. 1974. Percutaneous
penetration of some pesticides and herbicides in
man. Toxicology and Applied Pharmacology
28: 126-32.
Freed JR, Chambers T, Christie WN, Carpenter CE.
1985. Methods for assessing exposure to chemical
substances: volume 2 - methods for assessing
exposure to chemical substances in the ambient
environment. Washington, DC: U.S. Environmental
Protection Agency, Office of Toxic Substances,
Exposure Evaluation Division. EPA 560/5-83-
015.
Galey WR, Londdale HK, Nacht S. 1976. The in vitro
permeability of skin and buccal mucosa to selected
drugs and tritiated water. J. Invest. Dermatol.
67:713-717.
Harger JRE. 1979. A model for the determination of
an action level for removal of curene contaminated
soil. Memorandum to P.S. Cole, Executive Director.
Lansing, Ml: Toxic Substance Control Commission.
Howes D. 1975. The percutaneous absorption of
some anionic surfactants. J. Soc. Cosmet. Chem.
26:47-63.
McCullough JL, Snyder DS, Weinstein GD, Friedland
A, Stein B. 1976. Factors affecting human
percutaneous penetration of methotrexate and its
analogues in vitro. J. Invest. Dermatol. 66:103-
107.
Moschandreas DJ, Stark JWC, McFadden JF, Morse
SS. 1978. Indoor pollution in the residential
environment - vols. I and II. Washington, DC:
131
-------
Office of Air Quality Planning and Standards, U.S.
Environmental Protection Agency.
Nacht S, Yeung D, Beasley JN Jr., Anfo MD,
Maibach HI. 1981. Benzoyl peroxide: percutaneous
penetration and metabolic disposition. Journal
American Academy Dermatology 4:31-7.
Ott WR. 1981. Exposure estimates based on
computer-generated activity patterns. Paper
presented at the 74th annual meeting of the Air
Pollution Control Association. Philadelphia, PA.
Paper No. 81-57-6.
Patterson MR, Sworski TJ, Sjoreen AL, et al. 1982.
User's manual for UTM-TOX, a unified transport
model. Draft report. Oak Ridge, TN: Oak Ridge
National Laboratory. ORNL-TM-8182. IEG-
AD-89-F-1-3999-0.
Pennington, JAT. 1983. Revision of the total diet
study. In J. Amer. Dietetic Assoc. 82 (2).
Puffer H., Azen SP, Young DR, et al. 1981.
Consumption rates of potentially hazardous marine
fish caught in the metropolitan Los Angeles area.
California Department of Fish and Game. EPA
Grant No. R 807 120010.
Roberts MS, Anderson RA, Swarbrick J. 1977.
Permeability of human epidermis to phenolic
compounds. Journal Pharm. and Pharmacol.
29:677-83.
Schaefer H, Zesch A, Stuttgen G. 1982. Skin
permeability. New York: Springer-Verlag.
Scheuplein RJ, Blank IH, Brauner GJ, MacFarlane
DJ. 1969. Percutaneous absorption of steroids.
Journal of Investigative Dermatology 54(1):63-70.
Scheuplein RJ, Blank IH. 1971. Permeability of the
skin. Physiological Reviews 51(4):702-47.
Sutton TJ. 1973. Dermal toxicity and penetration
studies following topical application of three
histamine H2-receptor antagonists with a
comparison with an H1-receptor antagonist.
Toxicology and Applied Pharmacology 50(3):459-
65.
Tregear RT. 1966. Physical functions of skin. New
York: Academic Press.
USDA. 1980. Food and nutrient intakes of individuals
in 1 day in the United States, spring 1977,
nationwide food consumption survey 1977-78,
preliminary report no. 2. Washington, DC: Science
and Education Administration.
USDA. 1983a. Food consumption of households in
the Northeast, seasons and year 1965-66, report
no. 13. Washington, DC: Agricultural Research
Service. August 1972.
USDA. 1983b. Food consumption of households in
the North Central region, seasons and year 1965-
66, report no. 14. Washington, DC: Agricultural
Research Service. September 1972.
USDA. 1983c. Food consumption of households in
the South, seasons and year 1965-66, report no.
15. Washington, DC: Agricultural Research Service.
January 1973.
USDA. 1983d. Food consumption of households in
the West, seasons and year 1965-66, report no.
16. Washington, DC: Agricultural Research Service.
January 1973.
USDA. 1983e. Food consumption of households in
the United States, seasons and year 1965-66.
Washington, DC: Agricultural Research Service.
March 1972.
USDOI. 1973. Outdoor recreation: a legacy for
America. Washington, DC: U.S. Department of
Interior.
USEPA. 1980a. Dietary consumption distributions of
selected food groups for the U.S. population.
Washington, DC: U.S. Environmental Protection
Agency. Office of Pesticides and Toxic
Substances, Office of Testing and Evaluation. EPA
560/11-80-012.
USEPA. 1980b. Water quality criteria documents.
Federal Register, Vol.\45 No. 231, November 28,
1980.
USEPA. 1981. The exposure assessment group's
handbook for performing exposure assessments
(draft report). Washington, DC: U.S. Environmental
Protection Agency.
USEPA. 1985. Superfund public health evaluation
manual. Draft. Washington, DC: ICF, Inc. Prepared
for the Policy Analysis Staff, Office of Emergency
and Remedial Response, U.S. Environmental
Protection Agency. October 1, 1985.
USEPA. 1987. Exposure factors handbook.
Washington, DC: Exposure Evaluation Division,
U.S. Office of Toxic Substances, U.S.
Environmental Protection Agency. Contract No.
68-02-4254, Task No. 83.
Versar. 1984. Methods for estimating concentrations
of chemicals in indoor air. Draft final report. Versar
Inc. Washington, DC: Prepared for the Exposure
Assessment Branch, Exposure Evaluation Division,
132
-------
Office of Toxic Substances, U.S. Environmental
Protection Agency.
Versar. 1985. Exposure assessment for
perchloroethylene. Revised draft report. Versar Inc.
Exposure Assessment Branch, Exposure
Evaluation Division, Office of Toxic Substances,
U.S. Environmental Protection Agency.
133
-------
-------
Appendix B
Possible Exposure Assessment Data Requirements for Uncoltrolled Hazardous Waste
Sites and Index to Variable Terms
135
-------
Table B-1. Possible Data Requirements for Estimation of Contaminant Release and Transport and Exposed Populations
Type of Analysis
Type of Site
Area of Concern
Area Subclass
Parameter
Contaminant release
Contaminated surface soil Particulate release
(includes spills and leaks)
Wind erosion
Unpaved roads
• Soil erodibility index3
• Soil ridge roughness
factor3
• Field length along
prevailing wind direction
^Vegetative cover factor
jteConcentrationsof
contaminantsb
^Volume of contaminated
regionb
.««Silt content
j«Mean speed of vehicles
traversing contaminated
area"
j«Mean weight of vehicles
traversing contaminated
area"
j«Mean number of wheels
of vehicles traversing
contaminated aread
Volatilization
Excavation and transfer of
soil
Short-term release
Long-term release
Runoff to surface water
.e&tfSilt content0 p
•Mean wind speed
• Drop height
• Material moisture content
• Dumping device capacity
• Vapor concentration of
contaminants in soil pore
spaces'
• Depth from soil surface to
bottom of contaminated
regionb
AS Area of contamination11
• Depth of "dry"
(uncontaminated) zone at
sampling timeb
• Concentrations of
contaminants in soil and
in liquid phaseb
..a? Soil porosityb'c
j«Absolute temperatureb'e
• Time measured from
sampling time
• Soil erodibility factor9
• Slope - length factor
• Vegetative cover factord
Erosion control practice
factor"
W\rea of contamination
jsjsSoil bulk density0
.ujsTotal areal concentrations
of contaminants
Release to ground water
- See Chapter 3.6 of
Manual
(Continued)
136
-------
Table B-1. (Continued)
Type of Analysis
Type of Site Area of Concern
Area Subclass
Parameter
Landfill
Volatilization
NO internal gas generation
With internal gas
generation
j«Area of contamination
• Soil porosity0
^Effective depth of soil
cover
• Mole fractions of
contaminants in waste
^Absolute ambient
temperature6
^Absolute ambient
pressure6'11
^Soil bulk density0'1
contaminants in soilb
j»Volume of contaminated
regionb
j»Vapor concentration of
contaminants in soil pore
spaces'
• Area of contamination
Lagoon
Release to ground
water
Volatilization
Contaminant fate
Contaminated
surface soil,
landfill, lagoon
Migration into
ground water
Atmospheric fate
- See Chapter 3.6 of
Manual
jsd-iquid-phase
concentrations of
contaminants
j»Area of contamination
^Absolute ambient
temperature6
j«Volume of contaminated
regronb
- See Chapter 3.6 of
Manual
• Distance from site to
selected exposure point
j»Mean wind speed6
• Relative annual frequency
of wind flow towards point
x6
• Relative annual frequency
of stability class for wind
flow towards point x6
.^Stability classes
(A = unstable, F = stable);
according to Pasquill
classification system6
• Vegetative cover factord
(Continued)
137
-------
Table B-1. (Continued)
Type of Analysis
Type of Site
Area of Concern
Area Subclass
Parameter
Surface water fate
Ground water fate
Saturated zone
Unsaturated zone
Exposed populations
All
General
Contaminated surface
water
Contaminated ground water
^Combined effluent and
stream flow data
Intermedia substance
transfer rate'
j«Widtri of water body'
• Stream velocity'
j-asStream depth'
.^Slope of stream channel'
.easSoil hydraulic
conductivity1*
^(Hydraulic gradient1
..^Effective soil porosity"1
j««Average percolation or
recharge ratem
• Volumetric water content
of soil in unsaturated
zone'
.^Hydraulic loading from
manmade sources''"
• Precipitation rate"'0
j«Evapotranspiration rate
• Runoff rate''"
j«Average depth of
contaminated area"
^Evaporation rate0
• Location of population
.^Number of persons
jieAge/sex distribution
• Recreation patterns
(fishing, hunting,
swimming)
^(Commercial fisheries
present
^Drinking water intake
locations and populations
served
^Drinking water intake
locations and populations
served
f,n
b Some values can be obtained from existing literature.
c For calculation of long-term release ( > 70 years).
d Can be obtained from Soil Conservation Service (SCS) "Soils 5 File" data base.
e Estimated indirectly from site survey information.
f Can be estimated based on existing meteorological station data.
Can be calculated.
9 Can be obtained from SCS office or from existing literature.
Necessary only if diffusion coefficients for toxic components are not available from existing literature.
' Can be measured as an alternative to measuring soil porosity.
J Can be obtained from USGS data.
l< Can be calculated or estimated from Table in Manual.
Can be obtained from USGS or local university geology/hydrogeology departments.
m Can be calculated via equation in manual, or can be obtained from USGS, USDA, NOAA, or U.S. Forest Service.
" Needed to calculate average percolation/recharge rate when not measured at site.
0 Available from local or National Weather Service.
138
-------
Table B-2. Index to Variable Terms
Term
Used
E
1'
K'
C'
L'
V
EVT
k
s
SP
W
w
D p
H
D;
A
csi
p.
dsc
IVIj
T
MW,
MWa
2V,,
2V2
Pa
D'
MW
8
Definition
Potential annual wind erosion soil loss
Soil erodibility index
Soil ridge roughness factor
Climatic factor
Field length along the prevailing wind direction
Vegetative cover factor
Emission factor for vehicular traffic
Particle size multiplier
Sift content
Mean vehicle speed
Mean vehicle weight
Mean number of wheels
Number of days with at least 0.254 mm (0.01 in)
of precipitation per year
Emission rate of toxic component i
Diffusion coefficient of component i
Contaminated area
Saturation vapor concentration of component i
Soil porosity
Effective depth of soil cover
Mole fraction of toxic component i in the waste
Temperature
Molecular weight of contaminant i
Molecular weight of air
Molecular diffusion volumes of toxic contaminant
\/t) and air (V2)
Absolute pressure
Known diffusion coefficient of a compound with
molecular weight and molecular diffusion volume
close to that of the unknown (D,)
Molecular weight of the selected compound
corresponding to D'
Soil bulk density
Units
(mass/area/time)
(dimensionless)
(dimensionless)
(dimensionless)
(feet)
(dimensionless)
(kg/vehicle kilometer
traveled; Ib/vehicle mile
traveled)
(dimensionless)
(%)
(kph; mph)
(Mg; tons)
(dimensionless)
(dimensionless)
(g/sec)
(cm2/sec)
(cm2; areas; ha)
(100 in2)
(g/cm3)
(dimensionless)
(cm)
(g/g)
(K,C)
(g/mole)
(g/mole)
atm
(g/mole)
(g/cm3)
Equation(s) in
which term is used
2-1
2-1
2-1
2-1
2-1
2-1
2-2
2-2
2-2
2-2
2-2
2-2
2-2
2-3; 2-8; 2-9;
2-11; 2-15
2-3; 2-4; 2-5;
2-12
2-3; 2-8; 2-9;
2-11; 2-15; 2-
1 9; 2-21 ; 2-24;
2-25; 2-26;
2-30; 2-32;
2-33; 2-37
2-3; 2-7
2-3; 2-6; 2-12;
3-17; 3-34;
3-35
2-3
2-3
2-4; 2-7;2-10;
2-13; 2-16;
2-17
2-4; 2-5; 2-7;
2-10; 2-17
2-4
2-4
2-4
2-5
2-5
2-6; 2-25;
2-26; 2-27;
3-17
Source
calculated
site data and
literature
site data and
literature
literature
site data and
literature
site data
calculated
see text
site data, SCS
Soils 5 Fife
site data
site data
site data
Figure 2-3
calculated
calculated
site data
calculated
site data; SCS
Soils 5 File
site data
site data
site data
literature
see text
literature and
calculated
site data
see text
literature
site data; SCS
Soils 5 File
(Continued)
139
-------
Table B-2. (Continued)
Term
Used
P
P
R
Q*
Vy
kiG
MWH20
k|G,H20
cs
CB
t
d
D
Hi'
HI
h
td
K,
kiL
MW02
K|_, 02
EM
vc
CC
E
Y(S)E
a
vr
KP
L
S
C
Definition
Particle density
Vapor pressure of the chemical
Gas constant
Vapor concentration of compound i
Mean landfill gas velocity in the soil pore spaces
Gas-phase mass transfer coefficient of
chemical i
Molecular weight of water
Gas phase mass transfer coefficient for water
vapor at 25°C
The liquid-phase concentration of component i
Bulk contaminant concentration in soil
Time measured from sampling time
Depth of dry zone at sampling time
Related to the amount of contaminant i that goes
from liquid to gas phase, and then from gas phase
to diffusion in air
Henry's Law constant in concentration form
Henry's Law constant
Depth from soil surface to the bottom of the
contaminated region
The time at which all contaminant has volatized
from the soil
Overall mass transfer coefficient
Liquid phase mass transfer coefficient
Molecular weight of oxygen
Liquid phase mass transfer coefficient for oxygen
at 25°C
Average release of contaminant i
Volume of contaminated region
Concentration of contaminant i in soil
Total release rate of contaminant i obtained by
summing all above-listed releases of the
contaminant at the site
Sediment yield in tons per event
Conversion constant
Volume of runoff
Peak flow rate
The soil-erodibility factor. Obtained from the
local Soil Conservation Service Office
The slope-length factor
The slope-steepness factor
The cover factor
Units
(g/cm3)
(mm Hg)
(62.3 mm Hg-liter/
k-mol; 8.2 x 10'5
atm-m3/-mol-k)
(g/cm3)
(cm/sec)
(cm/s)
(g/mole)
(g/cm3)
(g/cm3)
(seconds)
(cm)
(cm2/sec)
(dimensionless)
(atm-m3/mol)
(cm)
(sec)
(cm/sec)
(cm/sec)
(g/mole)
(mass/time)
(cm3)
(g/cm3, kg/ha, Ib/acre)
(g/sec)
(metric tons)
(acre-feet, m3)
(ft3/sec, m3/sec)
(commonly expressed in
tons per acre per R unit)
(dimensionless)
(dimensionless)
(dimensionless)
Equation(s) in which
term is used
2-6
2-7; 2-37
2-7; 2-13; 2-16
2-8; 2-9
2-8
2-9; 2-10; 2-16
2-10
2-10
2-11; 2-14;
2-15; 2-19; 2-34
2-11; 2-14; 2-19
2-11
2-11; 2-14; 2-19
2-11; 2-12;
2-14; 2-19
2-12; 2-13
2-13; 2-16
2-14
2-14
2-15; 2-16
2-16; 2-17
2-17
2-17
2-18; 2-19; 2-29
2-18; 3-34; 3-35
2-18; 2-25;
2-28; A-3
2-18
2-20; 2-27
2-20; 2-21;
2-23; 2-24
2-20; 2-21
2-20; 2-24
2-20; 2-30
2-20; 2-30
2-20; 2-30
2-20; 2-30
Source
see text
literature or
estimated (see text)
see text
site data
see text
calculated
see text
calculated
site data
site data
site data
site data
calculated
calculated
literature
site data
calculated
calculated
calculated
see text
literature
calculated
site data
site data
calculated
calculated
see text
calculated
calculated
site data, literature
see Figures 2-4
through 2-6
see Figures 2-4
through 2-6
see text and Table
2-4
(Continued)
140
-------
Table B-2. (Continued)
Term
Used
P
Qr
Rt
sw
CN
Tr
Ss
DS
ec
Kd
PXi
PQi
B
N
Y(S)A
Rr
sd
Dd
LC
U
C0
QJ
b
Ks
K
c
Kw
DC
Dw
Uc
Uw
Definition
The erosion control practice factor
Depth of runoff
The total storm rainfall
Water retention factor
The SCS Runoff Curve Number
Storm duration
Sorbed substance quantity
Dissolved substance quantity
Available water capacity of the top cm
of soil
Sorption partition coefficient
Sorbed substance loss per event
Dissolved substance loss per event
Dissolved or sorbed loss per storm
event
Number of "average" storm events in
70 years
Annual soil loss in runoff
Rainfall and runoff factor
Sediment delivery ratio
Overland distance between site and
receiving waterbody
Contaminant Loading rate
Solubility of solid chemical
Volume loading rate
Soil specific exponential function
Soil hydraulic conductivity
Hydraulic gradient
Hydraulic conductivity of liquid
contaminant in site soil
Hydraulic conductivity of water in site
soil (same as K,)
Density of liquid contaminant
Density of water
Dynamic viscosity of liquid contaminant
Dynamic viscosity of water
Units
(dimensionless)
(in, cm)
(in, cm)
(in, cm)
(dimensionless)
(hour)
(kg, lb)
(kg, lb)
(dimensionless)
(cm3/g)
(kg, lb)
(kg, lb)
(kg, lb)
(dimensionless)
(tons)
(dimensionless)
(dimensionless)
(Ft)
(mass/time)
(mass/volume)
(volume/time)
(dimensionless)
(length/time)
(dimensionless)
(length/time)
(length/time)
(mass/volume)
(ma&volume)
[mass/(length x time)]
[mass/(length x time)]
Equation(s) in
which term is used
2-20; 2-30
2-21; 2-22;
2-24; 2-28
2-22; 2-24;
2-28
2-22; 2-23;
2-24
2-23
2-24
2-25; 2-27
2-28; 2-28
2-25; 2-28
2-25; 2-26;
3-17; 3-19
2-27
2-28
2-29
2-29
2-30
2-30
2-30; 2-31
2-31
2-32; 2-34;
2-37
2-32
2-33; 2-34
2-33; 3-13
2-33; 3-9;3-13
2-33; 3-9
2-35; 3-16
2-35; 3-16
2-35
2-35
2-35
2-35
Source
see text
calculated
National
Climatological Data
Center, Asheville,
NC; USDC (1961)
calculated
Table 2-6
National
Climatological Data
Center, Asheville,
NC; USDC (1961)
calculated
calculated
calculated
(see text)
literature
calculated
calculated
calculated
(see text)
National
Climatological Data
Center, Asheville,
NC; USDC (1961)
(see text)
calculated
calculated
site data
calculated
literature
calculated
Table 3-1 1
site data,
Table 3-8; 3-9
site data
calculated
site data
literature
literature
literature
literature
(Continued)
141
-------
Table B-2. (Continued)
Term
Used
PS
Ap
SH
*
d|
X
C(X)
Q
ay
°z
li
IT
C(X)
W(X)
CA(X)
*A
A,B,C,
D.E.F
C(CL)
Y(X)
C
ce
Qe
Ot
T,
MZ
W
u
dn
Definition
Permeation rate
Permeation constant for polymer
2nd permeation constant for polymer
Permachor value for polymer-
permanent pair
Thickness of liner
Distance from site to selected exposure
point
Atmospheric concentration of substance
at distance X from the site
Release rate of substance from site to
atmosphere
Atmospheric dispersion coefficient in
the lateral (crosswind) direction
Atmospheric dispersion coefficient in
the vertical direction
Mean wind speed
Tho \/£)liic> rii — "^ 141ft
1 1 lc ValUU (Jl O . I *r I o
Average atmospheric concentration of
substance at point X over long term
Relative annual frequency of wind flow
Atmospheric concentration at point X
during stability Class A
Relative annual frequency of stability
Class A for wind flow towards point X
Stability classes (A = unstable,
F = stable) according to Pasquill
classification system
Predetermined critical atmospheric
concentration level
Perpendicular distance from point X on
plume centerline to the C(CL) isopleth
boundary
Concentration of substance in stream
water
Concentration of substance in effluent
Effluent flow rate
Combined effluent and stream flow rate
Intermedia substance transfer rate
Length of mixing zone downstream of
effluent release to stream
Width of water body
Stream velocity
Stream depth
Units
[g-mil/
(100 in2 x day x
cmHg)]
[g-mil/
(100 in2 x day x
cmHg)]
(cc/cal)
(cal/cc)
(mils)
(length)
(mass/volume)
(mass/time)
(length)
(length)
(length/time)
, , . .
(Gtmensioniess)
(mass/volume)
(dimensionless)
(mass/volume)
(dimensionless)
(dimensionless)
(mass/volume)
(length)
(mass/volume)
(mass/volume)
(volume/time)
(volume/time)
(mass/time)
(length)
(length)
(length/time)
(length units)
Equation(s) in
which term is used
2-36; 2-37
2-36
2-36
2-36
2-37
3-1; 3-2; 3-3;
3-7; 3-8
3-1; 3-3, A-1
3-1
3-1; 3-3
3-1
3-1
3-1
I
3-2; A-1
3-2
3-2
3-2
3-1; 3-2
3-3
3-3
3-4; 3-5; A-2
3-4
3-4
3-4; 3-5
3-5
3-6
3-6
3-6; 3-7; 3-8
3-6
Source
calculated
Table 2-7
Table 2-7
Table 2-7; 2-9;
2-10
site data
site data
calculated
calculated
Figure 3-5
Figure 3-6
site data
calculated
site data
calculated
site data
site data
air quality criteria
calculated
calculated
contaminant
release analysis
contaminant
release analysis
site data
site data,
calculated
calculated
site data
site data
site data
(Continued)
142
-------
Table B-2. (Continued)
Term
Used
s.
g
W(X)
W(CL)
W(0)
K
e
v
• p w
q
V
P e
es
^(-15
e
HL
Pr
ET
Qr
EVAP
Get
vea
vi-y
Rd
vd
Koc
foe
KQW
xd
Td
Qd
Definition
Slope of stream bed
Gravitational acceleration constant
Water concentration of substance at
downstream distance X
Predetermined critical water
concentration level
Water concentration of substance
immediately below point of introduction
to stream
Overall aquatic decay coefficient
Exponential function
Interstitial pore-water velocity or
ground-water velocity
Average percolation or recharge rate
Darcy velocity
Soil Effective Porosity
Saturated Water Content soil (equal to
P.)
Wilting Point Moisture Content
Volumetric water content of soil
Hydraulic loading from manmade
sources
Precipitation rate
Evapotranspiration rate
Runoff rate
Evaporation rate
Correction factor for converting pan
evaporation rate to evapotranspiration
rate for turf grass
Correction factor for converting turf
grass evapotranspiration to that for
other vegetative cover
Retardation factor
Retarded velocity of hydrophobic
Partition coefficient for organic carbon
Fraction of organic carbon in soil
Octanol/water partition coefficient
Nomograph factor
Nomograph factor
Nomograph factor
Units
(dimensionless)
(32 ft/sec2)
(mass/volume)
(mass/volume)
(mass/volume)
(time-l)
(length&me)
(depth/time)
(length/time)
(dimensionless)
(dimensionless)
(dimensionless)
(dimensionless)
(depth/time)
(depth/time)
(depth/time)
(depth/time)
(depth/time)
(dimensionless)
(dimensionless)
(dimensionless)
(length/time)
(ml/g)
(dimensionless)
(ml/g)
(dimensionless)
(dimensionless)
(dimensionless)
Equation(s) in
which term is used
3-6
3-6
3-7
3-8
3-7; 3-8
3-7; 3-8
2-36; 3-7
3-10; 3-12;
3-18
2-32; 3-12;
3-13; 3-14
3-9; 3-10;
3-26; 3-27;
3-29; 3-30
3-10; 3-11;
3-28; 3-31
3-11; 3-13
3 11
3-12; 3-13
3-14
3-14
3 14; 3-15
3 14
3 15
3-15
3-15
3-17; 3-18;
3-27; 3-30
3-18
3-19; 3-20;
3-21; 3-22;
3-23; 3-24;
3-25
3 19
3-20; 3-21;
3-22; 3-23;
3-24; 3-25
3-26; 3-29
3-27
3-28; 3-31
Source
site data
calculated
water quality
criteria
calculated
literature, estimated
calculated
site data,
calculated
calculated
site data
site data, literature
site data, literature
site data,
calculated
site data,
calculated
site data
site data, calculated
site data, calculated
site data
Table 3-12
Table 3-13, see
text
calculated
calculated
calculated
site data, literature
literature
calculated
calculated
calculated
(Continued)
143
-------
Table B-2
Term
Used
Dx
Dy
m
ax
ay
Y
A
T,/2
v,2
1
v
vgw
cgw
Mc
cc
IEX
te
1
Bw
F
DEX
AV
PC
DA
C
(Continued)
Definition
Longitudinal Dispersion Coefficient
Transverse Dispersion Coefficient
Aquifer thickness
Longitudinal dispersivity
Transverse dispersivity
Coefficient for decay
Decay constant
Half-life
Volume of liquid chemical released
Average concentration of chemical
contaminant in released liquid
Volume of contaminated ground water
Average concentration of contaminant
in ground water
Mass of solid waste
Concentration Expressed as Mass
Fraction
Inhalation exposure
Duration of exposure event
Average Inhalation rate
Average adult body weight
Frequency of exposure event
Dermal exposure
Skin surface area available
Dermal permeability constant for
subject contaminant
Dust adherence
Contaminant concentration
Units
(lengthytime)
(Iength2/time)
(length)
(length)
(length)
(dimensionless)
(1/time)
(time)
(lengths)
(mass/lengths)
(lengths)
(mass/length3)
(mg)
(dimensionless)
(mg/kg/day)
(hours/event)
(m3/hr)
(70kg)
(number/lifetime)
(mg/kg/day)
(cm2)
(cm/hr)
(mg/cm2)
Equation(s) in
which term is used
'6-2V. '6-21;
3-28: 3-29;
3-31
3-28; 3-31
(Fig. 3-8)
(Fig. 3-8)
(Fig. 3-8)
(Fig. 3-8)
(Fig. 3-8)
(Fig. 3-8)
3-32
3-32
3-32; 3-33;
3-34; 3-35
3-32; 3-33
3-33
3-33
A-1
A-1; A-2
A-1
A-1; A-2; A-3
A-1; A-2; A-3
A-2; A-3
A-2; A-3
A-2
A-3
Source
calculated
calculated
site data
literature
literature
literature
calculated
calculated
site data
site data
site data
site data
site data
site data
calculated
estimated
Table A-3
Eq A-1
estimated
calculated
estimated
Table A-4
See text
144
-------
Appendix C
Data Management Forms
This appendix presents master copies of data
management forms designed for use when applying
the various analyses described in this manual. The
forms are intended to provide easy, consistent
organization of the results of each analysis
component in the human exposure assessment
process (qualitative analysis, quantitative contaminant
release analysis, etc.) for ready use in subsequent
analytical components. In addition, these forms will
also organize exposure assessment output in a form
most useful for conducting a risk assessment
(executed following and based on the results of the
exposure assessment) as well as the development of
a site Endangerment Assessment for enforcement
purposes.
These forms are included as master copies, that
should be photocopied for use in a given site
investigation. In many cases, a number of copies of
certain forms will be required to tabulate all results of
the exposure assessments. For example, Form No. 7:
Exposure Integration requires that the exposed
population segment be logged into the upper left
corner of the form, and exposure information for that
population segment be entered into the remaining
columns for each chemical to which the population is
exposed. If four distinct exposed population segments
are affected at the site, four copies of the form will be
required.
145
-------
Form 1: Qualitative Exposure Analysis Site Name:.
Date:.
Analyst:,
On-site Release Release Potentially Exposed
Chemical Source Likelihood Magnitude* Release Mechanism Receiving Medium Population Segment Exposure Mechanism
CD
-------
Form 1. Qualitative Exposure Analysis (Continued) Site Name:
Date:.
Analyst:.
On-site Release Release Potentially Exposed
Chemical Source Likelihood/Magnitude* Release Mechanism Receiving Medium Population Segment Exposure Mechanism
6.
7.
10.
"Code each source as to: (1) Likelihood of release and (2) Potential magnitude of release. Use H, M, L (high, medium, low) designation and provide a letter code for likelihood and
magnitude, each separated by a "/".
-------
Form 2. Quantitative Contaminant Release Data Site Name:=
Date:.
Analyst:.
1.
2.
3.
4.
5
7.
8.
9.
10-
Frequency of Short-
On-site Release Short-term Release term Release Rate Long-term Release
Chemical Source Receiving Medium Rate (units) (units) Rate (units)
-------
Form 3. Quantitative Environmental Fate Data
Site Name:_
Date:_
Analyst,
Chemical
Affected Medium
Short-term Environmental
Concentration (units)
Long-term Environmental
Concentration (units)
1.
7.
-------
form 4: Quantitative Exposed Populations Data
Site Name:_
Date;.
Analyst:.
Chemical
Affected Medium
Exposure Mechanism
Population Segment Number of Persons Potentially
(• denotes sensitive population) Exposed
i.
2.
en
o
3.
4.
5.
-------
Form 4: Quantitative Exposed Populations Data (Continued) Site Name:_
Date:_
Analyst:,
Population Segment Number of Persons Potentially
Chemical Affected Medium Exposure Mechanism (* denotes sensitive population) Exposed
6.
cr
8.
10.
-------
Form 5: Short-terrm Exposure Calculation Site Name:_
Date:.
Analyst:.
12 3 4567 8
Short-term Daily Exposed
Short-term Exposure Events per Time Body Mass Time Period Exposure (mg/kg/day) Segment
Chemical Exposure Mechanism per Event (units) Period (kg) (days) [3x4E5x6] (number of persons)
ui
ro
5.
-------
Form 5: Short-terrm Exposure Calculation (Continued) Site Name:.
Date:_
Analyst:_
12 3 4567 8
Short-term Daily Exposed Population
Short-term Exposure Events per Time Body Mass Time Period Exposure (mg/kg/day) Segment
Chemical Exposure Mechanism per Event (units) Period (kg) (days) [3x4E5x6] (number of persons)
6.
01
CO
§>_
9.
-------
Form 6: Long-terrm Exposure Calculation Site Name:.
Date:.
Analyst:.
12 3 4 56 7 8
Time Period Long-term Daily Exposed Population
Long-term Exposure Events per Time Body Mass (2.56 x 104 Exposure (mg/kg/day) Segment
Chemical Exposure Mechanism per Event (units) Period (kg) days) [3 x 4E5 x 6] (number of persons)
1.
01
-------
Form 6: Long-terrm Exposure Calculation (Continued) Site Name:.
Date:.
Analyst:.
12 3 4567 8
Time Period Long-term Daily Exposed Population
Long-term Exposure Events per Time Body Mass (2.56 x io4 Exposure (mg/kg/day) Segment
Chemical Exposure Mechanism per Event (units) Period (kg) days) [3x4E5x6] (number of persons)
6.
en
01
10.
-------
Form 7: Exposure Integration (Continued)
Site Name:.
Dale:.
Analyst:.
Exposure
Population
Segment
Chemical
Exposure Mechanisrr
Short-term
Long-term
Number exposed
6.
7.
en
-vl
9
10.
-------
------- |