5527
FINAL DRAFT REPORT
AN INTEGRATED GEOGRAPHIC APPROACH TO
DEVELOPING TOXIC SUBSTANCE CONTROL STRATEGIES
August, 1981
by
Joseph Fiksel: Project Leader
Marcos Bonazountas
Helen Ojha
Kate Scow
of
Arthur D. Little, Inc.
under
U.S. EPA Contract 68-01-6160
and
Randall Freed
Leslie Adkins
of
Versar, Inc.
under
U.S. EPA Contract 68-01-6271
Michael Alford: Project Manager
Toxics Integration Project
Office of Policy and Resource Management
U.S. Environmental Protection Agency
Washington, D.C. 20460
85225
An hurl.) Li; tie. inc.
-------
U,3. Environmental Protection Agency
-------
ACKNOWL EDGMENT
The authors wish to thank the following people, without whose
cooperation this report would not have been possible: Michael
Alford of the EPA, for his guidance and dedicated concern; Alvin
Morris, Dee Ortner, and their colleagues at EPA Region 3 for their
support and advice; Michael Slimak andMichael Callahan of the EPA
for their contractual assistance; George Harris of Arthur D. Little,
Gayanch Contos of Versar, and the many staff members of both firms
whose enthusiasm and hard work contributed to the success of this
project.
‘ \r hur L) i t e. .nc
-------
LIST OF TABLES
TABLE DESCRIPTION PAGE
2-1 Summary of Data Requirements and Possible Sources
of Data for a Geographic Toxics Study 2-11
2-2 Outline of Considerations Used in Identifying
Inportant Pathways and Media 2-22
3 — I Sources of Data on Land-Destined Disposal of Toxics 3-22
3-2 General Waste Disposal Data Sources 3—24
3—3 Residual Data Sources 3-25
3-4 Spill Data Sources 3-26
4-1 Features of Off-the-Shelf EPA Air Models for Toxic
Substance Exposure Assessments 4-15
4-2 Input Data Categories of Air Compartment
Simulations 4—17
4-3 Unsaturated Transport Soil Zone Models 4-20
4-4 Partial List of Saturated (Groundwater) Pollutant
Transport Models 4-21
4-5 Partial List of Unsaturated/Saturated (US) Soil
Zone Models 4—22
4-6 Representative List of “Watershed” Models 4-23
4-7 Soil Compartment Environmental Parameters (Exposure
Pathway Matrix) 4-26
4-8 Data Sources for Soil Compartment Environmental
Pathway Analyses 4-28
4—9 Selected Water-Body Models 4-31
4-10 Selected Multi-Media Models 4-36
4-11 A Typical Water Scenario 4-47
4—12 Selected Data Requirements of Pathway Analyses 4—49
4—13 Data Bases Overview; Pathway Analyses 4—51
5-1 Reviews of Human Exposure Assessment Methodologies 5-2
5-2 Specific Human Exposure Assessment Methodologies 5-7
5—3 Sources of Data on Human Intake Rates and Parameters 5—8
5-4 Commonly Used Intake Rates for Ingestion of Selected
Exposure Media 5-9
5-5 Inhalation Rates for Humans by Activity and Sex 5—12
5-6 Examples of Nonhuman Receptors to be Considered
in an Exposure Assessment 5-16
5-7 Sources of Data for Characterizing Human Receptor
Populations 5—21
Summary of Individual Intakes and Population Exposure 6—12
Factors for the Four Study Pollutants
7-i Stanc ards and Criteria for Tc dc Substances 7-6
7-2 Illustration of Alternative Control Strategies or
Lead Pollution in Pivers. Includinc indirect
Pat h iays
. rthur D Little, lnc
-------
LIST OF TABLES (Cont’d)
TABLE # DESCRIPTION PAGE
7-3 Illustration of River Loadings and Resultant Ambient
Concentrations 7-19
7—4 Illustration of Changes in River Loadings •Required
to Meet Ambient Targets 7-20
7-5 Treatment Processes Acting on Hazardous Components
in Waste Streams of Various Physical Forms 7-30
7-6 TSCA Section 6 Regulatory Authorities 7-45
7—7 Examples of Options for Implementing Toxic Substances
Controls in a Geographic Study 7-49
8—1 Estimated Resource Requirements for a Geographic 8—8
Study Excluding Field Sampling
V
\rthur L) little. Inc
-------
LIST OF FIGURES
FIGURE # DESCRIPTION PAGE
1-1 Flow Diagram for Geographic Study 1-6
1-2 Flow of Infor nation for Exposure Assessment 1-10
2-1 Pentagram Framework for a Geographic Study 2-2
2-2 Environmental Pathways of Toxic Substances 2-20
2-3 Results of Kanawha Valley Initial Scan 2-26
4-1 System Design Schematic for Environmental Modeling 4-3
4-2 Environmental Pathways of Toxic Substances 4-8
6-1 Human Exposure Routes - Lead 6—3
6-2 Relationship Between Sample Pollutant Concentration
arid Distance from Source 5-14
6-3 Exposure Characteristics of Nickel Emissions 6-15
6—4 Population Distribution for Ranges of Annual Carbon
Tetrachloride Intake Rates 6-16
7-1 Overview of Control Strategies Task: Problem
Identitication 7-4
7-2 Exposure Routes for Lead 7-12
7-3 Overview of Control Strategies Development Task:
Investigation of Technological Control Options 7-24
7-4 Overview of The Control Strategies Development Task:
Evaluation of Control Options 7-37
7-5 Percentage of Reaches Exceeding 1.9 ug/L vs. Minimum
Total Cost for Each Strategy - Mean Flow Case 7-39
8-1 Activity Flow Chart for Geographic Study 8-9
vi
rthur L) Htt etic.
-------
TABLE OF CONTENTS
CHAPTER DESCRIPTION — PAGE
1 INTRODUCTION 1-1
1.1 Background 1-1
1.2 Objectives of the Geographic Approach 1-3
1.3 Site Selection Requirements 1—4
1.4 Summary of Methodology 1-5
1.4.1 Overview of Geographic Study 1-5
1.4.2 Exposure Assessment Approach 1-8
Incorporation of Health Effects Data 1-11
l.6j Administrative Issues 1-12
2 PRELIMINARY PROCEDURES 2-1
2.1 Introduction 2-1
2.2 Study Boundary Definition 2-4
2.3 Data Acquisition 2-9
2.4 Pollutant Selection 2-15
2.5 Procedures for Initial Scan 2-18
2.5.1 Initial Scan of Sources and Emissions 2-18
2.5.2 Environmejital Pathways Scan 2-19
2.5.3 Exposure Route Scan 2-23
2.6 Output of Initial Scan 2-25
3 IDENTIFICATION OF SOURCES AND QUANTIFICATION OF
EMISSIONS 3-1
3.1 Introduction 3-1
3.2 Define Data Needs 3-1
3.3 Review and Assemble Existing Data 3-5
3.3.1 Airborne Emissions 3-6
3.3.2 Aqueous Discharges 3-11
3.3.3 Discharges to Land 3—21
3.4 Evaluate Adequacy of Existing Data 3-27
3.5 Design and Implement Program for Collecting New Data 3-29
3.6 Synthesize and Interpret Data 3-33
4 ENVIRONMENTAL. PATHWAY ANALYSIS 4-1
4.1 Objective and Scope of This Section 4-1
4.2 Introduction 4-1
4.3 Potential Pathways of Toxics 4-4
4.3.1 General Analytical Approach 4-4
4.3.2 Examination of Background Information 4-6
4.3.3 Quantitative Pathway Analysis 4-7
4.4 Mathematical Modeling 4-12
4.4.1 General Overview 4-12
4.4.2 Air Compartment Modeling 4-13
4.4.3 Soil Compartment P1odeling 4-18
Li.4.4 Water ComDartmt nt ! odeling 4-29
4.4.5 Multimedia Modeling 4- 5
4.4.6 Other Modeling Issues 4-39
Arthur I) I ittle, Inc.
-------
TABLE OF CONTENTS (Cont’d )
CHAPTER DESCRIPTION PAGE
4.5 Evaluation of Monitoring Data 4-40
4.5.1 Definition of Needs 4-40
4.5.2 Review and Compilation of Data 4-41
4.5.3 Evaluate Adequacy of the Data 4-42
4.5.4 Design and Implement Plans for New Data Collection 4-42
4.5.5 Synthesize and Interpret Data 4-42
4.5.6 Model Validation 4-45
4.5.7 Sensitivity Analysis and Scenarios 4-46
4.5.8 Monitoring Data Acquisition (Data Base) 4-48
4.6 Summary and Conclusions 4-52
4.7 References 454
5 RECEPTOR/EXPOSURE ROUTE ANALYSIS 5-1
5.1 Introduction 0 5-1
5.2 Exposure Routes and Exposure Media 5-3
5.2.1 Introduction 5-.3
5.2.2 Human Exposure 5-5
5.2.2.1 Introduction 5-5
5.2.2.2 Ingestion 5-6
5.2.2.3 Inhalation 5-10
5.2.2.4 Dermal Absorption 5-11
5.2.3 Exposure of Other Biota 5-14
5. 3 Receptor Population Characteri zati on 5-1 7
5.3.1 Human Populations 5-17
5.3.1.1 General Population 5-17
5.3.1.2 Special Subpopulations 5-19
5.3.2 Other Biota 5-20
6 EXPOSURE ASSESSMENT 6-1
6.1 Introduction 6-1
6.2 Purpose of an Exposure Assessment 6-1
6.3 Methodology 6-4
6.3.1 Organization of Input Parameters and Data 6-4
6.3.1.1 Pollutant Concentration Data 6-5
6.3.1.2 Standard Fate Models 6-5
6.3.1.3 Exposure Pathways Not Usually Considered in
Fate Models 6-6
6.3.1.4 Linkage of Fate Models to Receptor Distribution
Models 6-8
6.3.1.5 Occupational Exposure 6-10
6.3.2 Calculation of Individual and Population Exposure
Levels 611
6.3.3 Analysis of Exposure Data 6-13
6.4 Applications of Exposure Assessment Results 6-17
6.5 References 6-19
Arthur 1) i . ttI ihc
-------
TABLE OF CONTENTS (Cont’d )
CHAPTER DESCRIPTION PAGE
7 DEVELOPMENT OF CONTROL STRATEGIES 7-1
7.1 Introduction 7-1
7.2 Problem Identification 7-3
7.2.1 Interpreting the Results of the Exposure Assessment 7-3
7.2.2 Information Needed for Problem Identification 7-7
7.3 Identification and Selection of Feasible Control Points 7-9
7.3.1 Introduction 7-9
7.3.2 Purpose and Methodology of Control Point Selection 7-10
7.4 Determining a Set of Satisfactory Control Levels 7-17
7.4.1 Introduction 7-17
7.4.2 Purpose of Determining Satisfactory Control Levels 7-17
7.4.3 Methods for Determining Satisfactory Control Levels 7-18
7.5 Investigation and Selection of Technological Control
Options 7-22
7.5.1 Introduction 7-22
7.5.2 Developing a Set of Technical Control Options 7-25
7.5.3 Estimating the Cost and Effectiveness of Each Control
Option 7-31
7.6 Evaluation of Control Options 7-35
7.6.1 Overview of the Evaluation of Control Options 7-35
7.6.2 Selecting an Efficient Set of Control Options 7-36
7.6.2.1 Methods for Selecting an Efficient Set 7-38
7.6.3 Analyzing Environmental and Economic Impacts 7-40
7.7 Implementing Control Options 7-43
7.7.1 Introduction 7-43
7.7.2 Existing Control Authorities 7-44
7.7.3 Need for an Inventory of Control Mechanisms 7-48
7.7.4 Selecting Regulatory Alternatives 7-52
7.7.5 Identifying Non-Regulatory Alternatives 7-54
7.8 Conclusions 7-56
7.9 References 7-59
8 PRACTICAL CONSIDERATIONS 8-i
8.1 Feasibility of the Geographic Approach 8-1
8.2 Potential Limitations of the Methodology 8-2
8.2.1 Adequacy of Data 8-2
8.2.2 Validity of Models 8-3
8.2.3 Scope of Analysis 8-4
8.3 Study Timetable and Resource Requirements 8-
8.4 Conclusions 8-7
Appendix A Pollution Flow Model A-i
I Little, inc
-------
1. INTRODUCTION
1.1 BACKGROUND
The Environmental Protection Agency, through its Toxics Integration
Program, is currently investigating a number of alternative regulatory
approaches for dealing with toxic substances in the environment. Pilot
projects have been conducted to examine the feasibility of regulating
toxic pollutants on an industry—by—industry, chemical-by-chemical, or
geographic basis. The goal of these projects is to develop feasible and
cost—effective regulatory approaches that will rectify specific environ-
mental problems while minimizing the economic burdens of compliance. By
addressing multi—media environmental issues in an integrated fashion,
it may be possible to manage toxics exposure problems in a more efficient
manner than the traditional “corr nand and control” approach which separately
addresses each medium.
This report presents a preliminary framework for the geographic
approach to development of toxics control strategies. It represents a
general methodology which was developed in conjunction with a pilot tudy
of the Kanawha River Valley in West Virginia. Details of the pilot study
are presented in a separate report.* The underlying concept of the
geographic approach is that control strategies can be developed on a
site—specific or region—specific basis, possibly with assistance from
state and local governments and other groups, such as industry associations.
To pursue this type of approach, the EPA would require a site selection
procedure for identifying candidate geographic areas with significant
potential for environmental toxics problems. The feasibility of such a
procedure is currently being investigated by the EPA, but is not included
in the scope of this report. Thus, the general methodology described here
will assume that a specific study area has already been designated.
*
“An Integrated Geographic Study of Potential Toxic Substance Control
Strategies in the Kanawha River Valley, West Virginia”, Arthur D. Little,
Inc. , Final Draft, August 31, 1981.
Arthur D Little. !nc
-------
The concept of a geographic approach is not new to the EPA; several
earlier studies have attempted to assess the impact of toxic substances
in specific local areas and to investigate the need for controls. For
example:
• The EPA performed a study in 1978 to evaluate the levels of
pollutants in air and water for the Beaumont, Texas/Lake
Charles, Louisiana area, with emphasis on collection Of mon-
itoring data.
• EPA Region IV organized a geographic study of toxic substance
sources and pathways for the Memphis, Tennessee area in 1980,
partially in response to public complaints of pollution from
waste dump sites. The same Region is currently initiating
a carefully-planned study of the Louisville, Kentucky area,
with participation from governmental and citizen groups.
A good deal of useful information was obtained from these previous studies,
although it is not the intent of this report to review them in detail.
The present geographic methodology builds upon these experiences, but is
unique in that it addresses multiple chemicals, sources, pathways, and
receptors within a unified technical framework. The methodology is
designed to be predictive, in the sense that it identifies potentially
significant pathways from pollutant sources to receptors, and allows
quantitative evaluation of the cost/benefit tradeoffs of alternate
control strategies that seek to reduce these exposures. The integrative
multi—media approach ensures that control of a pathway in one medium
does not overlook the potential displacement of toxic substances into
other media which are not sufficiently controlled. By focusing upon a
specific geographic area, the EPA can achieve a richness of descriptive
detail that is not possible with a broader national scope of investiga-
tion, and will encourage appropriate local solutions to local problems.
I—’
Arthur D Litt’e, Inc.
-------
1.2 OBJECTIVES OF THE GEOGRAPHIC APPROACH
The geographic approach to regulation of toxics in the environment
provides an alternative to regulatory programs based on national or
industry-specific considerations. Under the geographic approach a selected
local area would be examined as a whole for potential environmental
problems. The investigation would include all sources of toxics in the
area and all possible routes of exposure for humans and other biota. If
problems are detected, control strategies would then be devised to mitigate
these problems in an efficient manner. The objectives of the geographic
approach may be stated as follows:
• To select a small number of sites across the U.S. which appear
to have the greatest likelihood of environmental toxics problems.
•. For each specific site, to select toxic substances of concern,
to evaluate their sources, inter-media pathways, and population
exposures, and thus to identify any existing or potential
problems.
• For each toxics problem designated, to evaluate the feasibility,
the costs, and the benefits of alternate control strategies,
which may exploit either Federal, State, or local authorities.
• To provide a comprehensive data base and analytic capability
for ongoing environmental surveillance and protection in the
study area.
The scope of the present methodology is limited to analysis of exposures
to toxic substances, without explicit quantification of health effects.
Also, attention is confined to exposure arising due to the presence of
toxics in the ambient environment; occupational and consumer use—related
exposures are not addressed. These are not inherent limitations to the
geographic approach, but reflect the current emphasis of the EPA Toxics
Integration Procram. Further discussion of these limitations in scope
is provided in Section .2.3.
—j
Arthur D Little. fnc
-------
1.3 SITE SELECTION REQUIREMENTS
In order for a geographic approach to be implemented on a regular
basis, certain realignments would be required in the conduct of EPA’s
existing programs and in the allocation of resources. Due to the inten-
sive technical and data reauirements of a geographic study, no more than
a few such studies could be in progress during any given period of time
within a Region. This implies the need for a priority-setting mechanism;
that is, a site selection procedure that would identify likely candidates
for geographic areas to be studied in detail. It will, be important to
ensure that this procedure is appTied in. a consistent and unbiased fashion
across different regions. Criteria for site selection would probably
include the following factors:
• Significant toxic levels in the ambient environment
• Presence of numerous potential sources of toxics (including
abandoned sites)
• Proximity Of substantial population groups to toxics sources
• Evi dence of heal th effects or envi ronme 1 tal damage
The development of a site selection procedure is a complex undertaking
whenever additional sites need to be selected. A preliminary outline
of such a procedure has already been developed under to Toxics Integra-
tion Program, and its feasibility is being explored.
due to the multiplicity of pollutants, source categories, and environ-
mental scenarios that must be considered. It is expected that the
procedure will involve several iterations, incorporating both subjective
assessments and computerized ranking methods to arrive at a final list
of sites. The States will no doubt play a leading role in this process,
and it will also be important to utilize the EPA Regions’ intimate
knowledge of each specific area within their authority. The site selec-
tion procedure would have to be flexible enough to accept a wide variety
of inputs, yet sufficiently well-defined that it could be repeated
1-4
Arthur D Uttle 1n
-------
1.4 SUMMARY OF METHODOLOGY
1.4.1 Overview of Geographic Study
The technical methodology for conduct of the geographic study is the
central concern of this report, and is described at length in the remain-
ing chapters. There are three major phases in the approach proposed
here, and they are depicted in simplified terms in Figure 1-1. These
three phases are sun narized individually below:
Phase 1 : Data acquisition, study boundary definition, initial
scan of significant enviornmental pathways, and selec-
tion of pollutants. (This phase is described in
Chapter 2.)
Phase 2 : Detailed exposure assessment, including quantification
of source emissions, modelling of chemical fate and
transport pathways, and quantification of receptor
exposure routes in various media. (This phase is
described in Chapters 3, 4, 5, and 6..)
Phase 3 : Identification of potential toxic substance problems,
collection of additional data, and evaluation of costs
and benefits of alternate control strategies. (This
phase is described in Chapter 7.)
The purpose of Phase 1 is to set priorities and lay the foundation
for the more detailed work in Phase 2. The outcome of the second phase
is a description and quantification (where possible) of potential
exposure levels by pollutant and receptor category. Moreover, the
origins of these exposures can frequently be traced back to the probable
pollutant sources by means of the environmental pathways analysis. At this
juncture certain problem areas may be identified that are due to excessive
environmental loadings or ambient levels. Several different courses of
action are possible in Phase 3:
Arthur D Little Inc
-------
FIGURE 1-1
FLOW DIAGRAM FOR GEOGRAPHIC STUDY
PHASE
1
PHASE
2
3
Development of
Control Strat i s
? HA SE
-
-o
-------
• Where clear violations of existing standards or regulations
have been detected, enforcement actions may be initiated.
• Where potential problems are suggested by the exposure
assessment, additional data (e.g., field sampling or health
effects data) may be collected to verify these findings.
• Where potential problems are of sufficient gravity, the
Centers for Disease Control (CDC) may be consulted regard-
ing the development of a public health protection plan
for imminent hazards.
• Where the possibility of long-term human health effects is
suggested, e idemiological studies may be initiated for
specific pollutants; the CDC can also be helpful in this
regard.
• For those problems not requiring imediate action, and for
which sufficient data are available, consideration can be
given to the relative advantages of different control
strategies.
This last step will be the focus of the Phase 3 methodology described
in Chapter 7. There will in general be a large array of control options
available to address a specific toxic substance problem, depending upon
the nature of the problem and the degree to which sources can be identi-
fied. State and local control options as well as Federal ones have been
considered, and a framework for selecting cost-effective strategies is
presented. However, due to the immense variety of situations that may
be encountered, the methodology is best understood by means of the
Kanawha Valley pilot study example. (See footnote on page 1-1.)
i—i
Arthur D Little. Inc
-------
Finally, once a geographic area has been thoroughly studied and the
necessary controls implemented, the study report would provide a compre-
hensive basis for ongoing surveillance and revision of control strategies.
It would be the responsibility of the EPA Regional Office to maintain
cognizance of changes in the relevant source configuration, receptor
distribution, or environmental conditions, and to coordinate with State
or local agencies in responding to these changes as necessary. Chemical
inventories, pathway models, and population exposure indices developed
in the course of the geographic study could be maintained and updated
periodically with a low level of effort, thus providing appropriate
tools for ensuring that adequate toxics controls are maintained in the
future. Many of the normal Regional activities could be adapted to
coordinate with this evolving geographic “profile”, so that redundancy
in collection or interpretation of data would be minimized. In this way,
the study would not become obsolete, but would remain a useful and pragmatic
instrument for carrying out EPA policy.
1.4.2 Exposure Assessment Approach
The exposure assessement portion of a geographic s tudy constitutes
the second Phase, as defined above. During this phase, a detailed
quantification is performed of the sources, pathways, and receptor expo-
sures for the toxic substances selected during Phase 1. A large part
of the data required for the exposure assessment will have been gathered
during the first Phase, so that the exposure assessment work will involve
mainly analytic efforts and computer modelling. However, additional data
collection may be necessary for specific inputs to various tasks, such
as meteorological information for the air pahtways analysis or population
data. An overview of the methodology is presented here, and then Chapters
3 through 6 describe each of the individual tasks.
In Phase 2, for those pathways and chemicals considered important
as a result of the initial scan, a more detailed data analysis is per-
formed. The analysis is guided by the pentagram franewOrk shown in Chapter
2, Figure 2—1. Five separate tasks are perfo iied, roughly corresDondinc
Arthur D Litde. Inc.
-------
to the five matrices shown in the pentagram framework. Starting from the
pollutant axis, these five tasks can be traced by moving counter-clockwise
within the framework. Figure 1-2 shows the flow of information among
these tasks.
• Task 1--Identification of Pollutant Sources (see Chapter 3)
For the list of pollutants being considered, an inventory
is performed of the various categories of sources, including
industrial, commercial, residential, and other sources where
these pollutants may be produced, used or found as by-products.
• Task 2 —— Quantification of Emissions (see Chapter 3)
The environmental emissions or discharges from each of the
source categories and for each of the pollutants considered
is quantified, where possible. In this way the total envi-
ronmental loading of pollutant emissions is established for
each of the receiving media considered.
• Task 3 —— Environmental Pathways (see Chapter 4)
For a selected number of pollutants that are discharged into
the various receiving media, their fate and transport in the
environment is investigated through modelling, monitoring
data or other means. As a result, the environmental levels
of these pollutants are estimated, where possible, for ambient
air, surface water, drinking water, biota, and other environ-
mental compartments that are pertinent to receptor exposure.
• Task 4 -— Receptor Exposure Routes (see Chapter 5)
For each of the receptor categories of concern, including
humans, aquatic organisms or other organisms in the food
chain, their exposure to the various environmental compart-
ments addressed in the previous task is quantified, where
possible. Exposure is expressed in terms of the extent nd
1—9
Arthur D L t:1e Inc.
-------
FIGURE 1-2
FLOW OF INFORMATION FOR EXPOSURE ASSESSMENT
POLLUTANT
uJ
PRESENCE
(TASK
1)
RECEIVING
MEDIUM
uJ
EMISSI
ONS
C l)
(TASK
2)
RECEIVING
MEDIUM
F-
-J
TOTAL
RELEASES
RECEIVING
MEDIUM
F=
UJ
ENVIRONMENTAL
PATHWAYS
(TASK 3)
COMPARTMENT
EXPOSURE
ROUTES
(TASK LI)
1
COMPARTMENT
F-
AMBIENT
LEVELS
•
-J
-
POLLUTANT
-
U.J
-
1—i
EXPOSURE
LEVELS
(TASK 5)
1—
A 1’) I
-------
frequency of contact with the environmental compartments
discussed above through inhalation, ingestion and absorption.
• Task 5 -— Exposure Assessment (see Chapter 6)
As a result of the previous four tasks exposure pathways
are established between sources and receptors. The final
task is concerned with estimating, where possible, the
quantity or concentration of each pollutant to which the
various receptor categories may be exposed.
1.5. INCORPORATION OF HEALTH EFFECTS DATA
Although the exposure assessment described above can be revealing in
terms of specific population exposures to toxic emissions, it may be argued
that the final criterion for control strategy selection should be related
to health effects rather than exposure. Since different substances may
have different effects and different levels of potency, the health impacts
of exposure to these substances may vary considerably. In particular,
it is possible to compare the potential long-term effects of various
substances in rough quantative terms, using the dose/response extrapolation
techniques that were adopted by EPA ’s Carcinogen Assessment Group. In
this way, the reduction in exposure that might be achieved by a particular
control strategy can be evaluated in terms of the corresponding risk
reduction that is predicted for the exposed population. Such a risk
assessment can provide the appropriate common denominator for balancing
exposures to different substances in different environmental media.
When incorporating health effects data, it is important to note the
limited accuracy of available risk estimation techniques. For example,
the following qualifications must be kept in mind when utilizing carcin-
ogenic risk estimates:
• The human carcinogenicity of most substances is not established,
and species differences may exist between human responses and
those of laboratory animals.
— •1
.Ar:hur D Ho
-------
• Estimation of equivalent human doses requires the use of scaling
factors whose accuracy is open to question.
• The shape of the dose/response curve at low doses is unknown,
and different dose/response models can yield widely differing
res ul ts.
Assuming that laboratory data can indeed be applied to humans, and assuming
a particular dose/response relationship, exposure assessment results can
be used to predict the potential incidence of cancer in an exposed pop-
ulation. Though there may be considerable uncertainty in such predictions,
they can serve as a useful guide to problem identification and policy
formulation.
1.6 ADMINISTRATIVE ISSUES
The purpose of this report is to present an appropriate technical
methodology for performance of a geographic study. It is important to
note, however, that for a geographic study to be successfully implemented,
it is essential that a suitable management plan be developed for carrying
out the technical work. Due to the complexity of the multi-media approach
and the potential concern of state and local interest groups, a consider-
able amount of planning and coordination will be necessary before launching
a geographic study in a particular area. Some of the main issues that
need to be resolved include:
• Authority and approval——the responsible agencies for performance
of the study must be identified, and approval must be obtained
from any groups (such as local governments) that nay be affected
by the conduct or findings of the study. In particular, clearance
must be obtained, if necessary, for use of restricted data.
1
Arthur D L ttie, !nc.
-------
• Technical resources--the participants in the study, both manager-
ial and technical must be identified, and sources of funding must
be established. Due to the complexity and sensitive nature of a
geographic study, it will be advisable to form a study team that
draws upon Federal, State, and local resources. This approach may
facilitate the acquisition of site—specific data, such as local
agency records and files.
• Public information——when a geographic study is initiated, local
governments, industries, and the general public will have to be
notified øf the study. From a. public relations point of view,
it is important that the EPA clearly articulate the goals of
the study - namely, to ensure the continuing protection of public
and environmental health. Rumors or misunderstandings may create
false beliefs about the existence of ii uninent hazards, which will
serve only to obstruct the conduct of a rational, scientific
study. Therefore, the initial planning and organization of the
study is crucial to its ultimate success and credibility.
Though it is not the intent of this report to address these administrative
issues, they will be no less important than methodological issues in the
implementation of a geographic aoproach. Adoption of this approach as an
ongoing EPA program could fundamentally alter the Agency’s allocation of
resources and institutional practices. Therefore the potential benefits
and obstacles associated with this approach need to be carefully examined.
i—13
Arthur D LHe, Inc
-------
2. PRELIMINARY PROCEDURES
2.1 INTRODUCTION
This section describes the procedures that are necessary during
Phase 1 of a geographic study. The purpose of Phase 1 is first to define
a geographic boundary for the study, and then to develop a rapid under-
standing of the potential environmental problems within the geographic
area by scanning the available data concerning emissions, ambient levels,
and exposures. Relevant data must be collected and interpreted in order
to determine what the significant exposure pathways might be, and to
identify any data gaps that need to be resolved. The framework for this
phase is provided by the pentagram structure shown in Figure 2-1. This
framework allowed simultaneous consideration of pollutants, source cate-
gories, receiving media, environmental compartments, and receptor categories
in order to identify potentially important pathways of exposure. This
framework also provides the basis for the detailed exposure analysis
performed in Phase 2. For the purposes of initial scanning, a qualitative
identification is made of pollutants within source categories, of emissions
from source categories to receiving media, of the fate of pollutants
within the various environmental compartments of concern, and of the
exposure of various receptors to these environmental compartments. Those
chemicals and pathways which are considered most significant on the basis
of the available data will then be subjected to more careful scrutiny in
Phase 2.
An important issue during Phase 1 is the availability of data for
the various analytic tasks of Phase 2. For most geographic areas there
will be significant data gaps that will limit the completeness and
accuracy of the exposure assessment. Data requirements and potential
data sources are discussed in Section 2.4, but the available data will
most likely have to be supplemented by a program of new data acquisition.
It is anticipated that new data will be required in one or more of the
following categories:
• Emission rates or environmental loadings of major point and
non—point sources for air, surface water, or land disposal.
2—1
- — i —-
-------
?CLLUTANTS
FIGURE 2-1
PENTAGRAM FRAMEWORK FOR A GEOGRAPHIC STUDY
00
Arthur D Little, Inc
-------
• Ambient concentrations of selected toxics in air, surface
water, groundwater, drinking water, sediment, or soil for
different seasons of the year.
• Concentrations of toxics in fish or animal tissue, in food,
in crops, and possibly in human tissue.
Thus a significant field sampling effort may be necessary during
the course of a geographic study, in order to assemble an adequate data
base. Though a long time period would be required to capture seasonal
variations, the sampling program could be conducted in parallel with
the exposure assessment. In fact, findings about potential high
exposure scenarios could influence the choice of sampling locations, and
conversely the discovery of unusual toxics levels could alter the focus
of the exposure assessment. The detailed design of such a sampling
program would depend upon site-specific circumstances, and therefore is
not addressed in this report. However, guidelines for the data require-
ments are presented in Section 2.4 and in the si ibsequent Chapters. An
efficient means of acquiring field data might be to incorporate the study
area within one of the national monitoring programs onducted routinely
by EPA, FDA or other agencies.
Arthur D L!ttle. Inc.
-------
Ver ar .
2.2 STUDY BOUNDARY DEFINITION
After a problem area has been selected for a geographic toxics study,
the study area boundaries must be delineated in order to allow for a syste-
matic analysis. The proper selection of study boundaries is critical to
the entire geographic study because the sources of toxics and exposed
populations that are included in the study area will directly affect the
findings. Study boundary delineation requires intuitive as well as objec-
tive judgment, because the information that would be useful is often not
available until after the boundaries have been tentatively selected. Study
boundaries will generally need to be modified for the different phases of
the exposure analysis and control strategies development because of
differences in the nature and goals of each phase. In practice, the study
boundaries determined initially may be modified later in the study as more
information becomes available. A tentative emissions inventory study area
will be defined prior to the initial scan, based on available information.
These boundaries may be modified on the basis of information collected in
the initial scan, and may again be changed to incorporate sources of toxics
discovered later in the data collection and analysis. The study areas for
the environmental pathway analysis and the exposure assessment will depend
on the results of the emissions inventory. Similarly, the area considered
for control strategies will depend on the results of the other analyses.
The important factors to consider when selecting study area boundaries are
presented below for each of the major phases of a geographic toxics study.
Sources zd Ertrzss-i ons Inventory
The objective in delineating a study boundary for the inventory of
sources and emissions of toxics is to encompass all areas containing
sources of toxics while excluding areas that are not potential sources
within the general area designated for study. In the ideal situation, the
study area would consist of a geographic area in which toxics produced are
self—contained and which receives no significant imports of toxics from
locations outside the study area boundaries. This situation is rare,
however; in practice, imports and exports of toxics must be estimated or
assumed negligible.
- ‘.4
-------
Ver ar .
A tentative study area boundary, delineated prior to the initial
scan, will depend on the information available to the investigators at the
start of the study. If the area was designated a problem area on the basis
of known sources of toxics (e.g., industries, waste disposal sites), the
tentative boundaries will be determined on the basis of locations of the
potential sources. If, on the other hand, the bases for concern are
ambient or epidemiological data suggesting toxics of unknown origin,
consider the transport modes of the toxics in order to include all
potential sources of toxics within the study area boundaries. During the
initial scan, toxics data for the entire tentative study area must be
sought, as well as other relevant data that might be the basis for
modification of the study area boundaries. Using the data collected in the
initial scan, refine the study area boundaries, if necessary, to better
suit the needs of the study.
In order to maximize accuracy and efficiency, physical factors that
affect the collection or analysis of data must be considered in delineating
study area boundaries. For instance, if preliminary data suggest that
non—point source runoff is a significant source of toxics in the study
area, the emissions inventory study area boundaries may have to be defined
by subwatershed, because non—point source pollutant loads are most
accurately estimated on a subwatershed basis, whereas other source/emission
analyses may be flexible. Similarly, if perusal of the initial ambient
data suggests that a number of undeveloped subwatersheds are not signifi-
cant toxics sources, delete these subwatersheds from the study area in
order to save time and effort. In addition to physical factors,
anthropogenic factors must be considered in the delineation of the source
and emissions study area; in cases where there are significant imports of
toxics via natural (e.g., air, surface water, groundwater) or man—related
(e.g., truck, train) transport, the geographic area for which data must be
collected will be larger than the designated study area.
-------
Consider political and socioeconomic factors when selecting study
area boundaries for the emissions inventory, but only after practical and
scientific requirements have been met. The aforementioned physical factors
that affect data collection and analysis, and the actual distribution of
potential sources of toxics are more important than political and socio-
economic factors, because toxics problems are rarely restricted to one
city, township, or county. In the cases where a political boundary meets
the scientific needs of the study, however, the use of this boundary may
facilitate data collection and analysis. For example, if the study area
boundary coincides with a county boundary, the researchers can save time
and effort by not collecting data on the excluded county.
E’nviroivnentai P tu Jay Analysis
The study area for the environmental pathway analysis will depend on
the areas designated for the emissions inventory and the exposure assess-
ment. The boundaries for this analysis are determined based on considera-
tion of the results of the emissions inventory, the initial scan of
envirorinental data (geological, topographical, meteorological, hydro-
logical), monitoring data, and population data. The emissions inventory
will indicate the areal extent of sources of toxics. Use the environmental
data in conjunction with the demographic data to make general projections
of the areal distributions of the populations of interest that might be
exposed to the identified sources of toxics. Ambient monitoring data, if
available, are particularly helpful in determining study boundaries because
they provide direct information on the environmental pathways of toxics
from the sources. Because the purpose of the environmental pathway
analysis is to trace the movement of toxics from source to receptor popula-
tions, the modeled study area must encompass both the emissions inventory
study area and the population exposure study area (discussed below).
Easu2’e As n
The appropriate study area boundary for the assessment of exposure in
a geographic study will not necessarily coincide with tne study area for
2—6
-------
Ver ar .
the emissions inventory. The objective in delineating the exposure study
area is to include all areas containing populations potentially affected by
toxics, while excluding from consideration the unexposed populations. To
some extent, the study area for the exposed population will be dictated by
the history of the particular study. If epidemiological data suggested a
toxics problem in a certain population, then the study area is best defined
on the basis of the geographic location of that population (e.g., a city
and its suburbs). Alternatively, if the toxics study was initiated because
of the existence of a number of potential sources of toxics in the general
area, the boundaries for the exposure assessment will be delineated only
after the emissions inventory has been completed and the data relating to
transport of toxics have been examined. As stated previously, tentative
study boundaries for the exposure assessment will be determined prior to
the environmental pathway analysis, because the pathway analysis study area
depends on the areal extent of populations that are expected to be exposed
to toxics. The use of available monitoring data is recommended, because
these data indicate areas where toxics exposure may occur. It is recom-
mended that the exposure study area be modified if the environmental path-
way analysis suggests that the exposure of some populations in the tenta-
tive study ‘area is negligible and/or the exposure of the populations not
included in the tentative study area is significant. For example, if the
modeling suggests that toxic concentrations in ambient air are high in a
residential area previously thought to be unaffected by a distant source,
expand the study area to encompass that residential area. Likewise, if the
modeling predicts that toxic concentrations are negligible in a recrea-
tional lake that was included in the tentative exposure assessment study
area boundary, the lake can be omitted from any further consideration with
regard to exposure in order to save time and human resources.
In practice, the optimal study area for analysis of population
exposure may be highly irregular, and may in some cases be represented by a
2-7
-------
Ver ar.
central area and several satellite areas. For instance, a town outside the
main boundaries of a study area will be included as a satellite” to the
main study area if its drinking water is obtained from a source within the
main study area. In the hypothetical case concerning a highly industria-
lized area adjacent to a river, the available data on toxics emissions and
environmental transport may dictate that the exposure study area include a
long narrow segment of river and adjacent floodplain downstream of the
industrial sources. Boundaries may also be refined to be made compatible
with census bureau delineations. It should be noted that, in some cases,
human or wildlife populations that are not permanent residents of a study
area must be included in exposure analysis; population and activity data
for commuters, vacationers, and migratory animals will have to be estimated
from data other than local census bureau statistics.
In some cases, it may be worthwhile to consider the existing politi-
cal boundaries in delineating study area boundaries for exposure analysis.
As in the case of e uissions inventory study area boundaries, however, use
municipal, county, or state boundaries only if they satisfy the technical
needs of the study. Consider political boundaries in delineating the study
area boundary in cases where the available census data correspond to these
bounda ri es.
Cont2’o 1 Strategi. .es
The study area boundaries for the development of control strategies
will generally encompass the geographic areas considered in the analysis of
emissions, fates, and population exposure and will depend on the results of
these analyses. Because the implementation of some control strategies will
depend on local, regional, and national political and socioeconomic
factors, the effective study area may correspond to municipalities,
counties, or states in which the toxics problem is located. Thus, the
geographic area to consider in the development of control strategies may be
as small as the areal extent of toxics sources and affected populations or
as large as a state or a region.
2-8
-------
2.3 DATA ACQUISITION
Data acquisition begins as soon as the toxics integration study is
underway. This subtask is crucial to the entire geographic study, because
the acquired data are the foundation for all subsequent analyses. It is
also the limiting factor for progress in other preliminary rocedures,
because the delineation of study boundary, pollution selection, and the
initial scan are all dependent upon the available data base. Ideally, most
of the data are acquired prior to the initiation of Phase 2 (detailed
expost’re assessment). In practice, however, data acquisition is an ongoing
process and late” data must be incorporated into the study whenever it
arrives.
The goal of the data acquisition subtask is to identify and acquire
all available information relevant to toxics sources and emissions,
environmental pathways, exposure, and toxics control in the study area. To
fulfill this goal, considerable general background information is necessary
in addition to data specifically related to toxics. For example, informa-
tion on the geology, hydrology, soils, meteorology, demography, and regula-
tions pertaining to the area is vital to the analysis of toxics problems
and the development of control strategies. It is recommended that site—
specific data be used wherever possible in the geographic study, because
these data are generally most accurate. When site—specific data are not
available, however, surrogate (i.e., regional or national) data must be
acqui red.
Because the data acquisition subtask is concurrent with the selection
of pollutants, the list of toxics to be considered in the study may not be
finalized until some of the acquired data have been examined. Therefore,
information relating to all possible toxics candidates must be sought until
the final pollutant list is known; thereafter, data acquisition should
concentrate exclusively on the toxics selected for the study. For the
2—9
-------
Ver ar .
selected toxics, it is crucial that the investigators familiarize them-
selves with general published information on the production, use,
emissions, environmental fate, and environmental and health effects in the
early stages of data acquisition. With this information, compile a list of
all possible sources (e.g., industry, commercial)/discharge modes (e.g.,
point source water, solid wastes) for each toxic. During the course of
data acquisition, investigate each potential source/discharge mode with
respect to the study area.
Most of the toxics information will be available through various
government agencies. Because the specific agencies responsible for
collecting and compiling the data used in a geographic study of toxics vary
from place to place, it is not possible to develop an exact “blueprint” to
follow in the acquisition of data. It is recommended that investigators
contact officials at the local, state, and federal level in order to
determine which agencies can provide the necessary information. A general
summary of the data requirements and data sources for each task of the
geographic study is presented in Table 2—1. More detailed descriptions of
the data needs and data sources for each phase are given in Chapters 3-7.
In all geographic studies there will be substantial gaps in the
available site—specific data that must be filled by surrogate (regional or
national) data. Even for areas where toxics have been relatively well
studied there will be gaps in the information on some toxics sources,
emission rates, environmental pathways, and exposure routes. While some of
the information can be obtained by acquiring new data (e.g., conducting a
monitoring program), this option is costly; consequently, surrogate data
will often be used in geographic studies. For this reason, it is
2—10
-------
Table 2— 1 . Summary of Data Requirements and Possible Sources of i)ata for a Geographic Toxics Study
Pollutant_Sources and [ missions
• Computerized surveys of chemical manufacturers.
, Surveys of menufacturers and users.
Computerized monitoring data.
C FliIlssions permits, registratIons, data, and
Inventories.
• Detailed land use informetion.
‘ Generic inforinalion on sources and emission.
i)emographjc data.
tivironmenla I lab
- Cliiieltc data: i.e., precipitation, wind direction,
wind speed, iemperaturo, pressure, etc.
• lly ) ologic data: i.e., surface and ground water
disiribution, qt.ian Ity, etc.
o Soil data: i.e., soil classification, geology,
soil permuability, etc.
land use data.
Chemical fate data: physical, chemical and
hiological properties of pollutants.
Monhtori(I(J dala: concunIra lions of pollutants
in envlrolmk)nta I cnirtmonts.
Federal : EPA Headquarters (computerized data bases and technical reports)
EPA Regional Offices (Air Quality Monitoring Branch, Water Quality
Monitoring Branch, I zardous Waste Task Force, Enforcement Branches,
Environmental Emergency Branch (Surveillance end Analysis Division),
Pesticides Branch. Also, U.S. EPA Water Qua lity Analysis Branch for
N(* P and PON Studies and Effluent Guidelines Division.
U.S. Army Engineer District
U.S. Department of Agriculture (Geological Survey, Office of Surface
Mining),
River Basin Commissions
State : department responsible for environmental affairs or natural resources,
health department, office responsible for solid and/or hazardous
wastes, office responsible for air pollution control.
agriculture department, mining agencies
Others : direct contact with sources
local soil conservation district office; county, township, or
municipal offices responsible for environmental affairs, natural
resources, health, and solid waste; chambers of commerce; published
literature
Federal : National Weather Bureau, NC A, USGS, USDA (Soil Conservation Service),
U.S. EPA Water Quality Analysis Branch (STORET/TOXET data bases,
REACH, stream gage data), EPA technical documents, Bureau of Mines.
State : Departments responsible for environmental affairs or natural
resources, departments responsible for geological and economic
surveys, soil conservation department, air pollution control, water
development authority
Others : universities, county and city offices, published literature
Class of Data
Possible Sources of Data
S
z
N)
— —a
— . 4
-------
Table 2- 1 . Summary of Data Requirements and Possible Sources of Data for a Geographic Toxics Study
(continued)
Exposure Routes
I Population (tile for humans broken down into regional
dlstrftiul ion, sex and age classes, special groups.
0 Occupational statistics.
Date on food consumption patterns, sources of drinking
water and t.npulations supplied, recreational patterns.
infornL tion on locations of areas with potentially
high exposure (beaches, playgrounds, industries fran
pruvious tasks).
uS.) intorii ation for specific products important In
terms of exposure.
• Dwumantalion of actual incidents of exposure.
• human ‘tissue and other biological media monitoring
surveys.
e health effects surveys.
Species end population surveys for fish end wildlife.
• Surveys of natural plant species and agricultural
statistics.
Endangered and threatened species lists — Federal and state.
: onlrol Siralegies
1e(hrai and slate air emissions standards.
Federal arid state water effluent standards
and gui doline .
Federal and state hazardous and solid waste
regu iat loris.
Federal and state ambient air standards.
1 -edaral arid state ambient water standards.
OSIIA standards.
ixposure cr1 tone.
Technological control options.
Federal : Bureau of Census, Department of Agriculture, Department of Interior —
other offices (endangered species, U.S. Fish and Wildlife, Bureau of
Land Management), HEW (NIOSH, CDC, FDA Fish and Crop i4,nitorlng), U.S.
EPA Water Quality Analysis Branch (fish kills), TSCA substantial risk
notices, USGS Water Quality Alert.
Regional EPA: Environmental Emergency Branch, Water Development
authority, Office of Special Programs, Exposure Evaluation Division
State : departments responsible for natural resources (parks and recreation,
‘water resources, wildlife resources), environmental health, community
health services, geological and economic survey, agriculture.
Others : state and local planning departments, local conservation groups (i.e.,
Audubon), universities, community health organizations, contact with
manufacturers or local retailers ADI Risk Assessment Methodology
Report, past ADL risk assessments, other literature.
Fedora! : EPA technical reports, Occupational Safety and Health Administration,
Food and Drug Administration, Code of Federal Regulations and Federal
Register
State : code of regulations
Class of Data ‘ Possible Sources of Data
7’
a
a
Others : universities, published literature, contact with generators of toxics
-------
Ver ar .
recomended that the EPA assemble generic data on sources, emissions,
environmental pathways, and exposure routes in the event that the geo-
graphic approach to toxics control is adopted. Much of this information
would be of use to the industry-by—industry and chemical-by-chemical toxics
integration activities as well. If such surrogate data is easily acces-
sible to personnel conducting geographic studies, the time spent tracking
such information would be kept to a minimum. Furthermore, knowledgeable
EPA personnel could assure quality control by selecting the “best” set of
surrogate data from all available data for use in geographic studies.
Aerial Reconnai ssance
The initial scan can be aided by the use of aerial reconnaissance and
related photographic imagery. Several data sources covering different parts
of the country at different times are available through the EROS Data Center
in Sioux Falls, South Dakota, the U.S. Army Corps of Engineers and the
U.S. Geological Survey. Photo—coverage of some areas may go back to 1945
and can be used to establish a historical imagery file for a selected site.
Such a file can be used to identify abandoned pollution sources such as
old dumps, landfills and areas of soil filling, industrial impoundments,
tanks and tank farms, drum-related sites, auto junkyards and stacks.
The Environmental Photographic Interpretation Center (EPIC), a field
station of EPA’s Office of Research and Development and Environmental Moni-
toring Systems Laboratory, Las Vegas, has developed a computer-aided image
interpretation capability. They have prepared many historical land-use
maps using photographic imagery for various site—specific studies. Their
approach is linked to USGS guidelines for Level I, II and III land-use
categories, and permits the use of high resolution U-2 overflight imagery
which is produced by NASA through an inter-agency agreement. Current
overflight coverage through a U—2 flight can be obtained for an entire
state at a fairly reasonable cost. The entire state of Pennsylvania was
recently done in this regard and the State of West Virginia is a tarQet
for an overflight proçram some time in 1981.
-------
The imagery data can be interpreted through microscopic and stereo-
scopic viewing of the imagery by persons expert in environmental, land-
use, and imagery analysis. The information is extracted and transferred
to overlaid map sheets, employing the USGS land-use codes. The result-
ing information for a single site can best be used as a comprehensive -
information base detailing overall development in and around an area.
Specific pollution sources can be analyzed in relation to land-use
changes over the period of time covered by the historical imagery.
Long-term trends in land—use can be mapped and used to predict the pres-
ence of potential pollution sources due to changing land-use patterns.
Such data collection technologies could be usefully employed in identi-
fying the boundaries of a geographic study area and in confirming the
presence of potential pollution sources not detectable through field
observation (e.g., tanks behind a stand of tall trees on private land),
or through national or state data bases.
2-14
Arthur D Litde. inc
-------
2.4 POLLUTANT SELECTION
The number of toxics likely to be present within any one geographic
area will probably be too many to allow individual consideration of each
one in a regional study. Therefore, it is important to have an efficient
and effective procedure for selecting the toxics that merit study.
The procedure should be well-defined and reproducible so that it can be
used at any location, but flexible enough to accept the variety of in-
formation needed to select the group of pollutants, This section is a
discussion of a recommended approach for pollutant selection.
The objective of this approach is to identify a group of pollutants
which meet at least one of the following criteria for a specific geographic
area:
• the most significant direct or indirect releases
of toxics in the geographic area including industrial,
municipal, commercial, domestic and natural sources.
• evidence of substantial ambient levels in air, surface
and ground water, or bibta.
• potential for signiftcant acute or chronic health
problems to humans and other biota due to:
-- toxicity;
—— likelihoodanddegree of human and other biotic
exposure via inhalation, ingestion of drinking
water and food, dermal contact, etc.
The final list selected will reflect the characteristics of the
geographic region under study. For example, in a heavily industrialized
area, the pollutants chosen are likely to be those released by production
and manufacturing processes, discharges from POTWs, and inadvertent
releases from fossil fuel combustion. In a rural agricultural area
the pollutant list is likely to include substances released from use
activities e.g., pesticides, contaminants in fertilizers.
2-15
Arthur D Little. Inc.
-------
The ultimate goal of the exposure assessment is to develop site-
specific control strategies based on the study’s conclusions. Therefore,
to streamline the process, the final list may contain certain pollutants
that are representative of a group of substance. For example, a recomended
control for an industry discharging a particular organic substance might be
implementation of biological waste water treatment; this control would also
be effective in removing other biodegradable pollutants present in the waste
stream. The selection of one pollutant which adequately represents an entire
group will require detailed preliminary knowledge of the local industries’
waste characteristiCs and treatment efficiencies. However, thi s approach
may overlook differences in exposure or toxicity between related substances.
The number of pollutants to select for study in a geographic exposure
assessment will depend on budget, fime and personnel constraints as well
as on the particular focus of the study. In some cases, it may be appro-
priate to collect some categories of data on a large number of pollutants,
particularly source emissions data which are derived from the same
general sources for most pollutants. The number of nollutants which can
be addressed in the environmental pathways analysis and exposure assess-
ment, however, may be a subset of the initial list due to the substantial
time and labor costs as well as the chemical-specific nature of the data
retrieval and modelling required by these tasks.
Presented below is an outline of a step-by—step methodology for
pollutant selection. The ordering of the steps may be reversed in some
cases, e.g., input from local Sources may enter at an earlier stage,
serving more as a preliminary screening than as a reviewing function.
1) Compilation of a preliminary list of pollutants based
on review of available government surveys of local
industries and other sources, state documents on sources
and emissions based on registration and pen it data, other
available Federal, region, state and local infoyination pub-
lished, unpublished, or based on personal conTnunication.
2- 3
ArthurD L rt1einc
-------
2) For each pollutant on the preliminary list, compilation of
general information of the following type:
• number of industrial and other direct sources;
• identification of other indirect sources based
on nonlocal information (e.g., Intermedia Toxics
List rationale, EPA Criterion Documents, EPA
Effluent Guidelines Documents);
• quantification of emissions if data are available;
• readily available ambient air and/or water
monitoring data for the geographic location
(e.g., from STORET, state air quality surveys)
• designation on EPA special toxics lists, e.g.,
Intermedia Toxics List, Priority Pollutant
List, Carcinogen Assessment Group’s “hit list”,
Regional Toxic Substance Policy Comittee
Screening results, and other lists.
• anecdotal local information on spills, reports
of problems and other episodes.
3) Review by EPA Region staff and any other interested
local reviewers.
4) Selection of a list of pollutants to be addressed in
the geographic study based on preceeding criteria;
the number of pollutants selected will be based on
the particular study’s level of effort.
It should be emphasized again that the criteria implemented in the
methodology for pollutant selection should not be rigid because of the
inadequacies and inconsistencies typical of readily available data.
The procedure should also allow the inclusion of quai tative infor-
mation which may exist only for a few of the pollutants on the preliminary
lis ,or which may be available cnly in selected geograchfc areas.
Ar h’ 9 L tt!e. !r c.
-------
2.5 PROCEDURES FOR INITIAL SCAN
2.5.1 Initial Scan of Sources and Emissions
The initial scan of sources and emissions begins as soon as the
information from the data acquisition effort becomes available. The
objective of the scan is to determine the relative importance of the
identified source/discharge modes and to identify other potentially
significant source/discharge modes that may have been previously dis-
counted. Some source/discharge modes previously assumed to be important
will be dropped from further consideration if emissions are insignificant.
The output from this effort will be a preliminary matrix of significant
sources and discharge modes, compilation of estimated emissions for each
source/discharge mode, and a rough map showing the location of the iden-
tified sources. This output will be used as input to the detailed source
and emissions inventory (Chapter 3). In addition, the output of the initial
scan of sources and emissions will be used as preliminary input to the study
boundary definition, pollutant selection, environmental pathway analysis,
and determination of exposure routes.
Before the initial scan takes place, the investigators should familiar-
ize themselves with the toxics to be studied. Of particular importance with
regard to sources and emissions is the acquisition of general information on
the production, use, and environmental emissions of the toxics. With this
information, a list of all possible sources of the toxic can be compiled,
together with the possible discharge modes. All of the source/discharge
modes included in this “master list” should be investigated with respect
to the study area at some point in the initial scan or during the quanti-
fication of emissions, in order to ensure that no significant sources of
toxics are neglected. In the event that the geographic approach to toxics
control is undertaken on a large scale, it may be advisable for the EPA to
assemble such a “master list” for each toxic of concern. This list and
supporting information could be distributed to groups conducting geographic
studies of toxics, thereby avoiding the duplication of time and effort.
7_12
-------
Ver ai .
The following types of information should be sought in the initial
data scan, as they will generally provide useful information on sources
and emissions of toxics:
• Detailed land use information.
• Location of all potential sources of toxics (using the “master list”
as a guide).
• A list of all industries in the study area.
• Qualitative information that can be used to confirm the presence or
absence of toxics in identified potential toxics sources.
• Quantitative data that can be used to estimate emission rates from
potential sources.
• Data on ambient levels of toxics in all media for sites outside the
study that can be used to estimate imports of toxics to the study
area.
Whenever possible, the data collected in the initial scan should be
site-specific (i.e., pertaining to sources and emissions in the study area).
When site-specific information is not available, however, regional data
should be compiled; generic (i.e., regional or national) data are recomended
if no other data are available. Some possible sources of data are described
in Section 2.3 above.
2.5.2 Environmental Pathways Scan
Environmental pollution is mainly man-made and originates from point
or non-point sources. Pollutants originating from these sources follow
various environmental pathways and result in pollutant concentration levels
in the three major physical environmental media: air, soil and water.
Figure 2-2 is a schematic presentation of potential pathways of toxic sub-
stances originating from a source and discharged into any of the three
receiving media (designated as compartments) of air, soil and water. In
this figure, sediment is part of the water in the soil compartment,
groundwater is part of the soil compartment, and biologic degradation can
take place in any compartment. Each major environmental compartment may
further export pollutants to a compartment of the same nature (i.e., out-
of-basin transport), aitnough such details are difficult to present in
this type of schematic figure.
2-19
-------
Near Fierd, Far Field Paths
• #‘s indicate priority of pathway
• #1: Direct Discharge Pathway
• #2-4: Indirect Discharge Pathway (primary, secondary, potential)
• Degradation of toxic substances can occur in any compartment
FIGURE 2. 2 : ENVIRONMENTAL P4THW.4YS OF lOX IC UESTAT CES
-------
Pollutant pathways can be classified into two major categories:
(1) Direct discharge pathways created by the direct release
of substances into the three major environmental compart-
ments; for example, source—to-air pathway; and
(2) Intermedia transfer pathways created by the migration
of substances between the three previously described
compartments; for example, the air-to-surface pathway.
In addition, direct discharge pathways may consist of a near-field and
a far-field subpathway (see Section 4.4). Potential pathways of toxic
substances have been numbered from 1 through 11 in Figure 4—2.
Individual and integrated modeling of these pathways is discussed in
Chapter 4.
Compartments can be prioritized by examining the characteristics
of the specific site. Intermedia transfer pathways can then be iden-
tified as having high, moderate, or low importance. The crucial com-
partments and pathways will depend on the nature of the sources (e.g.,
location, release mechanism and media), the pollutants present at a
site (e.g., soluble, reactive), the exposed population (e.g., size,
location), and the nature of the site (e.g., mountainous, coastal).
Table 2-2 outlines some specific considerations previously dis-
cussed. Examination of these factors for a site will lead to a conclu-
sion as to which media must be modeled, and which pathways and media
are likely to be of secondary importance. For instance, in the pilot
study of the Kanawha Valley* the major discharges were from industrial
sources by stacks (air) and river outfalls (water). Thus, air and water
were important media, and important pathways were air-to—surface—to—
water and water—to—air. Since the drinking water in this region came
*
see footnote, pace 1 1.
2-21
D L:rde. !nc
-------
Table 2-2
OUTLINE OF CONSIDERATIONS USED IN IDENTIFYING
IMPORTANT PATHWAYS AND MEDIA
1. Nature of Emission Sources
• Types of sources: point, non—point, area, line, volume
• Mode of emissions: continuous constant rate intermittent batch
operations, continuous operation-but variable emission rate
• Range in emission rates: significance of individual sources
vs. source aggregation
• Release mechanisms
• Release medium
2. Nature of Emitted Substances
• Toxic substance emitted directly
• Toxic substance precursor emitted
• Importance of physical, chemical, biological removal mechanisms
• “Order” of chemical removal mechanisms: inert, first—order
(simple exponential decay), or higher order
3. Nature of Exppsure Populations
• Distribution
• Near—field/far—field of sources
• Long—term (chronic) exposure vs. short-term (acute) exposure
• Statistical average or mean conditions vs. extreme “worst-case”
conditions
• Direct exposure to toxic in media vs. loss to other media and
pathways
• Routes of exposure, inhalation, ingestion, etc.
4. Nature of Site
• Geography: mountains, coastal plain, valley, etc.
• Special meteorological features: wind channeling in valleys,
sea breezes, extremes of temperature, rainfall
• Representativeness of available meteorological data
• Significance of “imported” or “background” levels of toxic sub-
stance with respect to boundaries of study area and air basin
• Soil types
• Biota types and distribution
2-22
Arthur D L :tie.!nc
-------
primarily from water resource supplies outside the study area, contamina-
tion of groundwater was of less concern. In another site, for example,
in a conununity with drinking water wells near a hazardous waste disposal
site, the soil and groundwater media and pathways might have been of
primary importance.
Once the major pathways and media are identified, the required
temporal and spatial distribution of the pathway analysis results can
be specified. These considerations will depend on the nature of the
sources (continuous, batch operation), the nature of the exposed popu-
lation (area of affected population, use patterns of media), and the
focus of the study (chronic versus acute effects). After the areas of
study and the type of results needed have been determined, the use of
actual models and analysis of existing data can be designed to satisfy
the study objectives.
2.5.3 Exposure Route Scan
It is unlikely that it will be possible to identify and quantify
every potential exposure pathway for toxics in a particular region
within the scope of a study. Therefore, one of the first steps in a
regional exposure assessment is to narrow the focus of the study to
encompass only the critical exposure pathways for each pollutant. The
final decisions will usually be based on limited, preliminary data for
the area; however, in most cases these data will be adequate to support
the initial screening.
Several types of data are useful in identifying exposures. The
particular set of pollutants selected for the regional study will deter-
mine whether exposure through product use (e.g., of pesticides) is
important and in what environmental media the pollutants are likely to
concentrate following release (e.g., sediment for substances with pro-
pensity for adsorption). The types of pollutant sources present--
industrial, comercial and domestic——will also influence the exposure
pathways which are most significant in a region. For example, sources
Arthur D Litt!e. Inc
-------
with ambient temperature releases from low stacks and vents are more
likely to have an impact on the immediately surrounding population than
are high temperature sources with tall stacks. Following specification
of the study area boundaries, some of the immediately apparent regional
characteristics, such as gross population distribution, the presence of
recreational areas, or the type of water supply, will also aid in
identification of critical exposure pathways. For example, in studying
trace metal emissions, remotely—located smelters may be of less interest
than coal—fired power plants sited in populous areas due to the lower
exposed population associated with remote locations.
Concurrently with the selection of critical exposure routes, a
decision should be made regarding which receptor categories to consider
in the study area and at what level of detail they should be grouped.
This choice is once again dependent on the variables discussed above.
A quick survey of the types of receptors (populations by species),
habitat or community, etc.) present in the area will indicate the
presence of sensitive or endangered conmunities or habitats, economic-
ally or otherwise important receptors (crops, game species), large frac-
tions of potentially sensitive receptors (elderly, children, pregnant
women), as well as other information.
Once the critical exposure pathways and receptors have been identi-
fied, the next step in the assessment is to gather and evaluate more
detailed quantitative information for these elements to use in esti-
mating exposure levels. Chapter 5 discusses a general methodology for
developing intake and population data.
2-24
D Lit e. iflC
-------
2.6 OUTPUT OF INITIAL SCAN
The results of the initial scan will be an identification of the
selected pollutants, sources, environmental media, and receptors to be
considered in the subsequent analysis. These results can be conveniently
displayed within the pentagram framework, as illustrated in Figure 2-2,
extracted from the pilot study of the Kanawha Valley (see footnote,
p.1-i). Four pollutants were selected in this study to be investigated
in detail for possible linkages between source emissions and receptor
exposures. These were lead, chloroform, carbon tetrachioride, and
vinyl chloride. Emissions of these substances are indicated for four
major source categories: industrial, POTW, coal mines, and POT14’s.
The dots in Figure 2-3 indicate that lead emissions were considered
in all four categories, chloroform emissions were considered in both
industrial and POTW sources, and the other two chemicals were identified
only in industrial emissions.
The next segment of the pentagram, moving counterclockwise, indicates
the receiving media that were considered for each source category. Indus-
trial emissions were addressed for air, water, and land; POTW emissions
for water only; coal mine emissions for water and land, and transportation
emissions for air only. It is important to note that these restrictions
reflect priorities that were established as a result of the initial
scan; there were in fact other less significant routes of toxics entering
the environment that were not addressed in the detailed Phase 2 analysis.
This is true of the next segment also, in which environmental fate and
transport mechanisms were considered. As indicated by the dots in Figure
2—2, attention was focused on transfers from air and land to surface
water; from surface water to drinking water, biota, and air; and also
deposition from air to land. Of course, the mass fraction of pollutant
remaining in each receiving medium was a primary consideration.
The next segment of the pentagram indicates the exposure routes
which were given highest priority for further investigation. Exposure
of humans was considered with respect to all five environmental
2-25
Arthur D L :r]e. Inc
-------
POLLUTANTS
FIGURE 2-3 : RESULTS OF KANAWHA VALLEY INITIAL SCAN
2-26
A k ._iTh T-nr
ttfli iL 1
-------
compartments—-air, surface water, drinking water, land, and biota. For
fish and wildlife, the only exposure routes considered were via surface
water and other biota. Finally, the fifth segment shows that all three
receptor categories may be exposed to any of the four pollutants, since
a multiplicity Of pathways exists from sources to receptors. Thus,
each dotted cell in the pentagram defines a unit of technical analysis
that will be pursued in the s.ubsequent Phase 2 effort. This device
serves both to display the output of the initial scan and to guide the
conduct of the exposure analysis.
2-27
Arthur D L tde. Inc.
-------
ver ar .
3.0 IDENTIFICATION OF SOURCES AND QUANTIFICATION
OF EMISSIONS
3.1 INTRODUCTION
The cornerstone of an effective exposure assessment is a thorough
identification of all significant sources of the toxics of concern and
quantification of emission rates via each discharge mode. These components
form the basis for all the succeeding phases of the exposure assessment.
In addition, because most control strategies are oriented to control of the
source, accurate characterization of emissions is vital to the overall
success of the geographic approach.
The pentagram framework for exposure assessment, described in Chapter
1, provides a conceptual framework in which identification of sources and
quantification of emissions occupy the first two grids in the pentagram.
In practice, the two components are addressed concurrently; all sources
with non—zero emission rates are potentially significant. Because of the
close association of the source identification and emissions quantifica-
tion, they are discussed together in the following pages.
The principal subtasks required to complete a sources and emissions
inventory are:
1. Define data needs
2. Review and assemble existing data
3. Evaluate adequacy of existing data
4. Design and implement program for collecting new data
5. Synthesize and interpret data
Each of these elements are discussed in detail in the following subsec-
tions.
3.2 DEFINE DATA NEEDS
The initial subtask in the sources and emissions inventory is to
define the minimum set of data that will be required to provide the needed
-------
t’er ar
levels of resolution and accuracy. As the foundation of all the succeeding
tasks in the exposure assessment and the most likely point in the source—
exposure chain for control strategy development, the source and emissions
inventory must be as accurate as possible. Given the resource constraints
that afflict all scientific studies, however, some balance must be struck
between the optimal level of accuracy and the level that can be considered
adequate to ensure valid results.
Because of the intrinsically site—specific nature of the geographic
approach and anticipated variance in resources available to conduct geo-
graphic studies at different sites, it is not reasonable to set across-
the—board standards for the minimum data set required. Nevertheless, the
major considerations in defining data needs will be the sane for all sites.
These are:
1. Scope of exoosure scenarios . The primary feature of the geo-
graphic approach is that it engenders site—specific analysis of
exposure via ambient (e.g., air and water), and drinking water
scenarios. The antient scenario includes exposure related to
waste disposal, transportation (spills), and several other com-
ponents such as atmospheric or riverine imports. The three other
scenarios considered by EPAs Exposure Evaluation Division (EED,
Office of Pesticides and Toxic Substances), occupational, con-
sumer, and food, can be integrated into the analysis but the
marginal benefits of including these scenarios may be negligible.
EPA’s jurisdiction in controlling occupational and consumer
exposure is limited. Locally grown foodstuffs, such as homegrown
vegetables and fish, are probably worth considering as potential-
ly important vehicles of exposure due to locally controllable
toxics problems; food “imported from outside of the study area
may not be effectively addressed by a geographic study. There-
fore, it is anticipated that sources of ambient exposure
(industrial discharges, commercial emissions, etc.), drinking
water exposure (ground water contamination, haloform reaction,
etc.), transportation—related exposure (souls), and waste
-------
Ver arc
disposal exposure (publicaly—owned treatment works, landfills,
etc.) will be the primary focus of the emissions inventory. Food,
occupational, and consumer scenarios may be partly or fully
included depending on the nature of the toxics studied, resources
available, and prospects for interagency coordination.
2. Age of data . Depending on how static or dynamic emissions rates
are for given sources, some date must be set as the limit for
accepting data as representative. This may be different for
different sources. For instance, for an industry which installed
new water treatment systems in 1979 due to BAT, the oldest
acceptable data may be for 1980; for abandoned mine discharges,
which might be suspected to have relatively constant discharges,
the oldest acceptable data could be for 1974.
3. Site—specific data vs. regional or national data . Site—specific
data are obviously preferred for geographic toxfcs studies but
will often be unavailable or depauperate. In such cases, region-
al or national data on sources of toxics and average emissions
rates could sometimes be used as a substitute, supplement, or
quality control check. Substituting such data entails a risk of
diluting the site—specificity of the approach but may be more
cost-effective than sampling or making engineering estimates. The
decision here must be based on the sensitivity of the overall
exposure results to the emission rates in question; the rela-
tive costs of generating site—specific data as opposed to
researching surrogate data; and the expected variability in
emission rates with respect to time and location, it is antici-
pated that there will be large data gaps for many or most
geographic study areas; because direct sampling may not be
possible or cost—effective, it is recommended that EPA compile
-------
Ver ar .
some general catalogs of emissions rates for toxics that will
commonly be studied. These catalogs could supply “faliback”
values in the absence of site—specific data or be used as i basis
for comparing site—specific emissions to emissions occurring
elsewhere for quality control and general reference. Primary -
information sources for regional or national emissions rates
include exposure assessments and mass balances performed by EPA’s
EED (OPTS) and Monitoring and Data Support Division (Office of
Water).
4. Variability in emissions rates . Many anthropogenic and natural
phenomena controlling emissions rates are highly variable with
respect to time and space. As examples, many industries dis-
charge in “batches” due to discrete process steps; urban runoff
occurs in conjunction with rainfall or snowmelt and the “first
flush” usually carries the bulk of the pollutant load. To
adequately characterize toxic loads, the distribution of emis-
sions rates must be known with respect to any correlated vari-
ables (e.g., rainfall, number of batches per year, atmospheric
inversions, flow, season, etc.). When defining the data set, the
variability in emissions rates from different sources needs to be
considered as a prime factor in determining the adequacy of data
in representing the range of conditions which obtain at the site.
“Worst case” and “average” scenarios can only be valid if these
parametric relationships can be described.
5. Minimum number of samples required to represent emissions . The
final consideration is the classic quality assurance problem:
how many samples are needed to provide a reasonable representa-
tion of actual conditions? Unfortunately, the answer to this
question is usually determined more by what the budget allows
than by what the statistician desires. Primary factors in
addressing this problem are related to the preceding considera-
tions, i.e., are corroborating data or surrogate (failback) data
-------
ar .
available fran regional or national studies?; what is the
temporal and spatial variability in the emission rates?; are the
data current?; etc. In some cases, mass balance calculations or
other indirect means of estimating emissions can be considered as
a substitute for direct sampling.
These five considerations for defining the required data set all
entail a balance between the rigors of scientific defensibility and the
constraints of manpower and funding limitations. A first cut at defining
the required data set must be based on the judgment of the study team and
the information collected in the initial scan. This target may need to be
modified as work progresses if it becomes evident that the original defini-
tion does not strike the necessary balance. Further development of guide-
lines and policies for data adequacy would be helpful in assuring the
quality of future geographic studies; it is recommended that these guide-
lines be developed prior to large—scale adoption of the geographic
approach.
It is vital to include the personnel involved in the other exposure
analysis tasks (i.e., environmental pathway analysis, receptor/exposure
route analysis) as well as those involved in control strategies evaluation
when defining the data needs. This helps avoid delays and
misunderstandings later on.
3.3 REVIEW AND ASSEMBLE EXISTING DATA
The next subtask in the source and emission inventory is to review
and assemble existing data. For each toxic on the list of pollutants to be
studied, the discharges to all environmental media must be quantified to
the extent possible. A logical framework to organize the emissions data is
to consider all discharge modes (e.g., to air, water, and land) for each
family of sources (e.g., industry, residential, commercial, imports,
mining). In most cases, emissions data are organized by discharge mode for
each source, rather than by pollutant. For example, an industry’s water
discharge pennit might require monitoring of five toxics and data on all
five would be available from the same reports. This organization of the
data reflects the legislative mandate for data collection as well as EPA’s
own internal structure, wnich is aenera iy media—specific.
-------
%1Vi& s . •.
WL’U inc.
Procedures for reviewing and collecting data on sources and emissions
to air, water, and land are discussed in Sections 3.3.1 — 3.3.3. It should
be noted that, although toxics emissions via spills are discussed only with
respect to land, the same data base can be used to evaluate discharges to
air and water.
3.3.1 Airborne Emissions
The primary goal in compiling a suitable data base for airborne
emissions is to compile emission rates for all quantifiable sources within
a defined study area. Because these data are used for exposure assessment
in a geographic toxics study, it is also necessary to obtain additional
specifications associated with the emissions, such as exit velocity, exit
temperature, inner stack diameter, stack height, and area source specifi-
cations. Meteorological data, including representative wind speed, wind
direction, atmospheric stability, ambient temperature, and mixing height,
are also needed to perform the dispersion modeling required for the
exposure assessment.
Emissions data can be obtained from emission inventories developed
for source—specific data, assumptions based on per capita considerations
and generic emission rates for source categories. In an area that has a
large number of emission sources, prioritization of source categories will
be necessary to make the inventory assessment manageable. Although it may
not be possible to obtain data on each source (or in some cases each cate-
gory of sources for heavily industrialized metropolitan areas), through
careful consideration of the sources accounting for an acceptably large
amount of the expected totals, meaningful results can be obtained in a
cost—effective manner. In attempting to thoroughly review all potential
data sources, a point of diminishing returns will be reached. Considerina
the degree of uncertainty in exposure assessment, the emphasis should be
placed on assuring that the sources expected to be the major emissions
within the study area are adequately characterized.
3—6
-------
vcr ar .
A general approach recommended for obtaining a suitable data base of
airborne emissions is described below. For each category, source specific
and generic approaches are identified, as applicable. The types of
applicable data, and their limitations, are explained as follows:
Source specific data — Under the authority of the Clean Air
Act, states regulate the emission of airborne pollutants.
Emission inventories are compiled that include major sources of
criteria pollutants (SO 2 , N02, total suspended particu-
lates, hydrocarbons, CO, ozone and lead), and also emissions
of hazardous pollutants (asbestos, beryllium, mercury and vinyl
chloride) and specific hydrocarbon compounds that could be
considered toxic. It is important to note that there can be a
wide range in the confidence of these data since they can be
obtained in a number of ways, including the use of stack
sampling, mass balance considerations, emission formulas or
simply best engineering judgments.
Generic data — There may be a large number of relatively small
sources which, when taken as a whole, can be a significant
factor in a toxics study. In this case, in lieu of source
specific data, estimates often canbe made using per capita
considerations such as emissions associated with degreasing,
printing, dry cleaning, architectural coating, etc. In some
cases, these estimates can be made based on employee totals
within Standard Industrial Classification codes rather than
general population data. Generic emissions are generally
treated as area sources, and have a considerable amount of
uncertainty. A drawback of generic data is that the maximum
impacts from the relatively large sources within these cate-
gories are not individually reviewed. Generic pointsources are
-another possibility. For sources such as gas stations, a
distribution of point sources can be made in an area based upon
the number of employees in each area, per capita data, sales,
etc. This approach allows for a better evaluation of impacts
near the sources. Availability of data on airborne toxics
emissions is aenerally very limited. Nevertheless, some
procedures can be used to maximize the utility of existing
data. These procedures are discussed separately, for various
sources, below. The relative importance of each of the sources
will, of course, vary from pollutant to pollutant.
Availability of data on airborne toxics emissions is generally very
limited. Nevertheless, some procedures can be used to maximize the utility
of xisting data. These procedures are discussed separately, for various
sources, bel ow. The rel ati ye importance of each of the sources wjii, of
course, vary from pollutant to pollutant.
-------
ver ar .
Industrial
Industrial sources of airborne emission of toxic pollutants
include those emitted from stacks/vents and from fugitive releases (i.e.,
emissions not released from a conduit). To obtain information concerning
these emissions, the following should be performed:
• Review state emission inventories mentioned by the state air
quality control agency. Obtain the following data on a
pollutant specific basis for each registered source that is
quantified:
- latitude/longitude
- exit velocity
- exit temperature
- inner stack diameter
— stack height
— adjacent building heights (building wake
considerations)
— continuous or batch releases
— maximum and annual average emission rates
— hours per year of operation
• Review air quality permits for major sources identified in
the study area to obtain operation limitations and additional
data on emissions, including limits on production rates;
hours of operation, control efficiencies, fuel restrictions,
etc.
• Contact the environmental affairs departments of major
industrial sources and utilities to attempt to obtain
necessary data. It is also advisable to contact major
sources in the study area to confirm that data being used in
the data base are consistent with current operations.
• Computerized data bases such as the National Emissions Data
Systems (NEDS), the Hazardous and Trace Emissions System
(HATREMS), the Storage and Retrieval of Aerometric Data
(SAROAD) systems and numerous other data sources are
theoretically useful in identifying major sources and areas
of high atmospheric concentrations. At this time, however,
there are two major problems in using these data bases for a
toxics study. The data are mostly associated with criteria
pollutants rather than a wide ranqeof toxics. In addition,
the data bases are generally incomplete. It is recommended
that these data bases be reviewed, but only as a supolemen-
tary measure.
-------
Vert ar .
Resid —r ia l/Corivnerc a1
Emissions of toxic pollutants from the residential and commer-
cial sectors can be significant when taken, as a whole, even though indi-
vidual contributions may be small. Because these sources are not usually
addressed in inventories, as the industrial sources often are, quantifi-
cation will generally involve a rough approximation. In addition, In
evaluating exposure, these sources are often treated as area sources, which
further adds uncertainty to the analysis.
A reasonable approach to the quantification of these sources
includes the following elements:
Residential
• evaluate fuel use for home heating (e.g., natural gas, oil,
coal).
• estimate emissions of toxics based on available emission
factors (e.g., Shih etal. 1980, U.S. EPA 1977b).
• uniformly distribute emissions over major residential areas
within study area, or base distribution on residential
density.
Commercial
• Identify the source categories that are expected to emit
significant quantities of toxic pollutants of interest to the
study. Possible categories include:
— gasoline marketing
— degreasing
- drycleaning
— architectural surface coating
— graphic arts
— asphalt sources
• Obtain the most recent pooulation data and project current
levels.
-------
Vcr ar .
• For source categories emitting significant quantities of
toxic pollutants of interest to the study, quantification of
emissions can be approximated by per capita factors.
• Determine if production data for relatively large sources in
the commercial categories can be quantified. If so, these
emissions should be subtracted from the per capita derived
totals, and modeled as discrete sources.
• Distribute area source emissions on the basis of the density
of commercial establishments within the study area. A
simplified approach would be to uniformly distribute them
over the central business district.
Mining
If product dust contains pollutants of interest to the study,
the emissions assessment must consider emissions from blasting, handling,
transfer, and storage of the products of the mines in the study area. Most
data on emission factors appear to be available for coal. There are a
number of references available that contain emission factors from mining.
Region VIII of the U.S. Environmental Protection Agency has summarized
several reports on the subject (U.S. EPA 1979q). Mineral processing
facilities are a potential source of toxic airborne emissions that should
also be considered. Mining and mineral processing facilities are covered
by reporting requirements similar to those pertaining to industry in
general. Use the same procedure as that outlined under hh lndustrialu to
obtain information on this source.
T2’ansportation
Transportation sources include automobiles and trucks, rail-
roads, ships and aircraft. In general , the major source of emissions from
this sector are road vehicles. Although the other sources should be
considered in geographic toxic studies, the primary emphasis would usually
be on gasoline and diesel fueled road vehicles. Emissions of toxic
pollutants from automobiles include lead, benzene, toluene, asbestos and a
wide variety of organic compounds. The composition of the exhaust from
Qasoline and diesel fueled road vehicles should be considered in order t
determine if these emissions are significant to the study. The followir g
describes a enerai approach to characterizing these emissions:
-------
cr arr.
• From the state highway administration for the state in which
the study area is located, obtain the most recent annual
vehicle miles traveled (AVMT) data representative of the
study area. Factoring may be required on the basis of popu-
lation to extrapolate to the study area in question. In lieu
of this information, fuel marketing data from tax records or
per capita fuel use estimates may be used in determining AVMT
in the study area.
• Emission rates (e.g., grams/km) may be obtained from refe-
rences, such as Black etal. (1977).
• Define the area over which the AVMT will be distributed,
given particular attention to the areas of maximum vehicular
traffic.
• If justified on the basis of traffic flow within and through
the study area and resources allocated to the study, major
routes may be treated separately for the emissions assessment
and dispersion modeling analysis. Additional information
such as road width, average speeds travelled, and maximum
usage rates should be obtained if this approach is deemed
necessary.
.Z nports of To ios wid Rio genic Production
In addition to the above sources, some toxics may be advected
into the study area (imported) from other areas (e.g., PCBs) or may
actually be produced biologically within the study area (e.g., phenol).
Although it is anticipated that in most cases these sources will be
negligible with respect to the others, these sources bear investigation.
The best source of information would be actual monitoring data at points
near the periphery of the study area (imports) or at the center and edges
of populations suspected of synthesizing or metabolizing toxics. Clearly,
such data are rarely available and can usually be regarded as low priority
data gaps.
3.3.2 Aaueous Discharges
Ac ueous discharges can be divided into non—point source and point
source discharges. These are studied in fundamentally different ways, and
so are discussed separately in the following paqes.
3—li
-------
Vcrt ar .
Non—Point Source Water Discharges
The methodology for assessing the Non-Point Source (NPS) water
discharge inputs of toxic substances to a geographical study area is
dependent on the resources available for the project. Unlike point source
discharges which are discretely identifiable and often do not vary con-
siderably with time, NPS discharges are diffuse and subject to great
variations dependent upon the meteorological conditions. In Its simplest
form, the methodology for assessing NPS water inputs to the geographical
study area is dependent on site specific data for rainfall, runoff
(hydrology), land use, and water quality concentrations that are averaged
over some time period, such as a season or an entire year. In its complex
form, the NPS input methodology must account for the ambient meteorological
factors that affect runoff, water quality, and resulting pollutant loads.
These factors include rainfall intensity, duration, wind velocity, solar
radiation, antecedent moisture conditions, and time of year. In addition,
land use characteristics such as basiii and channel slopes, soil types,
impervious areas, stream channel characteristics and the basin’s anthro—
pogenic activities must be considered. The complex methodology would
generally require the use of a sophisticated computer modeling package
(e.g., U.S. EPA SWMM model), which would require pre—calibration with site
specific data before the model could be run to generate predicted pollutant
loads. Considering the limited data base currently available on toxic
substances in NPS water discharges, this approach is probably unwarranted
unless the resources are made available to fill the existing data gaps.
Therefore, the methodology addressed in the following discussion is much
more practical and simple in scope, and designed to provide general esti-
mates of the possible magnitudes of NPS toxic loads. The methodology
consists of five steps.
• Define Study Area by Subwatersheds
The first step required in the NPS assessment effort is the
definition of the study area according to surface water
drainage patterns. Consider only those basins that drain
direct y into the study area. Ideally, basins identified
should be small enough that sources of ooliutants may be
readily identified according to some land use activity 3r
oattern.
3-12
-------
Ver ar .
This is particularly important in areas where the land use is
particularly diverse. Not only will the use of small
drainage basins permit the identification of pollutant
sources, it will also define the study area according to a
source control pattern if control measures are required.
In some cases where geographical study areas occupy the mid
section of large river basins, it will be necessary to
consider the NPS inputs of major tributaries and upstream
reaches under the source category of imports. This method
involves the use of river discharge records and in—stream
water quality to estimate toxic pollutant import loads, and
is discussed in greater detail with point source water
discharges.
• Collect Land Use Data
Once the study area has been defined according to surface
water drainage basins, the collection of the specific land
use data is required. Generally, NPS discharges are grouped
according to the following land use categories:
• Agriculture
• Silviculture
• Urban
• Construction
• Mining
Each of these categories could be further subdivided (e.g.,
urban—residential, industrial, canmercial); however, detailed
land use information is usually limited. Moreover,
considering that water quality data are usually also limited,
the allocation of considerable manpower and resources towards
a detailed land use characterization is usually unwarranted.
Total acreage for each of the land use categories chosen
should be obtained for each of the identified drainage basins
included in the study area.
• Determine Runoff Yield
The next step is to acquire data on runoff. Ideally, a
previous study effort may have already conducted instream
sampling and water quality monitoring according to land use
characteristics in the study area (such as studies conducted
under Section 208 of the Water Pollution Control Act).
Efforts of this type will frequently generate surface runoff
yields for same time frame (e.g., gallons/acre/year) accord-
ing to the major land use categories. Total basin seasonal
-------
or yearly flows are calculated by multiplying the area yield
flow rate for each land use category by the area in each land
use and summing accordingly. Alternatively, total flows for
each land use category may be allowed to stand alone to
permit toxic load calculation according to land use rather
than according to individual subwatersheds. In the absence
of area yield flows, use predictive runoff formulas that
incorporate data on rainfall volumes and land use charac-
teristics. The ilrational formula, for example, is commonly
used for average runoff predictions, particularly in urban
areas. The rational formula is:
Q = CIA
where Q = flow, C is a unitless runoff coefficient, i is
rainfall intensity over time of concentration, and A is the
area of the watershed. Values for the variables i and A are
available from many sources, including the U.S. Geological
Survey and the National Weather Service, and can be easily
estimated from rainfall records and topographic maps. The
runoff coefficient, C, varies as a function of the permeabi-
lity of the watershed surface. Values for C are listed in
standard hydraulic engineering texts such as “ The Manual on
the Desi9n and Construction of Sanitary and Storm Sewers ”
fASCE 1960). The accuracy of prediction can be increased by
the use of the Soil Conservation Service (SCS) formulas,
particularly for agricultural or other non—urban watersheds.
Detailed information on runoff prediction using the SCS
formulas is available in tbe “ National Engineering
Handbook-Hydrology ” (Mokus 197ZJ.
• Obtain Concentration Data
Obtain data on the levels of toxics in runoff so that pollu-
tant loads can be calculated. This step is usually the most
difficult. Site—specific water quality data are preferred;
however, non-point source toxic monitoring has been extremely
limited. It may be possible to locate heavy metal concentra-
tions and pesticide data from previous “208” water quality
programs in the geographic area. Toxic organic monitoring
data are virtually nonexistent, and only crude estimates of
pollutant loads are presently possible. In the absence of
site specific data for aqueous concentrations of toxic sub-
stances, generic data may be substituted in order to calcu-
late pollutant load estimates. The estimates generated for
generic data should be clearly identified as such, however,
and it should be understood that they may not represent
conditions in the study area. Recardless of the data
source it is important that averaae stormwater uality
3-14
-------
fgk •
we_I .ai
concentrations, rather than discrete or grab sample data, be
used. Use concentrations that have been recorded as a result
of flow—weighted sampling if available. This is because
toxics concentrations may vary by several orders of magnitude
across a storm hydrograph (particularly in urban areas that
exhibit a “first flush” phenomenon) and discrete concentra-
tions may significantly over- or underestimate a pollutants’
presence.
• Calculate Pollutant Loads
The basic method used to calculate loadings is simply to
multiply concentration of a given pollutant times the flow.
Pollutant load rate is thus expressed in units of mass per
unit of time. Ideally, NPS loads are estimated for runoff
from a series of typical storms occurring over the geographi-
cal study area during some time period s.uch as a season or a
year. Water quality concentrations that are seasonally
recorded are therefore used. This process, however, may be
extremely time consuming and considering the generally
limited data base, it may be best to calculate total average
yearly flow for each basin in the study area and multiply
this by the average pollutant concentration to obtain a
yearly loading rate for each land use category, as well as
for each subwatershed. Total annual pollutant load in the
study area is the sum of the individual subwatershed loads.
One alternative method of calculating NPS pollutant loads that
eliminates the need for water quality and average flow data is the use of
area yield pollutant loading rates according to land use. These data are
extremely limited and will only have been generated as a result of a
previous site—specific non-point source monitoring effort (such as a “208”
study). Area yield loading rates express individual pollutant loads
according to some measure of area and according to land use. For example,
the area yield loading rate of lead for a particular urban area may be
expressed as kilograms per hectare of surface area per year (kg/ha/yr).
Similar yield loading rates may also be available from agricultural,
silvicultural and mining areas. If accurate land use data are available,
calculate the pollutant load by simply multiplying the area of a particular
land use by tne area—yield rate. Pollutant loads may, therefore, be
expressed according to land use, or the individual land use loads may be
-------
kI.%
WLI II IflC.
summed to find the total watershed load, which in turn may be summed to
calculate the total load for the geographical study area. Data of this
type are not usually available, however.
Several other methods of estimating NPS pollutant loads are also
available and may be used for making qualitative assessments. For example,
if data on pollutant concentrations in runoff are not available, but flows
and concentrations have been measured at ambient stations upstream and
downstream of a study area, the following approach may be used: during dry
periods, the difference between the pollutant load in the upstream station
and the pollutant load in the downstream station can be assumed to result
from loading from point sources; this loading rate can be assumed to be
relatively constant. The upstream and downstream loads are then compared
during storm events. The difference in upstream and downstream loads,
minus the h*baselineu point source load, canbe considered to represent the
load from non—point sources. This method of calculating the NPS pollutant
load would have to be performed over several storm events of differing
magnitudes in order to det rmine an average load.
Although the methodologies discussed above are relatively crude,
the NPS data base for most of the priority pollutants is not sufficient to
warrant a more sophisticated approach. Frequently, only a qualitative
assessment is possible, based on land use activities, fate processes
affecting the residence and transport of pollutants in the aquatic environ-
ment, and meteorologic processes that may affect NPS loadings. The state-
of—the—art with regard to NPS studies and methodologies is currently in a
period of transition. As a result of ongoing sampling programs and
previous monitoring efforts (e.g., Section 208 studies), NPS pollution is
being given more consideration as a significant discharge mode. As the NPS
data base increases in size and the state—of—the—art advances, it is
expected that improved methods will be developed for assessing the magni-
tude of NPS pollutant loads in a given geographical area.
5at rc ; ‘ :- c rg s
The existing data base for assessing poir t source water dis-
char es of toxics in a geocraphic study area may often be more comolete
0
-------
Ver ar
and quantitative than the data bases for other discharge modes. This is
largely a result of the U.S. EPA National Pollution Discharge Elimination
System (NPDES) and effluent guidelines sampling efforts. In the near
future, data retrieval for NPDES will become even more streamlined with the
advent of a new cc iiputerized system for NPDES records: It must be empha-
sized, however, that the toxics on the list of 129 “priority pollutants”
have been given considerably more attention in previous sampling and moni-
toring programs than have other toxics; it is likely that a geographic
study of non—priority toxics would require significant amounts of new data.
A detailed discussion of the reccinmended methodology for assessing point
source toxic discharges is given below, followed by methods for evaluating
imports of toxics via waterways.
After completing the initial scan, the location of most sources
of toxics data will already be known and preliminary data gathering efforts
will have been completed. For point source water discharges, the bulk of
the pertinent site—specific data will be located at the regional EPA office
(with personnel responsible for enforcement), at the state office
responsible for water discharge permits, with the permit—holders or, in
some cases, with the U.S. EPA National Enforcement Investigation Center
(NEIC). The Effluent Guidelines Division of EPA headquarters will be the
most knowledgeable source for non—site—specific (i.e., generic) data.
As one of the first steps in assembling existing data, obtain a
list of all NPDES and state water discharge permit holders should be
obtained for the geographic study area. There are several computerized
data bases that will provide the NPDES list. Currently the U.S. EPA 1FD
(Industrial Facility Discharge) file is the most suitable, as it will
provide a list of NPDES permit holders for a whole USGS basin. If the
study area does not correspond to a basin, make a computer retrieval for
all basins in which the study area is located. In this case, permit holders
outside the study area can be eliminated from the list on the basis of
location (latitude/longitude). In addition to location, the IFD retrieval
provides useful information on the NPDES permit number, effluent flow,
Standard Industrial Classification (SIC) code, and receiving stream. In
the near future a new computerized NPDES file will be available, which
-------
may be more useful than the IFD file; this system will provide data on
pollutants requiring periodic monitoring in addition to most of the
information available in the IFD file. The list of state water discharge
permit holders in the study area may be obtained either from responsible
state personnel or by going through the state permit files. Depending on
the size of the study area and the organization of the files, however, the
latter option may be very time—consuming.
Once the list of all state and NPDES permit holders is assembled
it should be divided into several subcategories based on the likelihood
that individual sources contribute the toxics of interest in the study.
Based on discussions with knowledgeable federal and state personnel, a high
priority list should be compiled. Initially, the permit holders in this
category shbuld include all “major” dischargers (as determined by a ranking
system used by U.S. EPA), all 34 “primary” consent decree industries, and
all other dischargers suspected of discharging the toxics. By examining
available records, this list can then be pared down to those facilities
known to discharge the toxic, based on discharge monitoring reports (DMRs)
or other sampling results. The permit holders culled from this high
priority list that are suspected of being sources of the toxics of interest
can be placed in the second category along with other permit holders that
are suspected for any reason to be potential sources of the toxics. All
permit holders that are not potential sources of the toxic can be placed in
the third, low priority, category. As a general rule, only the first two
categories need be investigated further.
In order to assess the contribution of toxics from each of the
known or suspected sources, information on effluent flow, toxics
concentration, type of discharge, and operating schedule must be obtained.
Because the sources of information and calculations are different for the
“known ’ and “suspected” dischargers, these categories will be discussed
separately. Effluent flows for the known contributors of toxics will
generally be available from one or more of the following sources: state
and NPDES discharge monitoring reports (DMRs), IFO file, NPDES or state
discharge permit apolications, reports of NEIC investications. Usually the
3-18
-------
range of flows and the mean daily flow will be available on a monthly basis
for each outfall. The monthly records are preferable to all other data
because they are generally more up-to—date. The toxics concentrations may
be available from the DMRs, sampling results submitted to EPA and/or the
state by the dischargers to fulfill permitting requirements, or sampling
conducted by the agencies themselves. If available, the DMRs are the best
source of concentration data because they provide both a long-term
flow—weighted average and a range of concentrations. Information on the
type of discharge (e.g., batch, continuous), and the operating schedule
(e.g.,, 16 hour-day, 350 days/yr) can usually be obtained from permit
applications; otherwise, EPA and state personnel or reports may be helpful.
Once the flow, concentrations, and flow characteristics are,known, the
toxics loadings can be estimated on a mass/time basis by multiplying the
concentration by the flow times the appropriate conversion factors. Both
the average loadings and the range of loadings (e.g., worst case, ubestu
case) should be reported, if possible. The exact units will depend on the
type of modelling planned for the exposure assessment. Existing
source—specific data can be confirmed and new information obtained for the
dischargers in the hlknownu category by contacting personnel and/or
conducting effluent sampling, as discussed below for the ‘ suspected”
category.
Facilities in the flsuspectedu category, as mentioned previously,
will include those for which no site—specific toxics data are available
that belong to one of the 34 °primary’ industry categories or are otherwise
suspected of being dischargers of the toxics. Because this category, by
definition, includes facilities for which there are considerable data gaps,
these gaps must be filled in order to assess toxic loadings in plant
effluents. The best way to fill the data gaps is to contact plant
personnel. This should be done in an organized fashion beginning with a
letter and followed up, if necessary, with a phone call. For many of these
facilities the flow will be available through DMRs or permit applications.
The facility should be requested to confi flow estimates, as well as to
voluntarily provide data on effluent toxics concentrations collected in
-------
VCrNaFW
self—monitoring (if available), and information on raw materials, products,
and treatments—in—place. Obviously, the information requested should be
tailored according to the data needed and the type of facility.
Alternatively surrogate (national or regional) data can often be used to
estimate primary pollutant toxics concentrations and, if necessary, flow
characteristics. The Effluent Guidelines Division (EGD) of U.S. EPA is the
best source of surrogate information for the 34 “primary” industries,
because it has conducted effluent sampling of representative facilities for
these industrial categories. This information is usually available in
summarized form, by industry, in the “development documents” available
through EGD. In addition, EGD has conducted selective effluent sampling of
some industrial categories other than the 34 primary categories. Toxics
loadings can be estimated for the sources in the suspected category in the
same manner as for the known category. Loadings estimates, however, should
be clearly designated as to whether they are based on site-specific data or
surrogate data.
If the study area does not encompass the entire headwater portion of
a hydrologic basin, estimates must be made of toxics imports into the study
area via surface waters. Because these imports are conveyed in stream
channels, they are discussed here as point sources. Consider each aquatic
segment upstream of the study area as a potential point source of toxics
entering the study area at the point where the study area boundary -inter-
sects the stream/river. In order to evaluate the toxic loadings from
imports, collect ambient data on flows and toxics concentrations for all
tributaries to the study area. These data may be available through the
computerized EPA STORET file, or may be obtained directly from agencies
that conducted the sampling (e.g., EPA, USGS, state agencies, U.S. Army
Corps of Engineers). Once assembled, the data for each upstream segment
can be compiled in the following manner. Calculate a separate loading
(mass/time) for each sample taken, by multiplying the flow by the ambient
toxic concentration. Calculate average loading and the range of loads from
the individual data points. If the individual data ooints are not
available, multiply the average flow by the avera e toxic concentration;
however, this estimate will be less accurate than estimates using indi-
vidual data points.
3-20
-------
Ver ar .
3.3.3 Discharges to Land
Discharges to land include landfilling, disposal in hazardous waste
sites, sludge spreading, intentional application, and other activities.
Existing information on toxics discharges to land will usually rely on
RCRA manifests, treatment, storage, and disposal (TSD) permit applications,
and related material. Although the advent of RCRA hazardous waste
regulations has greatly increased the availability of data on land—destined
disposal, this media is still one of the least—quantified for toxics. The
sources of existing data can be discussed most logically with respect to
the environmental pathways they cover, i.e., waste disposal, residual
levels, or spills. Each of these are considered separately below.
Waste Disvosal
Waste disposal on land can be separated into two sources of toxics:
disposal of known hazardous waste and disposal of residential/commercial!
/industrial waste that may contain unknown amounts of toxics. For both of
these waste categories two types of data are required to categorize the
sources and emissions. First, the type of waste, its origin, quantity,
and form should be known. Second, the location and method of disposal
should be known. For known hazardous wastes the above data will often be
available through state and federal hazardous waste control laws, regula-
tions, and permits. For wastes with unknown amounts of toxics the data are
more difficult to acquire. While information on quantitites of waste,
generic waste types, and disposal of wastes is often readily available,
data on the specific quantities of toxics in “ordinary” residential!
commercial/industrial wastes are usually unavailable.
Contact regional EPA hazardous waste sections, State hazardous waste
disposal agencies, State health agencies, and local planning boards as the
primary information sources. The types of information available from each
of these organizations is listed in Table 3—1. This information should
provide qualitative data on the general composition of wastes or the types
of waste streams, and quantitative data on the total mass of wastes
-------
Table 3—1
SOURCES OF DATA ON LAND—DESTINED DISPOSAL OF TOXICS
Source Data Available
• U.S. EPA Regional Office - RCRA Part A hazardous waste
treatment, storage, or disposal
permit application forms
- RCRA Part A hazardous waste trans-
portation manifests
- RCRA Part B hazardous waste
disposal site inspection forms
- Investigative reports dealing with
a hazardous waste disposal site or
a source of hazardous waste
- computerized data base with all
RCRA application data
• State agency with responsibility — hazardous waste disposal permits
for hazardous waste regulation - hazardous waste site inspection
forms, monitoring records
- transportation manifests
• — knowledgeable personnel familiar
with the study area
- state investigative reports on
hazardous wastes in the study area
• State and local agencies with - may have data on risk to public
responsibility for public health from disposal of hazardous wastes
• Regional planning agency — waste disposal plan, may be part of
comprehensive plan
-------
generated at and hauled to different sites. Because the exact composition
(by percent weight, volume, etc.) of most wastes is not known nor is it
required under most existing hazardous waste regulations, high resolution
of the amounts of specific compounds in waste streams will usually prove
difficult or impossible. Stated differently, hazardous wastes are
generally defined as such due to the presence of a hazardous material;
regardless of whether the presence is in trace amounts or as a homogeneous
phase, it is reported the same way. In most cases, then, the available
data will only provlde a range of possible values for the “flow” of
hazardous wastes.
If the toxics of concern may be present in waste streams that are not
designated as “hazardous,” collect data on residential, commercial, or
industrial wastes. Volume, composition, and destination of the wastes is
required to characterize mass flow to the land compartment. Data sources
for these “non—hazardous” wastes are listed in Table 3—2.
In all cases, collect any available information on waste containers,
how long the sources have discharged the wastes, historical volumes of
discharges, and other pertinent information.
Regional or national data may be an acceptable surrogate for
site—specific data, depending on the target data needs defined in the
previous subtask. Mass balances, engineering estimates, or actual sampling
data should all be considered if actual site—specific data are lacking.
Res- dua 1
Residual toxics are those toxic pollutants that are found within the
soil as a result of past toxic discharges or geologic processes.
Sources of such residual toxics include pollutants in active or abandoned
waste disposal sites; toxics that have settled or precipitated out of the
air; and toxics that have been deposited by surface water and ground water.
Data on such residuals are often unavailable. Where information on
residuals exist, it is usually generic (i.e., not site—soecific) in nature.
Some potential data sources are described in Table 3—3.
-------
Table 3-2
General Waste Disposal Data Sources
Source Data Format
• State agency with responsibility — disposal site permit and inspection
for solid waste regulation forms
— documents with information on the
amount of toxics in solid waste
— personnel with knowledge of study
area
• Regional planning agency — solid waste disposal plan, may be
part of regional comprehensive plan
— existing land use data
• Local/Municipal planning — solid waste disposal plan, may be
agency part of comprehensive plan
- existing land use data
• U.S. EPA Headquarters and/or - reports on toxics in residential!
Regional Office commercial/industrial wastes.
-------
Ver ar .
Source
• Waste disposal data sources
(See Waste r sposaz. )
• Soil Conservation Service
• U.S. EPA Headquarters or
Regional Office
• State agency responsible for
ground water
• U.S. Army Corps of Engineers,
State conservation or natural
resource agency.
Data Available
— data on abandoned and active
disposal sites
— information on past disposal
practices
— soil reports and analysis
— reports on ground water contamina-
tion and analysis of toxics in
sediments.
— reports on ground water
contami nation.
— analysis of toxics in sediments.
Table 3—3
Residual Data Sources
-------
Ier ar .
Data on residual toxics, especially organics, are qenerally even more
limited than data on active waste disposal processes. Nevertheless,
residuals can be quite significant, especially for refractory compounds
such as PCBs or DOT, and act as a reservoir with potential for future
releases. For ephemeral toxics, residuals are not important and can be
disregarded.
Spills
Spills of hazardous wastes or toxics on land are often difficult to
quantify because of a lack of data for past events, and inadequacies in the
available data base. There are obvious reasons why many spills are never
recorded, or at least never made public. Reports on spills are often
sketchy as to details and may lack information on such activities as
cleanup and disposal of the spilled toxics. Some potential sources of data
are listed in Table 3—4.
Spills are intrinsically episodic and difficult to predict, so there
is little potential for using national or regional data for all but the
most frequent occurrences (e.g., loss during filling of oil tank trucks).
Relating the source and emissions inventory to later exposure assessment
tasks, most environmental models cannot realistically model spill events,
and the utility of spills data in making quantitative predictions of
exposure is not presently clear. Nevertheless, these occurrences can be
extremely significant vehicles for exposure, especially for high volume
industrial chemicals.
Table 3-4
Spill Data Sources
Sources Data Format
• EPA Regional Office Log of hazardous waste spills
• State agency with responsibility Log of hazardous waste spills, data
for hazardous wastes may be in separate company files
• State public health agency Record of toxic substance soii s
-------
3.4 EVALUATE ADEQUACY OF EXISTING DATA
As work progresses with subtask 2, (assembly of existing data) com-
pare the collected data with the data needs defined in subtask 1. It is
Important to periodically evaluate the adequacy of the existing data; do
not wait until the end of the data collection phase because most of the
measures employed to generate new data require considerable lead time.
This evaluation represents a checkpoint on data quality and quantity——are
there enough data, and were the original procedures used to produce the
data sufficient to build upon for the exposure assessment and control
strategy devel opment?
The gaps between the desired quantity and quality of data and the
available quantity and quality of data will often be large, particularly
for airborne and land—destined emissions. Several measures are available
to bridge these gaps (these are discussed in the next subsection), but most
are costly relative to the effort involved in collecting and assembling
existing data. Thus, when existing data are inadequate, the program
manager is faced with a difficult decision: to proceed with the analysis
and strongly caveat the results; to pour resources into further data
collection; orto abandon analysis of certain pollutants or sources.
Two tools are available to assist in this decision: sensitivity
analysis, and the environmental pathways models developed for the next
phase of the exposure analysis (discussed more fully in Section 4 and
Appendix A). Sensitivity analysis can be used to indicate how sensitive
the total emissions “budget” is to inputs from a given source or group of
sources. Inherent to this approach is the assumption that enough is known
about the source to assign some range of toxics loadings that will bracket
the actual loadings. The high end of the loading estimate is the critical
one, and should be enviror nentally conservative (i.e., do not underesti-
mate). Compare the estimated maximum load to the loads from other known
sources. If the total load is sensitive to the estimated maximum load from
the unknown or poorly quantified sources, further study is required. If
total load is not sensitive, but temporal or spatial considerations
-------
verxar .
indicate that the source may be an occasional acute problem or severe local
problem, further study may also be required. If neither of these condi-
tions obtain, it is unnecessary to acquire more data on the source(s).
A second and technically preferable tool is to use the models
developed for the environmental pathways analysis. Provided that adequate
ambient data are available, and that emissions are fairly well—defined for
some sources, the available source data can be input, and resulting model
predictions can be compared to measured ambient levels. If the input
emission rates account for the observed ambient concentrations, it is
reasonable to assume that other sources are negligible. If predicted con-
centrations are significantly lower than the actual ones, other sources may
be significant. Comparison of monitoring and modeling data is given more
complete treatment in Section 4.4, as a number of other inferences can be
made using this procedure. It is a valuable tool for evaluating the
adequacy of existing data, but has drawbacks in that complete ambient
monitoring data must be available; at least some emission rates must be
known, the modeling parameters must be previously calibrated and verified
to the extent possible; and it is expensive.
Both of the tools discussed above can be used to screen out sources
that are insignificant with reference to others. It is axiomatic that a
screening procedure based on comparison of relative values must have some
known values as a basis for comparison. In many cases, data for a given
pollutant will be inadequate or unavailable for all sources, and thus
neither of the above tools are useful. If this is the case, the data base
must be expanded.
There is one other situation in which further data collection is
unnecessary. This would occur if the emissions from a source do not enter
any environmental pathway that could result in exposure to humans or other
receptors. An example of this rare case would be where all of an indus-
try s emissions co to a secured waste disposal site with no potential for
volatilization, leaching, or otherwise re—entering the biosphere.
-------
Vrn ar .
To summarize, further data collection is not necessary if:
• the target criteria for data needs are satisfied by the existing
data, and
• ambient concentrations are accounted for by emissions from known
sources, or
• it is very likely that sources for which the data are inadequate
are negligible with respect to sources for which data are
adequate or
• although emissions are not quantified, it is known that there is no
potential for exposure to human or other receptors due to the
characteristics of the discharge mode.
More data collection is necessary if:
• the target criteria for data needs are not satisfied by the
existing data, and
• estimated maximum emissions from sources with inadequate data are
potentially significant in terms of total emissions or occasional
or localized exposure—related problems, or
• ambient concentrations are not accounted for by emissions from
known sources, or
• inadequate data are available for all sources of a toxic.
3.5 DESIGN AND IMPLEMENT PROGRAM FOR COLLECTING NEW DATA
It is anticipated that in most geographic studies, existing data will
not be adequate to characterize emissions for at least some sources (and
-often all sources). Designing and implementing a program for collecting
new data will thus become an important part of the technical approach.
Some measures to correct data deficiencies include:
1.- more exhaustive search for existing data
2. sample and analyze emissions or ambient media samples
3. develop engineering estimates or mass balance calculations
4. use available surrogate data (national or regional)
5. adapt analogs, using similar compounds witn known emission rates.
-------
Ver ar
Each of these are described briefly below:
E. haustive Data Search
Usually, the great majority of existing data will be available from
the government agencies and industrial contacts discussed in Section 3.3.
In some cases, however, universities, consulting firms, and others may have
conducted studies that are not known or available to the aforementioned
groups. Exhaustive searches of these potential sources may yield some
information. It is anticipated that the marginal benefits of an exhaustive
search will rarely outweigh the marginal costs with respect to a thorough
search of the previously identified data bases.
Sa’np7 e czn4 Analyze Emissions or Ambi ent Media
This is the most direct, technically satisfying, and expensive alter-
native. It entails extensive planning, coordination, and analytical time,
but, with proper design, provides the most useful data. Coordinate
sampling efforts with other elements of the exposure assessment; simultane-
ous analysis of emissions and ambient levels at critical points helps
establish a firm linkage in the source-exposure chain. The ambient
sampling program should be coordinated with any ambient sampling that is
needed for the evaluation pf monitoring data (Section 4.4). Use surrogate
data, mass balances, analogs, or other estimating techniques to help
determine where to sample. Like any other sampling program, it is
essential that quality control and quality assurance elements be
ificorporated into the procedures.
In addition to the typical air and water sampling schemes, consider
sampling media that act as information integrators. For instance, when
studying hydrophobic refractory compounds in aqueous discharges, sample
sediment in depositing zones; if spills are suspected from a facility but
normal discharges are reported to be of good quality, sample benthic
invertebrates downstream of the plant (since most of the organisms in the
community have a life—span of over a year, normal composition indicates
lack of chronic or episodic pollution).
-------
Sampling sediment and other environmental ‘sinks’ is an excellent
indication of the presence and relative amounts of a toxic. If insigni-
ficant amounts of a toxic are found in a ‘sink’ for the particular com-
pound, it may be safe to assume that the emission rates are insignificant
(depending on locations of sampling sites, sources, environmental fates,
persistence, etc.).
There are numerous references available on sample program design.
None of these, however, deal with the range of possibilities that are
encompassed by the geographic approach — all environmental media; scores of
point, non—point, and mobile sources; temporal variations; interagency
coordination; etc. This subject was only superficially addressed by the
current study and requires considerable development as a major component of
the geographic approach.
Develop Engineering Estimates or ss Balances
For some industrial, commercial, and natural processes, it is
possible to develop estimates of toxics emissions using known flows in
process streams and known or assumed partitioning into various waste
streams, products, or media. Although these calculations are helpful in
generating order-of—magnitude estimates, they cannot tfsually account for
temporal variability, which can be extremely important in terms of expo-
sure. The basic elements of mass balances and engineering estimates are to
identify all pertinent processes; describe the releases and products from
each process; and estimate flows and concentrations of toxics emanating
from each process.
These tools are useful for providing rough estimates of emissions of
toxics, but because the accuracy of the estimates is usually low, they are
best used to provide inputs to the sensitivity analysis described in the
preceding subsection, or to augment monitoring data.
Use AvaiZ bie 5urroq e
As previously mentioned, regional or national data on sources and
emission rates can be used as a substitute or auaiity control c ieck for
3—31
-------
inadequate site—specific data. Apply caution when using such data; toxics
emissions are obviously quite variable from area to area. Data from the
same geographical area are usually preferable to data from more distant
regi ons.
Because it is anticipated that surrogate data will be valuable as
“faliback” values, it is recommended that a general catalog of emission
rates be generated for seldom—studied sources such as urban runoff, mine
drainage, residential waste disposal, etc. This recommendation, first
mentioned in section 3.2, would allow either initial input of the cataloged
values (i.e., surrogate data would be initially defined as adequate to meet
data needs) or after the data gathering phase, if inadequate site—spetific
data were available, it could be used as a “faliback” or as a reference for
compari son.
Use sensitivity analysis or pathway modeling to evaluate surrogate
data; if the results are sensitive to the surrogate values, some sampling
and analysis is probably needed to verify these data.
Adapt Analogs
This approach involves extrapolating known data for one well-studied
compound to predict the emissions for a compound with inadequate data. In
this respect, it is another surrogate, but instead of using general geo-
graphic data for the same compound, it uses site—specific information for a
different compound. An example might be the following: suppose that, in a
given study area, airborne emissions of hexane are known, and are due to
its use as a solvent in an industrial process. Pentane, a chemically
similar compound, is an impurity in the solvent and its emissions are
unknown. If the percentage impurity of pentane in the solvent is known,
and the relative vapor pressures are known, the pentane emissions from the
industry can be estimated to originate from the same point as the hexane
emissions and at proportionately lower rates.
The use of analogous compounds is predicated on assumptions of simi-
larities of use and chemical properties. Like the engineering estimates,
-------
.u•
sSui inc.
mass balance calculations, and surrogate data, its use limits the confi-
dence that can be placed on the data. The analytic tools discussed in the
preceding section (sensitivity analysis/modeling) are helpful in evaluating
the significance of the estimates generated by using analogs.
3.6 SYNTHESIZE AND INTERPRET DATA
The final subtask in the sources and emissions inventory is to
synthesize and interpret the emissions data. The synthesis will generate a
table for each source and discharge mode (based on receiving media)
detailing the source locations, flows, concentrations, variability in
discharge, and other pertinent information. Again, since the emissions
data form the basis of most of the other exposure assessment and control
strategy work, consult with the personnel involved in those tasks to ensure
that the outputs from the source and emissions inventory are in a usable
format.
Because the source and emissions inventory is the first task of the
exposure assessment, the conclusions that can be drawn from it are, in a,
sense, preliminary. The final exposure conclusions cannot be reached until
the pathways analysis, receptor and exposure route analysis, and final
assessment are complete. Nevertheless, the source and emissions data alone
can indicate some important trends.
For each of the toxics studied, aggregate the emissions from each
type of source and discharge mode and compare them (e.g., 25% via
industrial! point sources, 30% via urban runoff, 10% via residential/air
emissions, 25% via aqueous imports, and 10% via industrial/land—destined
waste disposal). This kind of analysis can show where the greatest
environmental loadings originate and what the primary discharge modes are.
It may also be valuable for determining the effectiveness of existing
regulations.
Similarly, the data can be aggregated by location of sources. Use
this to identify ‘hot spots’ within the study area; the environmental
pathways task will generate more complete information on ‘hot spots’.
-------
rNar .
Data synthesis must involve a thorough description of all methods,
assumptions, and equations used to generate the sources and emissions
inventory.
Finally, the level of confidence in the data for each of the indi-
vidual sources and emissions rates should be evaluated and tabulated. The
level will vary depending on the adequacy of the data and the strength of
the assumptions used to generate and manipulate the data base.
-------
4. ENVIRONMENTAL PATHWAY ANALYSIS
4.1 OBJECTIVE AND SCOPE OF THIS SECTION
Environmental pathway analysis provides the link between quanti-
fication of source emissions and assessment of receptor exposure, by
estimating the ambient concentrations of toxic substances in various
environmental media. The objective of this section is to provide
prospective users of the geographic methodology with generalized pro-
cedures and techniques for performing environmental pathway analyses.
As such, this section provides general information for: (1) pathway
analyses in specific environmental media (air, soil, water)-—as well
as between these media; (2) factors to be considered in trying to
identify the most important pathways of toxics; (3) selection pro-
cedures for models and guidance for compiling the necessary data to
drive these models; (4) guidance for interpretation of model output
and model validation; and (5) additional issues that may arise during
modeling efforts.
It is not the intent of this section to provide modeling details
and material immediately applicable to perform a modeling effort.
Rather, it is assumed that experienced environmental scientists will
be responsible for mathematical modeling in a geographic study, and
that they will be acquainted with the techniques presented below.
Although it is not possible to provide a check list of “best environ-
mental modeling techniques covering all media, references are avail-
able (Bonazountas 1981) that present a thorough discussion of cur-
rently available models.
4.2 INTRODUCTION
successful aeographic study must quantify the relationship
between chemical releases into the environment and the actual amounts
Arthur D Litti , nc
-------
of these chemicals to which humans and other biota are exposed. Only
with detailed information as to pollutant concentrations in environ-
mental media and the distribution of these concentrations can the
issues of risks, human health, and environmental protection be ad-
dressed. Thus, the purpose of a pathways analysis is to describe the
processes by which pollutants traverse through the environment, and to
quantify the levels of pollutant at different points in each pathway.
Whether the concern is for human health or for environmental
impact, the concentration of the chemical compounds at user-specified
receptors or media of concern must be estimated. Concentration esti-
mates can be obtained by: (1) knowing the distribution of releases of
the material into the natural environment; (2) knowing the environ-
mental conditions influencing the fate (transport/transformation) of
the chemical compounds; (3) knowing the properties (physical/chemical)
of the material; and (4) employing techniques for analyzing infor-
mation gathered. These requirements are schematically shown in Figure
4-1.
Several techniques can be employed to investigate environmental
pathways. Analytic sampling programs, for example, may be designed to
measure actually occurring pollutant concentrations under a variety of
conditions. These data may then be used both to estimate both actual
levels, and to provide information about the processes and pathways
that may be important at specific sites. However, sampling programs
are costly to design and implement, and are also subject to biases due
to the occurrence of unusual conditions during the actual time that
samples are collected. Monitoring data from previous analytical work
may aso be used for these purposes, but such studies are subject to
tne same problems, with the additional limitation that these data are
freouently incomplete, may be outdated, and are rarely extensive enough
to provide a basis for evaluating an entire area.
Artht r D Lit Ie. n .
-------
I
/
/
— — —
-—— Media Concentrations
FIGURE 4-1
SYSTEM DEStGN SCHEMATiC FOR ENVIRONMENTAL MODELING
Calcolate Approximate
Dynamic Partitioning
/
/
/
\
— —
4-3
-------
1athematical computerized models of environmental processes are
frequently used to generate information unavailable by other means, or
to estimate data that are otherwise too costly to obtain. Models may
have the added benefit of allowing the design of analytical programs
that are more cost-effective, less subject to bias, and more attuned
to the dynamics of a local area. Thus, environmental models are use-
ful tools for identifying intermedia transfer mechanisms, generating
necessary but unavailable data, and simulating a wide range of real
and hypothetical environmental scenarios. Some important information,
such as sensitivity of concentrations to changes in environmental
parameters can also be obtained from the use of models. Mathematical
modeling efforts are discussed in the Sections 4.3 and 4.4 below, and
the incorporation of monitoring data is addressed in Section 4.5.
4.3 POTENTIAL PATHWAYS OF TOXICS
4.3.1 General Analytical Approach
Analysis of monitoring data may provide ambient concentration
information for specific conditions, but to quantify the pollutant
pathways in a region, environmental models are necessary. Thus, a
pathway analysis usually involves choosing or adapting environmental
models to be used to simulate pollutant behavior in the region of
interest. Selecting the most appropriate model can be difficult, be-
cause there may be several different single models, mathematical ex-
pressions, or multimedia models to describe the same pathway. There is
a tendency among scientists to use models they are familiarH with,
whether or not they are the most appropriate. Of course, in a given
situation no single model is best, and sometimes it is difficult to
differentiate the advantages and disadvantages of similar models.
Users have to always exercise judgment and apply their valuable ex-
perience to model selection and application.
4-4
Arthur D Little Inc
-------
The two major stages in performing an environmental pathway a-
nalysis are:
(1) A rapid examination of background information relevant
to the geographic specific study; and
(2) Performance of the quantitative analysis.
Stage 1--extends and amplifies the initial scan (see Section
2.5.2) and aims to:
(la) identify potential (probable) pathways of toxics;
(ib) evaluate available data information of a specific re-
gion;
(ic) identify what receptors in the region might be affected
or are of importance for further consideration; and
(id) identify mathematical model candidates to estimate
media concentrations.
Stage 2-—quantitative analysis involves:
(2a) collection of monitoring and other site data;
(2b) definition of the important pathways in the region;
(2c) selection of models to simulate these pathways;
(2d) compilation of input data for the models;
(2e) performance of the simulations;
4-5
Arthur D Little Inc
-------
(2f) analysis of model output/results; and
2g) output validation (with monitoring data) whenever fea-
sible.
4.3.2 Examination of Background Information
The rapid examination of background information can save, con-
siderable time if carried out properly. This effort requires a strong
interaction between team members of the study who: (1) perform the
pollutant source inventory, and (2) perform the exposure analysis.
Interaction with the first group may refine the modeler’s understand-
ing of spatial and temporal distributions of releases into the en-
vironment, and consequently will leadato an informed selection of
single medium models. Interaction with the second group will lead to
an environmental discretization (e.g., number and location of river
segments) and consequently to the optimal implementation of the se-
lected models (e.g., model output requested only at specific locations
or receptors of importance, or at locations where monitoring data will
be available).
Coordination with the exposure group, regarding what receptor
effects are likely to be of greatest concern, is also of importance
during Stage I. Examples of receptor effects are toxicity to fish via
the food chain in humans due to inhalation, chronic functional dis-
order in humans due to dermal absorption, lowered feeding rates in
fish due to gill absorption, or mutagenic effects in humans due to
drinking contaminated water. If the chemical substance is chosen for
a fate analysis, it is most likely that some effect of concern has
been already identified and can be used as guidance in constructing
the optimal set of building blocks that will ultimately constitute the
multimedia fate model.
4—5
Arthur D Little. Inc
-------
Finally, considering the fate properties and potential receptor
effects, during Stage 1, the user will gain a deeper understanding of
the critical environmental pathways, and can establish priorities
regarding Stage 2. For example, in dealing with a chemical whose main
effect is toxicity to benthic organisms, the pathway leading from the
sources through the air into the water and to the receptor would be a
logical priority in establishing a model. Considering relative con-
tributions of multiple release, modes, amounts, or dominant fate prop-
erties, certain pathways will be preferred to others in the rather
large array of possible comb”inations. Flow charts or graphical dia-
grams become useful tools in identifying and establishing the pathway
connections between sources and receptors needed for defining fate -
model approaches.
4.3.3 Quantitative Pathway Analysis
During Stage 2 the pathway analysis is performed. The situation
ini tediately preceding calculations of concentration estimates is per-
haps the most critical one of all: the choice of model logic used for
each medium of interest and the designation of intermedia transfer
processes. From the pathways diagram (Figure 4—2), the user can pro-
ceed to separately consider each medium and select the appropriate
calculation technique or model that characterizes both the behavior in
that medium and the transfers to other media from that medium. Judg-
ment and experience must be exercised in this step in order to choose
the appropriate time and space scales for the problem. A possible
complication might be that chemical transformations may effectively
terminate the pathway within a specific medium. Preliminary half-life
estimates for transfer and transformation, and a comparison among them
will adequately serve the purpose of indicating which fate properties
are important to consider.
The following paragraphs present five hypothetical examples ( [ ravamudan
etal. 1980) intended to illustrate the thinking required to analyze
and model the fate of a chemical. Examples deal with:
4—7
Arthur D Little. Inc.
-------
Near/Far Field Paths ——— This is not a real pathway.
. Near_.,. Far, —
#1 Direct discharge pathways.
#2-4 Intermedia discharge pathways (primary, secondary, potential).
Degradation of substances can take place in any compartment.
Out—of and into basin transfers are not shown.
Source: Bonazountas 1981
FIGURE 4-2 ENVIRONMENTAL PATHWAYS OF TOXIC SUBSTANCES
Arthur D Little In
-------
(1) Point source atmospheric emissions;
(2) , rea source atmospheric emissions;
(3) Point source water discharges (low vapor pressure);
(4) Point source water discharges (rapid hydrolysis);
(5) Non-point source water discharges (runoff source).
(1) First consider the case where pollutant A is pri-
marily emitted from a combustion point sources into the air.
Upon reviewing the properties of A, we find that it has a
relatively low vapor pressure and a low solubility. The
chief target of concern is human receptors exposed through
either the inhalation or ingestion pathways. The reasoning
that we go through leads us to two primary pathways: (a)
direct inhalation from the air of the material which has
adsorbed onto aerosol particles; and (b) an appearance of
material on soil or sediments by way of deposition to the
earth surface. The deposition is followed by runoff of
particles to which the material is adsorbed, in the case of
terrestrial deposition. Direct washout eventually leads
through aquatic pathways. In both cases, the material ends
up in sediments.
Microorganisms are known to bioconcentrate the material
through food chains leading to fish that are part of the
human diet. The environmental simulation will include a
variety of climatologic conditions in the region for point
source plume calculations of non-reactive pollutants. The
nonreactive designation is obtained by comparing the half-
life due to photo—oxidation with the time required for ad-
sorption on aerosol particles and deposition. If physical
fate mechanisms dominate, then the material is assumed to be
nonreactive. The mathematical models for water would be the
homogeneous well-mixed cell models because of the wide dis-
4-9
rrhur D Little ln(
-------
tribution of input frcm the various air to surface depcsi-
ticn processes and the long time scales of interest.
(2) As a second example, consider an area source emis-
sion into the atmosphere of some solvent material (pollutant
B) having a modest vapor pressure and a reasonably high
value of aqueous solubility. The area source associates the
material with an urban plume. This indicates that the air
model to be used would be a widely dispersed source model
that gives a plume spreading rapidly up to the atmospheric
mixing height but spreading horizontally only at a rate
dictated by turbulence in the direction transverse to the
wind. Pollutant B is assumed to have known serious effects
only on aquatic ecosystems. Since the discharges are pri-
marily into the atmosphere, the mechanism of transfer from
the plume to the surface must be part of the fate model.
Again, the pervasive use of the material requires that
a range of atmospheric environmental conditions be employed.
Both dry deposition and rainout will be considered as air to
surface transfer mechanisms because the material has a rea-
sonably high solubility. The Henry’s Law constant can be
used to determine an upper:concentration limit in the rain
water. The rainfall rate along with runoff quantities down-
wind of an urban area are used to establish the material in
the water. Concentrations thereby estimated provide an
input to exposure calculations.
(3) Another example may be water discharges from point
sources. The assumed pollutant C has a low vapor pressure
and moderate solubility. It has a relatively high partition
coefficient between the organic material and sediments. The
main concern regarding this pollutant is the concentration
of its residues in crops that are at the base of the human
4-10
Arthur D Little. Inc
-------
food chain. For discharges into aquatic environments, vari-
able stream flow conditions (high, low, medium) have to be
considered. The stream water may be used as irrigation
water and, therefore, bicta compartments (initially crops)
will be exposed to the material. Bioconcentration factors
for agricultural crops used for cattle feed and subsequently
in meat products give the ultimate concentrations needed.
(4) Consider pollutant 0 that is primarily discharged
into the water. An overwhelming fate property of this ma-
terial is its rapid hydrolysis. The point source discharge
calculation with a half-life appropriate to the hydrolysis
superimposed on it is an appropriate method for estimating
the zone of influence in this material. If human ingestion
through drinking water is the pathway of concern, then the
range of influence between discharges and inlets to drinking
water treatment facilities becomes a key parameter in de-
termining whether the fate calculation is even appropriate.
Various scenarios (see Exposure Analysis section; Chapter 5)
for these site specific conditions are drawn from situations
where the material is apt to be discharged. Assessments are
made based on these data.
(5) As a final example, consider pollutant E that is
primarily discharged into the water from a variety of runoff
sources. A chemical that is disposed of on land, or is
distributed over land by ininediate spreading, directs the
user to the soil compartment modeling section of this chap-
ter (see Section 4.4.4; Soil Compartment Modeling) and cal-
culations proceed to determine relative quantities of ma-
terial in soil moisture, surface runoff, groundwater runoff
and the soil itself.
4-li
Arthur D L tt1e. Inc.
-------
Either the full ooeraticn of the analytical models
described in Section 4.4 of this chapter or simple equilib-
rium partitioning may be employed to arrive at concentration
or transport estimates. If the main target receptor of the
material is a terrestrial ecosystem, then biotic considera-
tions become of importance by providina, for example, an
entry point to a food chain for the material. Also, the
possibility of biodegradation in the soil or in the food
chain at some point must be considered. Losses due to va-
porization from the soil and to leaching must be introduced
into the model calculations in order to arrive at an ap-
propriate mass balance of material.
4.4 MATHEMATICAL MODELING
4.4.1 General Overview
The following paragraphs describe the objectives, structure and
intended use of the three major model categories (air, soil, water).
Coordinating the brief discussion of sample chemical assessment given
above with these brief explanations will provide the user with a per-
spective on the optimal application of the various single medium or
multimedia models. Two other sunuTiaries briefly describe the materials
balance aspect of a fate analysis (Section 4.4.6.2) and the logic by
which an initial approach to the analysis may be determined (Section
4.4.6.1).
Previous modeling efforts have focused upon two distinct modeling
areas: (1) single medium pathway (or fate) models; and (2) multimedia
pathway models, both applicable to a variety of environmental condi-
tions and types of pollutants. Often, however, their applicability is
limited, due to data availability, complex physical boundary condi-
tions or insufficient mathematical description of processes (physi-
cal/chemical) involved. Multimedia models have been constructed in
4-1 2
Arthur D Little, In
-------
the past by sequentially linking (interfacing) single medium models
and using other equations to describe transfers between the media.
Generalized discussions and information regarding environmental
modeling can be found in numerous reports or publications (Miller
1978) and other exposure assessment documents (EPA 1980). EPA ’s
Catalogue of Environmental Models (EPA/MISS 1979) is a recent compre-
hensive compilation of selected models for water quality, surface
runoff, soil quality, air quality, and economic analysis.
Only a short description of the types of models available for
each environmental medium (air, water, soil; Figure 4-2) with a few
examples of each type are presented in this section, since a compre-
hensive “Multi-media Environmental Modeling Background/Catalogue”
(Bonazountas 1981) contains detailed state—of-the-art information.
4.4.2 Air Compartment Modeling
Air modeling efforts encompass both near-field and far-field
investigations (Figure 4—2). Briefly, models of urban scale are the
most advanced of the air quality models. They model the dispersion
process for area and point sources, for reactive and non-reactive
gases. Some allow for variations in meteorology; most are limited to
noncomplex terrain. These models are now being validated with field
data, and efforts are being made to simplify the required data input,
depending on the application, and to assess the models’ accuracy and
costs (Miller 1978). A considerable amount of air modeling effort is
being conducted at the EPA Environmental Sciences Research Laboratory,
Research Triangle Park, North Carolina.
There exists a library of EPA—developed computer programs which
all contain a basic source—oriented atmospheric dispersion model. The
features and assumptions of these programs should be critically ex-
amined with respect to the needs of the geographic study. This ii-
4—13
Arthur D Little
-------
brary of programs continues to grow and the capabilities of more re-
cent additions to the set may be flexible enough to treat source-to-
air and air-to-air pathways assessment for po lutants. Table 4-i
identifies selected programs of the library and their features, in a
geographically bounded exposure assessment it may be necessary to
analyze several toxic substances and/or several modes of exposure for
one or more substances, still within the framework of the air pathway.
Thus, it may be necessary to select arid apply more than one of the
algorithms listed in Table 4—1.
Air-to-surface pathway modeling can be performed either as part
of the air compartment simulation or separately. Additional infor-
mation is given by the “Multi-media Environmental Cata1ogue ’ (Bona-
zountas 1981).
Of all the compartmental simulations, the air compartment is
probably the least well defined of all because of the lack of distinct
boundaries. In attempting to postulate a well-mixed description for
chemical transport and transformation in the air, one must define
boundaries that take account of advection and diffusion. Consequent-
ly, the air compartment may not be a simple box, but rather a time-
variable control surface with permeable boundaries. Fortunately,
these difficulties are somewhat offset by the extensive inventory of
air quality models. For example, plume formulations and vertical dif-
fusion solutions may provide guides for defining the growth and motion
of an air compartment. Its size and variability will depend upon the
geometry of the source array and the chemical lifetime of the material
in question (Aravamudan etal. 1980).
The scope of work when performing an air compartment simulation
begins with an identification of the source characteristics (line
source, point source, area source, or urban area source), and proceeds
with the location and time dependence of the release mode. Physical
characteristics of the source such as mass flow and buoyancy are ob-
4-14
Arthur D 1Jtt e. Inc
-------
Table 4-1
FEATURES OF OFF-THE-SHELF EPA AIR MOL ELS FOR TOXIC SUBSTANCE
EXPOSURE ASSESSMENTS*
RAM ISC CRSTER PAL CDM FIIWAY
Source point, area point, area, point iinEifnt, area, point, line
line, volume line area
Source vary rate many options vary monthly vary hourly day/night none
Dynamics hourly average
Spatial Range near-field near-field & near—field near-field
& far-field f4r-field far-field near-field & far-field
Temporal Range hours to hours to hours to limited sequence seasonal! limited
1 year 1 year 1 year of hours annual sequence
of hours
-a
01
Removal Mechani sins
- chemical first-order first-order first-order none first-order none
- physical none dry deposi- none none none none
tion of
particulate
none elevated or
cut roadway
segiiien Is
Special site elevated elevated elevated none
Considerations terrain terrain terrain
impacts impacts impacts
*Thjs list of models is not comprehensive: these few were selected on
the basis. of their range of features, their flexibility for current air
cjuality regulatory applications, and appropriateness of model outputs.
-------
tamed from site specific information. If the material is chemically
reactive in the air, appropriate rate expressions have to be obtained
from the literature.
Another class of input information to the air compartment in-
cludes meteorological and topographical conditions which will affect
the advection and diffusion of the pollutant. This includes an ex-
animation of the temperature data, wind velocity data, and the stabil-
ity classification of the atmosphere. Gross topographic features
describing the type of terrain that may influence the conditions of
the atmospheric boundary layer are also specified. Table 4-2 lists
input data categories of an air modeling effort.
Meteorological data are archived at the National Climatic Center
in Asheville, North Carolina. Available data sources need to be re-
viewed to determine which station has collected the most representa-
tive data to input into the models. If none of the National Weather
Service stations adequately represent conditions withii the study
area, private, government or industrial data sources should be re-
viewed to consider the suitability of these sources. Monitoring
within the study area is another possibility; but, this option can be
quite costly.
Some important issues to consider in selecting the most repre-
sentative meteorological data base are:
• Topographic effects, such as valley flows and water/land
interfaces should be considered such that wind flow and
stability are representative of the study area.
• If National Weather Service data are used to estimate
stability, caution must be used if local influences
would result in substantially different stability con-
ditions than those predicted solely from insolation arid
4-76
Arthur D Little Inc
-------
Table 4-2
INPUT DATA CATEGORIES OF AIR COMPARTh1ENT SIMULATIONS
(1) Source and Pollutant Characteristics
* source type (point, line, area, urban)
* source location (x,y coordinates; r,e coordinates)
* pollutant type (reactive, non-reactive, gaseous,
particulate, buoyancy of effluent)
* release mode (continuous, instantaneous)
(2) Meteorologic and Topographic Characteristics
* atmospheric stability classification
* temperature, wind velocity
* cross topographic features
(3) Other Data
* dispersion data
* control parameters (output), type of simulation
* time resolution (1 hr, 24 hrs, average, extremes)
4-1 7
Arthur D L tt!e Inc
-------
wind speed. Topographic anomalies, urban effects,
differences in surface heat capacities, and other fac-
tors can influence stability, but are not accounted for
by the Pasquill-Turner stability classification system.
Wind sensors associated with the collection of the data
base should be located such that local influences from
building wake effects, trees, and terrain irregulari-
ties do not unduly modify flow.
The height of the wind sensors should be representative
of the layer into which the pollutants will be dis-
persed. If the heights are significantly different,
corrections based upon wind profile considerations
should be applied during processing of the data.
• The quality assurance of the data set, based upon cali-
bration documentation and data review, is an important
factor regarding its acceptability.
• A long-terni data set (i.e., greater than five years) is
preferable to a short-term data set because of year-to-
year variation of meteorological parameters.
4.4.3 Soil Compartment Modeling
Soil and groundwater models are used to model usaurce_to_soil,
“soiltogroundwater,” ‘groundwater-to-water ” and “soil-to-water”
pathways. The variety and complexity of mathematical models (i.e.,
set of equations) used in soil/groundwater application have increased
dramatically during the last ten years. The proliferation of models
is often bewildering to both the scientist who is trying to keep up
with the research literature and to the user who is trying to find the
best model. Ironically, although the number of model types is large,
4—18
Arthur D Little. Inc
-------
only a few basic processes can be modeled, so that the large number of
apparently different models are the results of a variatj of simpli-
fying assumptions used to reduce the general set of equations to a
solvable set.
Three types of “soil” models can apply to various parts of the
soil environment: (1) “unsaturated soil zone” models that simulate
conditions of a soil zone profile extending between the ground surface
and the groundwater table, and may include runoff and volatilization
processes; the Seasonal Soil Compartment model (SESOIL) is a model of
this type; (2) “groundwater” models that generally simulate the flow
and quality of specific groundwater aquifers; and (3) “watershed”
models that focus on the interactions at the soil surface such as
overland flow (runoff), and sediment transport and quality. The Agri-
cultural Runoff Management (ARM) model is of this type. These model
types can and do overlap.
The choice of a soil model requires finding the best compromise
between needed level of sophistication, available data, and needs and
resources (computer, time, and budget) of the project. Tables 4-3
through 4-6 present sdme of the available models for various aspects
(unsaturated, saturated, unsaturated/saturated, watershed) of soil
modeling, whereas the “Multi-media Modeling Environmental Catalogue”
(Bonazountas 1981) provides in depth discussions for the various path-
ways of a soil compartment.
Once the soil models have been chosen, the data for specific
model runs must be gathered. The actual data required will be spe-
cific to the model used, and the pathways involved. However, some
general data categories can be identified such as source data, cli-
matological data, geographical data, particulate transport data, and
biological data. Table 4—7 presents some parameters associated with
each category. Table 4-8 presents some sources, of data mentioned in
Table 4-7.
4-19
Arthur D Little. Inc
-------
PP! 1!p /co ’ii s..
Applied to irrigation retur,i how
quality s tudi es. i iic lies p I ant ron
uptake of water.
Miscible dlsiilaceu enl. in soils.
t)escr ibes two—ti tinoiis iona I traii.porl
of solutes under a trickle soIari:e.
Node I tog of leactiate and sot
Interactions in an aipa I fer.
Transport of ni tratc iii ti iifli (0 1 tasti
Applied to transport of 2.4-1) in
soils.
Compared results wi Lii observed I kid
data on chl or I tie I rans po it thur i
infiltration.
Applied to C I and 2,4-I) t,uive,iieiit
in two— layered soil cal into
Simulation of pol luLa,it. tiaisspttrt
in I.oiig Island, N.Y.
Applied to phosphorus transport in
soils; assumes constaiid (I ispoI , ion
coeFficient and k loot it; otoilo I I or
phosphorus adsorpt I on.
Describes vertical snhi*Ii± transport
in soils. Can he H oked to Ii) or
FE IIK)delS.
A seasonal soil mode I , S [ S0l L
(21) only)
TABLE 4-3
PARTIAL LiST
OF UNSATURATED (U) SOIL ZONE
POLLUTANT
TRANSPORT
MODELS
Model
!1i er
2)
Model 1 ) Solution Type 3)
jt yT ! .qq . fl .w• .
Type of
Type of
Chemical
Jnteract ions
ti-I
King and llanks (1973, 1975)
10 FUN Tr
—
Ad, lIe, Ce
(1-2 Smajstra)a 1. (1975) 11) 0
11-3 Bresler (1975) 20,C FUN
(‘-4
11-4 van Genuchten and Pinder
(1917) 10 FIN
[ 1-5 llildehrand and Htntnelbau
(1917); Ilildebrand (1915) 10 1 1 ) 14
11-6 Sal Im et at. ( 1916a) 11) 1DM
(1-1 (logs et at. (1976) ID 1DM
(I-U Selim et al. (1977) 10 1DM
11-9 Gureghlan t al. (1917) ID 1DM
(1-10 Shah al. (19/5) I i) FUN
11-11 llonaiountas et dl. (1919) 10 FIlM
i ll? ilnoazountas (1980) 31) 0
ID one—dimens tonal; 20 two—dimens tonal; 3D three—dimensional; A
1)1’) = I liii Fe Di F lerences Method; 1114 = Finite Element. Method; 0 = Other
lransimst; St = Steady-state
4)
I — layered
1 1\d udsitiptioti; Ce ration xdtaiuje (tutu Li— ion transport); tie decay
Tr
Ir
Tr
Ir
Tr
Tr
St
Tr
St
Tr
St
Areal
I. Ad, De, Ce
- Ad
- Ad,De
I Ad
I Ad lIe, Ce
I. Ad
1. Ad,De
- Ad, lIe, 0
(21) only); C Cross-sectional
-------
TABLE 4-4
PARTIAL LIST OF SATURATED (S) SOIL ZONE (GROUNDWATER)
POLLUTANT TRANSPORT MODELS*
l dsI SoI.t an Type .r Type sf’ O,anfcal
Pr1nc 4l tnyestiqat1 i 31a.pe*an1lI v 14 ipeS Plan SS6I atsrsctians an1$p
OO*I SOfl ( 74); 2 .A C Tr Ad Ad. ,f •nanstr$el W lan-
Adbrt$on and Ssrr.clouØ rsdloeattvs imstea tate 19s
1973) 5.* tivir Plate apelfr.r. 1 ID.
S Islad 20 ysan N$195t7 of
peflettan.
S-2 1I,lstra. and BOCA 0 StJTr Ad Ad. Cs CsanI*ri aiss,pttan and - anpe
of sly.!’) .
s- i Th jj. (1911); ______
£2 ( 9 ) 2 9.A P 0 1 St - Ad. Ad Osscribes . ._-*allr pelletfu.
f* ’an sa )t Ipecant.,.
2/30 0 Tr
S- OOIItSOn (1979) 2110 0 1 Ad. ). Ad. Os Three—ss t anat f Sr flu.,
taclellap telflty an stanlat.
r d aStir Ii ats
zu.s.
S-S
-------
TABLE 4-5
PARTIAL LIST OF UNSATURATED/SATURATED (US) SOIL ZONE
POLLUTANT TRANSPORT MODELS
2 3 Type of
Model Model Solutloà Type of Type of Chemical
‘ r !ricth L 1!iL! l ’i. . a!Q! WY. T! i ! _Th!1 J.L_. Jr 1 i
US-i Perez et al. (1914) 20,C FUN I - Groundwater pollution tram
agricultural sources.
US-2 Uzy et al. (1974) 2D,C 0 Tr Ad, Do Contaminant euovt’ment irenu
landfill.
IJS-3 Sykes (1975) 20 ,C FIN St LAn — t.,ntusiilna,it ,wiveinesit hunt a
latielt ii i
LJS-4 Quguid and Keeves
(1976, 1977) 20,C FIN Tr LAn Ad, Do Transport of rndinmsclieles rtt*ii a
waste—disposal site?. llIghL he
app) led Lu woc Lu S I tt,S.
US-5 Segol (19/6, 1977) 20,30 FIN Ir L,An Not appliud yet.
US-6 Van G.mucl.Le,, et al. (1977) 2 0,C FIN Tr L,An Ad, Do tuachate nIuVeuK nL front .t
hygwtthet ical l indi ill.
1)11) one-dimensional; 20 • two-dimensional; 3D three dia nsional; A — Area) (20 only); C • Cross-sectional (20 only)
Analytical; 11*1 Finite Differences Method; FIN — Finite Element Method; 0 • Other
ir iranslent; St • Steady-state
Layered; An anisotropic
Ad - adsorption; Ce cation exchange (emttl—iun transport). Do decay
Source: BonazounLas 1981
-------
TABLE 4-6
PARTIAL LIST OF “WATERSHED” MODELS
Model Carrier Exposure
Acronym — Vehicles Routes Basis and Summary Features
GWMTM1 Ground Mi determinable Based on convective-dispersive mass
Water from knowledge of transqort equation modified for 1st order
concentration decay. i-dimensional treatment. Surface
concentration can be constant or exponen-
tially varying. Vertical seepage constant.
Soil saturated or unsaturated.
GWMTM2 Ground All determinable Describes concentration distribution In
Water from knowledge of 2 underground dimensions. Advection and
concentration dispersion In 2 dimensions with 1st order
decay and an exponentially decaying
Gaussian boundary condition. Useful for
sanitary landfills wastewater lagoons,
and chemical dumps.
EPAURA Runoff All determinable Assumes accumulated pollutants are all
Water from knowledge of carried off in rainfall on an area of
concentration impervious surface.
EPARRB Runoff All determinable Assumes all rural areas have slope percen-
Water from -knowledge of tages allowing erosion to take place.
concentration Calculates delivered sediment to a water
body bas d on the universal soil loss
equation. Pollutant loads are outputed.
1
0
-------
TABLE 4-6 (contInued)
Model Carrier Exposure
Acronym Vehicles Routes Basis and Summary Features
tIPS Runoff All determinable Used to estimate nonpoint source pollutant
Water from knowledge of loads In urban and rural settings.
concentration
AGRIJN Runoff All determinable Simulates hydrology and channel pollutant
Water from knowledge of loads for agricultural watersheds. Uses
concentration the universal soil loss equation and
Horton’s equation to compute Infiltration
rates. Requires specification of soil
parameters.
4 .
ARM -I ! Runoff All determinable Mass balance. Assumes all runoff water
Water from knowledge of from locations in the watershed. Pollu-
concentration tant transformations are approximated by
a series of 1st rate expressions.
Arrhentus equation used to adjust rates
to different temperatures. Partitioning
between phases assumed to be Instantaneous.
HSPF Runoff All determinable Mass balance. Assumes all runoff water
Water from knowledge of from locations in the watershed. Pollu-
concentration tant transformations are approximated by
a series of 1st rate expressions.
Arrhenlus equation used to adjust rates
to different temperatures. Partitioning
between phases assumed to be Instantaneous.
Water body simulations.
-------
(t
TABLE 4-6 (continued)
Model Carrier Exposure
Acronym Vehicles Routes Basis and Suinnary Features
SESOIL Runoff All determinable Statistical seasonal (month, year) user
Sediment from knowledge of specified simulation, stochastic poisson
concentration and ganiua distributed functions. In-
depth chemistry, analysis, applicable to
both organic and inorganic compounds.
-I .
)
U i
Source: Bonazountas (1981).
-------
Table 4-7
SOIL COMPARTMENT ENVIRONMENTAL PARAMETERS
(EXPOSURE PATHWAY MATRIX)
PARAMETER EXPOSURE MEDIUM
(Variables)
Soil and Surface
Groundwater Water Air
• CLIMATE X X X
Evapotranspi ration
Temperature
Lati tucie
Sunlight
Plant Cover
Humidity
•Cloud Cover
Wind
Precipitation
• SOIL X X
Porosity
Density
Hydraulic conductivity
Permeability
Adsorption capacity
Organic carbon content
Clay content
• GEOGRAPHY X X X
Slope
Surface storage
Terrain
Receptor locations
• PARTICLE TRANSPORT X X
Wind
Flood flow
4-26
Arthur 1) Little Inc
-------
Table 4-7 (Continued)
PARAMETER EXPOSURE MEDIUM
(Variables)
Soil and Surface
Groundwater Water Air
• BIOLOGICAL ACTIVITY X X
Plant coverage
Plant types
Bioconcentration (both by plants & animals)
Biologi cal degradation
• SOURCES X X X
Emission rates
Release mechanisms
Patterns of operation (continuous, batch)
Locations
4-27
Arthur D Little Inc
-------
Table 4-8
DATA SOURCES FOR SOIL COMPARTMENT ENVIRONMENTAL
PATHWAY ANALYSES
Climatological Data :
National Oceanographic and Atmospheric A ninistration (NOAA)
Reports
Soil Data :
United States Geological Survey (USGS) Reports
United States Army Corps of Engineer Reports
Federal and Regional Agricultural Agencies
Landfill and Industrial Operators
Geography Data :
USGS Maps and Reports
Census Bureau Data
Particle Transpori ;
USGS Surveys
Army Corps of Engineers Reports
Biological Activity :
Agri cultural Agencies
Environmental Agencies
Uni versi ties
Sources :
Industry Representatives and Operators
Regional Discharge Permit Boards
Envi ronmental Agencies
4-28
Arthur U) Little Inc
-------
4.4.4 Water Compartment Modeling
Water quality models describe transport and transformation of
substances from point discharges and non-point sources, including
urban and non-urban runoff. Many models predict concentration levels
for water quality parameters such as dissolved oxygen, nitrogen, pho-
phorous, and pH; others can predict concentration levels of certain
chemicals in the water body. Most of the models can account only for
organic compounds (transport, dispersion), others can account for both
organic and inorganic (metal) compounds.
Models are structured according to the physical shape of the
waterbody (lake, river, estuary) and according to the type of dis-
charges (point source, non—point source). A few of these models are
mathematically structured around a mass balance equation describing
inflow, outflow and transformation of the chemicals in a given water
volume. Others are based on partial differential equations describing
flow and mass transport in fluid, the latter being from mass balances
of a given element.
Water compartment models are applicable to rivers, ponds, lakes,
estuaries and coastal nearshore waters. A simulation should be capa-
ble of providing estimates of either time-dependent or quasi—steady-
state aqueous concentrations. The compartment size and time scales
are established by the flow characteristics, the chemical degradation
processes, the discharge characteristics, the inter-compartmental
transfer times for each chemical/waterbody configuration, the exposure
assessment requirements, etc. Pollution input from other compartments
(e.g., air, soil, upstream water compartment) and pollutant transfer
to other downstream or adjacent compartments are part of a water com-
partment modeling effort.
4... 29
Arthur C) Uttle. 1n
-------
Two fundamental situations give rise to the family of water fate
models. One is a well-mixed water body, either in place (e.g., pond),
or in motion (e.g., segment of a stream undergoing time—variable boundary
conditions). The other relates to the problem of non-uniform concentra-
tion fields associated with point sources of discharge, and results in
the selection of a near—field (i.e., plume) or a far-field model.
Various models and additional information are presented in the
“Multimedia Environmental Modeling Background/Catalogue” (Bonazountas
1981) and in Table 4-9.
In general, water models that are strong in physics (i.e., hy—
drodynamic models) are weak in chemistry (i.e., water compartment
models) and vice versa. The choice of which model type to use depends
on the objective of the study. A selected water quality model should
in general be capable of:
(1) Simulating all the important aquatic environments (riv—
ers, estuaries, pools) within the region of study.
(2) Accounting for adsorption of pollutants on abiotic
surfaces, particulate matter and bed sediment.
(3) Treating movement of dissolved pollutants in water, as
well as movement of pollutant associated with sediment.
(4) Handling water-sediment exchange.
(5) Accounting for important loss processes, both transport
and transformation (e.g., volatilization, hydrolysis,
etc.).
Two useful water models, known for their diversity, strength in
chemistry and physics, and recent applicability are:
4-30
Arthur C) Little Inc
-------
TABLE 4-9
PARTIAL LIST OF WATER-BODY MODELS
Model Carrier Exposure
Acronym Vehicles Routes Basis and Suninary Features
EXPLORE-I Water All determinable Handles 1-dimensional flow in streams and
from knowledge of rivers, 2-dimensional flow In shallow lakes
concentration and estuaries. Capable of handling constant
or time-varying point or diffuse sources of
substances.
MS Water All determinable Sophisticated model with use In determining
CLEANER from knowledge of bloaccumulation of toxic substances.
concentration
DIURNAL Water All determinable Used to predict diurnal fluctuations during
from knowledge of periodic steady state conditions. Useful
concentration when algal oxygen production is related to
concentration of effective agent.
FEOBAKO3 Water All determinable Mass balance. Handles consecutive reactions
from knowledge of and 1st order kinetics. Assumes that steady
concentration state conditions apply.
PLUME Water All determinable Considers only mixing and dilution with no
from knowledge of water flow in a steady state stratified
concentration environment. 3-dimensional output.
Provides concentration data alOng plume
centerline.
-------
TABLE 4-9 (continued)
Model Carrier Exposure
Acron ym Vehicles Routes - is and Suu*nary Feat r __ .__..
QUAL-Il Water All determinable Simulates dynamic behavior of constitUentS
from knowledge of subject to dispersion, flow, nutrient
concentration cycles, and algal growth. 1-dimensional
for networks. Considers 1st order decay.
Only point discharges and constant inflows
are considered. Instantaneous mixing is
assumed.
SEM Water All determinable Handles only point source inputs to streams.
from knowledge of rivers, and shallow non-stratified lakes.
concentration 1st order decay. Simulates dilution,
advection, and temperature effects.
1-dimensional. Considers uncoupled chemical
reactions. Suitable for hand calculator.
ESOO1 Water All determinable Mass balance. Can be used for sequential
from knowledge of reactions of two substances having 1st
concentration order kinetics. Tidally averaged, steady
state model. Suitable for complex water
networks (100 junctIons. 50-100 sections).
OEM Water All determinable Real-time, link node model simulating
from knowledge of unsteady tidal flow and dispersion In an
concentration estuary. Two-dimensional flow. Hydraulic
and quality (pollutant concentration)
model components. Mass balance checks at
each junction. Predicts time varying
concentrations.
-I
4- ,
-------
TABLE 4-9 (continued)
Model Carrier Exposure
Acronym Vehicles Routes Basis and Suninary Features
TIM Water Mi determinable Derivative model of OEM. Handles up to
from knowledge of 4 consitituents with coupled or non-coupled
concentration reactions with 1st order decay. Used for
networks (300 junctIons, 300 channels).
HARO3 Water All determinable Mass balance. Multidimensional, steady
from knowledge of state model for two reacting substances.
concentration Handles 1st order kinetics. Incorporates
convective-diffusive mass transport with
source and decay terms.
Water All determinable Simulates near shore currents and exchange
from knowledge of processes. Sophisticated treatment of
concentration dispersion, advection, and dynamic plumes.
REDEQI Water All determinable Computes equilibria for up to 20 metals arid
from knowledge of 30 anions in a system. Includes coniplexation,
concentration precipitation, redox, and pH dependent
reactions.
RECEI V-lI Water All determinable Two-dimensional model representing advection,
from know’edge of dispersion and dilution. Used on networks.
concentration Can simulate coupled and non-coupled cheinic l
reactions. 1st order decay considered.
Assumes Instantaneous mixing.
-------
••1
(A)
TABLE 4-9 (continued)
Model Carrier Exposure
Acr ym Vehicl s Ro ! teS Bas1s and Suninary Featur j
EXAMS Water AU determinable A water body compartment modeL providing
from knowledge of user specified flexibility In application.
concentration Strong In chemistry of organic compounds.
No metal chemistry, no sediment transporta-
tion routines.
Sources: EPA/MII3S (1981); BonazountaS (1981).
-------
(1) EXAMS (Exposure Analysis Modeling System); a water
compartment model of EPA (1980), Athens Research Labor-
atory; and
(2) EXPLORE, a river basin water quality model developed by
Battelle Memorial Institute.
Basic parameters for the nation’s water bodies to run water qual-
ity models, such as pH, DO, etc, are reported in STORET, NASQUAN, USGS
and EPA data bases (see Section 4.5.4).
4.4.5 Multimedia Modeling
An overall model for the environmental fate of a substance in the
air, soil, subsoil/groundwater, water, and biotic environment of an
area can be constructed by sequentially interfacing single medium
models, such as those described in the previous sections. However,
the usefulness of such a “super-model” might be questioned, not be-
cause of the capability of this model to predict overall fate of pol-
lutants in the environment, but because no “general” model can be
developed that can account for all environments and all pollutants.
Limitations of multimedia models are mainly due to the limitations of
single medium models available and to the mathematics of interfacing
those models.
Multimedia environmental problems are receiving increasing at-
tention as more of them are recognized. A few multimedia models for
assessing such problems are available; all of them focused on poi-
lutant fate in a triple environment: air-watershed—water body. A few
of the existing models are presented in Table 4-10. Their applica-
bility, limitations, and deficiencies are briefly described in the
“Multi -Media Environmental Modeling Background/Catalogue (Bonazountas
1981).
4—35
Arthur 1) Little. Inc
-------
Table 4-10
PARTIAL LIST OF MULTI-MEDIA MODELS
Acronym Major Features — Description
(JIM “air-watershed-stream” The focus of the Universal Transport Model (VIM) has been
simulations on watershed phenomena that are important to trace contami-
nant mobility and pathways Including atmospheric dispersion,
deposition on land, entrainment and erosion in overland
flow, movement and exchange of soluble chemical species
with a moisture flux through the soil matrix, transport in
the stream channel, and exchange with the stream sediment.
CMRA “Overland Flow - Stream 1 ’ The Chemical Migration and Risk Assessment (CMRA) methodology
contaminant transport combines off-the-shelf overland and In-stream transport models,
statistical analysis routines, and risk assessment procedures
Into a single system, In order to predict the probability of
acute and chronic Impact on aquatic biota.
ALWAS “Air-watershed-water” The objective of this study was similar to the objective of
contaminant transport the VIM model 1 arid theoretically there is no conceptual
difference between the VIM and the ALWAS model. Both treat
atmospheric emissions, surface runoff and stream quality.
The scope of work, however, of AIWAS focuses in Interfacing
an atmospheric model, the nonpoint source (rIPS watershed
model and the exposure assessment modeling system (EXAMS)
model Into one single multi-media package. In addition,
AIWAS provide more accurate insight into the deposition
processes from air-to-surface (water, land).
TOHM “Air-watershed-water” The environmental Transport Model of Heavy Metals (buM) is
metal transport tailored to heavy metals, stack emissions of coal-fired
electric utilities, and semi-arid climate reaches. The
development of TOUM consisted of Interfacing an atmospheric
model, and a soil chemistry model.
-------
Table 4-10 (Continued)
Acronym t jor Features Descript
CONCEPTUAL l PestlcideH_in A conceptual model for the movement of pesitcides through
environment the environment developed by the EPA’s alternative cheiiiical
program. The study presents a conceptual model for the
movement and disposition of pesticides in the environment.
A multi-media model is built up conceptually from simple
modules representing basic processes and components of air,
soil, and water.
Environmental An Environmental Partitioning Model for Risk Assessment
Partitioning Model of Chemicals, encompassing air, soil, water, biotic and
partltlonln9 logics. Unpublished document (Aravamudan
et al. 1980).
Source: Bonazountas (1981).
-------
By counting the limited number of individual models presented in
the previous sections, for all potential pathways, it is easy to re-
alize that a missing feature from a key model (e.g., volatilization
from ARM) might be an adequate reason for a basic modeling package to
require modifications. Furthermore, as time progresses, new models
which are either directly applicable or applicable after modification
will become available. Such new models will replace older packages
and may simultaneously facilitate and improve simulations. Therefore,
the design of a multimedia package can never be considered as “final ”;
rather it has to be regarded as continuously evolving. A multimedia
model design interfacing (1) EXAMS, (2) SERATRA, (3) SESOIL, (4) two
or three groundwater models, (5) two or three air models, and (6) an
air-to-surface model has been proposed (Bonazountas 1981).
Two issues are worth consideration: (1) the temporal and (2) the
spatial resolution of both the single medium models employed by a
multimedia package and the multimedia package j se (Bonazountas
1981). These two issues frequently limit the general applicability of
multimedia model, in a sense that so many input data are required that
no user is willing to be subjected to this laborious and costly exer-
cise.
Finally, when dealing with the notion of designing or (applying)
a multimedia model, a developer is confronted with the optimal selec-
tion of single medium models, given: (1) study objectives, (2) data
availability (3) modeling expenses, time, computational costs (4)
mathematical constraints of single medium models employed, (5) expan-
dable logic and model capabilities of the multimedia package, (6)
calibration requirements of the multimedia package, and (7) clarity
for model users.
4-38
Arthur D Little. Inc.
-------
4.4.6 Other Modeling Issues
Each of the compartment models discussed above deals with chemi-
cal properties and transformation coefficients. Transformation co-
efficients/rates and other parameters may not be available in the
literature; therefore, model users may decide to compile these proper-
ties from a wide range of default values arranged by chemical class,
and allow for a sensitivity analysis to provide decision makers with
the relative importance of the process simulated in cases where no
site specific input information exist.
Another issue is the input of materials balance data. The ma-
terials balance can be defined as an array of the release rates of
chemicals from the industrial environment to the first point of entry
into the natural environment, with geographical and temporal distri-
butions specified. Two major components are encompassed by the ma-
terials balance; namely, the source identifications and the estimates
of the environmental loadings. Factors such as geographic scale,
industrial sector, environmental media, chemical class and time scale
for release rate all combine to determine the scope and focus of a
materials balance. In developing a materials balance approach, it is
fundamental that all major sources--natural and anthropogenic, de—
liberate and fugitive——that can lead to potential exposure of sen—
sitive geographic specific receptors be identified. Going beyond the
identification, the release rate of these sources must be character-
ized in coordination with the “source/emissions” group of this study.
4-39
Arthur D Uttle 1n
-------
4.5 EVALUATION OF 1ONITORING DATA
Monitoring data provide information on toxics that is useful in all
phases of a geographic toxics study, and the evaluation of monitoring data
is an essential part of such a study. Without a monitoring data base, the
problem identification and pollutant selection in the initial scan are more
difficult. A good monitoring data base is necessary to validate and guide
the sources and emissions inventory, the environmental pathways analysis,
the exposure assessment, and the control strategies development. Further-
more, a toxics study supported by solid monitoring data is more likely to
withstand legal and scientific scrutiny than a study without such data.
The evaluation of monitoring data should begin as soon as possible
after the initiation of a geographic study, because this analysis will pro-
vide guidance and feedback* to all subsequent phases of the study. A
systematjc approach that fully exploits the information available from the
monitoring data Is given below (with the exception of its use in the
preliminary procedures, which is discussed In Section 2). The major steps
in evaluating monitoring data are the same as for the sources and emissions
inventory.
4.5.1 Definition of Needs
In the initial scan an effort to acquire all available monitoring
data will have been made. Based on the current understanding of the
potential toxics problems, decide which environmental pathways, exposure
routes, and media are likely to be important with respect to the toxics
problems in the area. The monitoring data needs should be prioritized
based on these considerations. Ideally, a good monitoring data base is
desirable for all potential exposure routes and environmental media,
including ambient air, surface water, groundwater, soils and sediments,
biota, drinking water, and food. In practice, however, resource and time
constraints will limit the collection of new monitoring data and the
investigator must decide which data bases are most important in a parti-.
cular geographic area. For instance, if all the known toxics sources
discharge exclusively to surface water, monitoring data on toxics in
4-40
-------
surface water, sediments, aquatic biota, and drinking water (if ft comes
from the polluted surface water) may be all that is necessary.
The minimum monitoring data base necessary for each selected media
must also be determined. This will vary accordinq to the nature of the
toxics problem, and there are no fixed guidelines applicable to all
studies. The following questions should be considered:
• Will the media be modeled? If so, what is the minimum number of
sample locations necessary to adequately calibrate/validate the
model? Are there certain geographic locations where monitoring
data are needed?
• If the media will not be modeled but exposure will be assessed,
what is the minimum number of sample locations needed to adequately
characterize the environmental distribution and the degree of
exposure? Are there specific sites where monitoring data are
needed?
• Whet is the minimum number of samples and period of record needed
to adequately characterize the average concentration and the range
of concentrations for a particular sample site?
• What is the oldest acceptable age of monitoring data?
4.5.2 Review and Compilation of Data
After the minimum monitoring data needs have been established, all
available monitoring data are compiled, media by media. As a first step,
the number of samples, range of concentrations, and mean and median concen-
trations (if applicable) are compiled for each source/sample site. tn the
case of air and water, it may be useful to plot sources/sample sites on a
map of the study area and note the geographic distribution of toxics
concentration. This may elucidate trends (i.e., increase or decrease in
concentrations in the downstream portion of a river) and indicate toxics
1 hot spots’.
4-41
-------
At this stage, reconsider the previously defined toxics problem in
light of the monitoring data. Where possible, compare measured concen-
trations with existing environmental and health criteria. If the
monitoring data base is comprehensive there may be adequate evidence to
modify the source/discharge modes, environmental .oathways, or exposure
routes considered in the study.
4.5.3 Evaluate Adequacy of the Data
Once assen led, evaluate the monitoring data for each media with
respect to the minimum data requirements established according to Section
4.4.1. If the minimum requirements are met, the data can be synthesized
and interpreted as outlifled in Section 4.4.5. Otherwise, plans must be
made to acquire new data, as outlined below.
4.5.4 Design and Implement Plans for New Data Collection
If the minimum data requirements are not met, either revise the goals
or plan to acquire new data. If new data are needed, design a samplinq
program that fills gaps while meeting the monetary and time constraints of
the study. Coordinate sampling of ambient media to sampling of sources to
provide the temporal and spatial relationships that support the establish-
ment of a causal link. Developing a methodology for such a sampling
program is beyond the scope of this study; it is recommended that EPA
develop such a methodology, tailoring existing guidelines on sampling
design to the geographic approach. The use of surrogate (regional or
national) data is not recommended as a substitute for site—specific
monitoring data, unless it can be conclusively shown that the surrogate
data are representaive of local conditions.
4.5.5 Synthesize and Interpret Data
Once the complete monitoring data base has been assembled, e valuated,
and expanded (if necessary), the synthesis and interpretation begins. This
is avery important step because the monitoring data are the best tools for
evaluating the accuracy of the source and emissions inventory, the environ-
mental pathways analysis, and the exposure assessment. The actual proce-
dure followed in the data synthesis and interpretation depends on the moni—
torinq data base and the extent of modeling.
4-42
-------
If a given media was not modeled, use the monitoring data to check on
the sources and emissions inventory by: 1) qualitative comparisons and 2)
performing rough mass balance calculations based on the source and
emissions inventory and comparing these estimates with measured data. For
the qualitative evaluation, compare ambient concentrations with the known
sources of toxics to see whether emissions are reflected in ambient con-
centrations and whether any toxics detected through monitoring are not
accounted for by the emissions Inventory. The use of mass balance could be
used far an unmodeled river, for example, by estimating toxics loadiriq
from various sources in conjunction with known stream flow to develop a
rough (order of magnitude) estimate of expected toxics concentrations,
which can be compared to monitoring data. In the absence of modeling, use
monitoring data to develop exposure levels and to assist in the development
of control strategies.
For media that are modeled in the environmental pathway ai a)ysis, use
the monitoring data to verify the emissions inventory, the modeling, and
the exposure assessment. To verify the emissions inventory for a given
media, find the location of the receptor/model compartment closest geo—
graphically to each monitoring site. Compile the predicted and measured
concentrations for each pair of corresponding modeling/monitoring sites.
Compare these predicted and measured values for the entire study area. If
there are significant discrepancies between these data for one or more
sites that may be due to incorrect emissions estimates, use the model as an
investigative tool to site hypothetical undocumented sources of toxics or
generate emission rates from known sources that more closely account for
observed concentrations. Using this informationtry to find these sources,
which were missed in the initial data gathering effort. Alternatively, if
the resources of the study do not permit a computerized investigation with
the model, perform a qualitative ‘sensitivity analysis’, by carefully
examining the trends in modeled versus monitored toxic levels. Such an
evaluation may suggest the possible reasons for the observed discrepancies.
If the comparison of modeling results with monitoring data suggests
that there are significant discrepancies between the two that are du to
4-43
-------
Inc.
problems in the modeling, use the monitoring data to calibrate or validate
the model, if resources permit. tt should be emphasized, however, that
model calibration and validation can be expensive and that they generally
require a very sound monitoring data base in order to be accurate. For
example, the use of ambient monitoring data to enhance the predictive
capability of atmospheric dispersion models is justifiable only if the
monitoring data base is long—term (including at least one year of data),
intensive (based on hourly sampling), and based on numerous well—located
monitoring stations. In addition, comprehensive meteorological data nvst
be available. In model calibration an attempt is made to adjust model
estimates to better correspond with measured data. In this approach the
reason for discrepancies between modeled and monitored data is not
necessarily determined. A correction factor is developed using statistical
methods (e.g., regression analysis) and implemented to improve the agree-
ment with the monitoring data. Model validation, on the other hand, uses
statistical tools such as time series analysis and correlation analysis to
determine the reason for the discrepancy between modeled and monitored data
and to adjust the model accordingly. The implementation of a validation
includes the following steps: 1) comparison of measured and predicted
values using a statistical package and a regression analysis; 2) determi-
nation of the cause(s) of any discrepancy involving emission data,
meteorological data, or model; 3) correction(s) to the appropriate
component; 4) rerun.
Finally, monitoring data can be used alone or in conjunction with
modeling to pinpoint the problems that deserve consideration in the
development of control strategies. This is discussed in Section 7.2.1.
4-44
-------
i.5.6 Model Validation
Once the input data are complete, the models are run and the
results have to be analyzed. Various methods are available for these
analyses ranging from simple readings of output listings to sophisti-
cated computer analyses of output on computer storage devices (tape,
disk, data bases, etc.). The two major steps for analyzing/evaluating
model outputs are model validation and model sensitivity analysis.
The latter step is discussed in the next section.
Final model predictions should be validated for consistency with
any available monitoring data (see Section 4.5). However, a dis-
agreement in absolute levels does not necessarily indicate that either
method is incorrect, or that either data set needs revision. Field
sampling approaches and modeling approaches rely on two different
perspectives of the same situation. Field data give a concentration
at one point in time and space, models predict “average” concentra-
tions for some particular conditions. Thus, field and model results
may differ and still both be correct. Some possible reasons for a
discrepancy are:
• The field sample was taken from a spot with atypical
concentrations (e.g., a water sample may be 1.5 feet
from a source, and so give abnormally high readings).
• The sample was taken under atypical conditions (e.g.,
on the one day/month that it rained); model results
were calculated for average conditions, which rarely
occur (e.g., coastal water studied at a mean tide
level).
• The sample contained more than one phase (e.g., a water
sample contained some suspended sediment).
4-45
Arthur U) little Inc
-------
The extraction procedure for the sample was under or
over-efficient (e.g., it not only extracted all organic
pollutants from a soil sample but also dissolved the
soil).
Both types of data are important and convey different types of
information. These data can be compared if the differing approaches
are kept in mind.
4.5.7 Sensitivity Analysis and Scenarios
It is frequently worthwhile to perform sensitivity analyses to
determine the effect on the predicted concentrations caused by a change
in the input parameters. These sensitivity analyses are particularly
important when data gaps or uncertain input values exist. It may also
be useful to re-run the models to estimate the impact of various control
strategies on toxics distribution and concentrations in the environment.
Two main techniques are widely used to perform sensitivity analyses:
(1) model simulations; (2) analytic techniques.
Model simulations are performed by running and rerunning the
model, simultaneously varying the value of one or more parameters
following a “scenario logic. A typical scenario for a water compart-
rnent simulation is shown in Table 4-11. Model concentration pre-
dictions may be compared to monitoring data as described in the pre-
vious section.
Analytic techniques of linear systems theory (USDA 1973) and
optimization theory (Haimes 1977) may correlate sensitivity of model
input (e.g., effluent quantity) to model output (e.g., ambient con-
centrations) without performing multiple model reruns (simulations).
An example of such a methodology is presented in Chapter 7 (Develop-
ment of Control Strategies) of this report.
4-46
Arthur 1) Little Inc
-------
Water Scenarios: STREAMS
PHYSICAL PARAMETERS
Width, Meters
Average Discharge —
Cubic Meters/Second
Average Velocity, rn/s
Average Depth, rn
Low Q
Low Velocity
Low Depth
CHEMICAL PARAMETERS
in mg/i, Annual Data
Suspended Sediments
Dissolved Solids
Chloride
Total CaCO 3 hardness
SEASONAL DATA
Dissolved Oxygen:
February
April
August
November
Total Organic Carbon:
February
April
August
November
Table 4-11
A TYPICAL WATER SCENARIO
REGION: -
STREAM SIZES
Carge Med-Large Medium
Sinai I
Mean
X
20%
Value
80%
Value
OBSERVED
Minima
Maxima
pH
4-47
Arthur!) Little. Inc.
-------
4.5.8 Monitoring Data Acquisition (Data Bases )
Monitoring data can be obtained from: (1) site specific sampling
programs or (2) computerized data bases. Table 4—12 lists selected
data requirements for a geographic pathway analysis, while Table 4—13
lists widely used computerized data (direct, indirect) bases. Direct
data can identify areas where toxic pollutants have been found in the
environment and at what concentrations. Indirect data can be used to
identify areas where one might suspect that toxic pollution is oc-
curring, i.e., by locating concentrations of industries which are
suspected of contributing to toxic pollution.
4. -48
Arthur D Little Inc
-------
Table 4-12
SELECTED DATA REQUIREMENTS OF PATHWAY ANALYSES
Class of Data Possible Sources of Data
Pollutant Sources Reports ; TSCA inventory, SRI chemical dlrectQry,
Radian report on Organic Chemical Producers,
• Industrial surveys of manufacturers NEIC Kanawha Valley Study, NEIC follow-up report
and users. on 5 industries.
• RCRA inspection forms. Other : State offices, production and use reports,
risk assessments, other materials balance from
• Registration data. Chamber of Coninerce,
• Air quality emissions inventories.
quantification of Emissions Regional (EPA): Air Quality Monitoring Branch,
Water Quality Monitoring Branch, Hazardous Waste
• Air: stationary source emissions permits, Task Force, Enforcement Branches, Environmental
data from inspection/maintenance programs Emergency Branch (Survey and Analysis Division),
for non-point sources, compliance data. Pesticides Branch. (Also U.S. EPA Water Quality
• Water: NPDES permits, other permits Analysis Branch for NURP and POTW Studies and
(state and local. Effluent Guidelines Effluent Guidelines Division.) Also (USCOE)
Documents. U.S. Corp of Engineers.
• Land; licensing or permit information State : Water Development Authority, Natural
for transport and disposal of hazardous Resources Department (Water Resources, Air
and solid wastes. Pollution Control Covinission.
• Other: information on spills, pesticide Others : Ohio River Sanitary Cornisslon (ORSANCO),
accidents. Inspection reports. TSCA substantial risk notices, USGS Water Quality
Alert, county and city offices, local planning
agencies, contact with Industries, other
literature, county health offices, Bureau of Mines.
-------
Table 4-12 (Continued)
Class of Data _____ Possible Sources of Data
Environmental_Fate Federal*: National Weather Bureau, NOAA. USGS,
USDA ( o1l Conservation Service), U.S. EPA Water
• Climatic data: i.e., precipitation, wind, Quality Analysis Branch (STORET/TOXET data bases,
wind speed, temperature, pressure, etc. REACH, stream gages data), Bureau of Mines, LJSCOE.
• Hydrologic data: i.e., surface and ground- State : Natural Resources (Environmental Analysis,
water distribution, quantity, etc. Water Resources, Wildlife Resources, Parks and
• Soil data: i.e., soil classification, Recreation). Geological and Economic Survey,
geology, soil permeability, etc. Department of Agriculture (Soil Conservation
Department, Air Pollution Control Connission,
• Land use data. Water Development Authority.
• Chemical fate data: physical, chemical
and biological properties of pollutants. Other Organizations : University of W. Virginia,
Ohio River SanTtary Coninission (ORSANCO), county
• Monitoring data: concentrations of and city offices, literature (i.e., Versar fate
pollutants in environmental compartments. documents, past risk assessments, other).
-------
Table 4—13
DATA BASES OVERVIEW; PAThWAY ANALYSES
(1) DIRECT DATA
• AMBIENT MONITORING DATA
Water
- STORET - Storage and Retrieval of Water Related Data
- NASQAN - National Stream Quality Accounting Network
- Surface water monitoring
- Monitoring to detect previously unrecognized pollutants
in surface waters
Air
- SAROAD - Storage and Retrieval of Aerometric Data
Soil
- Ecological Monitoring System
Groundwater
- UIC/HWIS - Underground Injection Control!
Hazardous Waste Management Information System
• SOURCE SPECIFIC DATA
Effluent Monitoring Data (water
- Needs survey
- IFB organic data base
- Energy and mining point source category data base
Emissions Monitoring Data (air)
- NEDS - National Emissions Data System
- HATREMS - Hazardous and Trace Emissions System
(2) INDIRECT DATA
• INDUSTRY BASED DATA
- Data Collection Portfolio for Industrial Waste Discharges
- EADS - Environmental Assessment Data Systems
4-51
Arthur 1) Little Inc.
-------
4.6 SUMMARY ND CONCLUSIONS
The major steps in performing an environmental pathway analysis
are:
(1) Collection of monitoring and site data
(2) Defining the important pathways in the Region
(3) Choosing models to simulate these pathways
(4) Compiling data to drive the models
(5) Analyzing the results
The results from this analysis can then be used as input to the
Exposure Assessment Analysis (see Chapter 6), which evaluates the
degree of human and other biotic exposure to toxic substances in a
region, and as a basis for decisions regarding the environmental qual-
ity of the region. Therefore, an increased degree of collaboration is
required between team members of the pathway analysis group and team
members of both the pollutant source inventory and the exposure as-
sessment groups, in order to select, discretize and apply models opti-
mally.
Engineering judgment and professional experience are key factors
for a successful pathway simulation. Given the current number of
multimedia models it is advisable to perform multimedia simulation by
appropriately selecting and interfacing single medium models rather
than seeking immediate use of an existing multimedia package that can
be very costly (professional time, computer cost) and may not be di-
rectly applicable to a study.
-52
Arthur 1) LittLe Inc
-------
Study duration is a function of: (1) the study area (size, com-
plexity, location), (2) data availability , and (3) orofessional ex-
perience of performing organization (agency, consultant). A period of
3 to 5 months appears to be a reasonable estimate.
A pathway analysis requires an interdisciplinary team of:
(1) a senior task manager specializing in modeling of at
least two environmental compartments (e.g., soil and
water), who will be also effectively involved (as a
subtask leader) in one of them;
(2) three subtask leaders (project manager in one) for the
air, soil/groundwater and water compartments;
(3) three junior scientists for data compilation (environ-
mental scientist) computer progranning, model runs
(computer scientist) and output analysis (environmental
scientist); and
(4) secretarial and support staff
Large memory computer systems, computer graphic facilities, and
word processors are valuable tools for a cost effective performance,
since models would probably have to be run multiple times in light of
errors or new information discovered in input data. Therefore, graphs
would have to be reproduced and reports retyped in light of corrected
output.
4-53
Arthur D little nc
-------
4.7 REFERENCES
Aravarnudan, K., M. Bonazountas, A. Eschroeder, D. Gilbert, L. Nelken,
K. Scow, R. Thomas, W. Tucker, C. Unger (1980); “An Environmental
Partitioning Model for Risk Assessment of Chemicals,” Arthur 0. Little,
Inc., Cambridge, MA 02140 under EPA Contract 60-01—3857 for Office of
Toxic Substances, Washington, DC 20460.
Bonazountas, M. (1981); “A Multi-media Environmental Modeling Back—
ground/ Catalogue,” Internal Document, Arthur D. Little, Inc., Cam-
bridge, MA 02140.
EPA/MIBS (1981); “Environmental Modeling Catalogue,” EPA, Management
Information and Data Systems Division, Washington, DC 20460.
EPA (1980); “Handbook for Performing Exposure Assessment,” the EPA’s
Exposure Group; An unpublished document.
Haimes, Y.Y. (1977); “Hierarchical Analyses of Water Resources Sys-
tem,” McGraw-Hill Series in Water Resources and Environmental Engi-
neering, McGraw-Hill, New York.
Miller, C. (1978); “Exposure Assessment Modeling, A State—of-the Art
Review,” EPA, Athens Research Laboratory, EPA PB 600/3—78-065.
USDA (1973); “Linear Theory of Hydrologic System,” Technical Bulletin
No. 1463 (Stock O1Q0-02747), USDA, Washington, D.C.
4-54
Arthur D Uttle Inc
-------
5.0 RECEPTOR/EXPOSURE ROUTE ANALYSIS
5.1 INTRODUCTION
The purpose of the Receptor/Exposure Route Analysis is to define
critical exposure pathways for toxics in a specified region and describe
methods for estimating exposure through these routes, for both human and
nonhuman receptors. The results of this task combined with the output
from Task 3 serve as the primary input to Task 5, the Exposure Assess-
ment. The data prepared In this task are not, for the most part,
pollutant-specific but, in some cases —— such as the population charac-
terization -- are necessarily regionally specific.
The two discrete subtasks within Task 4 are: 1) selection of
algorithms for estimating individual Intake levels of pollutants for
each exposure pathway; 2) determInation of the regional distribution
of study area receptor populations, and characterization of the tem-
poral, behavioral and other patterns influencing their exposure through
each pathway.
Exposure assessment methodologies and handbooks have been devel—
aped for various government agencies and private organizations and,
although most are not regional in scale, they are useful resources in
developing any exposure assessment approach. Relevant sources are
listed in Table 5-1. The purpose of this section is not to reiterate
the general exposure assessment methodologies compiled in prior studies,
but to develop a region—specific approach and to relate it to existing
methods wherever appropriate.
There are several irmiediately apparent differences between a geo-
graphically-based exposure assessment methodology and other methodologies
intended to estimate national or generic exposure to a specific chemical,
industry, coimiodity, orother categories. A geographic assessment is
conducted on a small enough scale to develop detailed and accurate
receptor population data regarding the size of exposed populations and
the spatial relationships of receptors to points of pollutant release.
5—1
A I
-------
TABLE
REVIEWS OF HUMAN EXPOSURE ASSESSMENT METHODOLOGIES
Title Source
iuiaance for the Preparation of EPA Exposure Assessment Group (1980)
Exposure Assessments
Handbook for Performing Exposure EPA Exposure Assessment Group (1980)
Assessments (Preliminary Draft)
Report on Exposure Assessment Consumer Product Safety
Status and Principles 1 Coninission (CPSCJ (1981)
Proceedings of Workshop on exposure U.S. EPA (1980)
Assessment of Hazardous Chemicals 2
Integrated Exposure/Risk Assessment Monitoring and Data Support
Methodology (Preliminary Draft) Division, Office of Water
Regulations and Planning, EPA (1980)
1 Cited in EPA (1980).
2
By Radian Corporation (1980), EPA contract No. 68—02—3171.
5—2
Arthur fl I ittle Inc
-------
Aiso qeneralized expressions of annual pollutant intake rates can be
“calibrated ’ for the study area by adjusting for regional patterns of
ecreat on, food-consumption, occupation, transportation and other activi-
ties. These region-specific intake rates enable a more accurate estimate
of local expo iire than generalized averaqe intakes.
5.2 EXPOSURE ROUTES AND EXPOSURE MEDIA
5.2.1. Introduction
Following the initial scan and data retrteval effort (described in
Section 2.5), a set of critical exposure routes and receptors will have
been identified for the geographic area being studied. For each
combination of an exposure route and a specified environmental medium
(e.g., ingestion in drinking water), the general equation for estimation
of the individual pollutant intake via this pathway will be:
= ‘EX
where: = pollutant intake rate
1 EX = exposure medium intake rate
EX = pollutant concentration in exposure medium
The exposure medium intake rate may be expressed simply as a unit mass
per day (e.g. for ingestion), or with more detailed information, as:
‘EX = 1 EX . F 0 EX EEX . APEX
where: = unit exposure medium intake rate per exposure event
FEX = frequency of exposure events (number of events/time)
0 EX = duration of exposure event (time)
= extent of exposure (%)
EX = efficiency of absorption of pollutant for that
exposure route (%)
5-3
ArthurDUttleinc.
-------
The second equation is more applicable to exposure through dennal absorp-.
tion than through other exposure routes. The reasons are discussed in
Section 5.2.2.4.
The time period to use in estimating and grouping pollutant intake
values is determined by the eventual application of the exposure results.
Exposure medium intake rates may vary on a daily (e.g. for inhalation)
or seasonal (e.g. for recreational dennal absorption) basis. If the
variability has a significant influence on pollutant intake rates, then
exposure should be calculated in small enough time—steps to reflect
these differences.
The absorption efficiency term allows estimation of the effective
dose or the amount of pollutant which crosses the membrane of the exposed
tissue (e.g. the lung) and reaches a target organ (e.g. the liver). For
many pollutants this type of metabolic data is not available and conse-
quently 100% absorption is a comon preliminary assumption in exposure
assessments. For well—studied substances such as radionuclides, a
methodology for calculation of exposure levels has been developed for
specific organs including bone marrow, lungs, endosteal cells, stomach
wall, lower intestine wall, thyroid, liver, kidney, testes and ovaries
as well as for the total body (Moore et al. 1979). If individual
absorption coefficients have been measured for the transfer of a pollutant
into the lung or stomach or across the dermis, these factors can be
included as efficiency coefficients in the calculation of doses rather
than simply intake levels.
Intake can be expressed either as a pollutant mass per unit time,
as discussed above, or as a mass per kg of body weight per unit time.
The latter expression facilitates comparison to health effects data,
especially laboratory animal data, which are coninonly reported in
equivalent units. Similarly, depending on the route of exposure, intake
may be estimated on an annual basis to address chronic effects, or on a
smaller time scale for addressinq acute effects inclucinq lethality,
teratogenesis, reproductive and neurotoxic effects.
5-4
Arthur 1) Little Inc.
-------
It may not always be desirable to express exposure as a pollutant
intake level. For some types of exposures and receptors, such as fish,
health effects data are connuonly available in the form of concentrations
in the exposure media, for example as an LC 50 (a pollutant level in water
lethal to 50% of the experimental population). The relationship between
toxicity and the amount of pollutant absorbed by the gill in a unit perio4
of time is rarely determined. Similarly, the relationship between
pollutant levels in food and adverse effects on fish is not well under-
stood. Therefore in the case of aquatic species, it is appropriate to
report exposure levels as a concentration in water for easy comparison
to effects data.
The magnitude of exposure in a geographic area is a function not
only of the amount of pollutant to which an individual is exposed but
also of the size of the population exposed. Therefore, it may be useful
to express exposure on a population basis, thus accounting for the number
of receptors exposed to a particular pollutant level. The resulting quan-
tity is a population exposure factor which is the product of the individual
pollutant intake level per unit time multiplied by the population size
exposed. Exposure factors enable comparison of the magnitude of
exposure in different regions or for various subpopulation categories
and, consequently, helps identify the most significant local exposures.
The following subsections describe the methods coninonly used to
estimate intakes for each exposure route, and some of the input parameters
required to make the calculations. Also potential sources for input data
are identified. Each major exposure route is treated individually.
5.2.2 Humans
5.2.2.1 Introduction
The methods used to calculate exposure levels for hiinans are
relatively independent of the type of exposure assessment (e.g., regional
vs. national) requirino these data. Therefore, the methodologies and
5—5
Arthur D little Inc.
-------
models developed for other exposure applications are relevant to a
regional exposure assessment. Existing models are briefly discussed by
exposure route in the following section. The EPA Exposure Assessment
Group has compiled an initial draft of a Handbook for Conducting Exposure
Assessments which contains parameters and expressions for estimating
intakes currently being used in EPA work. A second draft will be published
in the near future. This document along with other references listed in
Table 5-2 and 5-3 provide more detailed documentation of relevant methods
and parameters, along with examples of their application to different
types of exposure assessments.
5.2.2.2 Ingestion
A conmon exposure route through which humans are exposed to pollu-
tants is ingestion of contaminated drinking water or food. Other less
typical ingestion-related exposures can occur through mouthing of soil
(e.g. by children), swallowing of saliva during exposure to high airborne
concentrations of pollutants, swallowing of water during tooth—brushing
or participation in water—contact sports. Only drinking water and food
are discussed in this section.
In a regional exposure assessment, drinking water supplies and
certain area—grown foods are examples of potentially important
exposure media which can be related to the local sources and emissions of
toxics. Questions may arise regarding the effectiveness of drinking
water treatment methods employed to prevent contamination of water supply,
or regarding the contribution of local foods to the total diet intake of a
specific pollutant. The environmental pathways leading to contamination of
these media are often complex and indirect, involving a series of intermedia
transfers. An example is the transfer of pollutants from a land disposal
site into a ground water aquifer used as a water supply. Without the
implementation of a fate model, it may be difficult to relate toxics
concentrations back to a specific source.
5-6
Arthur I) Little Inc
-------
TABLE 5-2
SI ECIFIC HUMAN LXPUSURE ASSESSMENT METHUDULO ilLS
Ti IL E
Methodology for Estimating
Direct Exposure to New
Chemical Substances
OFFiCE
Office of Toxic Substances,
EPA (III); k PA 560/13-79-008
EXPOSURE ROUTES AND PATHWAYS CONSIDERED
Inhalation, ingestion and dermal
absorption for occupational and
consumer populations
AII)OS-IPA: A computerized
Methodology for Estimating
Environmental Concentrations
and Dose to Man from Airborne
Releases of Radionucleldes
Office of Radiation Programs,
EPA; EPA 520/1-79-009
Inhalation, ingestion of foods and
drinking water, dermal absorption.
Transport from environment to food
was modelled.
(1
CUMEX-A Cumulative Hazard
Index for Assessing Limiting
Exposures to Environmental
F’ol lutants
Technical Support of Standards
for 111gb-Level Radioactive Waste
Manageiiient
volume C, Migration Pathways
Environmental Sciences
Division, Oak Ridge
National Laboratory, ORNI
Publication No. 1007
Office of Radiation Programs,
EPA; EPA 520/4-79-007C
Ingestion of water and food.
iransport from environment to
food and water was modelled.
Human Exposure to Atmospheric
Concentrations of Selected
Chemi ca Is
Office of Air Quality Planning
and Standards, EPA
Inhalation
Environmen tal Carcinogens
and Human Cancer
Mathematical Models for
AtniospheriC Pollutants
OffIce of Research and Deve1op
ment, Health Effects Research
Laboratory, PB-291—742
Inhalation, exposure of biota
to airborne pollutants
r*
Inhalation
Electric Power Research
Institute (EPRI), Palo Alto,
California
-------
TABLE 5—3
SOURCES OF DATA ON HUMAN INTAKE
RATES AND PARAMETERS
FASEB 1972. Biology Data Book. VolLine III.
FDA. Total Diet Studies. Con 1iance Program Evaluation. Bureau of
Foods, Washington, D.C.
National Center for Health Statistics (NCHS). 1977. Dietary intake
findings. Data from the National Health Survey. Series 11, rninber 202.
National Center for Health Statistics (NCHS). 1979. Dietary intake
source data. Findings of the Health and Nutrition Examination Survey
(HANES). DHEW publication no. (PHS) 79—1221.
Roddin, M.F., H.T. Eclis, N.W. Siddizee, and R. Lieberman. 1979. Back-
ground data for human patterns. Prepared by Stanford Research Institute
under Contract No. 68-02-2835. U.S. Environmental Protection Agency,
Office of Air Quality Planning and Standards, Research Triangle Park,
N.C.
Task Group of Committee 2 of the International Commission on Radiological
Protaction. 1975. Report of the Task Group on Reference Man. Pergamon
Press: New York.
USDA. 1972, 1980. Food and Nutrient Intakes of Individuals in the
United States. Agricultural Research Service, Washington, D.C.
U.S. EPA. 1979. Identification and Evaluation of Waterborne Routes
of Exposure from other than food and drinking water. Office of Water
Planning and Standards, U.S. EPA, Washington, D.C.
Yost, K.J. and L.J. Miles. 1980. Dietary consumption distributions
of selected food groups for the U.S. population. Prepared by Purdue
University under Contract No. 58-01-4709. U.S. Environmental Protection
Agency, Office of Toxic Substances.
5-8
Arthur £) Utt1eir c.
-------
• Beef
• Pork
• Laab, veal, gave
• Poultry
• Fish. shellfish
• Others
Milk nd Milt Products 2
Eggs 2
Legumes, .t4utS ana
Seeds’
Grain prouucts 2
Eats, oiis
vegetables’
• White Potatoes
• tometoes
• 1ar green
• oeeo yellow
• otners
t ever :oes
Sugar. sweets
Avenge: 54 g/d.y
Avenge: 20 g/day
Average: 2 glday
Average: 27 g/day
Avenge: 11 g/day
Average: 94 g/day
riverage: 270 g/day
Average: 21 9/day
Average: 26 gi day
Average: ZoO g/day
Average: 14 g/day
Average: 200 9/day
Average: 64 g/day
Average: 22 g/day
Mverage: 9 9/day
Average: a 9/Gay
AverageS 99 g/day
Average: 600 qiday
Average I40 g/day
Average: 23 9/day
Individual
Variability
Adult Range: 1.1 — 2.4 i/day
ChIldren: 1.1 — 1.7 1/day
Moderately active indivioual:
3.7 1/day
Under high t ratures: 2.8 —
.3.4 1/day
Aau)t range: 310 g/aay
Children: lOg - 150 g/day
Adult Range: 250q - 650 g/day
ChIldren: 400-800 5/day
Adult Range: ing - s5qiday
ChIldren: Sq - 22g/dsy
Adult range: 5-45g/Gay
Cnlldren: 20g - 65g/day
Adult Range: lSOg - 25Ug/day
Children: 40g - Z OOg/day
Adult range: log - 199/day
Children: ig - ug/day
AdultRange: 140g -l.BSg/day
Chiidren: 15g - 135 g/uey
Adult Kanne: 3Øfln - 840 g/day
Chisdren: SOg - VU 9/cay
Aault range: 105—190 9/day
Chiløren: 135g - 70 9/day
Aault range: 14g - £2g/day
Cnhluren: lOg - . 7 glcay
Not significant
Not significant
isot significant
Not significant
Not significant
Annual average
exposure; probable
higher intake of
locally grown prod-
uce during si er.
Annual average
exoosure; probable
hinher Intake
dur nq s u mner
Unknown
hot significant
TABLE 5-4
COI NI.Y USED iNTAKE RATES
FOil INGESTION OF ELtCTED
EXPOSURE I DIA
exposure
Drinking vater’
(total fluids)
Meat. Pnultry and
Fish 2
Typical
intake
Rates
Average: 2 I/day
Average: 200 ge/day
Probable
Seasonal
Variability
Annual average
exposure; higher
rate during si r
Annual •btYau*
exposure; probable
higher intake of
local fish during
eport of Task rouo on ilererence Man (l975
:veraae intake oe ,divioua in t day, Sorir. 1 77 JSUA 9 O
3. EPA reco inencs using ó.5 giday
c r c tion i ts e wifl ‘e ‘rciudec ur er r f idng vater t :ei fluids:
-------
The rate of ingestion or intake of a specific exposure medium is a
function of the particular medium or food group consumed. Table 5—4
surwnarizes coimnonly used intake rates for drinking water and selected
food groups, the potential for variability by age, sex, or region in
each rate, and the seasonality or temporalcomponent of these rates.
Since food consumption patterns vary regionally, both in terms of the
items and the amounts that people consume, a more localized approach than
the one described above may be desired. An FDA one—day survey of U.S.
consumption patterns by region, sex, age, income and season for a number
of food items was conducted in 1965—66 (FDA 1973), and again in 1977
(data still preliminary; FDA 1980). These data makeit possible to quan-
tify the consumption rate of numerous food groups and the size Of the
PoDulation consuming each food group in four major U.S. regions. -
Populations exposed to drinking water can be estimated from water
supply distribution information, permits, and laboratory records from the
state and local authorities and/or water supply companies. Populations
exposed through ingestion of specific food groups can be estimated from
generic data on consumption patterns for the general population, from
regional data developed by the FDA study described above, as well as from
contacts with local sources.
5.2.2.3 Inhalation
Intake of toxics may occur through inhalation of ambient air
contaminated by emission plumes at locations downwind from specific
industrial point sources. Exposure may also occur during the use of
volatile or sprayed consumer products and in the vicinity of non—point
applications or releases, e.g., downwind from pesticide treatments, dry
cleaning operations,or heavy traffic. The general population may also
be exposed to ambient “backgroun& levels not directly associated with
specific sources: Inhalation of mist, fog, and vapors in contaminated
air associated with contaminated water bodies, drift from cooling
towers, etc., iiay also contrioute to polluza t intake.
5-10
Arthur 1) Little Inc
-------
In comparison to other exposure media, air is a relatively uniform,
continuous medium to which all humans are exposed 24 hours per day. The
amount of pollutant taken into the body is a direct function of the rate
of inhalation volume of air inspired per unit time, which is in turn a
function of the metabolic rate. . Table 5—5 presents some rates coninonly
used in exposurè assessments. Higher rates may occur at high temperatures
or during strenuous exercise. Slightly lower rates may occur in people
with respiratory diseases and in the elderly. Exposure assessments
comonly employ either a single daily inhalation rate averaged over a
24-hour period or a daily rate combining a period of strenuous exercise
(high rate) with a sedentary period (lower rate).
Residential population data by census division, such as the
information collected in the census bureau surveys, Is comonly used to
estimate the size of receptor populations exposed to pollutants in a local-
ized area. Adjustments can be made to the population estimates to account
for receptor mobility over daily or longer time periods. A discussion of
the use of demographic data in geographic exposure assessments follows
in Section 5.4.
5.2.2.4 Dermal Absorption
Exposure to toxics through dermal absorption may occur during
swinmiing or other water-contact sports in surface water, during washing
and bathing in water supply, and during use of consumer products (liquids
and vapors) or contact with Items treated with such products. Intake of
any substance through ciermal absorption is a function of the extent,
frequency and duration of exposure. The extent parameter (E) indicates
the portion of the body (most con nonly measured as percentage of total
surface area) in contact with the exposure medium and hence the area at
which absorption through the skin will occur. A coninon range for this
variable is from 10% for hands to 100% for the total body surface (EPA
1979). For recreational activities suchas swiming and water skiing,
during which imersion is nearly total, E may be set at 100%.
5—11
Arthur I) ‘Little Inc.
-------
TABLE 5-5
INHALATION RATES FOR HUMANS BY ACTIVITY AND SEX 1
ADULT MALE ADULT FEMALE
Body weight (kg) 68.5 54.0
Experimental conditions light work light work
medi urn work medi urn work
heavy work heavy work
Respiratory frequency 11.7 (10.1 — 13.1) 11.7 (10.4- 13.0)
breaths/min) 17.1 (lb.7 — 18.2) 19
21.2 (18.6 - 23.3) 30.0 t25.0 - 5.3)
Tidal volume 750 (575 - 895) 339 (285 — 393)
1673 (1510 — 1770) 860 (836 - 885)
2030 (1900 — 2110) 880 (490 — 1270)
Minute volume 7.43 (5.8 — 10.3) 4.5 (4.0 - 5.1)
(liters/mm) 28.6 (27.3 — 30.9) 16.3 (15.9 — 16.8)
42.9 (39.3 — 45.2) 4.5 (17.3 - 31.8)
Source: FRSE3 s Biology Data Book, Volume III, 1972
5-12
Arthur D little Inc
-------
Frequency (F) and duration (D) give the total amount of time each
year during which exposure of an individual is oc uring. Frequency is the
number of events in a unit time period an exposure activity takes place, and
duration is the time period (usually hours) for each activity. The
product of these terms gives hours of individual exposure per year.
For derinal absorption, either one of two methods is used for
estimating the pollutant intake. The more reliable method incorporates
a measured permeability rate constant for dermal absorption of the
particular pollutant -— expressed in units of g/cr&’ /hr —- into the
equation, assuming low pollutant concentrations and an equivalent dermal
absorption across all areas of skin (U.S. EPA 1979). These measurements
are usually made in vitro with intact skin and may not accurately
represent in vivo behavior.
The second, less precise, method for estimating dermal exposure is
to assume pollutant diffusion across the skin at a rate equal to the
rate of diffusion for the solvent, in this case water, which equals
0.2 to 0.5 mg/cm 2 /hr (U.S. EPA 1979). This method may underestimate the
exposure of some substances, especiail.y those that are highly Tipophi lic
or have a high octanol/water partition coefficient.
The identification of the population exposed via dermal absorption
does not usually involve geographic subdivisions, as in the case of sub—
populations êxposèd to air. Recèptôrs are individuals who through their
behavior or activity patterns, rather than as a function of where they live,
come into temporary dermal contact with exposure media during specific
activities such as recreation, use of products, bathing, and others.
Survey data on the recreational use of beaches, use of home pesticides,
etc., and information on the population served by a water supply would be
the type of information required to quantify these groups.
Poflutant concentration data measured in water and in consumer
products are avaflable from several sources. Monitoring data for water
5—13
Arthur D Little Inc
-------
supplies are usually available from the state and county drinking water
offices. Surface water data can be found in the STORET water quality
data base, at the EPA regional office, at state environmental departments
and sometimes from private organizations. Information on pollutant
concentrations or exposure levels associated with the household use of
selected pesticides has been developed by the Special Pesticides Program,
Office of Pesticide Programs, in conjunction with their RPAR 1 program.
The methodologies developed to estimate exposure are applicable to other
substances as well.
5.2.3 Other Biota
In a regional exposure assessment it may be important to identify
the exposure and risk of fish, wildlife, plants and other nonh gnan
species to toxic pollutants for several reasons. Certain sensitive
species may provide an early warning signal for a local toxics problem,
as indicated by mortality or accumulated toxics levels in tissue.
Protection of crops, sport fisheries, game, and livestock are of interest
because of their economic value to man. Most importantly, enviromental
quality is a function of the diversity, abundance and productivity of
indigenous species in a coninunity, so the preservation of the existing
ecosystem, or at least part of it, is desirable. In some cases, the
exposure of abiotic receptors, such as buildings, works of art, or other
inanimate objects may be of concern in a geographic exposure assessment.
Identification of critical nonhuman receptors is very region-
specific and dependent on the species present and the number of individual
habitats to consider. In general, far less extensive population surveys
are conducted for nonhuman populations than for humans. Therefore, during
the screening for species of concern in a specific geographic area, some
of the factors which may be considered include presence in the area of
endangered or threatened species, agricultural land use (either for
crops or raising livestock), or presence of sensitive or critical
i Rebuttable Presumption Against Registration
5-14
Arthur D little. nc
-------
habitats (e.g., wildfowl nesting or fish spawning area; resting/feeding
areas along migration pathways). In industrial or highly populated areas,
a potentially critical receptor category may be the microbial populations
in POTWs and industrial secondary treatment processes.
Table 5-6 lists sane of the species categories which may be con-
si dered in a geographic assessment and the critical exposure routes of
concern for each group. Since there are numerous pathways in a specific
area, many Interrelated, which link a toxic from source .to receptor, the
focus may be limited to only critical routes. Critical routes are those
either involving a large mass of pollutant or leading to exposure of sensi-
tive or economically important receiptors, or of large numbers of receptors.
Ecosystem models have been developed for both generic ecosystems
(e.g. freshwater lakes, prairies) and specific locatlo,is (e.g. the Upper
Mississippi, Rocky Mountain Arsenal). Ecosystem models are useful for
understanding the impact of exposure on an entire, Integrated biotic
system and to evaluate indirect Impacts (such as reduction in an important
food source, inhibition of nutrient cycling). Several major problems
limit the use of such models, including nonlinearity, limited data for
many of the processes, the spatial and temporal resolution of different
processes, and lack of a general theoretical basis to support inclusion
of some of the parameters influencing ecosystem dynamics. Physical
system models are generally at a far more advanced state than biological
models because they do not have many of the problems described above.
An ecosystem model of the geographic area being studied would enable a
detailed and organized assessment of the local impact of toxics. However,
these models are only available for limited locations and are probable
too expensive and time-consuming to develop within the course of
geographic assessment.
5—15
Arthur D Little Inc.
-------
TABLE 5-
EXAMPLES OF NONHUMAN RECEPTORS TO BE CONSIDERED IN AN EXPOSURE ASSESSMENT
Cuticle absorption, root
uptake
Cuticle absorption , root
uptake
Ingestion , dennal absorption,
Inhalation
1
Exposure Media
Surface (natural
systems), Impoundment water,
sediment other blota
Surface water, sediment, oth
biota
Surface water , sediment, oth4
biota
Surface water, fish, plants ,
and other biota
Soil, fallout , rain, irri-
gation water, pesticides
Soil , fallout, rain, pestlci
Surface water, water supp] ,
locally grown feed and foddei
pastures, medicinal applica-
tions, pesticide-treated pro
ducts
Surface water, other biota ,
pesticide drift
Surface water , sediment, sol
P01W Influents, wastewater
Aquatic Coiiwiiunities:
• rish and other
vertebra tes
• Invertebrates
• Producers and pd-
mary consumers
• Waterfowl and other
hi rds
w Terrestrial Communities:
• Crops
• Indigenous plant
Ct)HflflUfl I ties
• Livestock
Important Exposure Routes 1
Absorption (gill and dermal)
Ingestion
Absorption, (mantle and dermal)
and Ingestion
Absorption , ingestion
Ingestion , dermal ‘absorption
(oil)
Warni water species,
salmonids, amphibians
Molluscs, crustaceans
Algae, macrophytes,
zoopi ankton
Ducks, geese, gulls,
predators such as
osprey, eagles
Leafy vegetables, root
crops
Wetlands, forests
Dairy and beef cattle,
chickens
(D
-4
• Wildlife Songbirds, hawks, deer Ingestion , dermal absorption,
Inhalation
Microorganisms
• Natural Communities Bacteria (e.g., nitri- Absorption
flers), fungi
• Waste-water treatment in POTWs Absorption
Populations
The underlined items are those most commonly considered In exposure assessments
-------
5.3 RECEPTOR POPULATION CHARACTERIZATION
5.3.1 Humans
5.3.1.1 General Population
An important part of a geographic assessment is the development of
a detailed and up-to-date human demographic data base for the area being
studied. These data can provide the basis for estimates of subpopulations
associated with different exposure routes. In many exposure assessments
it is coimion to use an average population density for the total U.S. or to
simply distinguish between rural and urban densities. In a geographic
exposure assessment in which site-specific data on pollutant releases,
environmental fate and ambient levels are measured or estimated, it .is
important to have equally detailed population data.
Availability of data from the 1980 census permits detailed analysis
of populations exposed to pollutants. By the early fall of 1981 population
breakdowns by age, sex, and housing units should be available for use
in an exposure analysis. The data are grouped into five major divisions
for each state, county, tract, enumeration district/place, block group
and block.
Census tracts tend to cover fairly large areas in each state, con-
taining on average from 2,500 to 8,000 residents; tracts are used both to
subdivide densely populated areas and to cover large rural areas which are
sparsely populated. These larger rural tracts are further broken down
into enumeration districts and places. Enumeration districts on average
do not exceed a population of 1,000. Places are dense concentrations of
population, primarily towns, which appear as distinct highly localized
areas within a tract. The majority of medium-sized towns in a state are
identified by place codes. Larger cities are also identified by a place
code but are further disaggregated into a number of smaller tracts.
5-17
Arthur D Uttle Inc
-------
The lowest level of detail in very densely populated areas is
provided by the block groups and blocks. In general blocks comprise an
area bounded by four streets and are aggregated into block groups. To
create special study areas for individual cases block groups and blocks
can be subtracted from one tract and added into a population division
that is entirely within an exposure assessment study area. Block groups
can also be used to subdivide area tracts which are so large that they
cannot be reasonably represented with a single exposure or concentration
value.
In order to understand the exact location of each census division,
detailed maps are provided by the Census Bureau which sh exact boundary
locations; these often include streets, train tracks, power lines, polit-
ical boundaries and topographical features such as streams and rivers.
When final versions of the data have been released, population centroids
expressed in longitude and latitude will be provided to allow approximate
boundary placement without maps. This will facilitate any computerized
analysis of these data that is sensitive to locations.
A geographic region’s population can be assumed to be static ——
that is, not dependent on mortality, mobility or birth patterns -— for
short-term exposure periods (e.g. <1 year). For the longer periods of
exposure which are of interest in a chronic study (e.g. from approximately
one year to a human life—span), time—dependence can be included in
the demographic model. Therefore, variations can be accounted for in
the total population size and distribution due to birth, death and
migration processes as well as changes in the distribution of age,
sex, racial, sociological and other classes. On a geographic scale, it
is likely that detailed information is available from the Census Bureau
arid land use studies to estimate historical population fluctuations as
well as to predict future trends. In addition, general mathematical
models have been developed to estimate population dynamics, for example
those described in ?ielou (1969) and Pollard (1973). Population
fluctuations over a shorter time period, such as changes with the r ovemeri:
5 -13
Arthur I) Little Inc
-------
of the working-age population, may also be important to include in a
geographic study. Hone and Stern (1976) have published one approach
to this problem.
5.3.1.2 Special Subpopulations
Within the general population of any given region, it may be
important to quantify special subpopulations associated with specific
types of exposure. Sources for data on these groups are sunvnarized in
Table 5-7. The subpopulations can be divided into at least two
categories: sensitive groups and specific exposure-related groups.
The sensitive subpopulations are those that because of their age
or health characteristics (e.g., very young or old age, poor health)
have the potential to be at greater risk when exposed. These groups can
be identified within a geographic region and, therefore associated with
an exposure level for that area. School children can constitute up to
20% of a regional population and spend as much as 8 to 10 hours daily at
the school grounds. School locations and associated attendance can be
obtained from local, county or state boards of education. Elderly
persons in nursing homes also represent an increasingly important
segment of the population. Data on nursing home facilities are coninonly
maintained by State health service agencies and county or local boards
of health, including geographic location and size of facility. In the
absence of such sun nary data, the yellow pages of the telephone book are,
at the least, a source of identification of nursing homes. Sensitive
populations located at hospitals (general and specialized care) can be
estimated through information from State health service agencies or boards
of health, who can provide data on hosoital locations and level of activity,
including: number of beds, average length of stay, statistics on types
of disorders, outpatient visits and emergency room activity.
Specific exposure—related subpopulations may be estimated from
survey inf rrnation on activity or use rates as well as from census
bureau statistics. Recreational activity in a region occurs in state,
5-19
Arthur D Uttle inc
-------
county, or town/city parks and recreation centers and at privately main-
tained recreational facilities (tennis, golf, swinining, fishing, hunting
clubs). Information on attendance (by month of the year) and location of
recreational facilities can be obtained through many of the public agen-
cies listed in Table 5-7 or from the private facility operators. Data on
the use of specific products of significance in geographic exposure may
be more difficult to obtain and require contact with the manufacturer or
with local vendors of the product. Employment—related exposure popula-
tions may be estimated from data available through the U.S. Bureau of
Census, County Business Patterns, and through state employment security
divisions. Often, available information provides location of a firm or
plant and associated employment. Because employment information at the
local level is occasionally sensitive (due to business confidentiality)
and unavailable, assumptions can be made using employment totals for a
coninunity and using land use maps to estimate geographical distribution
patterns.
5.3.2 Other Biota
Population data for flora and fauna present in a geographic area
provide a basis for assessing the impact of local pollutant concentra-
tions within various subregions in terms of total ninnbers of individuals
(e.g., fish) or acres (e.g., for crops) exposed and whether or not
sensitive age classes, reproductive stages, species, or other groups are
exposed. Receptors can be considered individually by species or as
entire corrinunities, depending on the eventual application of the exposure
estimates, e.g., whether pollutant effects on a particular species or on
the species diversity index for a particular habitat are of interest.
Table 5—6, presented earlier, listed examples of nonhuman
receptors that may be of interest in a regional exposure assessment.
Populations of these receptors within a region can be estimated from
agricultural statistics, from USDA extension offices or local universities,
and from contacts with local fisheries and wildlife resource offices,
5-20
Arthur D Little Inc
-------
TABLE 5-7
SOURCES OF DATA FOR CHARACTERIZING HUMAN RECEPTOR POPULATIONS
Receptor Category Source of Data
General Population U.S. Census Bureau, Council
Regional Subpopulations of Governments; local planning
agencies (county, coninunity)
Sensitive Subpopulations County and Corrinunity Boards of
Education; State Department of
Health Services; County Board
of Health; Health department
at local universities
Recreational Subpopulatlons State, County, and Coimnunity
Recreational Coninissions; local
Chamber of Coimnerce; State
Department of Natural Resources
Occupational Subpopulations State Division of .Employment
Security; State Manufacturing
Directory
Subpopulations by Food and USDA Nationwide Food Consumption
Water Consumption Patterns Surveys; State Office of Drinking
Water; county and coninunity
Drinking Water Divisions;
water supply companies.
General Information Phone books, local universities;
local industries, recreational
facilities and health care
facilities.
5—21
Arthur D Little Inc.
-------
U.S. Fish and Wildlife Service, National Park Service, U.S. Forest
Service and other agencies. Information on crops and domesticated
animals is usually better documented, more readily available and
accurate than information on natural populations. For the latter, data
consist primarily of results from mark and recapture or other sampling
surveys, for small areas and limited time periods. Due to the variability
in types of coninunities within a region, population data representative
of all habitats may not be available, and rough estimates based on
comparable biological communities in other areas may be enployed.
5-22
Arthur D Little. fr c
-------
6.0 EXPOSURE ASSESSMENT
6.1 INTRODUCTION
The Exposure Assessment is the fifth step in the analysis of
regional exposure to toxics. In this step, the data and methods
developed in the previous tasks are linked together in order that the
relationship between local sources of toxics and the exposure of humans
and other biota living in that locale can be examined. Through estimation
of the degree of exposure rather than mere consideration of pollutant
concentrations in environmental media, a more detailed analysis is possible
including:
• evaluation of whether or not the toxic ends up in
significant exposure media or at locations at which
exposure takes place;
• estimation of the amount of toxic to which a receptor
is exposed in a unit time period, including a cumula-
tive total for chronic exposure;
• consideration of the influence of receptor behavior
patterns or environmental conditions on receptor
exposure.
The ultimate goal of evaluating a toxics problem in a particular
area is to identify and evaluate potential adverse impacts on humans,
fish and wildlife, plant coninunities and other species. Therefore,
determining the nature and extent of toxics exposure is a logical first
step toward assessing these impacts.
6.2 PURPOSE OF AN EXPOSURE ASSESSMENT
The overall purpose of the Exposure Assessment is to estimate the
magnitude of exposure of receptors, both human beings and other biota,
to toxic pollutants distributed throughout the environment of a speci-
fied region. This includes specific consideration of all potential
exposure media such as air, drinking water, surface water, sediment,
0—
Ar±ur D Uttie. :
-------
soil, plant and animal food products and consumer products as well as
a combined or total exposure through all media, if possible.
Exposure is ideally estimated at all points of potential contact with
receptors in the pathway leading from a chemical’s points of release to
its final equilibrium distribution in the enviroment. Figure 6—1 thus—
trates the environmental pathways of a pollutant and locations of potential
exposure. The input concentration data required for these estimates
can be actual monitoring concentrations and/or estimated levels predicted
by environmental fate models. The advantage of monitoring data is that
they give known levels found in exposure media; however their relationship
to sources and the extent to which they are representative of regional levels
are uncertain. Estimates from fate models have the advantage that their
relationship to sources can be investigated through sensitivity analyses;
however they are often difficult to validate.
For certain exposure pathways, the spatial distribution of differen-
tially exposed receptors within the study area 4 s of interest. This
information is useful for populations exposed through inhalation or dermal
absorption from surface water, two exposure pathways which usually have a
spatial variability in their associated pollutant concentrations. If a
number of water supplies (e.g., private wells) serve one study area popu-
lation, the distribution of the associated receptors may also be tmportant.
On the other hand, exposure media that are channelled to a central loca-
tion before redistribution among receptors (such as most foods) are not
likely to exhibit spatial variability within a reglon therefore informa-
tion about percentage of the population that is exposed will be sufficient
to quantify the receptor population.
Ideally a regional exposure assessment will represent the probable
exposure of most of the local population for all times of the year and
under all environmental conditions typical for the area. Since an
assessment of this detail can be very difficult and costly to perform,
esoecially if it requires the use of comolex fate models, exposure
estimates may be limited to average seasonal conditions (e.g., summer--
6-2
Arthur D Uttle Ir
-------
FIGURE 6—1 : HUMAN EXPOSURE ROUTES - LEAD
Source: Per. :ak, 2. etaL(1981).
6—3
Arthur I) Little. Inc
-------
low flow -— low rainfall, etc.) and to specific “worst case” meteorological
or hydrological conditions for that area (e.g., 7 day—lO year low flow;
no wind).
Chapter 5 described the algorithms used for calculating intake and
the receptor population data needed to estimate exposure Chapters 3 and 4
described methods for the development of source and emissions data and
for estimating pollutant concentrations in exposure media. Essentially
the output of each task serves as input to the subsequent task. Quali-
tatively, information from each of the tasks also helps to define and
refine the elements of other tasks. In anticipation of the eventual
linkage of the tasks, the ways of expressing the data (units, time steps,
etc.), especially input and output data, should be made consistent at the
outset of a geographic study.
The following section describes a general methodology for conducting
an exposure assessment, discusses the input requirements, and presents
different ways to analyze and apply the results of the exposure assess-
ment. The methodology is by no means uniquely correct, but is presented
as a suggested framework for organizing regional exposure assessments.
6.3 METHODOLOGY
The Exposure Assessment can be roughly broken down into three
general steps: organization of existing parameters and data and inter-
face 0 f models; calculation of individual exposure and population
exposure levels; and analysis and applications of exposure data. Each
step is discussed below fri detail.
6.3.1 Organization of Input Parameters and Data
The input requirements for the regional exposure assessment are
the following:
• pollutant concentrations in environmental media
(described in Chapter 4)
• pollutant intake computational methods (Chapter5)
• receptor 7oDulatiOn data (Chapter 5)
6-4
ArthurD Uttleinc
-------
6.3.1.1 Pollutant Concentration Data
Pollutant concentrations are either in the form of monitoring data
measured in the study region or estimated levels based on the predictions
of fate model simulations of the area. For monitoring data, both mean or
median and maximum values should be used in estimating exposure; the former
as representative of an exposure typical of the majority of the populatton,
and the latter as an upper limit on regional exposure. Monitoring data
should be collected from sites at which exposure is likely to occur and
the data should represent variability in pollutant loadings, as well as
seasonal and other variability. Monitoring data from most regions, how-
ever, are usually limited, and the use of pollutant fate and transport
models will undoubtedly be required to simulate pollutant concentrations
in exposure media.
Two modelling approaches can be used to estimate pollutant concen-
trations. A multimedia fate model including transport of pollutants
between and the fate and distribution in all envirorm ental ccmpartnents
is one method; however these models are often complex and expensive.
Simpler single compartment models can also be implønented provided
coefficients are included to estimate transport into important
exposure media through processes such as deposition from air to
ground and plants, uptake by plants from soil, uptake by livestock from
water and feed. Both modeling approaches are discussed in the Environ-
mental Pathways Analysis (Chapter 4).
6.3.1.2 Standard Fate Models
The model(s) selected in the Environmental Pathways analysis
(Chapter 4) should be responsive to the requirements of the Exposure
Assessment analysis; therefore, there shouTd be initial input from the
exposure team in the choice of models, boundaries, time—steps, location
of receptor sites and other factors. From the point of view of the
Exposure Assessment, fate models should
• be sensitive enough so that exposure of near—field
populations can be quantified;
6—5
Arthur D Iittte Inìc
-------
• produce pollutant concentration estimates at reasonable
time intervals so that exposure estimates can be expressed
in a form compatible with health effects data;
• give pollutant concentration estimates within a bounded
area and at spatial locations reflecting receptor dis-
tn buti on patterns;
• estimate concentrations representative of local
pollutant release patterns, of average seasonal and
worst-case meteorological conditions for the study
region, and of the physical and chemical variability
within each medium modeled.
In general, the final output of an environmental fate model will
be a concentration or mass of pollutant in a volume or mass of
environmental media, either at equilibri in or as a function of time.
In some models, a concentration frequency distribution for a pollutant
can be generated as a function of local environmental conditions, release
patterns, or other variables. This distribution is useful in estimating
the probability of exposure to different concentrations or in generating
a realistic estimate of average and maximtnn exposures. For regional
models, the spatial distribution of various concentrations or ranges
within a specified bounded area can also be given.
6.3.1.3 Exposure Pathways Not Usually Considered in Fate Models
In a regional exposure assessment, it may be desirable to estimate
concentrations in environmental exposure media not included in standard
fate pathways models. The following briefly discusses some of these
cases and possible approaches for their inclusion.
Home Gardens : Certain toxics accumulate in vegetables grown in urban
hone gardens or agricultural areas via uptake from contaminated soil (by fall-
ou at previous waste disposal sites, by soil erosion and transport from
6—6
Arthur D Little lnc
-------
industrial sites), from atmospheric deposition, or from irrigation with
contaminated water. If a soil/air intermedia model is implemented,
then estimates can be made of soil concentrations at a particular site
and thus, if the process of plant uptake of the chemical is understood,
estimates of plant concentration can be developed. Also plant surface
concentrations can be estimated from a rate of atmospheric deposition.
The AIRDOS-EPA model for radionuclides describes a method for estimating
plant uptake from deposition and soil. At this time, both types of
estimates are uncertain for most toxic chemicals due to a lack of
empirical data on plant uptake of these substances.
Fish : Most surface water models do not estimate pollutant concen-
trations in biota. The EXAMS model does include prediction of concentra-
tions in phytoplankton but not in any species Ingested by hunans. The
water’column concentration output from these models can be used to
roughly estimate tissue levels in fish if the equilibrium bioconcentratlon
factor (BCF)* is known for that pollutant. The Criteria and Standards
Division of the EPA, Office of Water, has estimated BCF’s for many of
the priority pollutants. These are published in the Criteria Documents;
other data are also published in journals such as Enviromental Health
and Contamination, Residue Reviews, Journal of the Water Pollution
Control Federation, Environmental Pollution , and others.
Drinking Water : In the absence of drinking water monitoring data,
pollutant levels in surface water supplies may be estimated by retrieving
ambient concentration data (e.g. from STORET) at locations upstream from
the drinking water plant intake or, if available, at the intake itself.
The effectiveness of the plant’s treatment methods in removing any pollutants
present in the influent can then be evaluated and predictions made of the
* Ratio of fish tissue concentration to water column concentration.
6—7
Arthur D Little Inc
-------
pollutant concentrations at the tap. The STORET User Service or staff
members of the Monitoring and Data Support Division can provide information
on these methods. Pollutant levels in groundwater are more difficult
to estimate. However, with the use of a soil pollutant transport and
transformation model (such as SESOIL), the impact of a pollutant on the
water quality of an underlying aquifer can be assessed If local soil
conditions and geology are known and ground water concentrations estimated,
For all types of water supplies, the rate of formation of chlorinated
organics during chlorine treatment may be modeled If judged to be an
important pollutant source for the area.
6.3.1.4 Linkage of Fate Models to Receptor Distribution Models
One of the problems likely to be encountered in a regional exposure
assessment is the interface of the environmental fate models with receptor
distribution models or exposure site data. The dimensions and bounds of
fate models are usually dictated by local topography and meteorology and
by emissions patterns. Population subdivisions are based on factors such
as population density, political boundaries and sometimes natural boundaries.
Ideally, an a priori methodology should be decided upon in order to
systematically coordinate these two sets of data. The methodology will
be determined by the type of fate model employed.
The output of a compartment box model, such as EXMS,. is an average
concentration for a specified area. In the initial decision-making con-
cerning model bounds, the location of exposure sites should be accounted
for to ensure their placement within model compartments rather than
straddling boundaries. For atmosoheric box models, which are likely to
have too few boxes to accot nod te the desired number of Census Bureau
6-8
Arthur D Uttle. Inc
-------
divisions, the interface may be more difficult. At least two options
are available:
• the initial division of the study area into model boxes
may be based on census bureau delineations, so that
one fate box is equivalent to a population district
or same fraction of a district.
• Fate box boundaries may be determined independently
from population data. Subsequently, populations
within these box areas may be estimated from census
bureau and land uie data. There would be a greater
uncertainty associated with using these estimates
instead of directly using census bureau district
populati ons.
Additionally there may be other approaches developed by government
agencies involved in atmospheric exposure modeling, such as EPA’s office
of Air Quality Planning and Standards.
Gaussian plume models generally report concentration output at predeter-
mined locatiOns or receptor points. The placement of one set of receptor
points should largely be the responsibility of the exposure team and be
based on population trends. For example, a single receptor point may be
specified for small census bureau divisions such as city blocks. Larger
divisions, such as entmieration districts, may require several receptor
points; the concentrations reported at these points may be averaged or
weighted and averaged to give a mean level for the districts. It may
be important to locate receptor points at specific sites with a high
population density or sensitive locations such as apartment complexes
or hospitals. These receptor points may represent at the same time,
depending on their location, the census bureau district in which they
are found.
6—9
Arthur D Uttle Inc
-------
6.3.1.5 Occupational Exposure
For certain toxics, workplace exposure may be more significant in
terms of magnitude of pollutant intake levels than is environmental
exposure. For example, exposure levels resulting from dennal absorption
or inhalation of substances like chlorinated solvents and pesticides
may be two or more orders of magnitude greater than the levels to which
the general, non—occupational population is exposed. In a geographic ex-
posure assessment, which predominantly uses very detailed and localized
information, it may be important to estimate the occupational exposure of
the human receptor population in order to place the environmental con-
tribution to total exposure in a better perspective. For Instance, there
may be a geographic area in which the majority of the work force is em-
ployed by a major industry with high on-site airborne concentrations but
low atmospheric emission rates. If exposure outside of the workplace is
negligible compared to occupational exposure, then emission controls may
do little to reduce exposure to an acceptable level, and control strategy
development should address reduction of onsite levels. In addition, there
may be a trade—off between occupational and environmental exposure. For
example, increasing ventilation inside a plant to reduce employee exposure
may increase the exposure of the general population living downwind from
the site. Therefore, information on the exposure levels and number of
receptors associated with each exposure pathway would be required to
select and evaluate the most effective control strategy for this situation.
Occupational exposure assessments have been primarily conducted by the
Occupational Safety and Health Administration (OSHA) and the National
Institute of Occupational Safety and Health (NIOSH). The assessments are
usually based on extensive on—site monitoring data and not on model
predictions. Certain offices within the EPA, sucti as the Office of
Presticide Programs and the Office of Radiation Programs, may also esti-
mate occupational exposure. The present methodology does not attempt to
describe or evaluate the various approaches used in estimating occupational
exposure. In a geographic exposure as3essment requiring occupational
-i 0
Ar nur U) little hc.
-------
exposure information, the first step would be to contact the organizations
listed above and other relevant groups (e.g. the state, industry organ-
izations) for specific approaches and information.
6.3.2 Calculation of Individual and Population Exposure
Levels -
Once the input data and computational algorithms are organized and
the interfaces between fate and exposure models completed, the exposure
calculations can be made. Exposure levels are calculated for each
exposure pathway (e.g., dermal absorption from surface water) and for
each subpopulation or subregion (e.g., swininers at a beach lninediately
downstream of an effluent pipe) designated for that pathway. The results
of the exposure calculations can be expressed in at least two different
ways:
• typical or maxim n individual Intake (e.g., ug/day of
pollutant)
• population intake or population exposure factor
(e.g., pg/day of pollutant multipltêd by popu-
lation exposed)
Intake is presented as a rate, usually as a daily, seasonal or annual
exposure. The selected time frame is determined by the intended
application of the exposure results (e.g., whether they are to be used
in assessing the health impact of acute or chronic exposure), a decision
which is, in turn, a function of the particular pollutant’s emission,
fate, toxicological and exposure characteristics. For example, long—
term exposure to an intermittently released pollutant with a short en-
vironmental half-life (e.g., a haloether) will be of less relevance
than exposure to a relatively persistent pollutant (e.g., a chlorinated
hydrocarbon). In the latter case, annual exposure will be important
to quantify.
These results can be presented in tabular or graphic form, broken
down by exposure pathway or subpopulation. Table 6—1 is an example of
6-il
Arthur U Uttie. c
-------
TABlE li-I
S(IIMARY OF IUUIVlDUt t INTAKES AND POr(JLATION
EXPOSURE FACTORS FOR TUE FOUR STUDY P011UTA TS
Vinyl Carbon
Chloroform Chloride Tetrachlor lde Lead
[ JR INI IU( WATER COHSUI4PT JON
indivi IuaI iiitake (s gfday): 12 0 180 NE 2 NE <30
maximum population exposure Not calculated NE NE Not calculated
factor (!g! i y iffl ) based on prelimi- based on limited
1000 nary data data.
FISh CONSUMPTION
indivlilual intake (gig/day): n,ean,inax. 0.02/5.3 0.0001/0.02 0.05/15 9.6/580
population exposure factor 0.8 0.04 1.6 343
t i ti/d x.
1000
110111 PROIUJCI COISUI1PT ION
individual intake (iig/day): range NE NE NE 0.06-1.0
population exposure factor 4.1-6.8
(ig/day j pulat1on)
i000
SIJRFACI WfITLR A 13cORPT ION
individual Intake ( q/day): ‘man 15 2 x 10” 5 x 103 0.02
population exposure factor ‘O.01-tu.4 0.1 <0. 10.2 <0.)
AIR INHALATION
See Table 5-31.
1 Assumptlons for these values are given In the preceding study.
2 N( Not estimated due to lack of data.
Source: Kanawha Valley Regional Exposure Assessment. July, 1981. 6—12
-------
one method of presentation. For exposure pathways which have regional
differences, e.g., inhalation of air, exposure levels may be graphed to
illustrate their relationship to sources or local topography. Figures
6—2 through 6—4 illustrate a means of displaying these data fora spe-
cific area.
The level of detail to be used In the breakdown of regional siAb-
populations will depend on the sensitivity of the model and the environ-
mental behavior of the pollutant. For example, it may be possible to
aggregate regional receptor subpopuIations into larger and larger groups
with increasing distance from an atmospheric source due to a concurrent
decrease or stabilization in pollutant concentration. Usually, a pre-
liminary fate model run and sensitivity analysis Is required before It
is possible to make these decisions regarding elimination of receptor
points.
6.3.3 Analysis of Exposure Data
Receptor populations are likely to be exposed to toxics through
more than one exposure pathway at a time. Therefore, individual exposure
levels may be combined into a total exposure for intake of a pollutant
through ingestion of different substances, dermal absorption from surface
water and water supply, and inhalation at different locations in the
study area (e.g., work, recreational areas, home, coninuting routes). If
the pharmacokinetics of absorption for each route are understood for the
pollutant being modeled, then these different exposure can be aggregated
into a single total exposure. Otherwise the exposures for each route
must be presented separately and compared.
Daily and annual exposure profiles or scenarios can then be developed
for various subpopulations in the study area. The subpopulations selected
6—13
Arthur D Utt e. nc
-------
30
S. --
.!
.
;
j.
.t ,
r,i t
F I
. :F
I
F
.
:
.
I
E
— — — —
± =
i I Fi
- - - -- -
— — — —
. [ T,
f_ =:
II $
E E
— — — — — — — — — — — —
-
: F
IF
E ÷E
- --
— — b.
IF
E’ F
- - ,- - -- - ---
— —J—.. — — — — —
.,
F
,
- -,- --
— — —
L:
:
::
: :
I
i
:c
::
I .
F
ii
::::::i:
1
I
: : : :c
I
F I F
• ‘ I F
::... .... t.II: ‘_
I ‘ IFI f
—,- —‘.-———
i I 1
: : : : : :1: :r ! _cL: : : :
;..i
::c
I
: ::
1
i IF
:i i
I
——
: I : ::
J
: :c:
: : : :
F
:::
: : :
I
::
: ::
I
F I Ii
Ii I
::i:::::
II
—
F
:ii1: : : :
Ill I ’
:c:
::T:
i
:::c 1:Cirr’:—: 3:::::CI ,::c::::f:i:::::C::C:::
L If I!
- F - - - — - - - - - - -. - - - - -
f I 1. F 1 I
. 5
I I_. 1IIj__.__I_..j___..!_._j__I ?_.__.L_!__.
IF .FIFI I IF 1 F
± : ± : :± ± 1: ii
F 1 ii
T .;,
I II I . . 1 ‘I I
• •1u.. •• —
1 7 158 159 160 161 162 163 164 165 -RECEPTOR NUMBER
.3
2.5
4.5 -MILES FROM SOURCE
FIGURE 6—2
SOURCE: .A..L. (1;E )
RELATIONSHIP BETWEEN SAMPLE POLLUTANT CONCENTRATION AND
DISTANCE FROM SOURCE
Arthur D Little. Inc.
- -t± f- •
—S.
20
10
5
4
3
2
-------
S
\
JP 5-3 : EXPOSUE :cTE : T:: :
)
I
curc :
-------
/) [ ,1 I’!’)1 EX/ )8!’J) TO N! i1/1t, iNJ’1, i ’1 :‘i’ ?JfiI’rIC V — —C1r t )N ‘L’ ?’4! ?iI’.1l !!F;
J —1
&
I a
A
a
*
I *
I *
I A
a
*
*
*
() *
1’ U- a
11 I a
14 I *
it I *
.‘ T a
iu—I A
N I A
I £
I *
2 0- I A
I *
I *
I * * * *
I * A ¼ * *
lu -I A A * * * A
A A A * * *
I * * * A * * k * * *
I * * * * * * * * * * * *
* * * * * * A * * * * * * * *
— ——* — —* —* —* _* — —* —
I I 1 1 I I I I I
2 6 8 10 12 16 18 20
ANNUAL IN2’ 4JCE RATES(ldk’/YEAR)
FIGURE 6-4 . POPULATION DISTRIBUTION FOR RANGES IN ANNUAL CARBON TETRACULORIDE
INTAKE RATES
-------
will depend on the characteristics of the study area, e.g., recreatfonal
use, presence of agriculture, transportation patterns, climate, etc.
Some examples for humans of possible scenarios include:
• typical exposure for the general population by sex,
age group
• occupational exposure
• coninuters
• users of certain home products
• s iiner vs. winter exposure
• recreational exposure
• exposure of residents in the:vicinity of a
hazardous waste disposal site.
6.4 APPLICATIONS OF EXPOSURE ASSESSMENT RESULTS
The estimates of regional human exposure by route and subpopulation
can be used directly, without comparison to health effects data, in a
number of applications to evaluate potential toxics problems in an area.
• Comparison of regional exposure levels to national
“average” exposure levels;
• Comparison of the impact of different routes or
pathways of exposure such as drinking water vs.
inhalation;
• Comparison of locally attributable exposure vs.
imports from outside (such as in foods);
• Comparison of daytime to nighttime exposure or
identification of seasonal variability in exposure;
• Comparison of source-proximate (near—field) subpopulationS
to the rest of the population (far field);
0—
Arthur D L t:ie
-------
• Identification of the regional toxics source s) respon-
sible for the greatest exposure, both in terms of
population affected and magnitude of individual intake;
• Preliminary Identification of large r sensitive sub-
populations or high-exposure areas in the region;
• Comparison of exposures to different pollutants released
by the same sources or through the same environment l media.
Another potential application of the human exposure data developed
for a specific region will be to evaluate the Impact of the exposure
levels in terms of health risk. The more detailed the exposure analysis,
the more accurately Is an assessment of risks associated with regional
exposure. Human health data may be specific to selected subpopulatlons
or conditions (such as dosing) so that the level of detailed required
in the exposure assessment may Include population breakdown by age, sex,
race, as well as other groups such as pregnant women, people with respira-
tory ailments, and so forth. The exposure estimates should reflect
individual dosage over small time steps and the cumulative dose at any
time, as well as the duration of exposure, fluctuations in dosing, and
perhaps, presence of and interactions with other substances and other
influential factors. However, the precision obtained through much ex-
posure estimation may be more than outweighed by the uncertainty asso-
ciated with health effect estimates, particularly in the case of low—dose
extrapolation (e.g., for carcinogenesis).
6-18
Arthur C) Uttle Inc
-------
6.5 REFERENCES
Arthur D. Little, Inc., A Pilot Study - International Caninission on
Radiological Protection (ICRP). Report of the task group on reference
manual, Oxford, England: Pergamon Press; 1975.
Arthur D. Little, Inc., Integrated exposure/risk assessment methodology.
Preliminary draft. Contract 68-01-3857. Washington, D.C.: Monitoring
and Data Support Division, U.S. EPA; 1980.
Consumer Product Safety Cxinission (CPSC) as cited in U.S. EPA; 1980.
Food and Drug Administration (FDA). Total diet studies. Con 1iance
program evaluation. Washington, D.C.: Bureau of Foods; 1973 and 1980.
Moore, R.E., Baes, C.F., McDowell—Bayer, L.M., et al. AIRDOS—EPA: A
Computerized methodology for estimating envi roninental concentrations
and to man form airborne releases of radionuclides. EPA 520/1-79—009.
Washington, D.C.: Office of Radiation Programs, U.S. Environmental
Protection Agency; 1979.
Navada, N.L. Environmental carcinogens and human cancer: estimation of
exposure to carcinogens in the ambient air. EPA 600/1-79—002. Research
Triangle Park, NC: Office of Research and Development, U.S. Environmen-
tal Protection Agency; 1979.
Perwak, J. et al. Exposure assessments of priority pollutants: Lead.
Interim Draft. Contract EPA 68—01-3857. Washington, D.C.: Monitoring
and Data Support Division, Office of Water Planning and Standards, U.S.
Environmental Protection Agency, 1980.
Pielou, E.G. An introduction to mathematical ecology - Toronto: Wiley-
Interscience; 1969.
Pollard, J.H. Mathematical models for the growth of human populations.
Sidney, Australia: Macguarie University, Cambridge athe the University
Press; 1973.
United States Department of Agriculture (USDA). Food and nutrient intakes
of incividuals in the United States. Washington, D.C.: Agricultural
Research Service; 1980.
U.S. Environmental Protection Agency. The exposure assessment group’s
handbook for performing exposure assessments. Preliminary draft, not
formally released by EPA. November, 1980.
U.S. Environmental Protection Agency. Identification and evaluation of
waterborne routes of exposure from other than food and drinking water.
EPA 440/4-79—016. Washington, D.C.: Office of Drinking Water, U.S.
Environmental Protection Agency; 1979.
U.S. Environmental Protection Agency. Proceedings of Workshop on Expo-
sure/Assessment of Hazardous Chemicals, as cited in U.S. EPA 1980. By
adian Corporation, EPA contract Lo. 68—02—3171; 1980.
Arthur D L ttIe Inc
-------
8. PRACTICAL CONSIDERATIONS
3. 1 FEASIBILITY OF THE GEOGRAPHIC APPROACH
From a technical point of view, the geographic approach allows the
use of detailed local information to arrive at highly specific conclu-
sions regarding toxic substance multi-media problems, and then permits
the development of appropriate control strategies. This approach Is
feasible provided that sufficient data are available in the various
pathways of concern. In some cases, a preliminary data-gathering and
field sampling program may be necessary in order to lay the groundwork
for a geographic study. In other cases, the need for additional data
may be recognized only after a first Iteration of the exposure assessment
methodology. Apart from the problem of data sufficiency, the other
elements of the geographic approach have a sound technical basis. Though
there is some uncertainty with regard to the outputs of fate and transport
models, as well as exposure models, these outputs provide a sufficient
d. ‘ree of resolution to permit control strategy development.
A ‘iengraph : study, especially in the control strategies development
t sk, will require special administrative coordination so that represent—
dtives rrom relevant Federal, State and local programs and interest
.;roups can have access to the on-going work. Some routine forums, or
. d hoc tasks forces may be required to provide continuing interaction
among these parties. Public relations offices in each agency will have
t provide some of this coordination.
More importantly, means of cotmnunicating interim study results to
relevant parties will be very important in promoting meaningful inputs
from these parties, especially where control strategies other than
command/control are introduced. Similarly, means of soliciting
inputs from affected groups, industry, coninercial, local authorities
or conm unity groups will be essential if a working cooperation is to be
pursued. The Federal EPA can provide much needed guidance in this area.
8-1
Arthtir t) ittle Inc
-------
The most difficult administrative problems for control strategies devel-
opment are likely to be associated with managing local problen through
local authorities. !lere, the amount of technical expertise required to
provide input on assessments of the technical feasibility of alternative
control options will be quite high due to the site—specific nature of
such issues. If many studies are under way, the agencies may have a
critical deficit of such expertise available either in-house or through
its contracted services. This shortage of technical expertise will
affect not only the problem diagnosis and control strateqy selection
phase but also the follow-up program phase. It may be advisable in
some cases to consider developing a stock of locally available expertise.
In such a case, some local people with interest and relevant professional
background or experience might be brought into the study project and
prepared to manage some of the requl red follow-up programs. Such a
group could at the minim’in serve as an interface between the study
project and continuing efforts to implement resulting controls. This
prospect, however, could introduce another layer of administrative
complexity, and the efficiency of such a strategy should be explored.
8.2. POLENTIAL L1MITATI 4S OFTHE t’ETHODOLOGY
.1. Ade iacy uf Data
The most important limitation that will be encountered in
any fut ire implementation of the geographic approach is the
adequacy of the available data. This particularly affects the
inventory of sources and emissions, which is the foundation of
the entire exposure assessment methodology. Development of
quantitative emission estimates on a source—specific basis
requires detailed information which simply may not be available.
The use of assumptions, approximations, or national averages
to characterize emission rates can be a useful approach, but
will tend to weaken the credibility of the results. There is
no unified set of default values that may be used in the absence of
specific data. Moreover, there will inevitably be gaps in the
identification of pollution sources, due to non—uniform reporting
8-2
Arthur D UttLt Inc
-------
practices and ongoing changes in the study area. new facilities or
modifications in old ones may not be recognized, and existing records
may omit certain critical items of data. The source categories for
which data gaps are expected to be most acute are as follows:
• hazardous waste disposal
• toxics emissions to air
• non-point loadings to surface water
• point source discharges to water for industries other
than the primary ones or for non-priority pollutants.
Information for the source emissions quantification will rely mainly
on data furnished by industries in compliance with regulatory requirements;
as these requirements change, the data base will vary accordingly. In
general, it will not be possible to provide frequency distributions for
s urce emissions. The level of precision will usually be confined to
point estimates and ranges of variation, thus limiting the statistical
a curary of the subsequent exposure assessment tasks.
1n .dequacy of data can limit the accuracy of the exposure
assess nt in other ways. The absence of sufficient field
, onitorinu data for air, water, soil, or biota can make it
more ditficult to analyze environmental pathways for toxic
pollutants and to validate the results of modelling efforts.
•nce r..r delling can only be performed for selected pollutants
and selected media, an exposure assessment must rely upon
monitoring data to ascertain the ambient levels of many of the
.llutants under consideration. Furthermore, the estimation
of exposure levels for specific receptor groups requires a
considerable amount of data concerning recreational patterns,
dietary habits, and popu1 tion distribution, which may be
scarce or nonexistent. These types of data gaps will result
in exposure estimates that may be incomplete for certain
substances in certain media, or that may have high levels of
u icertai nty.
8-
Arthur I ) I ink. Inc.
-------
Finally, in the control strateoy development phase, the
data problems indicated above will often make it difficult to
present credible quantitative conclusions for the purpose of
justifying exposure problems or the need for control options .
In some cases it may not be possible to establish a substantial
link between sources and receptors upon which environmental
benefits must be predicated. However, even when the environ-
mental pathways are well understood, thereare additional data
concerns which may limit the usefulness of the control strategy
analysis. For example, the costs and efficiencies of emission controls
will often be difficult to evaluate on a site—specific basis,
due to confidentiality restrictions and peculiar features of
indivi ia1 facilities. Cost estimates based on generic industry
data may distort considerably the actual situation in the
study area.
8.2.2. Validity of Models
A c:ritical link In the exposure assessment methodology is the
analysi of environmental pathways of toxic substances. Therefore,
an important limitation of the general methodology has to do with
the limitations of the single medium models to be employed. In general,
these limitations may be due to: (1) data availability, (2) inappropriate
model s lection, (3) omission of important pathways, and (4) lack of
appropriate model validation.
The types of data required for iathway modelling include environmental
data (e.g., pollutant import to basin), chemistry data (e.g., chemical
Properties of pollutants) and monitoring data (e.g., ambient concentrations).
(The latter was discussed in Section 8.2.1.) It is not appropriate,
for example, to drive any river model without a proper set of flow
records, nor to attempt validation of the model output with inadequate
monitoring data.
With regard to model selection, the choice of a sinQie medium model
requires a compromise between level of sophistication, data availability,
f4
Arthur I) little Inc
-------
and needs and resources (computer, time, budget) of the project. Selecting
for example, a numerical finite element groundwater model and a lump—san
simulation watershed model might be an inappropriate decision due to mathe-
matical complexity and pragmatic difficulty. The selected models will
inevitably require some simplifying assumptions such as complete mixing
in rivers or average long-term meteorological conditions. tittlng
certain important pathways (e.g., waste disposal site to groundwater to
river) because of lack of data may also be an issue of concern. Such
pathways should preferably be discussed, and perhaps modeled with simpli-
fied algorithms, so that their omission can be justified.
Finally, model calibration and model validation are an important part
of the general methodology application. Non-validated modelling efforts
can have only qualitative use; therefore, final model predictions should
be validated with any available data. However, a disagreement in
absolute levels does not necessarily indicate that the selected method was
incorrect or that data sets employed are suspect. Rather, it indicates
need for recognizing the methodology limitations and for perfonning sensi-
tivity analyses upon the data sets in question. Engineering judgment and
professional experience will be required to assess the range of uncertainty
in model predictions and the importance of discrepancies between those
predictions and monitoring data.
8.2.3. Scope of Analysis
An important limitation of a geographic study may arise from the
scope under which thestudy is conducted. The present methodology has
selected exposure as a quantifiable endpoint of the analysis; thus health
effects of pollutant exposure have been excluded from scope. Since healt
effects are not site-specific, it can be argued that information about
health and environmental imports could be incorporated with the results
of the geographic study in order to assess the benefits of various control
st -ategies. However, consideration of health effects as an independent
item of information implies that exposures will not be weighted according
tc the relative potency or severity of effects for different pollutants.
8-5
Anhurl) t iu.k Iiic.
-------
rhough toxicity was considered inthepollutant selection (Section 23),
the subsequent exposure assessment and control strategy analysis did not
explicitly address the different risks associated with the pollutants in
uesti on. Unfortunately, the present state—of-the—art of risk assessment
does not permit accurate quantification of risk; in fact, the only effect
for which quantitative estimates of risk are possible is carcinogenesis,
and even in this case the range of error is enormous. Thus the inclusion
of risk estimates would greatly decrease the accuracy and credibility of
the results. Moreover, explicit statements about risk to local populations
might be inflaninatory and could create unnecessary difficulties for Federal
and State authorities by suggesting unconfirmed problems.
Another potential limitation associated with the scope of the study
is the variety, of possible exposure routes that are considered. The present
methodology has focused heavily upon exposure to toxics in the ambient
environment, including focu and drinking water, but has paid little or no
attention to exposures that may arise in occupational, coimnercial, or
residential settings. This emphasis is due mainly to the limits of the
EPA’s jirisdiction. However, the total exposure of individuals to a
particular toxic substance may be an important criterion in the develop-
ment of regulatory strategies. For example, if only 5% of average per
capita exposure to a substance is due to environmental pathways, then
reduction of environmental emissions by 90% will have an inconsequential
effect upon the per capita risk associated with that substance. Thus,
a broader scope of analysis may be warranted, if only to gauge the true
significance of environmental emissions. It might be advisable for the
EPA to coordinate with other Federal agencies, such as OSHA or FDA, in
the context of the geographic approach, so that a more complete perspec-
tive is obtained concerning toxic substance exposure.
vc— 1 “ !‘
-------
8.3. STUDY TIMETABLE AND RESOURCE EQUIREME Th .
The general methodology outlined in this report has focused upon
the functional activities required to perform a geographic study. A nin—
istrative questions such as staffing, public relations, anddivision of
authority have not been dealt with here. From a purely technical point
of view, however, it is possible to lay out a master project flow chart,
showing the timing and relationships of the various activities described
in the previous chapters. This chart is displayed in Figure 8-1. The
three main phases of a geographic study are depicted with dotted lines,
as defined in Section 1.4. In addition, the- interaction between a field
sampling program and.the study tasks is shown. The EPA has ha d extensive
experience in designing such sampling programs, so that the incorporation
of field work into the geographic study will present no hidden technical
obstacles. It is anticipated that the field work could be performed in
parallel with Phase 2 of the study, creating little or no time delay.
Modifications to the study outputs based upon field results could normally
be incorporated toward the end of Phase 2. However, if the field work
produc some surprise results whicn alter the significance of certain
environmental pathways 1 some revi’.iun of the initial scan and exposure
assess ient scope might be required. In a situation where the initially
available data were totally inadequate, it might be advisable to perform
the field work prior to completing the initial scan. In such a case,
the indicated parallelism would not be possible, and the duration of the
study would be extended by as much as nine months.
The anticipated requirements of a geographic study in terms of
professional time, computer and other expenses, and calendar time have
been estimated in very rough terms. These estimates are shown in Table 8—1;
note that field sampling cost estimates, though they have not been included,
could easily exceed the costs of the study itself. Because of the varia-
bility in potential site characteristics, and because of the limited
experience to date (only one pilot stuay), it must be recognized -that
actual costs could differ considerably from these estimates. They are
E-7
rU ur - 1:
-------
K J L .1- 1 : 4ct vity Flue Chart • jr uqra; flh. Study
2
3
4
5
6
7.
9
10
U
12
13
14
15
E1a s ’1 Initiate Study - 1
• Co-ordim te State and Loc 1 Participatlont
zo
Terminate ‘Study
- ReDort and li’ p nd
8-8
-------
TABLE 8-1
.ESTIMATED RESOURCE REQUIREMENTS FOR A GEOGRAPHIC STUDY
EXCLUDING FIELD SAMPLING
for one quarter of the pollutants
* for half of the pollutants
PHASE
TASK DESCRIPTION
Professional
Ti rue
flian-Hours L
20 POLLUTANTS
40
POLLUTANTS
Expenses
includin
Computer
Duration
( Months )
Professi anal
Ti me
( Man-Hours )
Expenses
includin
Computer
Duration
( Months )
1
State and Local CoordInation
Boundary Definition
Selection of Pollutants
Data Acquisition
Initial Scan of Pathways
Phase 2 Planning
HASE 1 TOTAL
200
200.
400
2,000
1,000
200
T )O0
.
$20,000
1
**
1
**
3
1
6
200
200
400
2,000
1,000
200
4,000 -
.
( $2O,0O0
1
1
**
3
1
6
2
Management and Reporting
Source Identification
Quantification of Emissions
Pathway Analysis*
Receptor Analysis
Exposure Assessment
1,001) .
500
1,500
.4,000
1,000
1,000
$20,000
,
,
1
2
4
**
2
1,000
500
2,500
6,000
1,000
2,000.
$40,000
1
2
4
2
PHASE 2 TOTAL
9,OOO
$40,000
9
13,000
$60,000
9
3
.
Management and Reporting
‘Problem Identlficatlont
Selection of Control Options
Evaluation of Control Options
1,000
500
1,000
2,000
$10,000
**
1
2
2
1,000
1,000
2,000
3,000
.
$20,000
**
i
. 2
2
Strat! y Development
PHA! 3 W11 E
500
5 ,OO0
$25,000
1
— 6
1,000
8,000
$35, 00 0
1
5
GRAND TOTAL
18,000
$75,000
21
25,000
$105,000
21
**can proceed in parallel with oth r t ”ks
-------
provided only as a guideline, subject to refinement as the geographic
approach continues to evolve. Two different estimates are given, assuming
that 20 and 40 pollutants are treated; clearly there are economies of scale
in the latter case. Personnel requirements are expressed in man-hours, so
that different conversion factors can be applied for cost estimation
In both cases (20 and 40 pollutants) it is assumed that pathway modelling
is performed for half the pollutants addressed; this may be somewhat high,
since in many cases either the necessary data may be absent or the prior-
ities established may dictate a smaller modelling effort. Likewise, in
Phase 3 it is assumed that the number of problems identified for further
study is one-quarter of the total number of pollutants; this assumption
is fairly arbitrary since actual occurrence of problems is difficult to
predict. The duration of both options is the same under the assumption
that increased man-power would be available for the 40-pollutant case.
The size of the project team could vary, but if full-time people were
assigned, about eight people would be required for the 20-pollutant case
and about twelve for the 40-pollutant case. The mix of personnel is
assumed to be about 2O managerial, 60% technical/scientific, and 20%
clerica i/support.
8.4__C0NCLIJS ONS
A geographically-oriented exposure assessment methodology has been
developed which is suitable for tue evaluation of intérmedia tuAjC
problems, and provides a basis for the development of cost-effective
control stratenies. Focusino on a we1l-d fined local area oermlts a
detailed inventory of sources and nathways of toxics exposure across all
environmental media. Moreover, the Investigation of feasible control
options can include state or local authorities in seeking the most
efficient solution to an existing or potential problem. Implementation
of the geographic approach will require the development of an appropriate
site selection mechanism for identifying local areas that merit this
type of intensive investigation.
a•l0
Arhur ) t k 1fl(
-------
The general methodology consists of three phases of work: an initial
scan period during which data are collected and priorities are established.
an exposure assessment phase during which detailed technical analysis is
performed of receptor exposure to selected toxics, and a control strategy
phase during which means are sought for reducing specific exposure levels.
Although the methodology is oriented toward quantification of exposure
rather than risk, it would be a simple matter to extend the analysis by
attaching risk estimates to each exposure level. However, such risk
estimation may introduce a substantial amount of uncertainty, due to the
difficulty of predicting human responses based on laboratory animal health
effects data.
The key to successful application of the geographic methodology is
the establishment of links between source emissions and receptor
exposures, via a description of the envirorinental fate and pathways of
toxic substances. This goal may be hampered by the absence of sufficient
data or by the difficulty and expense associated with pathway modeling.
However, explicit quantification of these links can provide a firm,
rational basis for decision-making relative to alternate control strategies.
Even when the source-pathway-exposure chain is not completely understood,
a systematic identification of optional control mechanisms will permit
a thorough evaluation of the costs and potential benefits of controlling
the substances in question.
8-11
Arthur 1)1 i k h
-------
APPENDIX A -- POLLUTION FLOW MODEL
In a geographic study, the relationship between source emissions and
ambient levels of toxic substances can initially be investigated through
the use of environmental pathway models. However, for repeated sensitivity
analyses these computer models become unnecessarily expensive to run.
Instead it is possible to take advantage of the linearity in the model
results, ind to replace the computer simulation with a set of algebraic
eqt atioris that relate emissions to anthient levels. This approach not
only reduces the cost of multiple analyses, but also permits the
determir tion of required emission levels as a function of target ambient
cor centr tions. . suimnary of the algebraic approach is presented below:
Suppose the area of study consists of N well-defined cells , or
reoions.. into and ouj of which a p lutant may flow. Suppose there are
4 sources of the rollutant in the a’-2a of study, each source contributing
a fixed ass of pollutant (perhaps ‘t ro to each cefl in each time step.
S pose irther that the proportior’ of the pollutant present in a given
cell at the beginning of a time step vhich has flowed to another cell by
the end of the time step is constant.
Define :
X; ( )
X(i) = ,x 2 (1) , x, (i) = mass of pollutant contained in cell j
at the beginning of time step i.
xN(i)
I,* I C. I,
A = a 2 , 1 a 2 , 2 a 2 ,N a proportion of mass in cell j
at beginning of time step that
aNi aN2 aN,N flows to cell i by end of time
step.
I. —l
I ! t...
-------
U. ’
U = u 2 , U 1 = mass of pollutant discharged by source
during one time step.
a 1 , 2 B1,M
a a ...
B 2,1 2,2 2,M , =proportion of mass discharged
• . (3 from source j that is
P1,1 ‘N,2 N,M discharged into cell i.
It follows, from our assumptions and these definitions, that the mass
of pollutant.contained in each cell at the beginning of time step n Is
given by:
X(n) = AX(n—1) + BU.
After many time steps, the mass of pollutant in each cell will be
great enough so that the constant proportion flowing out will equal the
proportions from other cells flowing in plus the fixed amount being
discharged to the cell directly. The mass of pollutant in each cell will
remain constant; the system has reached a steady state X;
Xl
= X 2 , = steady-state mass of pollutant in cell 1.
It follows, that:
= [ (I—A) BJU (1)
where I is the identity matrix. Given proportions and and discharges
U 1 , therefore, we may calculate the steady-state X directly.
A -2
Arthur D l ‘;kln .
-------
-------
Alternatively, we may calculate the quantity BU given a “target” X:
BU (I-A)T (2)
The quantity BU, it a ay be seen, gives the total mass discharged intn.
each cell by all sources:
l,.lul + a 2 , 1 2 # ... +
8 1 ,u + U+...+
= , , . 1 2,22
al,Nul + 2NU 2 + •* +
Sin le River Model
For a simple river model, we have:
X 1 (fl) = a 1 , 1 X 1 (n—l) + u _ 1 x 11 (n—1) + u 1 .
That is, the mass of pollutant in cell i at the beginning of time step n
is the amount remaining from the beginning of time step (n-i), pits the
amount flowing in from the cell upstream, plus the amoUnt discharged
directly into the cell,
In the steady state,
X. +U
1,1 i 1,1 — 1 — l 1’
and it follows that:
XM + ( ..T 1 = y 1 X 11 + & U 1 ,
a. 1
where = ( 1: -) - cl—x 1 j1
Knowing the steady-state solutions and for two different discharges
tJ and U , we may calculat. the y and S , and hence the a and a
2 i,i
dnd then iay use the general equations (1), and (2) for further analysi
of X, U, nd B.
Arthnrl)t titleinc.
------- |