-------
                                     EPA/540/1-881001
                                OSWER Directive 9285.5-1
                                          April 1988
Superfund Exposure Assessment
                 Manual
          U.S. Environmental Protection Agency
            Office of Remedial Response
              Washington, DC 20460

-------
                                    Notice
This report was prepared under contract to an  agency of the United  States
Government.  Neither the United States Government nor any of its employees,
contractors,  subcontractors,  or  their  employees  makes  any warranty,
expressed or implied, or assumes any legal liability or responsibility for any
third party's  use of or the results of such use of any  information, apparatus,
product, or process disclosed in this report, or represents that  its use by such
third party would not infringe  on privately owned rights.

-------
                                    Table of Contents

 Chapter                                                                              Page
List of Tables   	      vi
List of Figures  	      VIM
Foreword   	   ix
Executive  Summary  	      xi
Acknowledgments   	      xii

1      INTRODUCTION   	      1

       1.1    Purpose   	      1
       1.2    Background  	      1
       1.3    Scope                                      	    1
       1.4    Use   of  the   Manual  	    2
       1.5    Timeframe of Analysis  	     4
       1.6    Analysis of Exposure Associated with Remedial Actions   	    4
       1.7    Organization of the Manual  	     5

2     CONTAMINANT RELEASE ANALYSIS  	     7

       2.1   Introduction   	      7
       2.2   Contaminant  Release Screening  	     8
             2.2.1    Contaminants in Soil  	     8
             2.2.2  Contaminants  Above-Ground                                             10
       2.3   Quantitative  Analysis  of Atmospheric Contamination	     10
             2.3.1    Fugitive Dust Emission Analysis  	      10
                    2.3.1.1  Beginning Quantitative  Analysis  	      10
                    2.3.1.2 In-Depth Analysis    	      14
             2.3.2   Volatilization Emission Analysis  	      14
                    2.3.2.1  Beginning Quantitative  Analysis  	      14
                    2.3.2.2 In-Depth Analysis    	     21
             2.3.3 Long-Term and  Short-Term Release Calculation                          22
       2.4   Quantitative Analysis of Surface Water Contamination   	    22
             2.4.1    Beginning Quantitative Analysis   	     23
                    2.4.1.1  Dissolved and  Sorbed Contaminant Migration   	    23
             2.4.2   In-Depth Analysis    	     25
             2.4.3 Long-Term and  Short-Term Release Calculation    	    27
       2.5   Quantitative Analysis of Ground-Water Contamination    	    29
             25.1    Beginning Quantitative Analysis   	     29
                    251.1   Leachate Release Rate 	     29
             2.5.2 In-Depth  Analysis    	     31
             2.5.3 Long-Term  and Short-Term Release Calculation     	    31
       2.6   Soil Contamination   	     31
            2.6.1    Beginning Quantitative Analysis   	     31
            2.6.2    In-Depth Analysis    	     31

3     CONTAMINANT FATE ANALYSIS  	     35

       3.1   Introduction   	     35
       3.2  Contaminant  Fate Screening   	     36

-------
                             Table of Contents (Continued)

 Chapter                                                                                Page

             3.2.1   Atmospheric Fate  	      36
             3.2.2   Surface Water Fate   	      38
             3.2.3   Soil and Ground-Water Fate    	     40
             3.2.4  Biotic Fate                                                              40
       3.3   Quantitative Analysis of Atmospheric  Fate	     42
             3.3.1   Screening Analysis   	      42
             3.3.2   In-Depth Analysis	      46
                    3.3.2.1  Intermedia Transfer                                             46
                    3.3.2.2  Intramedia Transformation  Processes	     47
                    3.3.2.3 The Effects of Terrain                                            48
             3.3.3   Computer Models  	      48
             3.3.4  Short- and Long-Term Concentration Calculations                          48
       3.4   Surface Water Fate Analysis   	     53
             3.4.1   Beginning Quantitative Analysis   	     53
             3.4.2  In-Depth  Analysis.                                                       55
                    3.4.2.1 Intermedia Transformation Processes	     55
                    3.4.2.2  Intramedia Transformation Processes  	     56
                    3.4.2.3  Computer Models  	      56
                    3.4.2.4  Short- and Long-Term  Concentration Calculations                 57
       3.5   Quantitative Analysis of Ground-Water Fate   	    57
             3.5.1   Discussion of Ground Water Modeling  	     63
                    3.5.1.1  The Contamination Cycle                                        63
                    351.2 Ground Water Flow Conditions	      64
                    3.5.1.3  Multiphase Flow  	      65
                    3.5.1.4  Contaminant Flow and Hydrodynamic Dispersion   	    65
                    3.5.1.5  Transformation and Retardation  	     66
                    3.5.1.6  Higher Velocity  Transport   	     68
             3.5.2  Ground-Water  Modeling Equations and Nomograph                         68
                    3.5.2.1    Calculating Ground Water Velocity   	     68
                    3.5.2.2 Calculating the Velocity of Infiltrating Rainwater  	    69
                    3.5.2.3  Corrections  for Viscosity and Density  	     73
                    3.5.2.4  Retardation  Effects  	      73
                    3.5.2.5  Contaminant Velocity   	     75
                    3.5.2.6  Nomograph  Technique  	     77
                    3.5.2.7  Extent of Plume   	      77
                    3.5.2.8  Use of Monitoring Data  	     82
                    3.5.2.9  VMS Model   	      82
             3.5.3  In-Depth Methods and  Models                                            83
             3.5.4   Short- and Long-Term Concentration Calculations	      93
       3.6   Biotic Pathways  	      93
             3.6.1   Estimation Procedures  	      93
                    3.6.1.1  Aquatic  Animals	    94
                    3.6.1.2  Terrestrial Animals  	      94
                    3.6.1.3  Terrestrial Plants   	      94

4     UNCERTAINTY IN THE ANALYSIS   	      95

       4.1   Sources of Uncertainty   	      95
             4.1.1   Input Variable Uncertainty  	      95
       4.2   Modeling Uncertainty  	      96
             4.2.1   Model Simplification   	      96
             4.2.2 Averaging  Hydraulic Conductivities   	     96
             4.2.3 Dispersion Simulation	      97
             4.2.4 Numerical Models  and Analytical Models   	     97
             4.2.5 Chemical Degradation Simulation   	     97
             4.2.6 Model Operational  Parameters   	     97

                                               iv

-------
                            Table of Contents (Continued)

 Chapter                                                                            Page

            4.2.7 Source Shape  	      98
            4.2.8 Steady State Modeling.   	      98
            4.2.9 Number of Dimensions Addressed by the Model 	     98
      4.3 Scenario Uncertainty  	      98
      4.4 Approaches for Dealing with Uncertainty   	
            4.4.1 Sensitivity Appraisals   	      98
            4.4.2 Monte-Carlo Simulations              	     99
            4.4.3 Using Monitoring Data to Calibrate the Model  	     99
      4.5 Level of Uncertainty Appropriate for Exposure  Modeling  	    100
      4.6 Risk Communication   	     100

5     REFERENCES   	     103

APPENDIX A   Analysis of Exposed Human Populations and
               Exposure Calculation and  Integration   	    113
APPENDIX  13   Possible Exposure Assessment Data Requirements for Uncontrolled
               Hazardous Waste Sites and  Index to Variable Terms   	    135
APPENDIX C   Data Management Forms  	     145

-------
                                      List  of Tables
 Number                                                                             Page
1-1      Technical Resource Contacts for Superfund Exposure Assessments   	    5
2-1      Potential Contaminant Release Mechanisms  	     8
2-2      Environmental Variables and Model Parameters for
            the Wind Erosion Equation   	      13
2-3      Diffusion Coefficients of Selected Organic Compounds   	     18
2-4      "C" Values for Permanent Pasture, Rangeland, and Idle Land   	     26
2-5      "C" Values for Woodland   	      26
2-6      Runoff Curve Numbers   	      27
2-7      Parameter Values for Permeation Equation (at 25°C)  	     32
2-8      Polymer Categorization  for Permeation of Water  	     32
2-9      Permachor Values of Some Organic Liquids in Polyethylene and PVC  	    32
2-10    Water Permachor Value for Dry Polymers  	     33
3-1      Key to Stability Categories  	      45
3-2      Resource Requirements and  Information Sources: Atmospheric Fate Models	    49
3-3      Features of Atmospheric Fate Models  	      51
3-4      Data Requirements for Atmospheric Models  	     52
3-5      Resource Requirements and  Information Sources: Surface Water Fate Models  ....    58
3-6      Feature of Surface Water Fate Models  	      61
3-7      Data Requirements for Surface Water Models   	     62
3-8      Representative Values of Saturated  Hydraulic Conductivity  	     70
3-9      Saturated Hydraulic Conductivity Ranges for Selected Rock  and Soil Types  	    70
3-10    Representative Values for Saturated Moisture
            Contents and Field  Capacities of Various Soil Types  	     70
3-11    Representative Values of Hydraulic Parameters
            (Standard Deviation in Parentheses)   	     71
3-12    Suggested Value for Cet Relating Evaporation from a US Class A Pan
        to Evapotranspiration from  8 to 15-cm Tall, Well-watered Grass Turf     	    72
3-13    Crop Coefficients for Estimating Evapotranspiration   	     73
3-14    Resource Requirements and Information Sources:
             Unsaturated  Zone  and Ground-Water Fate Models   	     84
3-15    Features of Unsaturated Zone and Ground-Water Fate Models    	     88
3-16    Data Requirements for Unsaturated Zone and Ground-water  Models    	    91
                                             VI

-------
                              List of Tables (Continued)

 Number                                                                            Page

A-1      Regional Census Bureau Offices   	     118
A-2     U.S. Home Fruit and Vegetable Garden Use, 1977  	    119
A-3     Summary of Human Inhalation Rates for Men, Women, and Children
            by Activity Level (m3/hour)   	     123
A-4     Permeability Constants for Various Compounds  	    124
A-5     Typical Daily Soil Ingestion Rates for Children by Age Group   	    129
B-1      Possible Data Requirements for Estimation of
             Contaminant Release and Transport and Exposed Populations   	   136
B-2     Index to Variable Terms   	     139
                                             VII

-------
                                      List of Figures

 Number                                                                               Page

1-1      Overview of the Integrated Exposure Assessment Process  	    2
2-1      Contaminant Release Decision Network: Contaminants in Soil  	    9
2-2     Contaminant Release Decision Network: Contaminants Above-Ground    	    11
2-3     Mean Number of Days Per Year with >  0.01 Inches of Precipitation
           (i.e., "wet days")   	      15
2-4     Slope Effect  Chart Applicable to Areas  A-l in Washington,
           Oregon, and Idaho, and all of A-3    	     24
2-5     Soil  Moisture-Soil Temperature Regimes of the Western United States                  24
2-6     Slope Effect Chart for Areas Where Figure 2-5 is Not Applicable   	     24
3-1      Environmental Fate Screening Assessment Decision Network:  Atmosphere   	    37
3-2     Environmental Fate Screening Assessment Decision Network:  Surface Water  	    39
3-3     Environmental Fate Screening Assessment Decision Network:
            Soils and Ground-water   	      41
3-4     Environmental Fate  Screening Assessment Decision Network: Food Chain   	    42
3-5     Horizontal Dispersion Coefficient as a Function of Downwind Distance
            from the Source   	      43
3-6     Vertical Dispersion Coefficient as a Function of Downwind Distance
            from the Source   	      44
3-7     Area Within Isopleths for a Ground-Level Source   	     47
3-8     Nomograph for Solutions of Time, Distance, and Concentration
           for Any Point Along the Principal Direction of Ground-water Flow   	     78
A-1      Exposed Populations Decision Network   	    115
A-2     Quantitative  Exposed Population Analysis  	    117
                                             Viii

-------
                                  Foreword
The Super-fund Exposure Assessment Manual presents an integrated method
to help Remedial Project Managers and their  contractors define the three major
components involved in assessing human population exposure to contaminants
released from uncontrolled hazardous waste sites:

   1.  Analysis of toxic contaminant releases;
   2.  Determination of the environmental fate of such contaminants; and
   3.  Evaluation of the nature and magnitude of exposure to toxic
      contaminants.

This report provides guidance for the development of exposure assessments
using monitoring data (which may provide the most dependable basis for
evaluating some existing exposure levels), as  well as modeling techniques to
predict exposure over time.

-------
                                      Executive  Summary
The  analytical  process  outlined in the Superfund
Exposure Assessment Manual provides  a framework
for the assessment of exposure to contaminants at or
migrating from  uncontrolled hazardous  waste  sites.
The  application  of both  monitoring  and  modeling
procedures  to  the  exposure  assessment process is
outlined. This  process  considers  all  contaminant
releases  and exposure routes  and  assures  that an
adequate level of analytical detail is applied to support
the human health risk assessment process.

The analytical process is structured in five segments:

    1.  Analysis of contaminant  releases from a
       subject site into environmental media;

   2.  Evaluation of  the  transport and
       environmental fate of the contaminants
       released;
    3.  Identification
       characterization
       populations;
 en u meration ,  and
of  potentially  exposed
   4. Integrated exposure  analysis; and

   5. Uncertainty  analysis.

The  Superfund  Exposure  Assessment Manual
supports  the development  of exposure assessments
that are consistent from  site to site,  and provides a
means  of  documenting that  each  site  receives
adequate  evaluation.  The procedures  presented
reflect current  (at  the time of publication) state-of-
the-art  methods for conducting an  exposure
assessment. However, it is important for the analyst
to recognize  that  exposure  assessment is a
developing  science. Although the  overall  protocol for
conducting  exposure assessments at Superfund sites
will not  change significantly over time and the basic
parameters needed as input to the analysis are  not
likely to change, alternative analytical methods  may
be developed for many parts of the assessment. The
methods presented in this manual  can  serve as a
benchmark against which such new methods can be
compared.
                                                 XI

-------
                                Acknowledgments
This  document was developed  by  EPA's  Office of Emergency and  Remedial Response (OERR).
Dr. Craig Zamuda of OERR's  Toxics Integration  Branch was the EPA Project Officer. Additional
guidance  was  provided by Peter  Tong  and  Mary-Virginia Wandless  of  the  Toxics Integration
Branch.

Assistance was also provided by the following people:
      Bob Ambrose
      Doug Ammon
      Brint Bixler
      Robert Carsel
      Richard Daley
      Carl  Enfield
      Tom Evans
      Kevin Garrahan
      Mark Garrison
      Steve Golian
      Karen Hammerstrom
      Seong T. Hwang
      Joe Keeley
      Ashok Kumar
      Steve Ostrodka
      Zubair Saleem
      Paul  Schumann
      James Spatarella
      Richard L. Stanford
      Sherry Sterling
      David Tetta
      Louis J. Thibodeaux
      Jawed Touma
      Georgia Valaoras
      Paul K.M. van der Heijde
      Larry Zaragoza
ORD (Office of Research and Development)
Clean Sites, Inc. (formerly USEPA)
CH2M Hill (formerly USEPA)
ORD (Office of Research and Development)
OWPE (Office of Waste Programs Enforcement)
ORD (Office of Research and Development)
ORD (Exposure Assessment Group)
ORD (Exposure Assessment Group)
USEPA Region III
OERR (Office of Emergency and Remedial Response)
OTS (Office of Toxic Substance)
ORD (Office of Research and Development)
Oregon Graduate Center
University of Toledo
EPA Region V
OSW (Office of Solid Waste)
OSW (Office of Solid Waste)
Versar, Inc. (formerly USEPA)
Roy F. Weston, Inc. (formerly USEPA)
OWPE (Office of Waste Programs Enforcement)
EPA Region X
University of Arkansas
OAQPS  (Office of Air Quality Planning and Standards)
OWPE (Office of Waste Programs Enforcement)
Holcomb Research Institute
OSWER (Office of Solid Waste and Emergency Response)
Versar, Inc. assisted OERR in the development of this document in  fulfillment  of Contract Nos.
68-01-6271, 68-03-3149, and  68-01-7090. The Versar  project team included  H.  Lee
Schultz, Walter A. Palmer, Mark L. Mercer,  Ruth A. Dickinson, Gary Whitmyre, Alan F. Gleit, Gina
H. Dixon, and Van Kozak (currently Texas Department of Agriculture).
                                          XII

-------
                                           Chapter 1
                                          Introduction
1.1 Purpose

The  Superfund Exposure  Assessment  Manual
provides Remedial Project Managers (RPMs) with the
guidance necessary  to  conduct  exposure
assessments that meet the needs of the Super-fund
human health  risk evaluation process.  Specifically,
the manual:

1.  Provides an  overall description  of the integrated
   exposure assessment  process as it is applied to
   uncontrolled hazardous waste sites; and

2.  Serves as  a source  of reference concerning the
   use of estimation  procedures  and computer
   modeling techniques  for the analysis of
   uncontrolled sites.

This manual provides guidance for the assessment of
human population  health risk  only.  Guidance for
ecological risk assessment  will  be provided
separately.

1.2 Background

The  Comprehensive Environmental  Response,
Compensation, and Liability Act  of 1980 (CERCLA -
42  USC 9601  et.  seq.),  as  amended  by  the
Superfund Amendments  and  Reauthorization Act of
1986  (SARA),  was  enacted  to  provide  the Federal
Government with the authority to  respond to  releases
or threatened  releases  of hazardous substances,
pollutants, or  contaminants into the  environment. As
prescribed in  the revised  National Contingency Plan
(see  47 FR 137, July 16, 1982), all  sites designated
for in-depth evaluation are included on the  National
Priorities  List.  These sites are evaluated  for  remedial
action through the application  of a  Remedial
Investigation, which defines the  nature and extent of
contamination,  and a Feasibility  Study, in which
potential  remedial  alternatives  are developed  and
analyzed.  Guidance  for conducting these two major
components of the  remedial response process is
provided  in USEPA (1985a  and 1985b,  respective-
ly - currently  under revision). As discussed in  that
guidance, a  part of the Feasibility Study is the
development of  a risk assessment that projects those
health  impacts resulting  from the  uncontrolled  site.
The risk assessment is based on the results of a site
exposure assessment, which evaluates:

1. The type and  extent of contaminant release from
   a site to environmental media;

2.  The environmental transport and transformation of
   contaminants  following release; and

3. Implications of the resulting  contact with exposed
   populations.

Section 110 of SARA  mandates that  health
assessments  be  conducted by the Agency for Toxic
Substances and Disease Registry for all  sites  on the
National Priorities List. These health assessments can
be based on the results  of site-specific exposure
assessments. The exposure assessment,  therefore, is
an analytical tool that is used to comply with the
mandates of CERCLA.
1.3  Scope

This  manual provides guidance for the use (but  not
the  acquisition)  of field data  in  the exposure
assessment process. It  does not serve as an  all-
encompassing guide to the use of computer models
in the site remediation process, or direct the analysis
of health risks that  result  from  predicted  exposure.
This  manual is  intended to be used  in conjunction
with  other  related guidance,  such as that  for  the
acquisition of field data. As detailed in USEPA
(1987a), field sampling Data Quality Objectives
(DQOs)  establish  a  phased sampling strategy
designed to guide  the efficient acquisition of field data
for  site-specific exposure and public  health
assessments, and provide sampling  plan guidance
addressing the  location of sampling  points.  Field
operating procedures for obtaining  and handling
samples have also been  developed  (USEPA  1987b).
Other references,  (USEPA 1986a,  1986b, 1987c, and
1987d),  address the  utility,  applications, and
limitations  of  computer models for predicting
contaminant concentrations  and  transport through
various environmental  media.  The process  for

-------
developing  a human  health  risk assessment for
Superfund sites has been detailed in USEPA (1985c).

When conducting a comprehensive risk assessment,
the analyst will need to refer to all of the above-cited
guidance. While  none of these guidance  manuals
stands alone, taken as  a whole, they provide an
overall,  integrated approach to  analysis of site
contamination and health risk.
1.4 Use of the Manual

This  manual is  used to  apply state-of-the-art
exposure assessment  procedures  to  the unique
analytical needs of uncontrolled hazardous waste
sites. The ultimate  goal  of  human  exposure
assessment at Super-fund sites is the determination of
the type and magnitude  of potential exposure  to
contaminants  present at and migrating from the site.
To achieve this goal, many sites may require a mix of
qualitative  and quantitative  exposure analysis. The
latter may  range from simple analytical  techniques
(e.g., contaminant release  or dispersion  estimation
equations)  to  more complicated  computer  modeling
approaches.

The  general procedure for  conducting an  integrated
exposure  analysis  is illustrated in  Figure l-l.  This
procedure is based on EPA's published Guidelines for
Exposure Assessment  (USEPA  1986c)  and  other
related  guidance  (USEPA  1985d-i)  and  is an
adaptation of that process to the analytical problems
posed by  abandoned hazardous waste  sites. As
previously mentioned, target chemicals are selected
as part of the  human health  risk assessment process
(see  USEPA  1985c). Once these  chemicals are
chosen, the exposure assessment proceeds through
the following stages:
  . Contaminant Release Analysis
   Each on-site release point is  identified for every
   target  chemical, and the level of release (mass
   loading)  to  each  environmental  medium  is
   determined.  Determination of contaminant
   release  may  be  made  either  by   direct
   measurement (monitoring)  of such  releases  or
   by estimation. Although difficult to achieve for  all
   media, monitored release values provide a more
   sound  basis for projection of contaminant
   migration  later in  the  exposure assessment
   process than do  modeled estimates. When it is
   not  possible to obtain measured release rates,
   estimates  can  be based  on measurements  of
   contaminant  concentrations  in  pertinent  source
   media (e.g., estimates of contaminant release to
   ground water based on measured concentrations
   in  contaminated soil).  The  results  of the
                                                       Figure  1-1.
               Overview  of the integrated  exposure
               assessment  process.
                    Evaluation of Contaminant
                    Properties and Selection of
                       Target Chemicals1
                       Contaminant Release
                       Analysis (Multimedia)
                       — Monitoring Data
                      —  Modeling Estimates

                           (Chapter 2)
               Contaminant Transport and Fate Analysis
                      — Exposure Pathways
                   — Environmental Distribution
                       and Concentrations
                       • Monitoring Data
                      • Modeling Estimates

                           (Chapter 3)
                       Exposed Populations
                            Analysis

                          (Appendix A)
                        Integrated Exposure
                            Analysis

                          (Appendix A)
                        Uncertainty Analysis

                           (Chapter 4)
  'Part of Human Health
  Risk Assessment Process.
  Refer to USEPA (1985c).

  contaminant release analysis provide  the  basis
  for evaluating the potential  for  contaminant
  transport,  transformation,  and  environmental
  fate.
2.  Contaminant Transport and Fate Analysis
  This analysis describes  the  extent  and
  magnitude  of environmental contamination  (i.e.,
  contaminant concentrations  in  specific
  environmental  media).  When  possible, direct
  measurement of  contaminant  concentrations  is
  preferred, and collection  of samples  during site
  evaluation  will  provide  a clear  basis for
  determining exposure potential  for some
  exposure routes.  However, the human health
  risk  assessment process  also requires projection
  of potential exposure over a lifetime (see Section
   1-5), which  can only be  accomplished using
  estimation procedures.

-------
3.  Exposed  Populations Analysis
  The results of contaminant transport  and fate
  analysis allow the analyst to evaluate populations
  contacting  chemicals emanating from  the site.
  Analysis of exposed populations  involves  the
  identification, enumeration, and  characterization
  of  those population  segments  likely to be
  exposed. The goal of this analysis  is not only to
  delineate those populations coming into  contact
  with contaminants emanating from the site,  but
  also to determine  how and with what frequency
  and duration such  contact occurs.

4.  Integrated  Exposure Analysis
  In  this step,  the  individual  chemical-specific
  exposure estimates for  each  exposure route
  (i.e.,  inhalation,  ingestion, and dermal  contact)
  are developed.  For each exposed population, all
  exposures to  each hazardous substance  are
  identified. In cases in which a  population group
  experiences more  than one exposure by a given
  route, exposures are  summed to  develop  a
  cumulative exposure value  for  the  route
  involved.  For  example, persons who  reside in
  the vicinity of  a Superfund site may experience
  dermal exposure to a given contaminant directly
  on  site as  well as  directly through  basement
  seepage, and exposures via both of these routes
  should be  summed for  exposure  integration
  purposes.

5.  Uncertainty Analysis
  The exposure  assessment concludes  with an
  analysis of uncertainty.  In  this analysis each step
  in  the assessment is  reviewed to identify  any
  uncertainties  involved and  to evaluate their
  separate and cumulative  impact on assessment
  results. Uncertainties may result from the use of
  default values  for analytical  input parameters,
  from the use of simplified estimation procedures
  as  opposed to  more rigorous computer analysis
  or  monitoring-based analysis, from  an  inability
  to  define exposed populations  with  confidence,
  etc. The uncertainty analysis provides necessary
  input  for remedial  decisionmakers who must
  evaluate the results of the exposure assessment
  with regard to their implications for  potential risks
  associated with  the  uncontrolled site and
  appropriate  remedial alternative selection

This  manual  is intended  to be  used in  conjunction
with various other guidance to conduct Superfund site
Remedial Investigations and Feasibility Studies.  The
use of this manual is particularly linked to the  Public
Health Evaluation Manual. The two are intended to be
used  as  two parts of the same  process: the analysis
of health  impacts  resulting  from  uncontrolled
hazardous  waste sites.  In  conducting a  Superfund
evaluation of exposure and  public  health impact, the
analyst initially applies the indicator chemical selection
process"  outlined in the Superfund Public Health
Evaluation  Manual to select the chemicals on which
the site analyses will focus. Once the chemicals have
been selected, the  analytical  framework of the
Superfund  Exposure Assessment Manual  is applied.
Following  completion  of the exposure  assessment,
the analyst returns to the Superfund  Public  Health
Evaluation  Manual  for guidance  in  determining the
degree of  human health risk for each exposed
population.

The user of this manual should  understand that these
analytical  procedures are intended  to  be applied
site-specifically. No two sites will be exactly alike  in
terms of the extent and complexity of contamination,
of  contaminant migration, or of  potentially  exposed
populations. Therefore,  the specific  analytical
procedures  to be  applied in  all Superfund exposure
assessments cannot be fixed in general. Instead the
approach  and methods applied  to conducting  an
exposure  assessment must  be tailored to address
existing  site conditions. In  some  situations
contaminant releases or exposure  routes  may  be
adequately addressed  by applying only screening
procedures.  In  other cases  more  complex,
quantitative evaluation will be  necessary.

The Superfund Public Health Evaluation Manual
(USEPA 1985c) lists five factors affecting the degree
of analytical complexity for site analyses:

1.  Number and identity of chemicals present;

2.  Availability of appropriate standards and/or toxicity
    data;

3.  Number and complexity  of exposure  pathways
    (including  complexity  of release  sources  and
    transport media);

4.  Necessity for precision of the results; and

5.  Quality and quantity of available monitoring data.

Simplified  analyses may be used in the following
instances:  only a small number of chemicals must be
evaluated;   environmental  standards or criteria  for
chemicals under study are available; a small number
of  exposure  pathways  are present; release  and
transport processes are relatively simple: or there is a
limited  need for detail  and  precision  in  the
assessment  results (e.g.,   screening  studies).
Conversely, sites  that have  many contaminants for
which  no  environmental  standards or  criteria are
available,  that exhibit multiple exposure pathways,
that have complex contaminant release and  transport
processes  in  effect,  or that require analytical  results
in  great detail and  precision  will require more
  -.Selection of indicator chemicals will be required only at those
  sites where the number of contaminants present is too large to
  individually evaluate exposure to each.

-------
complex, quantitative analytical  methods. Most  sites
will fall somewhere between these two extremes.
    obtain and  review the original source
    documentation  cited for analytical components.
Procedures presented in this manual  for conducting
quantitative  analyses include  both  simplified  "desk
top"  approaches  for  developing  order-of-magnitude
estimates and more resource-intensive,  in-depth
approaches. Computer modeling and site monitoring
are included.  Generally, it  is appropriate  to  apply
simplified analysis to all pertinent  exposure  routes at
the beginning of quantitative evaluations so that those
causing  greatest concern  can  be  identified  for
subsequent in-depth  analysis.

It is important to understand that analysis of  exposure
and  resultant  health  impact is often a  complex
process  in which  selection  and  application  of  the
most  appropriate  analytical tools,  as well as  the
insightful interpretation of their results,  can be critical.
The  U.S. EPA encourages  ongoing  communication
between site analysts and experts in various  exposure
and  health impact  assessment  fields. Thus,  when
questions  arise regarding the utility of a  particular
model  or  mathematical  solution, it is  recommended
that the  analyst review the pertinent  sections
described  in  this  manual  or  contact the Toxics
Integration Branch of the Hazardous Site Evaluation
Division of the Office  of Emergency  and  Remedial
Response (FTS 475-9486).  In  addition,  Table 1-1
lists  specific EPA  contacts  who can  provide  insight
into particular site  assessment problems.

In developing this  manual, an  attempt was  made to
compile analytical methods appropriate for assessing
exposure  to chemicals  migrating from  uncontrolled
hazardous waste facilities. There are limitations  to the
application  of these analytical  tools  and to the
interpretation of the results obtained,  including:
1. While some of these tools  have been developed
    specifically for  application  to  Superfund  sites,
    others  were  originally developed  for  different
    purposes  and  had to be  adapted or directly
    applied  to  evaluation of conditions  present  at
    uncontrolled  hazardous waste sites.  The analyst
    must  be  careful  in  interpreting  the  results
    obtained from application of these tools and must
    consider their inherent  uncertainties.

2. This manual  assumes that the  analyst  has a
    strong technical  background in engineering or the
    sciences. This background  is essential to ensure
    that analyses are carried  out in a  technically
    sound  fashion  and that interpretations of the
    results obtained  are realistic.

3. It was not possible  to include discussion  of  all
    technical limitations and caveats pertaining  to
    each analytical tool or procedure reviewed in this
    manual. It may  be  beneficial for the analyst to
4. Results obtained  through application  of  these
    tools must be interpreted based on conditions at
    the site being analyzed. These  tools are provided
    to aid the analyst in making  decisions, not to
    make  decisions for the  analyst.  When  possible,
    models used in analyzing a given site should  be
    verified with  field  monitoring data that  test and
    validate model predictions at that site.

5.  The  approach  to  conducting  exposure
    assessments  outlined in  this  manual  is
    conservative  as are  human  health risk studies.
    However, the analyst needs to be sensitive to and
    to compensate, at  least qualitatively, for the
    additive  effect  of  multiple  conservation
    assumptions. The  degree of conservatism should
    not  be so extreme  that  the conclusions drawn
    from the analysis are unrealistic.

1.5 Timeframe  of  Analysis

Quantitative exposure  assessments generate
estimates  of the long-term (chronic daily  intake) and
short-term (subchronic  daily intake) exposure  to
contaminants.   The  output of  each  analytical
component (contaminant release, environmental fate,
etc.) must be expressed  in the same long-term and
short-term form.  Long-term releases are  defined  as
the release rates of each contaminant migrating from
the site averaged  over an assumed 70-year human
lifetime.  Short-term  contaminant  releases are
defined  (USEPA 1985c) as those  that occur over a
short period (usually 10 to 90 days) during the first
year following site investigation.

1.6 Analysis of Exposure Associated with
Remedial Actions

The analytical tools  presented in this  Superfund
Exposure  Assessment Manual are  those  appropriate
for  analyzing  exposure associated  with the  baseline
condition (i.e.,  the  uncontrolled  site  prior  to
implementation of any remedial action). It should  be
noted, however, that waste treatment processes used
as  part of a remediation strategy can themselves
contribute  significant releases of  contaminants to the
environment.  Stripping  volatiles from wastewaters,  for
example,  generally  involves  artificial acceleration  of
the natural volatilization  process,  resulting  in forced
transfer  of the volatile contaminants from  wastewater
to air. Thus,  analysts  must evaluate  the  engineering
design of  each remedial alternative to determine the
level of contaminant release  associated with  its
implementation. The user of this  manual should refer
to Farino et al. (1983)  for  a discussion of methods to
estimate  wastewater treatment air  emissions.  When
incinerating toxic wastes other than those containing
PCBs, Destruction  and  Removal  Efficiency (ORE)

-------
   Table 1-1.   Technical Resource Contacts for Superfund Exposure Assessments
                                 Office
                 Commercial
                 phone number
                                                                                         FTS phone number
   I.   U.S. Environmental Protection Agency:

   Office of Air Quality Planning and Standards; Research Triangle Park, N.C.            (919) 541-5381         629-5381

   Office of Toxic Substances; Washington, D.C.                                  (202)382-3886         382-3886

   Office of Research and Development, Exposure Assessment Group; Washington, D.C.   (202) 475-8919         475-8919

   Office of Research and Development, Hazardous Waste Engineering Research
   Laboratory; Cincinnati, Ohio                                                (513) 569-7418         684-7418

   Environmental Research Laboratory; Ada, Okla.                                (405) 332-8800         743-2011

   Environmental Research Laboratory; Athens, Ga.                                (404) 546-3134         250-3134

   Center for Exposure Assessment Modeling; Athens, Ga.                           (404) 546-3585         250-3546

   II.   Centers for Disease Control:

   Agency for Toxic Substances and Disease Registry; Atlanta, Ga.                    (404) 454-4593         236-4593

   III.   International Ground Water Modeling Center:

   Holcomb Research Institute, Butler University; Indianapolis, Ind.                     (317) 283-9458
requirements  can  be  found in  40  CFR  264.343
(Environmental Protection  Agency Regulations for
Owners and Operators of Permitted Hazardous Waste
Facilities; Subpart  0 - incinerators).  For incineration
of wastes  contaminated with PCBs, the analyst can
refer  to  40  CFR 761.70  (Polychlorinated  Biphenyls
(PCBs)  Manufacturing, Processing, Distribution  in
Commerce, and Use Prohibitions - Incineration).

Well  engineered  remedial  alternatives  planned for
uncontrolled  hazardous waste sites are not expected
in themselves  to  cause additional releases of toxic
contaminants  to  ground-water  systems. Even  if an
unexpected  spill  of toxics  occurs  when  remedial
action is taken, contaminant release should be  slow
enough  to  allow  spilled substances to be isolated
prior  to  their  reaching  the  saturated  zone.  Short-
term  release of contaminants to air may occur while
excavating contaminated  soil and loading it for
removal  from the  site. In such  situations, the analyst
should refer to USEPA (1983a), for release  equations
for  material  transfer.

The  effectiveness  of  contaminant  control,  however,
may  vary among different remediation  technologies.
To  evaluate post-remediation  control  effectiveness,
many of the analytical procedures presented  in this
manual  may be useful. For example,  reductions  in
contaminant  releases  can be  estimated by
recalculating releases  using  altered (from the baseline
case) site-specific input variables based  on the
remedial action under consideration. Alternatively, one
can obtain a  rougher approximation  by  applying the
expected remedial action  percent control  (based on
engineering experience)  to  the  source release
estimates calculated for  the baseline  case.  In
addition, the analyst should  refer to USEPA (1985j)
for  a detailed discussion of both simplified analytical
methods and numerical modeling approaches that can
be  used to estimate remedial effectiveness.

1.7 Organization of the  Manual

The following chapters of this manual detail methods
for  evaluating exposure to chemicals migrating  from
Superfund sites. The body of  the  manual provides
guidance for the qualitative  and quantitative evaluation
of  contaminant  release,  migration,  and  fate  in the
environment, along with that for  evaluating uncertainty
in the  analysis.  Procedures  for conducting exposed
populations analysis and  for developing an  integrated
exposure  analysis are provided  in appendices to this
report.

-------
Preceeding Page Blank
                                                 Chapter 2
                                    Contaminant Release Analysis
    2.1 Introduction

    This  chapter  provides guidance  for the  analysis of
    contaminant  releases from  uncontrolled hazardous
    waste sites. The goal of this analysis is to determine
    contaminant  release  rates to specific environmental
    media over time. The following sections address the
    release  of contaminants  to  air,  surface  water,  and
    ground  water  from wastes placed both  above-
    ground and  below-ground. In particular, guidance is
    provided for the evaluation of the following categories
    of contaminant releases:

    1.  Air releases:
       a. Fugitive dust resulting from:
              Wind erosion of contaminated soils
              Vehicular   travel   over   contaminated
              unpaved roadways

       b. Volatilization releases from:
              Covered landfills (with and without internal
              gas generation)
              Spills, leaks, and  landfarming
           - Lagoons

    2.  Surface water releases: contaminated runoff

    3.  Ground-water releases:
       a. Landfilled  solids (lined or unlined)
       b. Landfilled  liquids (lined or  unlined)
       c. Lagoons  (lined or unlined).

    Contaminant  release  analysis is conducted  in  two
    stages  -   screening  of contaminant  release
    mechanisms and quantitative analysis. The screening,
    which is a qualitative evaluation of site conditions,
    identifies each potential  contaminant release source,
    determines the environmental media affected by each
    release,  and  broadly defines the possible extent of
    the release.  The following  section  is designed to
    establish  a consistent  basis for the qualitative
    screening of contaminant release  from site to site.

    Once the  potential sources of  on-site  contaminant
    release  have been screened, those requiring  further
    evaluation  are  quantitatively analyzed. This  may
    involve  the  application  of  a  simplified "desk-top"
    estimation  approach, or a  more in-depth  method
    such  as computerized  modeling or additional  site
    monitoring.  The  goal of this analysis is to  generate
release rate estimates (mass per unit time) for each
source of contaminant release.  Release rate  values
are necessary as input for subsequent environmental
fate  analysis (see  Chapter 3). Individual on-site
releases of each  contaminant are  summed  to
generate an overall,  medium-specific release rate for
each  chemical migrating from the site. Short-term
(worst-case)  release rates are  developed, as are
long-term rates (averaged over 70 years).

The simplified estimation  procedures that follow allow
the analyst to make release approximations based on
chemical- and  site-specific  factors.  However,  these
calculations do  not take into account the full range  of
variables that  affect on-site  contaminant release.
These approaches  (with  one  exception) assume
steady state conditions. They do  not directly address
the reduction  in contaminants present  (due  to release
losses), or the associated reduction in  release loading
over  time corresponding with the  decreasing
contaminant reservoir.*

When possible, monitoring should be  used to quantify
rates  of contaminant  release. In  some cases,
however, this may not be feasible because methods
to directly measure releases from certain settings are
still being developed. Moreover, it may not  always be
possible  to monitor  contaminant  releases  under the
conditions of  concern (e.g.,  dust  releases under high
wind conditions, surface  water runoff  releases during
storm  events,  etc.). It may often be necessary  to
estimate  release  rates in the exposure  assessment
process.  All of the release rate estimation procedures
presented here,  however,  require some  monitored
values as input. (Examples are measured contaminant
concentrations  in  soil,  soil  characteristics.)  The
analyst should  be aware of the  need to  develop a
monitored data base that is adequate to support the
needs  of the contaminant release analysis  portion  of
the exposure  assessment.

In general, the procedures to  estimate  the rate  of
contaminant release are complete. When  analyzing
   Estimation of the variation in the level of release over time is
  calculated  separately.  See  Long-Term  and  Short-Term
  Release calculation subsections in this chapter.

-------
wind  erosion  releases,  however, the  analyst  should
consult other  published guidance that addresses the
application of the  wind  erosion  equation  in  various
regions of the country.  Depending on the location of a
particular site, one of the following three manuals will
be necessary:

        Craig and Turelle (1964): Great Plains
    -  Haynes  (1966):  Northeast
    -  Skidmore and  Woodruff (1968):  entire
        United States.

2.2 Contaminant Release Screening

The manner of waste placement at an  abandoned site
determines whether  contaminant release*  occurs  by
any or  all of  the mechanisms summarized in Table
2-1. In  contaminant release screening, the likelihood
of release  from each source, the  nature  of the
contaminants involved,  and the probable magnitude of
their  release  (relative  to  other  on-site sources)  are
considered.

Figures  2-1 and  2-2 present the decision networks
that  guide contaminant  release  screening analysis.
Figure  2-1 deals with  contaminants in or under the
soil   and  Figure 2-2  addresses  above-ground
wastes. Any  release mechanisms  evident  at the site
will  require  a  further screening   evaluation  to
determine  the  likely  environmental  fate  of the
contaminants involved (see Chapter 3).

2.2.7  Contaminants in Soil (see Figure 2-1)
The following  numbered paragraphs help to interpret
and  apply the  steps  of  the contaminant  release
decision  network presented  in Figure 2-1. Each
paragraph refers to a particular numbered  box in the
figure.

1. Most  uncontrolled  hazardous  waste  sites  will
    exhibit some degree of surface  or  subsurface soil
    contamination. This contamination  may  be the
                 result  of  intentional  waste disposal underground
                 (landfilling) or in surface soils (surface application
                 or  landfarming),  or  it  may  be caused  by
                 unintentional waste releases from spills or leaks.

             2. Landfilled wastes may become mobile  if they are
                 not contained in impervious  containers, or if the
                 containers are  leaking.  Release  of such  wastes
                 may contaminate subsoils, ground water (through
                 percolation),  or air (through volatilization).

             3.  Landfilled wastes  will  be  covered  with  soil;
                 however,  soil  cover will  not necessarily  isolate
                 wastes from  the environment. If the cover  can be
                 penetrated by  rainwater  or  run-on, wastes  can
                 be leached from the  landfill cells  and contaminate
                 subsoils,  ultimately reaching  ground  water.
                 Similarly,  the soil cover  may  not  be deep enough
                 to prevent the  migration of volatile contaminants
                 into  the  atmosphere.  Estimations are that  60
                 percent of hazardous waste is in  liquid (sludge)
                 form  (USEPA  1980a).  Infiltrating  rainwater  can
                 increase the migration rate of liquid or semiliquid
                 materials by increasing the  hydraulic head
                 affecting them, as well as by the  leaching of toxic
                 components. Such factors  as erosion  or extreme
                 drying (and  cracking) can  reduce the  ability of a
                 soil  or clay  cover  to  maintain  the isolation of
                 wastes. Also,  contaminated  soil  may  cover the
                 waste  cells  themselves.  When  evaluating the
                 potential for landfill  releases, current  conditions,
                 along  with the  long-term integrity  of  the  landfill
                 and  its soil  cover,  should  be evaluated.  If the
                 landfill soil  cover  does  not assure  long-term
                 For the  purposes  of this  manual,  contaminant  "release" is
               defined as any process that results in migration of contaminants
               across the site  boundary. Within this context, volatilization,
               generation of surface runoff, or leachate, are considered to be
               release mechanisms. Contaminant transport equates with those
               processes that  carry  released contaminants  to points  distant
               from the site.
              Table 2-1.    Potential Contaminant Release1 Mechanisms
                         Process
       Media directly affected
      (media indirectly affected)
                                                                                 Timeframe
               Volatilization

               Overland flow2

               Direct discharges

               Leachate generation4

               Fugitive dust generation5

               Generation of surface runoff

               Combustion3
Air

Soils, surface water (ground water)

Soils, surface water (ground water)

Soils, ground water

Air

Soils, surface water (ground water)

Air
Chronic

Chronic, episodic

Chronic, episodic

Chronic

Chronic, episodic

Chronic, episodic

Episodic
               1See Section 2.2 for a definition of contaminant "release" as used in this manual.
               Impoundment overflow/failure, drum leakage, etc.
               Includes on-site treatment releases (e.g., wastewater/runoff treatment, incineration).
               4Buried wastes, wastes stored above ground (leaks), land application, lagoons.
               Contaminated soils,  particulate wastes.

-------
      Figure  2-1.      Contaminant release decision network: contaminants in soil.
                                                                                                         Are Toxics Present In:
                                                                                                                 Soil?
t
Are Toxics La ndfilled?
15

*
Are Toxics Spilled, Leaked,
or Surface Applied? n


t
Is Site Accessible?
E
                        Does Soil Cover
                       Prevent Percolation
                        of Precipitation? f
CD
         Consider Long-
         Term Integrity
          of Soil Cover  [3]
                 Is Leaching
             Release to Subsurface
             Soils or Ground Water
                   Possible?
         Consider Long
         Term Integrity
          of Soil Cover Ig
     Is Leakage of
 Containerized Liquid
Waste to Ground Water
       Possible?
Is Volatilziation"
Release to Air
  Possible?
   Is Runoff Release
   to Soils, Surface
Water, Ground Water or
 Air (Via Volatilization)
      Possible?      |
Does Soil Cover
Prevent Vapor
Release to Air? R
.
Is Surface
Soil
Contaminated? R
_
E

Is Soil'
Cover
roding? fj
Is Release to
Ground Water
(Leaching) Possible?


Is Releas
or Surfa
(Runoff)

a to Soils
:e Water
Possible? [5
Is Fugitive Dust
Release to Air
Possible? |e


Is Volatilization
Release to Air
Possible?


                                                                                       Go on to Environmental
                                                                                    Fate Analysis for Contaminants
                                                                                      Released Via Each Existing
                                                                                    or potential Release Mechanism

-------
    isolation of the wastes, one should evaluate  the
    leachability  and  volatilization  potential of  the
    landfilled wastes.

4. At some hazardous wastes sites, toxic  materials
    may  have  been  purposefully  incorporated into
    surface  soils to promote  their  microbial
    destruction. In such cases, toxic components may
    still  remain in  the soil. At most  sites,  however,
    surface soils have become contaminated because
    of hazardous  material  spills  or leaks during
    manufacturing,  processing,  storage,  or  transfer
    operations.  In these situations, the potential  for
    release of contaminants  in surface  soils through
    four  mechanisms should be  evaluated. These
    mechanisms are:  (1) release of  volatile
    components to the atmosphere  (via evaporation);
    (2) release of  toxic  particulate matter (via wind
    erosion); (3) surface  runoff -related releases; and
    (4)  percolation of contaminants  or  leachate  to
    ground water.

5. The percolation  of contaminated runoff may
    contaminate surface  soils and  underlying ground
    water. Surface water  systems  may  be similarly
    degraded by contaminated runoff inflow. Runoff
    may  also  serve  as a source of volatilization
    release  to air,  although  releases  from
    contaminated soils  would  be  expected  to
    constitute  a greater threat  than  that  from
    contaminated runoff.  Hydrophobic  wastes may
    contaminate surface  waterbodies by adsorbing
    onto soil material that can be eroded from the site
    and enter a waterbody  in surface runoff. In a
    waterbody,  sediment  transport is much slower
    than  is  water movement, and  contaminated
    sediments  may  remain in the  vicinity of the
    contamination source for a long  time.

6. Under high  wind  conditions,  wind  erosion may
    carry solid  particulate wastes or soil particles with
    sorbed hydrophobic toxic materials from the site.

7. If the site  is  accessible, direct contact with
    contaminants  may  occur.  Also children  may
    ingest some contaminated soil  during  play.  Such
    ingestion  may  result  from "pica"  behavior (i.e.,
    intentional  eating  of soil  by very young  children)
    or from normal hand to mouth contact.
2.2.2 Contaminants Above-Ground
The following numbered paragraphs help to interpret
and  apply the steps  of  the  contaminant  release
decision network  presented in Figure 2-2. Each
paragraph refers to  a particular numbered box in the
figure.

1. Wastes  can  be  stored  above-ground  in
   lagoons/ponds,  in containers (drums, tanks), or in
   piles. Unless containers effectively isolate wastes
    from the environment, above-ground storage
    can  result  in  the direct introduction  of
    contaminants into  air,  soils, surface water,  or
    ground water.

2.   Lagoons  may  introduce hazardous  materials  to
    the environment  through a number of pathways.
    Erosion  or  overflow resulting from heavy  rainfall
    can breach  the lagoon and result in the outflow of
    liquid wastes  that  contaminate surface soils,
    ground  water,   and surface  waterbodies.  In
    addition,  unlined lagoons  may introduce toxics
    directly into ground water via percolation through
    the  lagoon  bottom.  Also,  because  lagoons are
    uncovered,  the release  of  volatile  compounds  to
    the atmosphere is a common  problem.

3.  Wastes stored  above-ground in containers  may
    not be effectively isolated  from the environment.
    Over time, container corrosion and leakage occur.
    Leaked  wastes  will contaminate  soils  in the
    storage area; may  percolate  to ground  water;  or
    may  contaminate surface  runoff, which, in turn,
    can extend  the area of  soil contamination  or can
    enter local surface waterbodies. Leaked  materials
    may also evaporate into  the atmosphere.

4.  If  the  site  is accessible  to the  public,  direct
    contact  with contaminants  may occur. Also,
    children may ingest contaminated  soils, either
    inadvertently or as a result of  pica behavior.
2.3  Quantitative  Analysis  of Atmos-
pheric Contamination

2.3.1   Fugitive Dust Emission Analysis
Emissions  of  contaminated  fugitive  dusts  (airborne
wastes or  contaminated soil  particles)  originating at
uncontrolled hazardous waste sites can result from a
combination of such factors as (1) wind erosion of
wastes  and  contaminated soils,  and  (2)  vehicles
traveling over contaminated, unpaved roads.

Methods for analyzing such contaminant releases are
presented below.

2.3.1.1 Beginning Quantitative Analysis
The following procedures are useful in estimating total
fugitive dust  releases  likely to result from the  two
factors cited  above.  Once total  suspendible  dust
generation  levels  have been  calculated using  these
equations, one can project the  amounts of hazardous
substances expected  to  enter the  atmosphere in
fugitive dust using either of the  following approaches:

*es Multiply  the amount of dust  generated by  the
    weight  percent of the  toxic substance  in  soil or
    waste.  This approach does not take into account
    factors  relating to such aspects as particle  size or
    adsorption potential, which  can affect the amount
                                                 10

-------
Figure 2-2.     Contaminant release decision network: contaminants above-ground.
                                                                            Are Toxics Present in:
                                                                           Lagoon, Above-Ground
                                                                                  Storage?        ry
                                                                                Yes
                                                                                          No
              Is Lagoon Overflow
                or Breaching
              a Potential Threat?
                           No
        Is Direct Overland Flow
          of Toxics to Soils or
       Surface Water of Indirect
      Percolation to Ground Water
     	Possible?	
Is Volatilization
Release to Air from
Lagoon Possible?
•

•
Yes
i

i
No
[2


     Is Lagoon Lined
      (i.e., Preclude
       Leakage)?
  Is Leakage a
Potential Threat?
  Do Containers
  Isolate Wastes
  (i.e., Preclude
   Leakage)?   [3
                        Is Site
                      Accessible?
                                                                                              L    J.
                                                                                              Yes
    Is Soil,
Ground water,
  or Runoff
Contamination
  Possible?
                                                                                              Yes
                                                                       Is Percolation
                                                                     to Ground Water
                                                                         Possible?
                                                                       Yes
                                                                                                                                       Yes
                                                                                                                                                 No
 Is Gaseous
Release to Air
  Possible?
                                                                                                         No
                                                                                                                   Yes
Consider Direct
 Contact with
 Contaminants
                                                                                                                             No
                                                                   Go on to Environmental
                                                               Fate Analysis for Contaminants
                                                                 Released Via Each Existing
                                                               or Potential Release Mechanism

-------
    of  contaminant  actually entering  the  atmosphere
    as  dust.

    Multiply the estimates for total dust generation by
    percentages  (by  weight)  of the  substances of
    concern  in actual fugitive dust samples  obtained
    with on-site  air monitoring. This  approach takes
    into account those chemical-specific and site-
    specific  factors that affect release  of
    contaminated dust in the field.
(1) Wind Erosion Analysis*
Wind  erosion  of agricultural  soils  and,  by
extrapolation, other disturbed soils,  depends  upon a
variety of factors.  These  include surface roughness
and  cloddiness;  surface soil  moisture  content,  kind,
amount (and orientation, if applicable) of vegetative
cover:  wind  velocity; and the amount of  soil  surface
(length) exposed to the  eroding wind force. The U.S.
Soil  Conservation  Service  (SCS)  has developed a
method to estimate wind erosion  based on a series of
graphs relating  variables  presented  below. The
graphical method for calculating  wind erosion based
on the functional relationship of these variables is  not
presented in this manual;  instead,  the  analyst is
directed to the Skidmore and  Woodruff (1968) source
document.
  E = f(l', C', K', L', V)

where
(2-1)
E  =   potential annual  average wind  erosion  soil
       loss,  (mass/area/time).
       soil credibility index, (dimensionless).
       climatic factor, (dimensionless).
K' =   soil ridge roughness factor, (dimensionless).
L' =   field length along the prevailing wind direction,
       (feet).
v  =   vegetative cover factor, (dimensionless).

Multiplying  E times the contaminated area will yield a
release rate in units of mass per time.
Table 2-2  identifies  the factors that determine the
values of the five variables used in the SCS equation.
Note that the vegetative cover factor (V)  specifically
applies  to  crop  residues, and  care must  be taken
when extrapolating to the cover conditions present  at
uncontrolled waste sites. For Remedial Investigation
and  Feasibility  Study estimation purposes, one can
use a "zero pounds per acre" vegetative cover value.
This  assumes  a worst-case situation (from  a
vegetation-related wind  attenuation  perspective)  and
   Applied to  nonadhering, noncompacted contaminated soil or
  waste materials.
provides  a  conservative  estimate of contaminated
fugitive dust release.

A series  of publications issued  by the  U.S.
Department  of  Agriculture provides  directions  for
applying this equation to a  site-specific situation.
Craig and  Turelle (1964)  present  estimation
procedures  for  the Great Plains;  Haynes (1966)
addresses the Northeast; and Skidmore and Woodruff
(1968) offer procedures for the entire nation.

Although it  is  strongly  recommended that  site-
specific soils data  be  obtained  for  each site under
evaluation, it is  not necessary  to  do so  in  order to
obtain parameter values for use with  the wind erosion
equation (or other fugitive  dust generation equations).
Instead, when necessary,  soils data  can  be  obtained
from the local Soil  Conservation Service office. SCS
has on record a  range  of pertinent soils data for sites
across the  country where soil  surveys  have  been
conducted. In addition,  SCS maintains  an extensive
computerized soil  properties  data  base called  the
Soils  5 File. This data base lists  estimated soils data,
based on surveys of surrounding soils properties, for
areas where surveys  have not been conducted to
date.  These data are readily available from local SCS
officials. Users of this manual should consult SCS to
obtain more detailed information  regarding the nature
and accessibility of  information contained in  the  soil
surveys and the Soils 5 File.

The SCS wind erosion  equation is one of a number of
approaches  for  estimating  particulate release from
abandoned  hazardous  materials facilities. One such
source (Cowherd et al. 1985) is  specifically designed
to guide  rapid (less  than 24 hours) evaluation of the
potential degree  of  particulate  emission  from
uncontrolled hazardous  waste  sites. This  method
uses an emission factor approach to  estimate release
and   procedures  adapted  from  computerized
dispersion models  for approximating concentration
isopleths.  Concentration estimates  and Bureau of the
Census  data are  used  to identify the exposed
population and  estimate the level  of exposure. This
approach includes the three  key  components  of
exposure  analysis:  release rate estimation,
contaminant migration  analysis,  and   population
exposure determination.  However,  Cowherd et  al.
(1985)  caution  that their method  is designed  for
emergency evaluations or  as  a preliminary
assessment tool,  which may  then be used  in
undertaking a more  detailed  investigation.
Nevertheless, the degree  of accuracy attained using
this method is  consistent  with simplified  quantitative
estimation  procedures. This  approach  provides  the
analyst  with estimates of  short-term  (worst-case,
24-hour)  release and exposure estimates, as well as
long-term (average  annual) estimates.*

The  SCS  wind erosion  equation  is designed  to
provide  annual  erosion losses  only,  and cannot be
                                                  12

-------
Table 2-2. Environmental Variables and Model  Parameters for the Wind Erosion Equation
 Equivalent SCS wind erosion equation primary wind erosion variables
                        Parameters
Soil erodibility index, I (function of soil particle size distribution; read
    from a table)
Knoll erodibility, ls (function of knoll slope steepness; read from a
    graph)
Surface crust stability, Fs
Soil ridge roughness, K,., (function of height, width, and spacing of
    clods and furrows)

Annual average wind velocity, v (read from map)
Surface soil  moisture, M (estimated using Thornthwalte's (1931)
    precipitation-evaporation index)
Distance across field, Df (field width in direction of primary erosive
    wind)
Sheltered distance, Db (calculated from barrier height upwind of field)
Quantity  of vegetative cover, R'  (mass of standing or fallen vegetative
    residue per unit area)
Kind of vegetative cover, S (factor related to  erosron-reducing
    effectiveness of residues from different crops)
Orientation of vegetative cover,  KQ (factor relating erosion reduction to
    standing vs. fallen crop residues)	
Soil and knoll erodibility, I' (equal to I x I,)
Disregarded-crust  is transient
Soil ridge roughness factor, K (estimated by comparison to a set of
    standard photographs Included in SCS wind erosion equation
    users' manuals)

Local wind erosion climatic factor, C'  (may be calculated but commonly
    read from maps of C')

Field length, U (the difference between  Df and Db)
Equivalent vegetative cover, V (the product of R', S, and KJ - can
    often be assumed = 0 for abandoned waste sites (see text)
Source: Smith et al. 1982.
reliably altered  to generate short-term estimates.**  In
addition,  it  cannot  be  used  with  data delineating
climatic extremes for  a  given location,  but must be
based on  average  annual  climatic data. Instead, for
exposure assessment  purposes the short-term
release, estimated using the wind erosion equation, is
assumed  to  equal the average release  over the first
year following site investigation.

The user of this manual should review Cowherd  et al.
(1985) and compare that method with the  SCS wind
erosion  procedure  before  selecting  an  analytical
approach  for  estimation  of particulate  contaminant
release and  related exposure.  The analyst can  also
refer to USEPA (1983b),  Farino et  al. (1983) Sehmel
(1980), and  Smith et al.  (1982)  for a  review of other
possible approaches.

As  noted in  USEPA (1983b), the SCS wind equation
computes  the  total  wind  erosion  soil loss resulting
from  the combination of surface creep, saltation, and
suspension.   Although appropriate for studies  of
agricultural soil loss,  in  exposure evaluations the
analyst is  generally concerned  only with that fraction
of the  total soil loss that  consists  of particles  of
suspendible, wind transportable,  and  inhalable  size.
When the wind erosion equation is used to estimate
contaminated fugitive  dust  exposure situations, the
    Note: EPA (1985c) defines short-term concentrations to
  equate with a  10- to 90-day period.  Thus, the 24-hour
  maximum  exposure  may not adequately  represent  subchronic
  exposures.
    Personal communication between  Lee  Schultz (Versar Inc.)
  and Thomas George (US. Soil Conservation Service), July 24,
  1985.
total  soil loss results  obtained from the  wind erosion
equation must be  adjusted (reduced) to reflect only
that  portion of the total  soil loss that is suspendible
and transportable over significant distances by wind.

Considerable  discussion  of the  cut-off  point  for
suspendible soil  particle size  exists  in  the  literature
(Sehmel 1980, Smith  et al.  1982, and USEPA 1983a.
b). As  a group,  particles  <  100  urn   aerodynamic
equivalent diameter include  those  that  can  be
suspended by and transported in the  wind and those
that can be inhaled (see Miller et al. 1979 and USEPA
1986d for a discussion of the extent to  which various
particle sizes  penetrate  the  human  respiratory
system). Particles in the 30  to 100 pm diameter range
will often settle within  a   few hundred  feet of the
source  (USEPA  1983a), while those  particles <  30
pm in  diameter  can be  transported   considerable
distances downwind.  To estimate inhalation  exposure,
only  the inhalable  fraction  of suspended particulates
{<  10 pm in diameter) must  be considered.

For  particles in the  2- to 20-um size range,  the
particle size distribution of the parent soil determines
the size  distribution of suspended particles (Smith  et
al.  1982). Therefore, that proportion which is  < 10 urn
in  diameter  can be determined based on the soil size
distribution of the parent soil.  It can be assumed that
this proportion of the  total soil loss, as  calculated  via
the SCS wind erosion equation,  is lost to suspension
and is available for inhalation.

(2) Unpaved Roads Analysis
The  following  equation (USEPA  1983a)  can be used
to  estimate  fugitive  dust  releases  associated with
vehicles traveling on contaminated unpaved roads.
                                                       13

-------
            / s WSpWW\a7/w\a5/'365-Dp\
   BvT-«">(ii)(»)(7)   d)   (-sr)
                      or in metric form
                           0.7 ,   , 0.5,
                                    (2-2)
             / s \/SpW W\   /w\   /366-Dp\
   E___ = k<1.7)( — )( — )( — 1   I—)    	-
    VT       \12/\48/\2.7/   \4/   V  365  /
where

EVT  =

 k  =
s =

Sp =
w  =
w  =
Dp =
emission factor for vehicular traffic, (Ib/vehicle
mile traveled: kg/vehicle kilometer traveled)
0.45  =  particle  size  multiplier for  particles
<  10  pm (i.e.,  particles  that  may  remain
suspended once  they  become airborne  and
which  can be inhaled into the respiratory
system).*
silt  content  (of  road surface material),
(percent)."
mean vehicle speed, (mph;  kph).
mean vehicle weight, (tons;  Mg).
mean number of wheels.
number of days with at least 0.254 mm (0.01
in)  of precipitation per  year  (see Figure 2-3).
The  emission factor (EVj)  can be  multiplied  by a
"vehicle kilometers  traveled  per  time"  value to
generate  a "dust release  per time value." Short-
term (maximum  release) estimates can be made by
using a  reduced value of  "Dp"  in  the  equation to
reflect assumed  drought conditions at the site. Figure
2-3 reflects the  range  of average  "Dp"  values  for
locations  in  the U.S.  Consultation with  the  local
National  Weather Service office may provide locale-
specific  insight into what "Dp" values should be  used
to represent dry years  at the site.  Long-term
(average) releases  can be estimated  by  using the
annual  average value for "Dp." USEPA (1983a)
states that this  equation  is valid for situations that
comply with the following source conditions:

x   Road surface silt  content = 4.3 - 20  percent;
eses Mean  vehicle weight  =  3-157 tons  (2.7-142
    Mg);
esx Mean  vehicle   speed  =  13-40  mph  (21-64
    kph); and
^^ Mean  number of  wheels =  4-13.

For  an  overview  of the  utility and  limitations
associated  with the application of emission factors to
particulate  release estimation problems,  the user of
  See EPA (1983a) for "k" values used when release of
  specific particle size groups other than < 10 pm is desired.
  "soil silt content can oe estimated from SCS Soils 5 File data
  by subtracting the "percent clay" value from the "percent
  material passing No. 200 sieve"  value. (Personal
  communication between Lee Schultz (Versar Inc.), and Keith
  Young (U.S. Department of Agriculture, Soil Conservation
  Service), Washington, DC,  May 1, 1984.)
this manual can refer to USEPA (1983a, b), Farino et
al. (1983) Sehmel (1980), and Smith et al. (1982).

2.3.1.2 In-Depth Analysis
For  contaminated  fugitive  dust emissions,  in-depth
analysis will  consist  of  monitoring  and  modeling
activities.  Generally,  air sampling  will  be conducted
downwind and  upwind  of the uncontrolled hazardous
waste site. The difference in particulate loading
obtained at the two (or more) sampling locations will
quantify the particulate mass loading  attributable to
the site  alone  (assuming that air sampling stations
can  be sited  to eliminate  interference from  other
sources).  Using these  data,  either simple dispersion
equations or computerized  air dispersion  modeling*
can be  used to back-calculate the emission level at
a  "virtual  point  source."   The use of dispersion
modeling to back calculate  emission  levels, however,
is  often quite  unreliable because  of the difficulty in
obtaining  accurate  ambient  monitoring  and
meteorological input data.

The  virtual point  source is a  hypothetical  source
located  upwind of the subject site that has a
hypothetical release  rate which would  result  in  the
contaminant  concentrations observed at the
uncontrolled hazardous waste site (area source). The
virtual point source release rate can then be used in
subsequent contaminant transport analysis  for  the
subject  site. The user  of this manual should refer to
USEPA (1983c) and  Seely et al. (1983) for a detailed
presentation of ambient air sampling strategies and
procedures appropriate for  abandoned hazardous
waste facilities.

2.3.2  Volatilization Emission Analysis
Volatilization  of  contaminants  at  uncontrolled
hazardous waste  sites can  occur at  the following
sources:

    (1) Covered landfills - without internal gas
       generation;
    (2)  Covered landfills  - with internal gas
       generation;
    (3)  Spills,  leaks,  landfarms - concentrated
       wastes  on  the  surface or adhered to  soil
       particles below  the surface; and
    (4) Lagoons -  wastes dissolved  in  or  mixed
       with water.

In the  baseline situation,  one or  more  of  these
sources  will contribute  to  the overall  air loading
originating  at the site,  and will  need  to  be  controlled
through  remedial action.

2.3.2.1 Beginning Quantitative Analysis
This section presents simplified analytical procedures
for estimating  releases from the above source
categories.  Because the chemical  properties of a
given  substance  largely determine the volatilization
rate,  the  equations  presented  require input  of
                                                  14

-------
Figure 2-3.   Mean number of days per year with >0.01 inches of precipitation (i.e., ' 'wet days") (USDC 1968).
                                                                   /      '--xrJ^i
                        Alaska
                   0   200  400
                    100  300

-------
quantified property values. These  data are available
for many chemicals that  may  be  present  at
uncontrolled hazardous waste  sites, and  are found in
various chemical  reference  texts. In cases where
chemical data are  missing, the analyst must estimate
the property  values.  This  section  provides equations
for estimating certain  requisite chemical  properties.
Comprehensive  guidance for chemical  property
estimation is  provided in reference materials  such  as
Lyman et al.  (1982).  Readily accessible computerized
systems  are  available to predict a range of pertinent
chemical properties.  The computerized Graphic
Exposure Modeling  System  (GEMS), and  its
subsystem  CHEMEST, is an example.  The  EPA
Office  of Toxic Substances in  Washington, D.C. has
developed and is managing this system.  Essentially a
computerized  version of Lyman et al. (1982), it can
be rapidly  accessed  to estimate the  chemical
characteristics necessary for volatilization estimation.

The user of this manual can  refer to Farino  et  al.
(1983) for a detailed  review and evaluation  of existing
equations   for estimating  volatilization  from
uncontrolled  hazardous waste  sites.  This report
presents a survey of available  air  release  models  for
volatile substances  and  a critical  analysis of the
applications and  limitations of each.

(1) Landfills Without  Internal Gas Generation
Equation 2-3 can  be  used  to  estimate  volatile
releases from  covered  landfills containing toxic
materials alone, or  toxic  materials segregated  from
other landfilled nonhazardous wastes. Equations 2-4
through  2-7  are  used  to calculate certain  input
variables that are  required to apply Equation 2-3.
Farmer  et  al.  (1978) developed  an  equation  to
estimate the  effectiveness  of various landfill  cover
types and depths in  controlling volatile releases. This
equation, based on  Fick's First Law of steady  state
diffusion, assumes that diffusion into  the atmosphere
occurs at a plane surface  where  concentrations
remain constant. It  ignores biodegradation, transport
in water, adsorption, and  production  of landfill gas.
Diffusion of the  toxic vapor through the soil cover is
the controlling factor. It also assumes that there is a
sufficient mass of  toxicant in the  landfill so that
depletion of  the  contaminant will  not reduce the
emission rate.

Equation 2-3, simplified  by Farmer et al.  (USEPA
1980b),  incorporates a number of assumptions (see
Farino et al.  1983  for a complete discussion), such as
completely dry soil  (worst  case)  and zero
   Although computerized  dispersion modeling can be used to
   obtain contaminant release  rates, it  is primarily a tool  for
   determining contaminant  atmospheric fate. Thus, refer to
   Chapter 3, Environmental  Fate Analysis,  for detailed
   discussions of  air dispersion models applicable to uncontrolled
   hazardous waste facilities.
concentration  of volatilizing  material  at the  soil
surface.  Shen  (1981) converted Farmer's  simplified
equation for calculating  the vapor flux  rate to a form
that provides  a toxic vapor emission  rate  by
multiplying the  basic equation by  the  exposed
contaminated surface area. In addition, Shen modified
the equation to  allow calculation of the  volatilization
rate of a  specific component  of  the overall  toxic
mixture by  multiplying by the weight  fraction of the
component  in the mixture. However, as  pointed out
by  Farino  et al.  (1983), a  more accurate  approach
would  be to multiply by the  mole fraction  of the toxic
component in  the buried mixture. Thus,  Farmer's
equation, as modified by  Shen (1981) and  Farino et
al.  (1983) is:
                     M.
     . = D.C.A( ?*•>) —
     1     1  SI  \  t  / A
                     asc
                                        (2-3)
where
A
p
 sc
    =   emission rate of component i, (g/sec).
    =  diffusion coefficient of component i  in  air,
        (cm2/sec).
    =   saturation  vapor concentration of component
        i,  (g/cm3).
     exposed  area, (cm2).
    =  total  soil porosity,  (dimensionless).
    =   mole fraction of toxic  component i  in the
        waste,  (gmole/gmole).
    =   effective depth of soil cover,  (cm).
Note that total soil porosity,  rather than air-filled soil
porosity, is  used  in this  equation. The presence  of
water in a soil cover will tend to decrease the flux rate
of a volatile  compound by effectively decreasing the
porosity,  and also  by  increasing  the  geometric
complexity of the soil  pore  system  (because water
adheres to  soil particles), thus  effectively increasing
the  vapor  path  (USEPA  1980b).  Farmer  et al.
suggest, however, that when using  their equation  to
design  a  landfill cover,  the total porosity value  be
used (USEPA 1980b),  thereby designing for the worst
case (i.e., dry conditions). In  most instances, it  will be
appropriate  to apply  this same  worst-case logic  to
the  analysis of volatilization  release  from  landfilled
wastes, assume that  landfill  cover soils are dry, and
use  a value for  total  porosity in  Equation 2-3. It  is
recognized,  however, that  there  may  be  situations
where it can be shown that cover soils  exist in a wet
condition more often  than  in a  dry  one. In  these
cases,  the air-filled soil porosity  (Pa)  may  be more
appropriate,  and  this  value  can be substituted for  Pt
in Equation 2-3 when  analyzing volatilization release.

If not provided in existing literature, D,,  a compound's
diffusion coefficient (required for the above equation),
can  be calculated  by Fuller's  Method  (Perry  and
Chilton 1973):
                                                   16

-------
        0.001T
               1.75
                    MW.   MW
   D.=

where
          T
    MWi;MWa


         Pa
  EVi;SVa
                          )1/3]2
                                          (2-4)
                  absolute temperature, (°K).
                  molecular weights of toxic
                  substance and air (28.8),
                  respectively, (g/mole).
                  absolute pressure, (atm).
                  molecular  diffusion  volumes of
                  toxic  substance and  air  (20.1).
                  This is the sum of the  atomic
                  diffusion  volumes  of   the
                  compound    components,
                  (cm3/mole).
To  estimate short-term  (maximum)  release  rates,
use  a value for the temperature that reflects the
expected  summer maximum temperatures. Annual
average  temperatures should  be  used  to  initially
estimate  long-term  (average)  release rates. This
initial  estimated long-term  release value will  be
revised as described in Section  2.3.3 to develop final
long-term  release  estimates.

Relevant atomic diffusion volumes for use  in
estimating D, are (Perry and Chilton 1973):
                        Aromatic ring      = -20.2
                        Heterocyclic ring = - 20.2
 C =  16.5 Cl =  19.5
 H =  1.98 Br =  35.0
 O  = 5.48 F =  25.0*
 N =  5.69 S  =  17.0
Table 2-3  presents  diffusion  coefficients  that  have
been  calculated for a variety of compounds, some of
which may be present at abandoned sites.

An alternative method (Shen 1981) for approximating
DJ involves the identification of a compound listed in
Table 2-3 that has a molecular weight and  molecular
diffusion  volume  (calculated)  similar  to those of the
toxic  substance  under  evaluation.  The  unknown
diffusion coefficient can then be calculated using:
    D. = D'f
          VMW.
                                           (2-5)
where
    D,  =  diffusion coefficient of the compound to
           be estimated from the known D'.
    D'  =  diffusion coefficient of a compound that
           can be found in the table, the molecular
  ' This value is from Shen (1981).
                                                         MW
                                                         MW,
           weight and atomic diffusion, volume of
           which are  close to that of the unknown.
           =  molecular weight  of  the selected
               compound D'.
           =  molecular  weight of the compound to
               be estimated.
Total  soil  porosity, Pt,  can be  calculated  as follows
(USEPA 1980b):
                                          (2-6)
                                                     where

                                                         Pt
                                                         8
                            (dimensionless).
                                _3\.
= total soil porosity,
=  soil bulk density,* (g/cmj): Generally
   between 1.0 and 2.0 a/cm
= particle  density,  (g/cm ):  usually 2.65
   g/cm3 used for most mineral material.
For estimation,  Pt  can  be  assumed to  be
approximately 0.55  for  dry,  non-compacted soils,
and about 0.35 for compacted soils. This same value
(0.35)  is also appropriate for use as a  generic air-
filled  soil  porosity  (Pa)  when analyzing  the
volatilization release  from soils with a  high  moisture
content (Shen 1981). Alternatively, the local  Soil
Conservation  Service office  can be  contacted to
obtain  site-specific  estimated air-filled  soil  porosity
values  for specific locations.

Saturation  vapor  concentration,  Csi   can  be
determined by (USEPA 1980b):
                                                           C .=
                                                            SI
           pMW.

            RT
                                                                                                (2-7)
                                                     where
       R

       T
    = saturation vapor concentration of
       component i,  (g/cm3).
    =  vapor pressure of the chemical," (mm
       Hg).
    =  mole weight of component i, (g/mole).
    = molar gas  constant,  (62,361  mm Hg-
       cm3/mole-°K).
    = absolute temperature, (K).
                                                     Again,  use  maximum summer  temperatures to
                                                     estimate  short-term  release  and  annual  average
                                                     temperatures to initially estimate long-term release.
                                                        Values for soil bulk density for specified locations can  be
                                                       obtained from the U.S. Soil Conservation Service, Soils 5 File
                                                       data base.
                                                       " If the vapor pressure of a chemical under consideration is  not
                                                       available in standard reference texts, estimate it as described in
                                                       Lyman et al. (1982).
                                                  17

-------
Table 2-3. Diffusion
Compound
Acetaldehyde
Acetic acid
Acetone
Aniline
Benzene
Bromoethane
Bromoform
Carbon tetrachloride
Chlorobenzene
Chloroethane
Chloroform
Chloromethane
Cyclohexane
Dichloroethane
Dichloroethylene
Dicchloropropalene
Dimethylamrne
Ethanol
Ethyl acetate
Ethylamine
Ethylbenzene
Fluorotoluene
Heptane
Hexane
Isopropanol
Methanol
Methyl acetate
Methyl chloride
Methylethyl ketone
PCB (1 Cl)
Pentane
Phenol
Styrene
Tetrachloroethane
Tetrachloroethylene
Toluene
Tricyhloroethane
Trichloroethylene
Trichlorofluoromethane
Vinyl chloride
Xylene
Coefficients of Selected Organic Compounds
Atomic Diffusion
Molecular diffusion
Formula weight volume at 10°C
C2H40
C2H402
C3H60
C6H7N
C6H6
CH3Br
CHBr3
CCI4
C6H5CI
C2H5CI
CHC13
CH3CI
C6H12
C2H4CI2
C2H2CI2
C3H6CI2
C2H7N
C2H60
C4H802
C2H7N
C8H10
C7H7F
C7H16
C6H14
C3H80
CH40
C3H602
CH2CI2
C4H80
C12HGCI
C5H12
C6H60
C8H8
C2H2CI4
C2CI4
C7H8
C2H3CI3
C2HCI3
CCI3F
C2H3CI
C«H,ft
44
60
58
93
78
95
118
154
113
65
120
51
84
99
97
113
45
46
88
45
116
110
100
86
60
32
74
85
72
189
72
84
104
168
166
92
133
131
138
63
106
46.40
51.88
66.86
118.55
90.68
57.44
53.48
94.50
128.40
62.40
76.89
57.94
122.76
75.96
106.96
100.38
52.55
50.36
92.80
52.55
151.80
154.36
146.86
126.72
37.82
29.90
72.34
59.46
87.32
235.32
106.26
96.16
137.84
1143.96
111.00
111.14
97.44
93.48
100.00
58.44
131.60
.11758
. 10655
. 09699
.07157
.08195
.09611
.09655
.07500
. 06769
.09789
. 08345
. 10496
.07139
.08557
. 07442
.07519
.11161
.11297
.07991
.11161
.06274
.06262
. 06467
.07021
. 12004
. 14808
.09054
.09610
.08417
. 04944
.07753
.07919
.06620
.06858
.06968
.07367
.07496
.07638
.07391
. 10094
.06742
coefficients (cm2/sec)
at 30°C at 50°C
. 13249
.12007
. 10930
. 08065
. 09234
. 10830
. 10880
.08451
.07627
.11031
. 09404
.11827
. 08045
. 09643
.08386
.08473
.12577
.12730
.09005
.12577
.07070
.07056
.07287
.07912
.13526
. 16686
. 10203
. 10830
.09485
.05571
.08737
. 08924
. 07460
.07729
.07852
.08301
. 08447
.08606
.08329
.11375
.07597
.14816
.13427
.12223
.09019
.10327
.12111
.12167
.09451
.08530
. 12336
.10517
. 13226
. 08996
. 10784
.09377
.09475
. 14065
. 14236
. 10070
. 14065
. 07906
.07891
.08149
. 08848
.15126
. 18660
.11410
.12111
. 10607
. 06230
.09770
.09980
.08343
.08643
.08781
.09283
.09446
.09625
.09314
. 12720
.08495
Source: Shen 1981
                                           18

-------
See Section 2.3.3 for directions for calculating a final
long-term release rate.
(2) Landfills with Internal Gas Generation
Thibodeaux (1981) developed a method for estimating
toxic vapor releases from   co-disposal  landfills.
These facilities  contain toxic  wastes  in  combination
with  municipal  or sanitary  wastes  that,  because of
their  considerable organic  content, generate  landfill
gases (e.g.,  H2, CH4,  C02).   In these  cases,  the
upward movement  (convective  sweep) of the  landfill
gas  becomes the significant controlling factor,  greatly
accelerating  the  upward  migration and  subsequent
release to the atmosphere of  the  co-disposed toxic
substances.  In  fact,  review  of Thibodeaux's work
indicates that the toxic gas migration accelerating the
effect of the  landfill gas is so great that both soil  and
gas  phase diffusion essentially become  insignificant.
The  following simplified equation is  recommended for
estimating  the volatilization  of  toxic substances from
co-disposal  landfills:
Ei = Ci*VyA


where
(2-8)
           = emission rate,  (g/sec).
    Ei      = concentration of compound i in the soil
              pore spaces, (g/cm3)
    v      = mean  landfill gas velocity in the  soil
     y        pore  spaces,  (cm/sec). Thibodeaux
              (1981) provides  an average  value of
              1.63 x 10-3 cm/sec for this factor.
    A  =  area,  (cm2).
 Recalculation of the  toxic  vapor release estimates
 presented in Thibodeaux (1981) using this simplified
 equation  yields results  within  = 1  percent of the
 values obtained using the full computation cited in the
 paper.  Thibodeaux  (1981)  notes,  however, that
 various site factors such as the presence of saturated
 soils will  tend to reduce the rate of  volatile chemical
 release from landfills. The degree to  which this model
 is able to accurately reflect contaminant release rates
 for gases, especially soluble gases, generated at sites
 with moist or wet soils is unknown.

 (3) Spills and Leaks
 Equations  2-9  and  2-11  will  estimate the  volatile
 releases  from fresh  and old (respectively)  chemical
 spills on soil.  Equations 2-10  and  2-12 through 2-
           14 provide means of estimating certain input variables
           required to solve Equations 2-9 and 2-11.

           As discussed  in  Farino et al. (1983)  one can  apply
           Equation 2-9 (adapted from Thibodeaux  and Hwang
           1982)  to  estimate volatile releases  resulting from
           spills or leaks where a contaminant  pool is visible on
           the soil  surface, or  where  soil  is  contaminated
           (saturated) from the surface down. The equation does
           not consider soil  phase mass transfer  resistance, and
           therefore  is  not appropriate  for  use when spilled
           contaminants have seeped into  surface soils (in this
           case,  use the  landfarming  equation that  follows).
           Similarly,  because it does not consider liquid phase
           resistance, it is  only useful for estimating releases of
           pure compounds. The original  equation presented in
           Thibodeaux and  Hwang (1982) has  been modified to
           include a  contaminated  surface area term,  thereby
           resulting  in the calculation  of a release rate rather
           than  a  flux  rate value:
Ei =  kiGCi*A

where
                                                                                                    (2-9)
               A
          =  emission rate  of chemical  i,  (g/s).
          =  gas  phase mass transfer  coefficient of
              chemical i, (cm/s).
          =  vapor concentration  of chemical i,
              (g/cm3).*
          =  area, (cm ).
           Hwang  (1982)  has developed the following simplified
           means  of estimating  a compound's gas phase mass
           transfer coefficient.
            k..
                                                              MW.
                   MW.
                                                                       0.335
                                                                           '/
                    V298
                                      1.005
                                                       (2-10)
           where
           MWH20;  MW, =
              klG, H20  =
                   gas  phase  mass transfer
                   coefficient of chemical i,  (cm/s).
                   molecular  weight  of water;
                   compound i, (g/mole).
                   temperature,  (°K).
                   gas  phase  mass transfer
                   coefficient  for water vapor at
                   25°C, (cm/sec).
   " For conservative analyses, one can assume that the actual
   contaminant vapor concentration in the soil pore spaces is the
   same as the equilibrium vapor concentration. In such cases,
   C can be used in place of C,. Direct measurements of C,,
   h&ever, may be developed during the site investigation. When
   such data are available, their use is preferred.
               For conservative analyses, one can  assume that the actual
             contaminant vapor concentration in the soil pore spaces is the
             same as the equilibrium vapor concentration. In such cases, Cj,
             can be used in place of C,*.  Direct measurements of C, ,
             however, may be developed during the site investigation. When
             such data are available, their use is preferred.
                                                    19

-------
When  estimating short-term  (maximum)  release
rates,  the  highest  (summer)  seasonal temperature
expected at the site can be  used in calculating the
gas  phase mass transfer coefficient. For  initial
estimation  of long-term release  rates, the  seasonal
average  temperature  should  be used.  Final long-
term release notes are developed  as discussed in
Section 2.3.3.

(4)  Landfarming
In cases where past spills,  leaks, or intentional
disposal  directly  onto  or into  surface  soils
(landfarming) have resulted  in contaminated surface
soils with liquids  in the pore spaces, Equation 2-11
can be used to estimate volatilization releases. This
equation  assumes that soil  pore spaces connect with
the  soil surface,  that soil  conditions are isothermal,
and that there  is no capillary  rise of contaminant. It
also  assumes  that there  is  sufficient liquid
contaminant in  the pore spaces  so  that volatilization
will  not deplete the  reservoir  of  contaminant to the
point where it affects the  rate of volatilization.
Modeling  the  release from soils with  sorbed
contaminants and no  free  liquids  requires another
model.

Two  models  for  predicting  the  time-varying
volatilization of  sorbed  contaminants  on soil  are
presented in USEPA  (1986e). The equation presented
here is adapted from Thibodeaux  and Hwang (1982),
which  presents  a volatilization  release estimation
equation  designed for application to active or planned
landfarms for petroleum wastes. Farino  et al. (1983)
determined it to be preferable to other approaches for
estimating volatilization release of chemicals spilled or
incorporated into  soils, because  it directly takes  into
account the contaminant loss over time. It  describes
vapor  diffusion  as being  soil-phase controlled,  and
essentially  assumes  that contaminant concentrations
in the  soil remain constant (until all contaminant is
lost to the air), and that contaminant release occurs
by  the  "peeling  away" of  successive  unimolecular
layers  of contaminant  from the surface of  the "wet"
(contaminated)  zone. Thus, over time  this process
results in a "dry zone" of increasing depth at the soil
surface,  and a  wet zone of decreasing depth  below
the  dry zone. The original equation has been adjusted
somewhat  for  use at  uncontrolled waste  sites,  and
has also been simplified as discussed in Farino et al.
(1983), by  assuming  that the oil layer diffusion length
value is  low (i.e., that the  spilled  contaminant  has
become  incorporated into  surface  soils and is not
present as a discrete film).
             2DCSA
where
 D
Cs

CB

 A
 d
             average  emission  rate of component i
             over time t, (g/sec).
             phase transfer coefficient, (cm2/sec).
             the  liquid-phase  concentration  of
             contaminant i in the soil, (g/cm3).
             bulk contaminant concentration in soil
             (g/cm3)
             contaminated surface area, (cm2)
             depth of dry  zone  at sampling  time,
             (cm).
             time  measured from  sampling  time,
             (seconds).
D (cm2/sec) is related to the amount of contaminant i
that goes from liquid to gas phase, and then from gas
phase to  diffusion  in  air. It can  be estimated  as
follows:
where

     D
    DI

    Pt
   Hi'
                                          (2-12)
         phase  transfer  coefficient, (cm /sec).
         diffusion  coefficient of component i  in
          air,  (cm2/sec).
         total soil  porosity,  (dimensionless).
          Again, use of total soil porosity in this
          equation  results in  a worst-case (dry
          soil) estimate  for  D. As  previously
          discussed, however,  in  some cases
          (i.e., where  soils are wet more often
          than dry) it may be more appropriate to
          use air-filled soil porosity (Pa) in place
          of Pt.  See text  addressing Equation  2-
          3 for a discussion of the  application of
          and values for these two terms.
         Henry's Law constant in  concentration
          form, (dimensionless).
Hi', the Henry's Law  constant in concentration form
(ratio  of the  boundary  layer concentration  of
contaminant in air to the boundary layer concentration
of contaminant in  "wet" soil) can be  determined  as
follows (Lyman et al. 1983):
where
         H.
         -^
         RT
                                       (2-13)
                                                          R
                                          (2-11)
          =   Henry's Law constant of contaminant i,
              (atm-m3/mol).
          =  gas  constant,, (8.2  x  1 0~5 atm-
              m3/mol-°K).
     T    =  absolute temperature, (°K).

Again,  use  summer  maximum  temperatures to
estimate short-term release and annual  average
                                                  20

-------
temperatures for the initial estimation of long-term
release. Final long-term release rates are developed
as discussed in Section 2.3.3.

Note  that Equation  2-11 assumes that  the
contaminant concentration  in the liquid  and  gas
phases in  the soil  remains constant  until all of the
contaminant has  been  released  to  air. Also, the
equation holds from time zero (the time at which the
soil  was sampled) to td  (the time at  which the soil
becomes dry, i.e.,  all contaminant has volatilized and
the  release  process stops).  The formula for
calculating  td (in  seconds)  is:
              'B
        2D
                                          (2-14)
where
     td
   CB
          = the time at which  all  contaminant  has
             volatilized  from the soil, (sec).
          = depth  from soil  surface  to  the  bottom
             of the contaminated region,  (cm).
          = depth  of dry  zone  at sampling time,
             (cm).
          D = phase transfer coefficient, (cm /sec).
          = bulk contaminant concentration in soil,
             (g/cm3)
    Cs    = contaminant liquid phase concentration
             (g/cm3)
(5) Lagoons
Mackay  and Leinonen  (1975) have  developed an
equation  for estimating  volatilization releases  of low
solubility compounds from  waterbodies  such as
hazardous waste  lagoons. This is  presented as
Equation  2-15. Equations  2-16  and   2-17  provide
means  of calculating  certain input parameters
required  by Equation  2-15. This  approach  assumes
that  conditions are steady state (i.e., no  constant
addition of contaminant), that  diffusion  is  liquid state
controlled, and that  it  occurs  from  a well-mixed
water phase to a  well-mixed  air phase across a
stagnant  water/air interface. As  pointed out  in Farino
et al. (1983), if it can be assumed that atmospheric
background levels of the contaminant of concern  are
negligible, (as would usually  be  the  case  at
abandoned hazardous waste facilities), then Mackay
and  Leinonen's basic equation  can be simplified to
the following form  (which includes an  area term to
convert flux rate to emission rate):
where
    K,
              (2-15)
          = emission rate, (g/sec).
          = overall  mass transfer  coefficient,
             (cm/sec).
                                                         Cs    = contaminant liquid phase concentration,
                                                                  (g/cm3)
                                                         A    = area, (cm ).
                                                     The  overall mass transfer  coefficient (K,)  can be
                                                     calculated via the following relationship:
                                                       1    E    RT
                                                      — = —   	                         (2-16)
                                                      K.   k.                                  V    '
                                                     where

                                                         KI

                                                        kiL

                                                         R

                                                         T
                                                                                                    -5
        iL   HjkiG

           = overall  mass transfer  coefficient,
              (cm/sec).
           = liquid  phase  mass  transfer  coefficient,
              (cm/sec). See Equation  2-17.
           = ideal  gas  law constant, (8.2 x  10
              atm-m3/mol-°K).
           = temperature,  (°K).
    H |     = Henry's  Law  constant of  compound  i,
              (atm-m3/mol).
    klG  =  gas phase  mass transfer coefficient,
            (cm/sec). See Equation 2-10.

 Hwang  (1982)  provides a  simplified method for
 determining a compound's liquid  phase mass transfer
 coefficient for use in the  above equation. To estimate
 kiL, use  the following equation:
                                                      k-r =
                                                       iL

                                                     where
                                                            MW.
         MW.
\ 298
                          kL'°2
                                                                                              (2-17)
                                                            kiL
                                                             T
                                                         kL,02
              =  liquid  phase  mass  transfer
                 coefficient, (cm/sec).
              =  molecular  weight  of  oxygen;
                 compound,.
              = temperature,  (°K).
              =  liquid  phase  mass  transfer
                 coefficient  for oxygen  at 25°C,
                 (cm/sec).
The value for kL,O2 can  be obtained from chemical
reference texts or can be calculated (the preferred
method) as described in  Farino  et al. (1983).

2.3.2.2 In-Depth Analysis
In-depth analysis of volatile release can be executed
in the same manner as that described for particulates.
Subtract the monitored upwind  (control) ambient toxic
vapor concentration from the monitored downwind
concentration. Use the difference between these two
values in an air dispersion  model to estimate the
release rate at a "virtual  point source" that would
correspond with the source of the measured
downwind  concentration.

The user of this manual should  again refer to USEPA
(1983c) and Seely et al. (1983)  for detailed
discussions of the planning and execution of air
monitoring studies. Refer to Chapter 3 of this manual
                                                 21

-------
for a description of air contaminant dispersion
modeling tools.

2.3.3 Long-Term and Short-Term Release Cal-
culation
Long-term release values (70 years) for lagoons with
dilute solutes can be estimated as follows:
  E. .=
   Ai
where

   EAi

    Vc
     e
     E
   V C.
     C 1

   ~70~
1-e
                   V C.
                    C I
                      -<2.2xl09>
(2-18)
      =  average annual release of contaminant
         volume of contaminated region, (cm3).
         initial  average  concentration  of
         contaminant i in site soils, (g/cm3).
         2.71828
         initial  combined  release  rate  of
         contaminant i,  (g/sec). Obtained  by
         summing all above-listed releases of
         the contaminant at  the  site.  For
         particulates,  convert the  average
         annual release from Equation 2-18 to
         mass  per second by dividing by 3.16 x
         107 seconds.
Note  that Vc and  C,  must be based  on the same
value.   (VcCj) is equal to  the total mass of
contaminant; it can be the total mass of contaminant
in a lagoon.

For  landfills and wind  erosion  of contaminated
particulates,  the  release  rate is  assumed  constant.
The  70-year average annual  release  rate can  be
calculated by  first ascertaining if contaminant will
remain after 70 years. If  so, then the release  rate
itself is the  70-year average annual release rate. If
not, then the 70-year average annual release rate is
the total initial mass divided by 70 years.

To estimate  long-term release from  contaminated
surface soils, Equation 2-14  (converted to  years by
dividing by 3.16 x  107) is first used to determine the
dry-out time. If  no contaminant is  expected to
remain after 70 years  (i.e.,  70 > td), simply
determine the total amount of contaminant present at
the time of site investigation  and  divide by 70 years
(in seconds) to get a  conservative long-term release
value  (i.e.,  ACs(h  -  d)/2.21  x  109).  If some
contaminant is expected to  remain after 70 years (i.e.,
70 < td),  use the following equation to  estimate
long-term  release:
  E,
   Ai

where
   AC /
.= —(CdJ
1    70 V
  + 4.4xl
-------
toxics can  be quantified  directly by  measuring
(sampling) the  source  material and  determining the
volume and  rate  of release.  Alternatively,  runoff
release estimation  procedures,  less  costly  than
monitoring or  modeling approaches,  can  also  be
applied to  uncontrolled sites.

In addition, surface waters may be contaminated by
inflows of ground  water through  bank seepage and
springs. In order to estimate the rate of such inflows,
one  must  conduct modeling of ground-water/surface
water linkages  (see  Chapter 3 for  a  discussion of
ground-water modeling options).

This section  reviews methods  for estimating  toxic
releases of uncontrolled  hazardous waste sites to
surface waterbodies. Note, however, that only the
surface  runoff component of release to surface  water
is addressed  here. Other sources must be  estimated
for each site based on judgment and experience.

2.4.1 Beginning Quantitative Analysis

2.4.1.1 Dissolved and Sorbed Contaminant
Migration
Many of the organic substances of concern found at
Superfund sites are  relatively  nonpolar,  hydrophobic
substances (Delos  et al., 1984). Such substances can
be expected to  sorb to site soils and migrate from the
site  more  slowly  than  will polar compounds. As
discussed in Haith (1980) and Mills et al. (1982),
estimates  of the amount of hydrophobic compounds
released in  site runoff  can  be  calculated  using the
Modified Universal Soil  Loss Equation (MUSLE) and
sorption  partition coefficients derived  from the
compound's octanol-water  partition  coefficient.  The
MUSLE allows  estimation of the amount of surface
soil eroded in  a storm event of given intensity, while
sorption coefficients allow the projection  of the
amounts of  contaminant carried along with the soil,
and the amount carried in dissolved form.

(1) Soil   Los Calculation
Equation 2-20  is  the  basic equation for  estimating
soil  loss.  Equations  2-21 through 2-24  are used to
calculate certain input parameters required to  apply
Equation  2-20. The modified  universal  soil loss
equation (Williams  1975), as presented in Mills  et al.
(1982), is:
  Y(S)E = a(Vrqp)a56 KLSCP
                               (2-20)
where

 Y(S)E

     a

    Vr
= sediment  yield (tons  per event,  metric
   tons per event).
= conversion constant,  (95 English, 11.8
   metric).*
= volume of runoff, (acre-feet,  m3).
                                               q     = peak flow rate,  (cubic feet per second,
                                                p       m3/sec).
                                               K    = the  soil  erodibility  factor,  (commonly
                                                        expressed  in tons  per  acre  per
                                                        dimensionless rainfall erodibility unit). K
                                                        can be obtained from the local  Soil
                                                        Conservation Service off ice.
                                                L    = the  slope-length factor,  (dimension-
                                                        less ratio).
                                                S    = the  slope-steepness factor,  (dimen-
                                                        sionless ratio).
                                                C    = the  cover factor,  (dimensionless ratio:
                                                        1.0 for bare soil;  see the  following
                                                        discussion  for  vegetated  site  "C"
                                                        values).
                                                P    = the  erosion  control practice  factor,
                                                        (dimensionless ratio:  1.0  for
                                                        uncontrolled hazardous waste sites).

                                           Soil  erodibility factors  are  indicators  of the  erosion
                                           potential  of given  soil types. As such, they are highly
                                           site-specific. K  values  for sites under study  can be
                                           obtained from the  local  Soil Conservation Service
                                           office.  The  slope length factor,   L,  and  the slope
                                           steepness factor,  S, are  generally entered into the
                                           MUSLE as a combined factor, LS, which is obtained
                                           from  Figures  2-4 through  2-6. The  cover
                                           management factor,  C, is determined by  the  amount
                                           and type of vegetative cover present at the site. Its
                                           value is  "1" (one) for bare soils.  Consult Tables 2-4
                                           and 2-5  to obtain C values for sites  with vegetative
                                           covers.  The factor, P,  refers to  any  erosion control
                                           practices used  on-site. Because these generally
                                           describe  the type of agricultural  plowing or planting
                                           practices, and  because it is unlikely that any erosion
                                           control  would  be  practiced at an abandoned
                                           hazardous waste  site,  use  a  worst-case
                                           (conservative) P  value  of 1 (one) for uncontrolled
                                           sites.

                                           Storm  runoff volume,  Vr, is calculated as  follows
                                           (Mills etal.  1982):
                                             Vr =  aAQr

                                           where
                                          (2-21)
     a    =  conversion constant, (0.083  English,
             IOO metric).
    A    =  contaminated area, (acres, ha).
    Qr    =  depth of runoff, (in, cm).

Depth of runoff, Qr, is determined by (Mockus 1972):

  Qr = (Rt-0.2Sw)2/(Rt + 0.8Sw)            (2-22)
   Metric conversions presented in the following  runoff
  contamination equations are from Mills et al. (1982).
                                                  23

-------
Figure  2-4.    Slope effect  chart applicable to areas A-1 in
             Washington,  Oregon, and Idaho, and all of A-
             3: MO  Figure 2-6  (USDA 1974 as presented
	in Mills et al. 1982).	
                        Slope Length, Meters

        20  3040  6080100150200300400600800
    40.0
    20.0
    10.0
     6.0
  2  4.0
  o
  
-------
values of uncontrolled hazardous waste  sites from
Table 2-6.

The  peak runoff  rate,  qp,  is determined  as follows
(Haith 1980):
          aArtQr
      Tr(Rt-0.2Sw)
                                         (2-24)
where

    qp
     a

    A
    Rt
    Qr

    Tr
             the peak runoff rate, (ft3/sec, m3/sec).
             conversion  constant,  (I.OI  English,
             0.028 metric).
             contaminated area, (acres, ha).
             the total storm rainfall, (in, cm).
             the depth of runoff from the watershed
             area, (in, cm).
             storm duration, (hr).
             water retention factor, (in, cm).
(2)Dissolved/Sorbed Contaminant Release
As discussed in Mills et al. (1982), the analyst can
predict the degree of soil/water partitioning expected
for given compounds once the  storm  event soil loss
has  been calculated  with  the  following equations.
First, the  amounts  of  adsorbed  and  dissolved
substances  are determined,  using  the  equations
presented below as adapted from Haith (1980):
S8 =
          +
    and
where

    S8
    ec
     P
    Ci

    A



    Ds
                        i) (A)
            (Kdp)/ec)](Ci)(A)
(2-25)
                                         (2-26)
             sorbed substance quantity, (kg, Ib).
             available water capacity of the top cm
             of soil (difference between wilting point
             and field capacity), (dimensionless).
             sorption partition coefficient, (cm3/g).
             soil bulk density, (g/cm3).
             total  substance concentration,  (kg/ha-
             cm,  Ib/acre-cm).
             contaminated  area,  (ha-cm  acre-
             cm). (Actually a volume; assumption  is
             contamination  in upper  1  cm   is
             available for release.)
             dissolved substance quantity,  (kg, Ib).
This model assumes that only the contaminant in the
top 1 cm of soil is available for release via runoff.

The soil sorption  partition  coefficient  for  a given
chemical can  be determined from  known values  of
certain other physical/chemical  parameters,  primarily
                                                     the chemical's  octanol-water  partition  coefficient,
                                                     solubility in water, or bioconcentration  factor. Lyman
                                                     et al.  (1982)  present regression equations that  allow
                                                     the analyst  to determine  sorption  coefficients  for
                                                     specified  groups  of chemicals (e.g.,  herbicides,
                                                     polynuclear aromatics).  If parameter values  required
                                                     by  the  appropriate equations  are  not available  in
                                                     chemical reference literature, they can be estimated
                                                     according to  procedures described  in  Lyman et al.
                                                     (1982).  Initially, the octanol-water partition coefficient
                                                     can be estimated based on the substance's molecular
                                                     structure.  If necessary, this  value  can be  used,  in
                                                     turn,  to  estimate  either  solubility  in water or
                                                     bioconcentration factor.

                                                     After calculating the amount of  sorbed  and dissolved
                                                     contaminant,  the total  loading to the receiving
                                                     waterbody  is  calculated as  follows (adapted  from
                                                     Haith  1980):
                                                     PXi=[Y(S)E/100p]Ss

                                                         plus
                                                     (2-27)
                                                                                               (2-28)
where

   PXi    =  sorbed substance loss per event, (kg,
             Ib).
 Y(S>E    =  sediment yield, (tons per event,  metric
             tons).
     P    =  soil bulk density, (g/cm3).
    Ss    =  sorbed substance quantity, (kg, Ib).
   PQi    -  dissolved  substance  loss  per  event,
             (kg, Ib).
    Qr    =  total storm  runoff depth, (in, cm).
    Rt    =  total storm  rainfall, (in, cm).
    Ds    =  dissolved substance quantity, (kg, Ib).

PXi and PQi can be converted to mass per volume
terms for use in estimating contaminant concentration
in  the  receiving waterbody  by dividing by the site
storm runoff volume (Vr,  see Equation 2-21).


2.4.2  In-Depth  Analysis
Releases  to surface waterbodies at  uncontrolled
hazardous waste  sites  can be  quantified   most
accurately  by direct  measurement (sampling  and
analysis)  of the contaminant flow. Alternatively,
upcurrent  and  downcurrent  sampling can  be
conducted  to determine the release level  at the site
that would  be used to  estimate the  ambient
concentration (i.e.,  the  difference between the
upcurrent and  downcurrent concentrations).  Either
simple  dispersion  equations  or  sophisticated
computer  modeling approaches (see Chapter 3) can
be  used to "back  calculate" the measured ambient
concentration to the "virtual point source."
                                                  25

-------
Table  2-4.     "C"  Values for Permanent Pasture, Rangeland, and Idle Land
                                                                      Cover that contacts the surface
Vegetal canopy
Type and height
of raised canopyb
No appreciable canopy

Canopy of tall weeds or short
brush (0.5 m fall height)




Appreciable brush or brushes
(2m fall height)




Trees but no appreciable low
brush (4 m fall height)




Canopy
cover0
(%)


25

50

75

25

50

75

25

50

75

Percent ground cover
Typed
G
w
G
w
G
W
G
W
G
W
G
W
G
W
G
W
G
W
G
W
0
0.45
0.45
0.036
0.036
0.026
0.026
0.17
0.17
0.40
0.40
0.34
0.34
0.28
0.28
0.42
0.42
0.39
0.39
0.36
0.36

0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
20
.20
.24
.17
.20
.13
.16
.10
.12
.18
.22
.16
.19
.14
.17
.19
.23
.18
.21
.17
.20
40
0
0
0
0
0
0.
0
0
0
0.
0
0.
0.
o
0.
0.
0.
0.
0.
0.
.10
.15
.09
.13
.07
11
.06
.09
.09
14
.085
13
08
.1 2
10
14
09
14
09
13
60
0.042
0.090
0.038
0.082
0.035
0.075
0.031
0.067
0.040
0.085
0.038
0.081
0.036
0.077
0.041
0.087
0.040
0.085
0.039
0.083

0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0.
80
.013
.043
.012
.041
.012
.039
.011
.038
.013
.042
.012
.041
.012
.040
.013
.042
.013
.042
.012
,041
95-100
0.003
0.011
0.003
0.011
0.003
0.011
0.003
0.011
0.003
0.011
0.003
0.011
0.003
0.011
0.003
0.011
0.003
0.011
0.003
0.011
Source:  Wischmeier 1972
aAII values shown assume:  (1) random distnbution of mulch or vegetation, and  (2) mulch of appreciable depth where it exists.
bAverage fall height of waterdrops from canopy to soil surface: m = meters.
CPortron of total-area surface that would be hidden from view by canopy in a vertical projection (a  bird's-eye view).
dG: Cover at surface is grass, grasslike plants, decaying compacted duff, or litter at least 5 cm (2  in) deep.
 W: Cover at surface is mostly broadleaf herbaceous plants (as weeds) with  little lateral-root  network near the surface and/or
     undecayed residue.
Stand condition
Well stocked

Medium stocked

Poorly stocked

Tree canopy per-
cent of area3
100-75

70-40

35-20

Forest litter per-
cent of areab
100-90

85-75

70-49

Undergrowth0
Managedd
Unmanagedd
Managed
Unmanaged
Managed
Unmanaged
"C" factor
0.001
0.003-0.01 1
0.002-0.004
0.01-0.04
0.003-0.009
0.02-0.09"
      Source:  Wischmeier  1972

      awhen tree canopy is  less than 20 percent, the area will be considered as grassland or cropland for estimating soil loss.
      bForest litter is assumed to be at least 2 in deep over the  percent ground surface area covered.
      °undergrowth is defined as shrubs, weeds, grasses, vines, etc., on the surface area not protected by forest litter. Usually
      found under canopy openings.
      dManaged - grazing and fires are controlled.
        Unmanaged - stands that are overgrazed or subjected to repeated burning
      eFor unmanaged woodland with litter cover of less than 75 percent,  C values should be derived  by taking 0.7 of the
      appropriate values in  Table 3-4.  The factor of 0.7 adjusts  for  much higher soil organic matter on permanent woodland.
                                                            26

-------
         Table 2-6. Runoff Curve Numbers
            Soil
           group	Description
                       Site type
Overall site"    Road/right of way     Meadow
                          Woods
                  Lowest runoff potential: Includes
                 deep sands with very little silt and
                 clay, also deep, rapidly perme-
                 able loess (infiltration  rate =
                 8-12 mm/h).
                  Moderately low runoff potential:
                  Mostly sandy soils less deep than
                  A, and loess less deep or less
                  aggregated than A, but the group
                  as a whole has above-average
                  infiltratipn after thorough wetting
                  (infiltration rate = 4-8  mm/h).
                  Moderately high runoff potential:
                  Comprises shallow soils and soils
                  containing considerable clay and
                  colloids, though less than those of
                  group  D. The group has below-
                  average infiltration after
                  presafuration (infiltration rate =
                  1-4 mm/h).
                  Highest runoff potential: Includes
                  mostly clays of high swelling
                  percent, but the group also in-
                  cludes some shallow soils with
                  nearly impermeable subhorizons
                  near the surface (infiltration
                  rate = 0-1 mm/h).	
    59
    74
74
84
30
58
45
66
    82
90
71
77
                  92
               78
              83
          Source: Adapted from Schwab et al. 1966.
          Values taken from farmstead category, which is a composite including buildings, farmyard, road, etc.
2.4.3  Long-Term and  Short-  Term  Release
Calculation
For  surface  runoff releases, the  long-term release
value can be calculated as follows:

I    Characterize an average storm event for the area
    in terms  of duration.  This  can  best  be
    accomplished  by  consulting  local  or regional
    climatological  experts,  or the National
    Climatological Data  Center  in Asheville,  North
    Carolina.  Then,  using  USDC  (1961)  determine
    the  amount  of rainfall corresponding to  the
    selected  duration  rainfall  event on  a  one year-
    return  frequency  basis.  Divide  this  amount  into
    the mean annual rainfall for the area to  obtain the
    average  number  of  average  rainfall events  per
    year.

I  Use these  data  and the  equations  presented in
    this  section  to  calculate  runoff  contaminant
    release  associated  with each yearly average
    storm.

I  Estimate the potential total  long-term release  for
    both dissolved and sorbed runoff loss* as follows:
  * This approach is overly conservative as it assumes that the
  contaminant concentration in surface soil remains essentially
  the same during the entire  70-year period.
             EAi = BN

           where


              EAi     =

                B


                N
                                     (2-29)
      long-term  release of contaminant  i in
       runoff, (mass/70 years).
      dissolved or  sorbed  loss  per storm
       event, (i.e.,  PX, or PQ,; see Equations
       2-27  and 2-28).
      number of "average"  storm  events in
       70 years.
                Determine the total amount of soil that will erode
               from the  site over 70 years.  This  can  be
               accomplished by applying the Universal  Soil Loss
               Equation  (USLE-Wischmeier and Smith 1978).
               This  equation,  from  which the  MUSLE  (see
               Equation 2-20) was developed, estimates annual
               soil  losses in runoff. The USLE takes the same
               form as the MUSLE, except  that  the storm
               event-specific volume  and  flow  rate  variables
               are  replaced by  a factor,  R,  the rainfall runoff
               factor. Therefore, the USLE  is:
           Y(S)A =  R,KLSCPASd

           where
                                     (2-30)
            Y(S)A    =  annual  soil  loss  in  runoff,  (tons/yr,
                         tonnes/yr).
                                                     27

-------
    Rr    =  the  rainfall  and  runoff   factor,
             expressing  the  erosion potential  of
             average annual  rainfall  in the  locality
             (can be obtained from  the  local Soil
             Conservation Service office),  (dimen-
             sionless).
    K    =  the soil-erodibility  factor,  commonly
             expressed in tons per acre per Rr unit
             (can be obtained from  the  local Soil
             Conservation Service office)  (in metric
             tons/ha/Rr unit),  (English K'1.292   =
             metric K).
    L    =  the slope  length  factor,  (dimen-
             sionless).
    S    =  the slope  steepness factor,  (dimen-
             sionless).
    C    =  the cover  factor, (dimensionless ratio:
             1.0  for   soil,  see  the  following
             discussion  for  vegetated  site  "C"
             values).
    P    =  the erosion  control  practice  factor,
             (dimensionless:  1.0 for uncontrolled
             hazardous waste sites).
    A    =  acreage of area,  (acres,  ha).
    Sj    =  the sediment  delivery  ratio,  (dimen-
             sionless).

The sediment delivery ratio (Sj)  can  be  estimated
using  the following equation:
Sd = Dd-0.22

where
(2-31)
    Da    =  the overland distance between the site
             and the receiving waterbody (ft).

Mills  et  al.  (1982)  note that  this equation  was
empirically derived from data for D  values from 0 to
800  feet,  and  caution  that  it  may require  further
testing, particularly in sites located in the Midwest and
Central U.S.

Note  that  in  certain  areas of the  Pacific  Northwest
and central western states, thaw and snowmelt may
contribute  most of the runoff erosive  force on  an
annual basis.  In such cases,  an  additional erosion
factor, Rs, must be added to the  rainfall and  runoff
factor, R, to calculate  the total R value  for use in the
USLE. Limited field  data have  indicated  that  an
approximate estimate  of Rs  may  be  obtained  by
multiplying 1.5 times the local average total rainfall (in
inches) for the period  December 1  through March 31
(Wischmeier and Smith 1978). However, the local Soil
Conservation Service office can  provide the overall R
value (Rr plus Rs).

•   Based on  site monitoring   data,  estimate  the
    average  contaminant concentration in the layer of
    soil  that must be  eroded  to  equal the  total
    estimated  amount of soil  lost over  70  years
    (based  on site  soil  sampling data  and  the
    calculated  vertical  depth  of soil that will erode
    over that time period).

•   Multiply  the  average  contaminant  concentration
    on site by  the site  area to calculate the mass of
    contaminant  present  in  that  amount  of  soil
    estimated  to  be eroded over 70  years.  This
    represents the maximum  amount of contaminant
    available for  erosion  losses over  the 70-year
    period.

•   Compare the  estimated  potential  contaminant
    runoff losses over 70 years with  the mass of
    contaminant  present (in 70-year erodible  soils at
    the site). If the estimated total loss  to runoff is
    less  than  the  amount  available,  divide  the
    estimated  total  70-year losses by  the  total
    volume  of stormwater runoff estimated over 70
    years.  This  will  approximate the  contaminant
    concentration  in  runoff  (both  dissolved  and
    sorbed).

•   If the total estimated  contaminant  runoff losses
    exceed  the  amount of contaminant present in
    70-year  erodible  site soils,  divide the total mass
    of  contaminant present  in  such  soils  by  the
    volume of  runoff  estimated to leave the site over
    70 years  to  develop  adsorbed and  dissolved
    contaminant loss  estimates in concentration form.
    In  either  case,  the  runoff  value  needed to
    estimate contaminant transport and dispersion in
    surface waterbodies can be estimated by  dividing
    the total volume  of runoff estimated to  leave the
    site over 70  years by the  number of seconds,
    minutes, etc.  in 70   years to estimate  runoff
    volume per unit time.

Many factors  can influence  the  actual degree of
contaminant loss in  given storm events.  Because
such  factors vary from locale  to locale,  no single
method will guarantee accurate estimates of short-
term  contaminant losses  in  runoff from  all sites.
However,  the  following  approach should yield
reasonable approximations of the magnitude of such
short-term  loss.  While  short  duration,  high intensity
storm events (thunderstorms)  clearly cause  significant
erosion, the water quality  effects  of such storms are
too  ephemeral  to adequately  reflect  short-term
releases  as defined (i.e., 10 to 90 days). Therefore, a
storm event is needed  that will  generate contaminant
releases  adequate to affect water quality  over a time
period approaching the 10-day  lower bound of the
short-term  time  frame. For  this  analysis,  a  1-year,
24-hour  storm  event has  been selected.  Data
quantifying the amount of  rainfall that corresponds to
the 1-year,  24-hour storm event (as well as similar
data for other storm return periods and durations) are
provided  in USDC (1961).  To  estimate  short-term
runoff release,  the  average  site  contaminant
concentration should  be estimated based on sampling
                                                  28

-------
data for the top cm of soil  only. This value  is then
used  in Equations 2-27 and 2-28 to  estimate runoff
losses on a single storm event basis.

Research based on the work of Haith  et al. (1980) is
currently  underway  at  Cornell University* to  develop
runoff  loading  factors for organic chemicals in soils.
After these factors  are devised,  the  analyst  will be
able to obtain average  loading values based solely on
a  chemical's  octanol/water  partition  coefficient  and
the geographic location under study. This will greatly
simplify the generation of long-term average  release
estimates.

Note that  in order to  estimate  long-term and short-
term contaminant concentrations in surface water, the
long-term  and  short-term  release values  are used,
along with average and minimum streamflow  data as
described in Chapter 3, Environmental  Fate Analysis.

2.5 Quantitative Analysis of Ground-
Water Contamination

Surface soils at uncontrolled hazardous waste sites
may become  contaminated with  toxic  materials as a
result of (1) the intentional  placement of wastes on
the ground  (dumping,  landfarming),  (2)  spills,  (3)
lagoon failure (overland flow), or (4) contaminated site
runoff.  Leaching of toxics from  a contaminated  soil
surface can  carry  contaminants  into subsurface
layers.

2.51 Beginning Quantitative Analysis

2.5.1.1 Leachate Release Rate
This section  presents simplified approaches  for
estimating contaminant release rates to ground water.
Such  estimation can be determined for dry landfills,
lagoons, or wet landfills, whether unlined or  lined  with
clay or flexible  membrane liners.

(1) Estimating Release Rate  from Facilities Lined with
Clay or Natural Soil
Release rate estimation involves the determination  of
both the contaminant  concentration in the leachate
and the volumetric flux of leachate. The determination
of  contaminant concentration is   made   using
equilibrium conditions (steady  state), whereas  the
volumetric flux can be  ascertained with instantaneous
time-varying models or with  steady state equations.

Modeling the  release  rate  of toxic constituents  can
thus be done   in terms of  either the instantaneous
time-varying  releases  or the  annual average  release
(i.e., steady state release rate  based  on an  annual
average). This section  discusses the determination  of
the steady  state release rate (annual average);  the
  *  Contact Douglas A. Haith, Cornell University, Ithaca, N.Y.,
  (607)256-2280.
equations  are  simpler than  the  computer  models
necessary  for  instantaneous  time-varying releases.
Analysts  interested  in performing  instantaneous
time-varying release rate determinations  are  referred
to Chapter 3,  where the HELP  and SESOIL models
are discussed.  HELP and SESOIL  are appropriate for
modeling  dry  solid waste  in  a landfill  or landfarm
situation; they  are  not  appropriate for modeling  the
release rate of  liquids from lagoons,  landfills,  or
landfarms.  Rainstorms come in  discrete  intervals
separated  by dry  periods. Using steady  state
equations  to  model  rainfall-induced leaching,
however,  assumes  that  1/365th of  the  annual
recharge  occurs each day. Although  this is  an
assumption, it  is felt to be  a useful  one for most
cases. Most abandoned hazardous waste sites have
received  liquids in the past; very  few  have received
only  dry  solids.  Hence,  the  question  of the
assumption of steady state  conditions  is relatively
moot. For the  bulk of  the  modeling situations (liquid
wastes), the steady state and  the instantaneous rates
are the same,  and since the  steady state equations
are simpler, they are the method of choice.

For lagoons, the analyst should use the concentration
of contaminant  in the lagoon as the concentration of
the contaminant leaving the  lagoon, since the
"leachate"  is the  waste itself. The waste leaves  the
lagoon by  percolating  through the clay  liner or  the
native soil,  or  it permeates   the  flexible membrane
liner (FML).

For landfills, the  analyst should use  the equilibrium
solubility  of the solid waste,  assuming that the
contaminant will  have fully  equilibrated with the
percolating  rainwater.  The use  of the  equilibrium
solubility  concentration  as the leachate concentration
is an  assumption, it is based  on a typical residence
time  of 21 years  for  rain  percolating  through a
covered (1  09-7  cm/sec) secure  landfill. The
assumption is that the  time used for determining  the
equilibrium  solubility of the  chemical is much shorter
than  the  residence time  in  the   fill. If the fill  is
uncovered  (or covered with a permeable cover),  the
travel time through  the  landfill may be too short for
the above  assumptions to  be valid. In these cases,
the analyst should calculate the travel time  and
compare it to the time used in  the solubility test. If the
travel time is not longer than the test time, the analyst
should estimate  the leachate  concentration  as a
fraction of the equilibrium solubility concentration.
Additionally, the above  assumptions assume a landfill
of only one waste stream,  if the fill has only  a  small
quantity of  the subject waste in it,  the contact time is
the time  for travel  through the  isolated  material.  In
these  conditions, the  leachate concentration will
typically be a fraction of the equilibrium solubility. The
analyst may wish, in  some instances,  to model the
solubility  of the contaminant within  a complex
leachate. In this case, the  solubility of a  hydrophobic
                                                  29

-------
contaminant can be increased by the organic fraction
of the complex leachate.

For  landfarms,  the  assumption  that adequate
residence time  is available for contaminants to reach
equilibrium  solubility may  not  be  viable,  and  the
analyst should  estimate the degree  of solubilization.
This can be done by dynamic modeling of the kinetics
of dissolution, or it can  be approximated  based on
experience and engineering judgment. Because of the
complexities of dynamic  modeling, this  approach
usually is not worth the slightly  increased  accuracy
gained, especially since other parameters may affect
the accuracy of the  final  answer.  Concentration is
typically  estimated  as a  fraction  of the equilibrium
solubility.

The  volumetric flux of contaminated water can be
calculated in two ways, one for solid wastes and  one
for liquid wastes.

(a)  For  landfilled solids,  the  only liquid present is
water percolating into the  fill. For uncovered landfills,
this  can range from the infiltration  fraction  of  the
rainfall, to the full precipitation  (if  no rain runs off of
the fill  before infiltrating), to larger flows of water if the
site  is exposed to  stormwater run-on  from  an
adjacent area.  For covered landfills, the  infiltration
fraction may be limited  by the  permeability  of  the
cover.  Typically in wet climates the cover permeability
is limiting, while in dry climates the permeability does
not limit percolation, and normal soil percolation ratios
can be used.

The  loading rate to ground water can be calculated
with the following equation:
where
                                          (2-32)
            contaminant loading rate, (mass/time).
             percolation  rate,  see  Equation 3-14
             for calculation of q, (length/time).
            area of landfill, (length squared).
             solubility   of  solid chemical,
             (mass/volume).
(b)  For  lagooned or  landfilled  liquids,  precipitation
has  a minimal influence on  leachate generation,  as
liquid waste will  percolate to the watertable under the
influence of gravity. The rate-determining  step is the
permeability of the liner or underlying soil (if there is
no liner). For liquids, the following form of Darcy's law
should be used to estimate the volumetric flux leaving
the site.
   =  K*i*A
(2-33)
                Q1     = volume loading  rate,  (volume/time).
                Ks     = Darcy's coefficient; for unlined  lagoons
                         use  native  soil hydraulic  conductivity;
                         conductivity (length/time) (see  Chapter
                         3 for sources of hydraulic conductivity).
                 i     =  hydraulic gradient,  (length/length).
                         Equations  2-33 will  handle  situations
                         where the liquids in  the  lagoon have  a
                         free  depth. In  many  cases  the depth of
                         the free  liquids is small, or it  is  small
                         with  respect to the distance  between
                         the  lagoon  and the  watertable (when
                         the  Ks  is for native soil).  In these
                         cases the term "i" can be taken as  1.
                A     = area of lagoon,  (length squared).

                This QJ is then used to  estimate mass  loadings
                with the following equation:
            Lc=C,*Qi

            where

                LC
                Cs
                                       (2-34)
                Q
       = contaminant loading rate, (mass/time).
       =  contaminant  concentration  in  lagoon
          fluid, (mass/volume).
       = volume  loading rate, (volume/time).
            Equations  2-33 and  2-34  model the  release  rate
            from a lagoon whether  the  flow through the vadose
            zone is saturated or unsaturated.  For unlined active
            lagoons, the flow is typically saturated all the way to
            the watertable.  For  clay-lined  lagoons, the flow is
            saturated through the liner and unsaturated between
            the liner and the  water-table (assuming no breaches in
            the liner).  Equations 2-33 and  2-34 are appropriate
            when  analyzing  lagoon  releases, but should not be
            used  for  spills or other  conditions  where  the
            chemicals on the surface do  not pond for a  long time.
            In  these conditions, the  assumption of saturated flow
            (through the liner or soil) may be violated.

            Equations  2-33  and 2-34 apply  to  liquids that  are
            mostly  water. For lagoons that contain organic fluids,
            however, the equations may need to be  corrected.
            For liquids with  a density or  viscosity that differs from
            water,  correct   Ks  for  this  different viscosity  and
            density by calculating  the  term  Kc,  using the
            following:
            K, = Kw * DC/DW * Uw/uc

            where
                                       (2-35)
                                       s term  =  hydraulic
                                          of  contaminant,
where
 Kc    =  corrected K
          conductivity
          (length/time).
Kw    =  hydraulic  conductivity of ground  water,
          (length/time).
 D    =  density  of  liquids:  c=contaminant,
          w = water, (mass/volume).
                                                   30

-------
    U   =  dynamic viscosity  of liquids:  c  = contam-
           inant, w = water,  (mass/length  * time).

and then  substituting Kc for Ks in Equation 2-33.

(2) Estimating Release Rate from  Facilities Lined with
Flexible Membranes
The release rate from an intact lined landfill or lagoon
can be calculated  for a small group of contaminants.
Failed  liners can  be modeled as a function of the
extent of the failure using the modeling  equations for
clay or natural soil-lined facilities. Although a flexible
membrane  (FML)  liner appears to allow no migration
through the barrier, it  may indeed be  penetrated by
organic compounds and contaminated water, although
the  rate of permeation  is understandably  small.  The
rate  at which  a  contaminant permeates  through  a
polymeric  material has  been  shown to be  dependent
upon  various properties of the permeant,  such as
size, shape, polarity, and other factors (Steingiser et
al. 1978).

Salame   and  others  proposed  the  use  of  a
permeability equation to predict the rate of  permeation
of liquids  and gases  through  various   polymers
(Salame 1961, 1973, 1985; Steingiser et al. 1978):
Ps = Ap0e-sH

where
                            (2-36)
    Ps


   AP


   SH


    0
 permeation  rate,  (g-mil/100
 in2*day*cmHg).
constant solely  dependent  on the type
 of polymers used,   (g-mil/100
 inn*day*cmHg).
constant solely  dependent  on the type
 of polymers used, (cc/cal).
 the polymer "permachor" calculated  for
 each polymer-permeant pair, (cal/cc).
Salame lists  values  for these  parameters  obtained
from his extensive experimental work. These values
are shown in Tables 2-7,  2-8, 2-9, and 2-10.

For permeation of water through FMLs,  polymers are
categorized into five  groups  based on the values of
the solubility  parameter as shown  in Table 2-8. This
grouping was  achieved   after  examination  of
experimental data for  about 70 different  polymers
(Salame 1985). The  solubility parameter provides an
indication of polymer  interaction with water, with  more
interaction occurring at  higher values  of the  solubility
parameter.  Examples of hydrogen  bonding for
polymer group  5  include hydroxyl  (OH) and  amide
(NHCO)  radicals as  in  nylon and  polyvinyl alcohol.
The polymer with hydrogen bonding  but  with the
value  of "delta" less than  11  does not belong to
group 5. Permachor values for some selected organic
liquids and for water are shown in Tables  2-9 and
                                        2-10,  respectively. The water "permachor" values
                                        for various polymers given  in Table 2-10 apply under
                                        dry conditions.  For  water permeation  under  wet
                                        conditions,  permachor  values  may be reduced by
                                        about 20 percent.

                                        The term P can be used to calculate the release  rate
                                        in grams/day. P is multiplied by the area of the liner,
                                        and then divided  by  its thickness. This  assumes  a
                                        normal water vapor pressure of 1  cm Hg at ambient
                                        temperature. The equation is:
                                        LC = P,*A*p/d!

                                        where
                                         (2-37)
                                            LC
                                            PS

                                            A
                                             P
                                            de
            contaminant loading rate, (mass/time).
             permeation  rate,   (g-mil/IOO
             in2*day*cmHg).
             area of liner, (in units of 100 in2).
            vapor pressure,  (cmHg).
            thickness of the  liner, (mils).
2.5.2 In-Depth Analysis
In-depth  analytical approaches  for  quantification  of
baseline contaminant release to  ground water involve
the use of computerized models. Refer to Chapter 3
of this  manual for a detailed discussion of the nature
and applications of such modeling tools.

2.5.3  Long-Term  and Short-Term Release
Calculation
For  toxic  substance release to ground-water
systems,  directly  calculate the short-term
(maximum)  release values from the measured surface
and  subsoil  contaminant  concentrations using the
tools  discussed in  this  section. Obtain long-term
(average) values by applying the procedure previously
outlined for  particulate releases  to air (see Section
2.3.3).

2.6 Soil Contamination

2.6.7 Beginning Quantitative Analysis
No estimation methods are presented for analysis  of
surface soil contamination. Site soils will  be  sampled
directly and the  degree   and extent of their
contamination delineated  during the  Remedial
Investigation.  Sampling and analysis may also  have
been conducted for subsurface soils.  In certain cases,
however, it may be  desirable to project subsurface
contamination  without conducting unsaturated  zone
sampling.  USEPA  (1987a)  covers soil sampling
strategies.

2.6.2 In-Depth Analysis
Surface soil monitoring, usually conducted during the
Remedial  Investigation,  constitutes in-depth
quantitative analysis. Subsurface (unsaturated  zone)
in-depth  analysis  will usually involve  application  of
sampling and modeling approaches. Sampling  and
                                                 31

-------
Table 2-7.   Parameter Values for Permeation Equation (at 25 °C)
                              Liquid organics in3
                                              Water in polymer categoryb
Parameter
A a-mil
100 in2 day cmHg
s (cc/cal)
0(cal/cc)
PE
1 x 104
0.506
Table 2-gc
NVC
1 x 104
0.23

1 2
11.5 10.2
0.16 0.135
Table 2-K
3
5.4 x 102
0.115
3C
4
25
0.035

5
(d)
0.099

aSource: Salame n.d.; Salame 1961.
bSee Table 2-8 regarding polymer category. Source: Salame 1985.
cSee the table indicated for these values.
dA = 0.33 exp (0.064 xO2), where o is the solubility parameter, (cal/cc)1'2.
Table 2-8.    Polymer Categorization for Permeation of
            Water
 Polymer group
Categorization
       1       Any polymer with 8 < 8.9a
       2       Any polymer with 8.9 < 8 < 10
       3       Any polymer with 10 < 8 < 11
       4       Polymer containing nitrile (CN) group with 10.5 < 8
 	5	Polymer with H bonding and 11 <  8	

 a8 = Solubility parameter (cal/cc)1/2
 Source: Salame 1985.
 analysis can provide  a  direct  quantification of  the
 degree  of  contamination  in subsurface soils.
 Alternatively,  computer models  (e.g.,  SESOIL;
 Bonazountas and Wagner 1981) are used to project
 the level of unsaturated zone contamination over time
 from surface placement of toxics.  Refer to  Chapter 3
 of this manual  for a detailed discussion  of computer
 models  that can be  applied  to the unsaturated  zone
 contamination estimation.
Table 2-9.    Permachor Values of Some Organic Liquids in
            Polyethylene and PVCa
                    In nonpolar polymer   In polar polymer
       Liquid
Acetic acid
Benzaldehyde
Benzene
2-Butoxy ethanol
Butyl acetate
Butyl alcohol
Butyl ether
Butyraldehyde
Capryllc acid
Carbon tetrachloride
p-chlorotoluene
Cyclohexane
Dibutylphthalate
Diethylamine
Ethanol
Heptane
Hexane
Methyl ethyl ketone
Methanol
Nitroethane
i-Pentyl propionate
i-Propyl amine
Trichloroethylene
o-Xylene
p-Xvlene
13.0
15.9
5.4
24.4
13.0
18.0
10.4
13.5
19.0
5.8
7,6
7.0
31.4
10.0
16.0
7.0
6.0
12.5
15.0
15.4
15.0
11.0
5.4
9.4
7,4
44.0
4.0
7.0
75.0
5.0
50.0
46.0
0.0
50.0
22.0
7,5
45.0
17.0
5.7
48.0
44.0
43.0
1.0
47.0
7,0
7,0
6.7
3.0
11.0
9.0
                                                            aPolyethylene and PVC are nonpolar and polar polymers,
                                                             respectively.
                                                            Sources: Salame n.d.; Steingiser et al. 1978.
                                                        32

-------
Table 2-10.   Water Permachor Value for Dry
              Polymers
                                          Permachor
	Polymer	value (0)
Polyvinyl alcohol                              TCU
Polyacrylonitrile                               109
Cellulose (dry)                                97
Polyvinylidene chloride                         87
Polycaprolactam  (dry)                         80
Polyacrylonitrile styrene  (70/30) (Lopac)        76
Polyacrylonitrile styrene/butadiene              75
(70/23/7) (Cycopac930)
Polychlorotrifluoroethylene                     71
Polyethylene terephthalate                     68
Polyvinylidene fluoride (Kynar)                 67
Polyacrylonitrile styrene/ = tibutadiene          65
(56/27/4/13) (Cycopa\c 920)
 Polyvinyl chloride                             62
 Polyoxymethylene (Delrin)                     57
 Polymethyl  methacrylate                       55
 Polyvinyl acetate (dry)                         45
 Polystrene/acrylonitrile  (74/26)                 45
 Polyethylene  (HD)                             40
 Polysulfone                                    34
 Polypropylene                                 33
 Polycarbonate (Lexan3)                        33
 Polystyrene                                    28
 Polyethylene  (LD)                              26
 Polyisobutylene                                17
 Polyethylene/vinyl  acetate  (85/15)              15
 Polybutadiene                                  8
 Polymethyle  pentene (IPX)                     8
 Polydimethyl  siloxane (dry)                     -4

 Sources:  Salame 1961; Salame n.d.;  Steingiser et  al.
 1978.
                                                              33

-------
 Preceeding Page Blank
                                             Chapter 3
                                   Contaminant  Fate Analysis
3.1 Introduction

This chapter provides  guidance for  evaluating the
transport, transformation, and fate of contaminants in
the environment following their  release from an
uncontrolled hazardous waste site.  The contaminant
release rate estimates described  in the previous
chapter  provide  the  basis for contaminant fate
analysis.  The results form the  basis  for subsequent
analysis of exposed populations and estimation of the
levels of exposure  incurred (see Appendix  A). The
goal of contaminant fate analysis  is  to  identify off-
site areas affected  by  contaminant  migration and to
determine contaminant concentrations in these areas.

The following sections address analysis  of
atmospheric fate,  surface water fate, ground-water
fate, and  biotic fate. Within each of those sections,
contaminant transport is addressed  (except for  biotic
fate analysis,  which does  not  involve  contaminant
transport). A screening analysis  is conducted  to
provide an initial qualitative  assessment  of
contaminant transport in  the  environment.  It  is
designed  to  (1)  identify  each transport  process
governing the  movement of  various contaminants
within and among environmental  media, (2) determine
the  direction  and roughly gauge the  rate  of
contaminant movement  from the site,  and  (3) identify
areas to which contaminants have  been or may  be
transported.  Screening  analysis  is  designed both to
provide  initial organization  and  direction for
subsequent in-depth analysis of contaminant
environmental transport, and to  provide  a consistent
basis for analysis from site to site.

When likely pathways of contaminant  migration  have
been identified by screening analysis, those pathways
requiring  further  evaluation  are  quantitatively
addressed. Like analysis of contaminant release, this
analysis  can involve either the use of "desktop"
analytical solutions or numeric methodology.

Simplified environmental  fate  estimation procedures
are  based  on the predominant  mechanisms  of
transport within each  medium,  and  they generally
disregard  intermedia  transfer or transformation
processes.  In  general, they  produce conservative
estimates  (i.e.,  reasonable upper  bounds)  for final
ambient concentrations and the extent of hazardous
substance  migration.  However,  caution should  be
taken to avoid using inappropriate analytical  methods
that  underestimate or overlook significant  pathways
that impact human health.

When  more in-depth  analysis  of  environmental  fate
is required, the analyst must select  the  modeling
procedure  that is  most appropriate to  the
circumstances. In general, the more  sophisticated
models are  more data-, time-,  and resource-
intensive.

The  following criteria should  be considered  when
selecting an  in-depth  environmental fate model or
method:

   Capability of the  model to  account for  important
   transport,   transformation,  and transfer
   mechanisms;

*  "Fit"  of the model  to  site-specific and
   substance-specific   parameters;

x* Data requirements of the  model,  compared to
   availability and reliability of site information; and

x  Form and content  of model output.  This  refers to
   the model's ability  to address important questions
   regarding human exposure  or environmental
   effects and to provide all data required as input to
   further analysis.

Information  regarding  the major environmental
processes  that  may  affect the  fate  of hazardous
substances in each  medium is provided. These
processes  include  transformation  and  intermedia
transfer mechanisms,  as well as the more  complex
transport mechanisms  that are not incorporated  into
estimation procedures.  By comparing  the list of
important  processes identified  for the  site  with  the
summary of model features presented  at the end of
each section, the analyst can select the  model best
suited to the requirements of the site.

The  Graphical  Exposure  Modeling  System  (GEMS),
developed by the EPA's  Exposure Evaluation Division
(EED), Office of Toxic Substances (OTS) is  a  set of
                                                 35

-------
computer  models that  is easily accessible  and has
the ability to produce sophisticated  analyses  of
environmental  fate. GEMS  consists of models
capable of assessing contaminant  fate in air, surface
water,  ground water, and  soil. These fate models
contain pertinent data files  (including  nationwide soil,
land use, and meteorological data, and data on many
major  river  systems, lakes,  and  reservoirs);  user-
input  data  manipulation and storage  capabilities;
statistical  processing  programs; and such  graphics
capabilities as presentation of results in map  form.

GEMS is designed to be  user-friendly.  Although
environmental fate  modeling  experience   is  highly
desirable,  personnel with no  computer  programming
background  can  also use the system because of  its
progressive  menu  and  user prompting formats.  At
each decision point, the user is presented  with a  list
of  possible  selections. When  specific  data are
required to activate a program, the system requests
each type of data  needed and the units required.  At
any point in  the procedure,  the user can request help
from the  system,  and  a clear explanation  of the
choices or steps  facing the user is  provided.

The GEMS  host computer  is a Vax-11/780, which is
located at the  EPA National Computer System  at
Research  Triangle  Park, North Carolina. The system
can be accessed and used  with the following terminal
types:  DEC UT-100 series, Tektronix  4014  series,
and ASCII.

Terminals  must  be capable of transmitting  or
receiving ASCII data in full  duplex  mode, using even
parity and  seven-bit  data word  length, with
communication rates of 300 or 1200 bits per second.
Most common  acoustic modems are compatible
(GSC 1982).*

Monitoring  data can also  be useful  in analyzing
contaminant transport and fate. Monitoring results can
provide, however, only a measurement of the existing
extent  of  contamination. In addition,  monitoring data
alone may not  allow the analyst  to discriminate the
contributions of  specific  sources to  measured
contaminant loadings. In  all assessments,  some
degree of modeling contaminant movement within and
among  environmental media will be  necessary  to
predict the associated  exposure  over a 70-year
lifetime.  Thus,  a   combination  of monitoring and
modeling techniques will be necessary  to conduct  an
analysis of contaminant fate for exposure assessment
purposes.

For in-depth guidance in  selecting and  running a
computer model to use  in  analyzing  contaminant
migration from  a  particular site, the analyst  should
review the following guides:
 -USEPA(1977a):
 USEPA  (I986b):
 -USEPA(1987d):
 -USEPA (1986a):
Guidelines  for Air Quality
Maintenance Planning  and
Analysis,   Volume  10
(Revised):  Procedures  for
Evaluating Air Quality Impact
of New Stationary Sources

Guideline  on  Air Quality
Models  (Revised)  1986 and
Supplement A (1987)

Surface  Water  Model Se-
lection Criteria

Ground  Water  Model Se-
lection Criteria
  * Contact personnel within the EED are Ms. Patricia Harrigan,
  Mr. Loren Hall, or Mr. Russell Kinnerson. They can be reached
  at EPA, Washington, D.C., (202) 382-3931.
 - USEPA (1985J):      Modeling Remedial Actions

In addition, it is recommended that the analyst obtain
the user's  manual  for any model  selected  before
attempting its application.

For contaminant fate in estuaries  and reservoirs, the
analyst should review Mills et al. (1982).

To  evaluate  the  retardation of contaminant plumes
composed of mixed wastes in  ground-water systems
the analyst is referred to  the following references for
detailed guidance:  Nkedi-Kizza et al. (1985), Rao et
al. (1985), Woodburn et al. (1986).

3.2  Contaminant  Fate Screening

Figures 3-I through 3-4 present  the  decision
networks for screening  contaminant  fate in  air,
surface water, ground water, and biota. Any migration
pathways (identified in the qualitative evaluation)  that
will  require  additional analysis are described in
Sections 3.3 through 3.6. These pathways will be
further evaluated to determine the likelihood of
population exposure as described in Appendix A.

In Sections  3.2.1  through 3.2.4, brief  guidance is
provided for the qualitative evaluation  of contaminant
migration pathways. The paragraphs presented below
are  keyed  to the  accompanying decision  networks
and  are intended to provide  further elaboration of
those boxes in the decision networks.

3.2.7 Atmospheric Fate
The following numbered  paragraphs each  refer to
particular numbered boxes in the  Figure  3-1.

1. The  atmospheric fate  of  contaminants must be
assessed whenever  it is  determined  that significant
                                                  36

-------
            Figure 3-1.     Environmental fate screening assessment decision network: atmosphere.
                                                                              Contaminant Release
                                                                             Screening Assessment
                                            Potential Volatilization of
                                             Contaminants from Site
                                                                   _E
                                                                                          Potential Release of Fugitive Dust/
                                                                                         Contaminated Particulates from Site
                                   Consider Direction and Rate of Contaminant
                                          Migration within Air Medium.
                                  Major Mechanisms: Wind Currents, Dispersion
                                                                            s.
                                                                                          Consider Direction and Distance of
                                                                                      Paniculate Movement with Wind Currents.
                                                                                    Major Mechanisms: Wind Speed, Particle Size.
                                                                                         Gravitational Settling, Precipitation.

                                                                                                           	El
CO
   Will Settleout and Rainout
  Potentially Result in Sufficient
   Soil Contamination to Bring
About Leaching to Ground Water?
                                                      Will Contaminants Potentially
                                                       Reach Agricultural, Hunting,
                                                            or Fishing Areas?
                                         Yes
                                                                                  J3
                                                                             Determine Probable Boundaries
                                                                               of Elevated Concentrations.
                                                        No
                                                                               Yes
Consider Contaminant Transfer
 to Ground Water. Assess Fate
       Associated with
 This Medium (See Figure 3-3)
  Consider Transfer of Contaminants
to Biota Used by Humans. Assess Fate
  with This Medium (See Figure 3-4)
                                                                     J
            Will
Contaminants Potentially Reach
    Surface Waterbodies?    .-
                                                                                                                                                            Yes
                                                                                                        Identify
                                                                                              Populations Exposed Directly
                                                                                             to Atmospheric Contaminants
                                                                                                   (See Appendix A)
                                                                                                                     Consider Transfer of Contaminants
                                                                                                                       to Surface Water. Assess Fate
                                                                                                                             Associated with
                                                                                                                       This Medium (See Figure 3-2)

-------
gaseous or airborne particulate  contaminants are
released from the site. The atmospheric  fate of
contaminants  released originally to other  media, but
eventually partitioning to the  atmosphere beyond site
boundaries, must  also be assessed whenever this
intermedia transfer is  likely to  be significant.

2. The predominant  directions of contaminant
movement  will be  determined  by  relative directional
frequencies of wind  over the site  (as reflected in
area-specific wind  rose  data).  Atmospheric  stability
and  wind  speeds  determine  off-site areas affected
by ambient concentrations of gaseous contaminants.
Usually, high  stability and low wind speed conditions
result in  higher  atmospheric concentrations of
gaseous contaminants close to the site. High  stability
and  moderate wind speeds result in  moderate
concentrations over  a larger  downwind  area.  Low
stability or  high wind  speed conditions cause greater
dispersion  and dilution of contaminants, resulting in
lower concentrations over larger areas.

For  particulate  contaminants  (including  those
adsorbed  to dust  or soil  particles), ambient
concentrations in the  atmosphere and areas affected
by  airborne  contaminants  are determined by
windspeed  and  stability  and  also  by  particle  size
distribution. High winds  result in  greater dispersion
and  cause particulates to remain  airborne  longer
(which may also increase  release  rates).  Low winds
and  high  stability  will result  in  rapid settleout of
particulates and in  a  more concentrated contaminant
plume  closer to  the  site. Larger  particles will settle
rapidly,  decreasing the  atmospheric  concentrations
with  distance from the site. Finer particles  will remain
airborne longer, and  their behavior will more closely
approximate  that of gaseous contaminants, as
described above.

3. Settleout and rainout  are  important mechanisms
of contaminant transfer from the atmospheric media
to both surface  soils and surface waters. Rates of
contaminant transfer  caused  by these mechanisms
are  difficult to assess qualitatively; however,  they
increase  with  increasing soil  adsorption coefficients,
solubility  (for particulate contaminants  or those
adsorbed  to particulates), particle size,  and
precipitation frequency.

Areas  affected  by  significant  atmospheric
concentrations of contaminants  exhibiting  the above
physical/chemical  properties  should also be
considered as potentially affected  by contaminant
rainout and settleout  to surface media. Contaminants
dissolved in rainwater may percolate  to ground water,
run off or fall directly  into surface waters, and adsorb
to unsaturated soils. Contaminants  settling to the
surface through  dry  deposition may dissolve  in or
become suspended  in surface waters, or may be
leached into unsaturated  soils and ground water by
subsequent rainfall. Dry deposition  may also result in
formation of a layer of relatively high contamination at
the soil surface.  When such intermedia transfers  are
likely, one should assess the fate of contaminants in
the receiving media.

4. If areas identified  as  likely to receive  significant
atmospheric  contaminant  concentrations  include
areas  supporting edible biota,  the  biouptake  of
contaminants  must  be considered  as a  possible
environmental  fate pathway.  Direct biouptake from
atmosphere  is  a potential  fate mechanism  for
lipophilic contaminants. Biouptake from soil or water
following transfer  of  contaminants to  these  media
must  also  be  considered  as  part of  the  screening
assessments   of these  media; for  example,
hexachlorobenzene  was found  to accumulate  in
plants  (Russell et al.  1971, Gillet 1980, Trabelka and
Garten 1982).

3.2.2 Surface Water Fate
The following  numbered paragraphs  each refer to
particular numbered boxes in the Figure 3-2.

1. The aquatic  fate  of contaminants  released from
the CERCLA site as well  as those  transferred  to
surface water  from  other media  beyond site
boundaries must  be considered.

2. Direction of  contaminant movement will usually
only be clear for  contaminants  introduced  to rivers
and streams. Currents,  thermal stratification  or
eddies,  tidal pumping, and flushing in  impoundments
and estuaries  render qualitative  screening
assessment of  contaminant directional  transport
highly  conjectural for these types of waterbodies.  In
most  cases,   entire waterbodies receiving
contaminants   must  be considered potentially
significant  human exposure points.  More  in-depth
analyses or survey data may subsequently identify
contaminated and  unaffected  regions  of  these
waterbodies.

3. Similarly, contaminant concentrations in  rivers  or
streams can be  roughly assessed  based  on rate of
contaminant introduction and  dilution  volumes.
Estuary or impoundment concentration regimes  are
highly  dependent on the transport  mechanisms
enumerated above.  Contaminants may be  localized
and  remain concentrated,  or disperse rapidly  and
become  diluted to  insignificant  levels. The
conservative approach is to conduct a  more in-depth
assessment and  use  model  results or survey data as
a  basis for determining  contaminant  concentration
levels.

4. Important  intermedia  transfer mechanisms  that
must be considered  where significant  surface water
contamination is  expected include transfers to ground
water  where hydrogeology of the area  indicates
significant surface-water/ground-water exchange;
transfers to biota where waters contaminated with
                                                 38

-------
           Figure 3-2.    Environmental fate screening assessment decision network: surface water.

                                                                              Contaminant Release
                                                                             Screening Assessment
                                                                        Potential Release of Hazardous
                                                                       Substance to Surface Water-body
                                                                                                    IT
                                                                    Consider Direction and Rate of Contaminant
                                                                          Migration Within Waterbody.
                                                            Assess Distance Downstream, or Areas of Lakes and Estuaries

                                                             Major Mechanisms: Currents in Affected Rivers or Streams;
                                                                  Dispersion in Impoundments; Tidal Currents and   .»_
                                                                              Flushing in Estuaries               I 2
Co
CO
                   Estimate Surface Water Contaminant
                            Concentrations

          Major Factors: Source Release Strength, Dilution Volume
                  Is Exchange of Water
              Between Surface Waterbodies
              and Ground Water Significant?	
              	Li
                                                                   JL
Is Water Used for Irrigation or Watering
 Livestock, or Does Waterbody Support
 Commercial or Sport Fish Population?.-
              Consider Transfer of Contaminants
             to Ground Water. Assess Fate in This
                   Medium (See Figure 3-3)
      Consider Transfer of Contaminants
      to Biota Used by Humans. Assess
            Fate Associated with
               This Medium
              (See Figure 3-4)
                                    Is Hazardous Substance Volatile?
      Identify Human
Populations Exposed Directly
     to Surface Waters
     (see Appendix A)
  Consider Transfer of Contaminants
to Air Medium. Assess Fate Associated
         with This Medium
          (See Figure 3-1)

-------
lipophilic  substances  support  edible biotic species;
and transfer to the  atmosphere where surface water
is contaminated by  volatile substances. High
temperatures,  high  surface-area-to-volume  ratios,
high  wind conditions, or turbulent stream flow also
enhance volatilization rates.

Contaminant transfer to bed  sediments represents
another significant transfer  mechanism,  especially in
cases where  contaminants are  in the form  of
suspended  solids,  or are dissolved,  hydrophobic
substances  that can  become adsorbed  by  organic
matter in  bed sediments. For the purposes  of this
manual, sediments and water are considered part of a
single system  because  of  their  complex
interassociation. Surface water/bed sediment transfer
is reversible; bed sediments often act as temporary
repositories  for  contaminants  and  gradually  re-
release contaminants to surface  waters.  Sorbed  or
settled contaminants are frequently transported with
bed sediment migration or flow. Transfer of  sorbed
contaminants  to bottom-dwelling,  edible  biota
represents a  fate  pathway potentially  resulting  in
human  exposure. Where this transfer  mechanism
appears likely, the biotic fate  of contaminants should
be assessed.

3.2.3  Soil and Ground-water Fate
The  following numbered paragraphs  each refer  to
particular numbered boxes in  Figure 3-3.

1. The fate of contaminants  in  the soil medium is
assessed whenever  the  contaminant release
atmospheric or fate screening assessments  results
show  that  significant contamination of soils is likely.

2. The most significant contaminant movement  in
soils  is a  function of liquid movement.  Dry,  soluble
contaminants  dissolved in precipitation,  run-on,  or
human-applied water will migrate through percolation
into the soil. Migration  rates  are  a function  of net
water recharge rates and contaminant solubility.

Liquid contaminants may percolate directly into soils.
Organic liquids may  alter soil permeabilities or may be
of lower  viscosity and/or higher density  than water,
resulting in percolation rates many times  greater than
that of water. Contaminants with high soil  adsorption
coefficients may bind to soils  and become relatively
immobile.

3. Important  intermedia transfer  mechanisms
affecting soil contaminants include volatilization  or
resuspension to the atmosphere and biouptake  by
plants and soil organisms.  These, in turn, introduce
contaminants to the food chain.

4. The fate of contaminants in  ground  water is
assessed whenever  site contaminant release
screening analysis indicates  direct introduction  of
contaminants to ground water  (e.g., through disposal
wells or fluid releases to an aquifer near the ground
surface), or whenever the screening assessments of
atmospheric, surface water, or soil contaminant fates
(as outlined  above) indicate  potential  contaminant
transfer to ground water.

5. The  qualitative assessment  of ground-water flow
is often  based on  the  assumption that subsurface
hydrologic gradients (which determine flow  directions
and  rates)  approximate  surface topography.  This
approach is unreliable and should be used only in the
absence of hydrogeologic data. Ground-water flow is
influenced  by many  factors  including  hydraulic
conductivity  of soils, hydraulic gradient,  presence of
subsurface impermeable  barriers, presence  of
discharge areas  (e.g.,  streams intercepting ground-
water flow) and  presence of fissures, cavities,  or
macropores.  Hydrogeologic  survey data (where
available)  provide  a  more  reliable basis for
contaminant transport  assessment  than  do surface
topographs.

6. Site and  surrounding  community survey  data
describing the location of wells are compared with the
expected subsurface contaminant plume boundaries
to identify locations of potential exposure  points.

7. Important mechanisms of contaminant transfer
from ground water  to  other environmental media
include  contaminated water  exchange  between
surface  waters  and ground water and uptake  of
contaminants by  edible biota.  The former mechanism
must  be considered whenever surface  waters are
downgradient from the  CERCLA site;  it  increases in
likelihood  with closer  proximity of these surface
waters to the site. Available hydrogeologic information
for the site and surroundings  should be  reviewed for
any indication  that the  aquifer underlying the site is
connected to surface waters.

The  second  major intermedia  transfer  mechanism,
biouptake,  may   occur  through  two  pathways: (1)
direct exposure   of plants and  lower trophic  level
animals  to contaminated ground  water in regions
where  the  ground-water  level is close to  or at the
soil surface (e.g., marshy  areas, areas  adjacent to
aquifer discharge  points),  and (2) biotic exposure to
ground water resulting from human  activities such as
irrigation or watering of  livestock with well water.

3.2.4 Biotic Fate
The  following numbered  paragraphs each refer to
particular numbered boxes in  Figure 3-4.

1.  A  screening  environmental fate assessment for
the biotic  medium is  performed  after the fate  of
contaminants in  the  atmosphere,  surface waters,  or
ground  water has been  assessed.  Starting  with the
expected distribution of contaminants in each of these
media,  potential  points of biotic  contact  with
                                                 40

-------
       Figure 3-3.     Environmental fate screening assessment decision network: soils and groundwater

                                                                                Contaminant Release
                                                                               Screening Assessment
                                                                                                              Release to Soils at
                                                                                                              or Surrounding Site
                                                                                                        Consider Rate of Contaminant
                                                                                                       Percolation Through Unsaturated
                                                                                                      Soils, Based on Soil Permeabilities,
                                                                                                       Water, or Liquid Recharge Rates.rj
                                          Release to Ground Water
                                                Beneath Site     _
                                       Will Contaminants
                                          Potentially
                                     Reach Ground Water?
                                   Does Contaminated
                                       Soil Support
                                    Edible Species?
                                     Consider Direction and Rate of
                                   Ground Water Flow, Using Available
                                   Hydrogeologic Data, or by Assuming These
                                   Will Approximate Surface Topography,
                                       No
             Are Contaminants Volatile?
               Are Contaminants in
              Fine Paniculate Form on—
             Sorbed to Particulates? I3
t

Could Contaminants Reach
A Surface Waterbody?
[7
\
t


Could Contaminants Reach
Any Wells Located
Downgradient? rr-

i '
F

1
Is Plume Sufficie
Surface to Allow Dir
of Contaminated
i

  Consider Transfer of Contaminants
to Surface Water Medium, Assess Fate
   in This Medium (See Figure 3-2)
                                                    M
                                                                                               !Ez
                                                  Is Well Water Used for Irrigation
                                                     or Watering of Livestock?
                                                  	             nl
       Identify
  Human Populations
  Directly Exposed to
Well Water (Appendix A)
                                                                                                       Yes
  Consider Transfer of Contaminants
To Biota Used by Humans. Assess Fate
    Associated with This Medium
          (See Figure 3-4)
  Consider Transfer of Contaminants
to Atmosphere. Assess Fate Associated
  with This Medium (See Figure 3-1)
       Identify
  Human Populations
Exposed Directly to Soils
     (Appendix A)     ;

-------
contaminated media and important  affected  biotic
species are identified.

2.   Important species are those used directly by man
(game animals, sport or commercial fish, crustaceans
and mollusks,  agricultural  crops  and livestock;
naturally-occurring  fruits,  herbs,  other edible
vegetation), and those that introduce contaminants to
species used by man  through  the  food chain  (e.g.,
livestock  feed  crops; or plants and lower  trophic-
level animals consumed by any of the animal groups
listed above).

3.  Assessed mechanisms of transport  in  the  biotic
medium  include  the  food  chain,  natural animal
migration,  or human  commercial activity. Food  chain
transport can  result  in high concentrations  of
contaminants in the  tissue of edible  species not in
direct contact with contaminated air or water. Human
commercial transport and natural migratory behavior
of  contaminated  species  can result in  a  wide
distribution of edible  species  or  tissue-containing
contaminants.

4.  Edible  tissue concentrations  are  a function of the
level and  type of biotic  exposure to contaminants, the
partitioning  of  contaminants  between organic  tissue
and substrate  media, the  biodegradability  of
contaminants,  organism-specific metabolic  charac-
teristics, and ecosystem characteristics.

3.3 Quantitative  Analysis of Atmospheric
Fate

3.3.1 Screening Analysis
The atmospheric  fate  of substances released  from
uncontrolled  hazardous waste sites can  be estimated
by  using  the following  equation to  estimate ground-
level atmospheric concentrations  of  pollutants  at
selected  points  on a centerline of a plume directly
downwind from a  ground-level source (Turner 1970):
                                             Figure 3-4.
   C(X) = -
           Q
         no a u
           y *
                                            (3-1)
             Environmental fate  screening assessment
             decision network: food chain.

            Ambient Contaminant Concentration
              and Distribution Estimates from
             Air, Surface Water, Ground Water
                 Screening Fate Analyses
                                                                 Potential Biotic
                                                                  Exposure to
                                                                 ContamlmBTte
                                                       Consider Biotic Species Within Areas of
                                                       Elevated Ambient Hazardous Substance
                                                        Concentrations as Potential Vectors
                                                            of Hazardous Substances     r
                                                     Consider Transport of Hazardous Material
                                                            Within Biologic Medium
                                                   Major Mechanisms: Human Commercial Activity,
                                                    Organism Migration, Movement of Hazardous
                                                          Material Through Food Chain.
                                                          Identify Edible Biotic Species
                                                           Affected Indirectly Through
                                                                 Food Chain               I
                                                         Assess Potential Edible Tissue
                                                         Concentrations, Distribution of
                                                           Contaminated Organisms
                                                                   Identify
                                                           Exposed Human Populations
                                                                 (Appendix A)
where
  C(X)

     Q
= concentration of substance  at  distance
   x from site,  (mass/volume).
= release rate of substance  from site,
   (mass/time).
= dispersion  coefficient in  the lateral
   (crosswind)  direction, (distance).
= dispersion coefficient  in  the  vertical
   direction,  (distance).
= mean  wind speed, (distance/time).
= the value  pi =  3.14.
 The  appropriate  dispersion  coefficients  can  be
 obtained  from Figures  3-5  and  3-6.  These figures
provide  values for 6y  and  6Z,  respectively, as
functions of  downwind  distance,  x, and stability
classes A though F.  These stability classes  are based
on the Pasquill stability classification system, where
Class A  is very unstable and  Class F is moderately
stable (Pasquill 1961). Table 3-1  presents a  brief
illustration of how stability classes are defined.
                                                    42

-------
Figure  3-5.    Horizontal dispersion coefficient as a function of downwind distance from the source (from Turner 1970).


 10,000-
  1,000-
                                                              P
 e
 
-------
Figure 3-6.    Vertical dispersion coefficient as a function of downwind distance from the source (from Turner 1970).
    1,000
       10-
                                                       B-^
                                                                             E --
                                                                                    Fj_.
        1.0'
            0.1
1                                  10

       Distance  Downwind,  km
                                                                                                                         100
  "Curves  designated A through  F  represent  dispersion coefficient functions for  atmospheric  stability classes A  through  F.  See  text
   for sources of atmospheric stability data.
                                                             44

-------
 Table 3-1.    Key to Stability Categories
                                       Night
                                   Thinly
 Surface wind                       overcast or
  speed at a Day incoming Solar radiation > 4/8 Low   < 3/8
 Height of 10 	(insolation)	  cloud     Cloud
m (m/sec)
<2
2-3
3-5
5-6
>6
Strong Moderate Slight
A A-B B
A-B B C
B B-C C
C C-D D
C D D
Cover
E
D
D
D
Cover
F
E
D
D
 The neutral class (D) should be assumed for all overcast
 conditions during day or night.
 "Appropriate insolation categories may be determined through the
 use of sky cover and solar elevation information as follows:

Sky cover
4/8 or Less or
Any Amount of
High Thin Clouds
5/8 to 7/8 Middle
Clouds (7000
feet to 16,000
foot base)
5/8 to 7/8 LOW
Clouds (less than
7000 foot base)

Solar
elevation
angle > 60°
Strong
Moderate
Slight
Solar
elevation
angle < 60°
but > 35°
Moderate
Slight
Slight
Solar
elevation
angle < 35°
but > 15°
Slight
Slight
Slight
  Source: USEPA 1977b

To  obtain the maximum hourly concentration, select
the calculational methodology for  coning  and fanning
plumes in USEPA (1977b). To obtain the  estimated
maximum concentration  for  a 3-,  8-,  or 24-hour
averaging time,  multiply the l-hour maximum  by  the
factors given in  USEPA (1977b).

To estimate   long-term  mean   atmospheric
concentrations,  obtain  STAR (Stability  Array) data
specific to the  site. These data provide  seasonal or
annual joint frequencies for each  stability class, wind
direction, and  wind  speed  category. Assume  an
annual average wind speed of 3  meters/second, and
calculate the  long-term  mean  atmospheric
concentration  for each  exposed   population by
applying  a  weighted average, based  on the relative
frequency of each  stability class and of  wind  flow
toward selected  exposure  points.   Equation  3-2
provides a  rough weighted average estimate (Turner
1970):

C   (x) = W(x) [CA(x)fA + CB(x)fB + Cc(x)fc

+ CD(x)fD + CE(x)fE + CF(x)fF]               (3 - 2)
where
C(x)
=  average concentration at point x over
   long term.
                                           CA(x)     = concentration at point  x during  stability
                                                        class A (from Equation 3-1).
                                               fA     = relative  annual frequency  of stability
                                                        class A for the specified wind direction.

                                          and subscripts  B  through  F represent the  various
                                          stability classes.

                                          Note that this  estimate is  a rough approximation
                                          because  it  is simplified  by the assumption  that the
                                          mean wind speed  is 3 meters/second for  all  stability
                                          classes. A more sophisticated estimate can be made
                                          by  incorporating site-specific  wind  speed frequency
                                          data,  and  performing  similar  weighted average
                                          calculation of ambient concentrations. This is a time-
                                          consuming procedure,  however,  and the  use  of
                                          computer-based  estimation  procedures may  be
                                          more cost-effective  if sophisticated estimates are
                                          required.  STAR data are available from the  National
                                          Climatic  Center (NCC), Asheville,  North Carolina
                                          (phone: (704)  259-0205) for all National Weather
                                          Service (NWS)  locations  in the U.S. The NWS Station
                                          that is most representative of the site should be used.

                                          The area  within  which  the   ground-level
                                          concentration of a  hazardous substance  is  above  a
                                          predetermined  critical concentration  (i.e.,  the plume
                                          isopleth)  can   be  described using the following
                                          procedures.  Calculate the  crosswind distance from
                                          any  point along the  plume  centerline (i.e.,
                                          perpendicular to the plume centerline) to the  isopleth
                                          boundary by Equation 3-3 (Turner 1970):
where


  C(CL)


    V(x)


   C(x)
                                                                                     (3-3)
                                                         predetermined  critical concentration
                                                         level, (mass/volume).
                                                         perpendicular distance from point on
                                                         plume centerline to the C(CL) isopleth
                                                         boundary, (length units).
                                                         concentration at plume centerline, x
                                                         distance from source, (mass/volume,
                                                         as calculated by  Equation  3-1).
                                                         lateral  dispersion  coefficient,  (length
                                                         units).
                                          Vary  the value for  x (downwind distance  from the
                                          source)  input  into Equations  3-1  and 3-3, starting
                                          at a point near the site* and increasing this value until
                                          the value for C(x)  (obtained  from Equation 3-1)
                                          equals  the  predetermined  critical  concentration
                                          C(CL). Values calculated for y  describe the isopleth
                                          boundary on either side of the plume centerline.
   Equations are  generally considered  applicable  to downwind
  distances of at least 200 m.
  W(x)    =  relative annual frequency of  wind flow
             towards point x.
                                                  45

-------
Estimate the area within a plume isopleth using Figure
3-7  which plots  the value  CfCLJ/j  (relative
concentration times wind speed versus  isopleth area,
for each stability class A through F).

All  of the preceding  simplified  equations  provide
atmospheric fate estimates based on several simple
assumptions, one of which requires special mention.
This is the assumption  that the hazardous substance
released from a site is in a  form  that can  remain
airborne indefinitely (i.e., either gaseous or consisting
of particles less than 20 microns  in diameter)  (Turner
1970).

In cases  where fugitive  dust blown  from the  site
includes  solid  hazardous substances (or soil
particulates  carrying  adsorbed  hazardous substance)
of greater diameter than 20 microns, relatively rapid
gravitational settling  of the larger  particles  occurs.
Consequently, much  of  the hazardous material
reaches the ground  before advection and dispersion
can transport and  dilute the plume as  described by
the above  equations.  Thus, areas   close  to  the
uncontrolled hazardous  site  may   experience
significant soil contamination,  and  human  exposure
points farther from  the site  may experience lower
atmospheric concentrations than  estimated by these
equations.  Hanna  and Hosker  (1980)  present  a
procedure for estimating the gravitational settling rate,
distance of travel from the  source,  and deposition  rate
of airborne particulates.

All of the  above simplified  procedures incorporate the
following additional assumptions:

•   Steady-state  condition,  i.e., windspeed is steady
    at rate u, and the hazardous substance release  is
    continuous, at  average rate Q.  Wind direction  is
    also  assumed to  be  steady;   short-term
    fluctuations are disregarded.

•   Longitudinal  dispersion  is negligible  (substance
    travels at wind speed in the downwind direction).

•   The  substance  is  refractory  (all  removal  and
    decay processes are disregarded).

•   The substance is distributed  normally, or
    according  to  a Gaussian   distribution, both
    vertically and in the crosswind direction.

•   The air environment is  homogeneous; wind
    speeds and stability are equal  at all  heights above
    the  ground, and  no obstructions to wind  flow or
    dispersion  exist  other than  at  the ground.
    Complete reflection  occurs at the ground/air
    interface.

3.3.2 In-Depth Analysis
Where  estimates of  ambient atmospheric
concentrations of  hazardous  substances  developed
by the preceding simplified  procedures  indicate  that
these concentrations pose  potential  health hazards,
more accurate, in-depth analysis of atmospheric  fate
may  be  required.  Numerous  computer models  are
available for  this purpose and  are  listed  in USEPA
(1986b).  These models  vary  in  sophistication  and
capability,   and  in  their  ability  to  incorporate
expressions describing  the  effect  of various
processes  on  the  atmospheric fate of  hazardous
substances. The most  important of these  processes
are briefly  described below.  Consider the  importance
of each of these processes to the atmospheric fate of
the  substances under analysis before selecting  a
computer model.

3.3.2.1 Intermedia Transfer
The  following  are the most important processes  that
affect the removal  of hazardous substances from the
air medium and their transfer to other sectors of the
environment.

(1) Dissolution
This is the  process whereby  hazardous substances in
the gaseous  state are dissolved  into water droplets
present in the atmosphere. This process, followed by
precipitation, distributes the  substance  over  the
surface media, and percolation  to ground  water may
follow. Direct dissolution may also occur between
gaseous  substances in the  atmosphere and surface
waters at  the air/water  interface.  Dissolution  is  a
constant,  reversible process,  the  amount of  haz-
ardous substance in  the  aqueous phase  is  de-
termined  by the partition  coefficient  of the substance
between the  gas and aqueous  phases. This partition
coefficient is  in turn a function of the vapor pressure
and  water  solubility of the  substance,  its
concentration in the air, and  temperature. See Lyman
et al. (1982) or Hanna and Hosker (1980) for methods
of estimating  this partition coefficient  and atmospheric
half-lives resulting  from dissolution/  rainout.

(2) Adsorption
Through the  process of adsorption, hazardous
substances in the  vapor  phase  become attached to
particulate  matter suspended in the  air (aerosols), or
onto  soil  particles at  the  air/soil  media  interface.
Suspended aerosols settle to surface media, thereby
removing  adsorbed  substances from the  air
environment.  The  adsorption  rate  of  a  particular
substance  is  principally a function of the number  and
surface area  of aerosols  per volume of air,  the
molecular weight of the substance  in  question, its
concentration in the air, and its  saturation  vapor
pressure.  Cupitt  (1980) provides a  method  for
estimating  atmospheric  contaminant removal rates
due  to adsorption to particulates and settleout.

(3) Gravitational Settling
This  mechanism is most important for  particulate
hazardous substances,  or  hazardous substances
                                                  46

-------
  Figure 3-7.   Area within isopleths for a ground-level source (Hilsmeir and G if ford 1962. as presented by Turner, 1970)


        10"
        10"
         107
         10"
         105
        104
        103
        10*
           10'
10"
                                                                           10'
                                                                10'
  *Curves designated A through F represent functions for atmospheric stability classes A through F. See text for sources of atmospheric
   stability data.
adsorbed  onto  suspended  particulates,  if the
particulate  matter  is more than 20 |jm in diameter.
These particles settle to the  surface media  at a rate
that is a  function  of their density,  shape,  and
diameter, and  of  wind speed (Hanna and Hosker
1980).

(4) Precipitation
Precipitation itself  is a major mechanism for removal
of particulate and  aerosol matter. Raindrops  require
particulates or  aerosols to serve  as  nuclei for  their
condensation  from  the vapor state  of water.
Moreover,  raindrops generally  remove particulates
                     and aerosols > 1.0 urn in diameter as they fall below
                     the cloud level.

                     3.3.2.2 Intramedia Transformation Processes
                     Many hazardous  substances are  subject to  decay or
                     transformation  to  other substances with  new
                     properties while entrained in the air environment. The
                     two most important of these processes are described
                     below.  While  the product of such  transformation
                     processes will  usually  have different  properties  from
                     those of the original  hazardous  substance, the new
                     substance  produced   may  also  have  hazardous
                     properties.  Cupitt (1980) provides estimates  of
                                                   47

-------
constants  that determine the rate  of each
transformation process  below,  as well as  of the
importance and  likely products of these processes,
for 46 hazardous materials. Hendry and Kenley (1979)
provide  rate constants and estimation procedures for
these processes.

(1) Photolysis
This is  the breakdown of substances  because  of
photochemical  reaction brought about  by solar
energy.  Photolysis can be direct, when the hazardous
substance  is  itself affected  by  solar  radiation,  or
indirect  when  the  hazardous substance reacts  with
other substances that have been  raised  to a  reactive
state  by solar radiation. Photolysis rates depend on
solar radiation  availability, the  light  absorption
coefficient of the  hazardous substances,  and  a
reaction yield constant (which describes the efficiency
of transformation of the hazardous substance with the
available sun energy).

(2) Oxidation
The  reaction  of substances  with oxidants  in the
atmosphere  can result in their  transformation.  The
two  most important atmospheric  oxidants are ozone
and  the hydroxyl radical.  Reaction  rate  constants for
oxidation are  chemical specific;  the  overall  rate  of
transformation  of a hazardous substance  by oxidation
depends on the  concentration of  the oxidant  and the
reaction rate constant.

3.3.2.3 The Effects  of Terrain
Features such as vegetation,  large buildings, urban
areas,  rough topography, hills, or mountains can all
profoundly affect the atmospheric fate  of  airborne
substances,  principally by altering the laminar  flow  of
transporting  wind currents. The effects  of terrain on
wind  currents  may  include  increased  turbulence,
downwash in the lee of large obstacles, or  localized
alterations in the direction  of  flow.  Because the
release  of substances from  hazardous  waste sites
usually  occurs at ground level,  the fate of these
substances is  especially susceptible to the effects  of
terrain.  Select a model  capable of  accounting for
these effects in  any case where these   listed terrain
features exist  between the site and points of  human
exposure.

3.3.3 Computer Models
Tables 3-2,  3-3, and 3-4  provide  general
information about computer-based models that could
be  appropriate  to in-depth  analysis  of  the
atmospheric  fate  of substances released from
CERCLA  sites. Table 3-2 contains resource
requirements,  references,  and sources  for each
model;  Table 3-3  summarizes  their features  and
capabilities;  and  Table 3-4 discusses  the data
requirements of  each. By comparing  the information
in these tables with identified site features, site  data
availability, final output  requirements,  and  resource
availability,  one can  select the most applicable and
cost-effective  model.

The Industrial Source Complex  (ISC) long-term
model  and  the  TOXBOX  area source  model are
presently integrated  into  the  GEMS  system. These
models  are  accessed under a  subsystem of GEMS
referred to  as the  GEMS Atmospheric Modeling
System  (GAMS). A brief description  of ISC  is
provided below.

The ISC  (Bowers  et  al.  1979) is  a  Gaussian
dispersion  model,  capable  of  estimating the
concentration and  deposition  rates of  gaseous and
particulate  pollutants around a point,  area, or line
source. Because  it  is  integrated into the GEMS
system, it  is especially useful for the analysis of the
atmospheric  fate of hazardous  substances. Based on
a  user-input release location  (in  the  form of
latitude/longitude  coordinates  or  zip  code), stored
climatological data from  the  nearest  meteorological
monitoring stations  are retrieved (GSC 1982).

The integration of  ISC with a population distribution
model  called  SECPOP  gives it the  capability of
expressing atmospheric fate of pollutants  in terms  of
numbers of  people affected at various  concentration
levels (this capability is discussed  in more detail  in
Appendix A,  Exposed Populations).

The ISC model can estimate the concentration  of
pollutants released from point,  area, or line sources.
Area sources are simulated by use of a virtual  point,
and  line sources  by a  series of points. Short-term
(hourly) or  long-term  (seasonal,  annual  average)
concentration estimates  can  be  developed,  and
gravitational settling can  be  simulated based on
user-input  half-life  data (GSC  1982).

ISC can be  used with IBM, CDC, or VAX computers.
The model  is implemented within  GEMS on EPA's
VAX 11/780 and can be accessed with a variety  of
user terminal  types.  (See Section 3.1  for access
instructions.)

3.3.4 Short- and Long-Term Concentration
Calculations
Long-term  average  ambient  air  concentrations  of
hazardous substances at human exposure points are
estimated  using the  long-term average release rate
over the time period of interest,  and  the  weighted
averaging  algorithm presented  as Equations  3-1 and
3-2. Annual average climatological data, or STAR
data including  long-term  frequencies of all
climatological parameters, should be used as input to
these equations.

Where  site-specific data  are unavailable, short-
term concentration levels  are  estimated  using the
maximum  short-term release  rate and  climatological
assumptions presented in Table 3-1. When  using
                                                 48

-------
             Table 3-2.    Resource Requirements and Information Sources: Atmospheric Fate Models
                               Model
                Description
  Resource Requirements, comments
                                                                                                                                                References, sources of documentation,
                                                                                                                                                              software
             Box Model
             Climatological  Dispersion
             Model (COM)
              Industrial Source Complex
co
              Ram
             CRSTER
xx Area Source.
xx Vertical dispersion or no vertical
   dispersion option.
xx Basic box model.
• Long-term seasonal or annual.
• Point or area  sources.
xx Gaussian plume  model.
• Simulates  nonconservative pollutants.
xx Can simulate turbulence over urban
   areas.
• Outputs long-term average
   concentrations at user-specified
   receptors.
xx Operates in both long-term and short-
   term modes.
x  Accounts for settling and dry deposition
   of particles; downwash, area, line, and
   volume sources;  plume rise as  a  function
   of downwind  distance: separation of point
   sources;  and limited terrain adjustments.
xx Appropriate for industrial source
   complexes, rural  or urban areas,  flat or
   rolling terrain, transport distances less
   than 50 kilometers, and one hour to
   annual averaging times.
  -, Steady-state Gaussian plume  model.
   Appropriate for point and area sources,
   urban areas, flat terrain transport
   distances less than 50 kilometers, and
   one hour to one  year averaging  times.
   May be used to  model primary,
   pollutants, however settling and
   deposition are not treated.
xx Steady-state Gaussian dispersion
   model.
xx Designed to calculate concentrations
   from point sources at a single location.
xx Highest and high-second high
   concentrations are calculated at each
   receptor.
xx Appropriate  for single point  sources,
   rural or urban areas, transport distances
   less than 50 kilometers, and flat or rolling
   terrain.
                                          xx Available through GEMS (see Section
                                             3.1).
                                          xx Requires stability array data.
                                          xx FORTRAN V program language; has
                                              been Implemented on the  UNIVAC 1110.
                                          j«22 K bytes storage required.
                                          x  Software available as part of UNAMAP
                                              package for $420.
r Integrated into GEMS (see Section 3.1).
< Source data:  location, emission rate,
 physical stack height, stack gas exit
 velocity, stack inside diameter, and stack
 gas temperature. Optional inputs  include
 source elevation, building dimensions,
 particle size, distribution-with
 corresponding setting velocities, and
 surface reflection.
< Meteorological data:  includes stability
 wind rose (STAR deck), average
 afternoon mixing height, average  morning
 mixing height,  and average air
 temperature.
< Available code on UNIMAP (Version 6).
< Source data:  point sources require
 location, emission rate, physical stack
 height, stack gas exit velocity, stack
 inside diameter and  stack gas
 temperature. Area sources require
 location, size,  emission rate, and  height
 of emission.
< Meteorological data:  hourly surface
 weather data  from the preprocessor
 RAMMET. Actual anemometer height is
 also required.
? Available on UNIMAP (Version 6).
? Source  data:  emission rate, physical
 stack height, stack exit velocity, stack
 inside diameter and  stack gas
 temperature.
? Meteorological data:  hourly surface
 weather data  from the preprocessor
 RAMMET. Actual anemometer height  is
 also required.
                                         Documentation:  Busse and  Zimmerman
                                         1976
                                         Software: Computer Products, NTIS,
                                         Springfield, VA.  22161
                                                                                     Documentation: Bowers et  al.  1979
                                                                                     Software:  Computer  Products, NTIS,
                                                                                     Springfield, VA. 22161
                                                                                                                                             Reference: Turner and  Novak, 1978.
                                                                                                                                             Reference:  USEPA  1977b.
                                                                                                                                                                           (Continued)

-------
           Table 3-2.    (Continued)

                              Model
                                                                     Description
                                              Resource  Requirements,  comments
                                             References, sources of documentation,
                                                          software
            Texas Climatological Model Control (TCM)*
            Texas Episodic Model (TEM)*
CJ!
o
            Model MPTER
            VALLEY"
x* Long-term (seasonal or annual).
•  Gaussian  dispersion.
•  Two pollutants per run.
•  Includes option for simulation of urban
   area turbulence classes.
•  Handles nonconservative pollutants.
•  Point or area sources.
•  Up to 2,500 receptor locations on
   downwind user-specific grid.
•  Outputs average  concentration data.
•  Steady-state model.
•  Point or area sources.
•  Short-term - 10 minutes to 24 hours.
•  Produces maximum and average
   concentrations over time periods selected
   by user.
•  User can select up to 2,500 downwind
   receptor points, according to an
   automatic or specified grid  array.
•  Handles nonconservative pollutants.
•  Up to 24 meteorologic scenarios can be
   input for a single  run.
*  Multiple point source algorithm useful for
   estimating air quality concentration of
   relatively non-reactive pollutants.
•  Appropriate  for point sources, rural or
   urban areas, flat  or rolling terrain,
   transport distances less than 50
   kilometers, and one hour to one year
   averaging times.

•  Short- or long-term.
•  Simulates plume  impact in  complex
   terrain.
x  Provides screening estimates of worst-
   case short-term concentrations.
•  Provides annual  average concentrations.
•  12-receptor grid.
xx Requires stability array data.
*  FORTRAN  program language; has been
   Implemented on Burroughs 6810/11.
•  Batch mode.
x  17 K bytes memory required.
•  Technical background in meteorology, air
   pollution useful.
   FORTRAN  program applicable to a wide
   range of computer types; has been
   Implemented on Burroughs 6810/11.
   Requires approximately 26 K bytes
   memory.
   Engineering, meteorology, atmospheric
   transport background useful.
   Source data:  location, emission  rate,
   physical stack height, stack gas exit
   velocity, stack inside diameter, stack gas
   temperature, and optional ground level
   elevation.
   Meteorological data: hourly surface
   weather data  from the preprocessor
   RAMMET. Actual anemometer height  is
   also required.
   May require careful analysis of output  by
   experienced air quality modeler.
Documentation:  Texas Air  Control Board
1980.
                                                                                                                                          Reference: Christiansen 1976.
                                                                                                                                          Documentation:  Pierce and  Turner  1980.
                                                                                                                                          Chico and Catalano  1986.
Reference: Burt 1977.
Software: Computer Products, NTIS,
                                                                                                   FORTRAN V program, applicable to wide Springfield, VA 22161.
                                                                                                   range of computers.
                                                                                                   Approximately 13 K bytes memory
                                                                                                   required.
            Sources: Bonazountas et al. 1982; USEPA 1979; USEPA 1982a.
            'These models are not EPA preferred models. They can, however, be used if it can be demonstrated that they estimate concentrations equivalent to those provided by the preferred
            models, e.g., COM, RAM, ISC, MPTER. CRSTER. for a given application.
            **Thus model is recommended for screening applications only.

-------
                        Table 3-3.    Features of Atmospheric Fate Models
en
                            TEH
                            TON'
                            COM
                            VALLEY*
                            ISC
                            BOXMOO
                            RAM
                            GRSTER
                            MfTER
                         Source: Bonazountes et al. 1982; USEPA 1979; USEPA 1982a.
                          **This model is recommended for screening applications only.
                         *** These models are not EPA preferred models. These models can be used if they can be
                           demonstrated to  estimate  concentrations equivalent to those provided by the preferred
                           models, e.g., CPM, RAM, ISC, MPTER, CRSTER. for a given application.

-------
                          Table 3-4.    Data Requirements for Atmospheric Models
en
                               TEH'
                               TCNT
                               COM
                               VALLEY'
                               ISC
                               BOXMOD
                               RAM
                               CRSTER
                               MPTER
                          Source:  Bonazountas el al. 1982; USEPA1979; USEPA 1982a.
                           • This model to recommended tor screening applications only.
                          " These models are not EPA preferred models. These models can be used H they can be
                            demonstrated to estimate concentrations equivalent to those provided by the preferred
                            models, e.g., CPM, RAM, ISC, MPTER, CRSTER, for a given application.

-------
site-specific  data,  the most stable atmospheric
conditions, lowest wind speed, and greatest percent
of wind flow toward the exposed population should be
used as input to Equation  3-1, along with maximum
release  rate  estimates for the duration  of  interest.
Usually, the population nearest the point or area of a
ground-level  release  experiences the highest  short-
term exposure.

As  indicated  in  Table 3-2, several  atmospheric  fate
models  have  the capability of producing  short-term
maximum and long-term average  ambient
concentration  estimates  where in-depth  analysis  is
desirable.
3.4 Surface Water Fate Analysis

The  environmental  fate  of  hazardous materials
entering surface waterbodies is highly dependent on
the type of waterbody. The three major classifications
are rivers and streams, impoundments, and estuaries.
Methods for estimating contaminant concentrations in
the first category are provided below.

As  mentioned in the introduction to  this chapter,
contamination of flowing waterbodies will  probably be
a  more  common  occurrence with  regard to
uncontrolled hazardous  waste facilities than  will
contamination of impoundments or  estuaries. Thus, in
this section guidance for estimating contaminant  fate
in  flowing  waterbodies is  presented. In  those  cases
where contaminant fate in  an impoundment or estuary
is  necessary, the analyst  is referred to  Mills  et al.
(1982)  for  guidance.

The Probabilistic  Dilution  Model is an analytical  tool
that can be  used  to extend the qualitative screening
analysis presented in the previous section and that in
some  cases may make application of the quantitative
analyses  discussed  in following  sections
unnecessary. This model  has  been  adapted  by the
U.S.  Environmental  Protection Agency, Office of
Toxic Substances, to  support  the exposure
assessment process  for  contaminants in  surface
water. The  model  is  based  on  the  fact that, in
general,  the most  important  process  affecting  a
contaminant's concentration in  a surface waterbody is
the degree  of its dilution. Thus,  the  model uses
streamflow data for a given subbasin  and  contaminant
loading data (from the  contaminant  release analysis
discussed in Chapter 2)  to predict the number of
times  per year a  given contaminant concentration will
be exceeded.  For contaminants  that  have health-
based  concentration standards (or for  which  health-
based  concentration  cut-off values can  be
calculated),  the model  can be  used to predict the
annual  number  of occurrences  (days)  that
unacceptable health risks may  result  for persons
using  the affected waterbody.  This model can be
applied to  the  Superfund exposure  assessment
process as  an  extended screening  tool to highlight
contaminant releases to surface water that actually
require detailed environmental  fate (and subsequent
exposed  populations) analysis.  Contact  the USEPA
Office  of Toxic Substances,  Exposure  Evaluation
Division  (Pat Kennedy, (202)  382-3916)  for  more
detailed information on accessing the Probabilistic
Dilution Model.
3.4.1  Beginning  Quantitative Analysis
The  following  equation  (adapted from  Delos et al.
1984) provides a  rough estimate of the concentration
of a  substance  downstream from  a point source
release into  a  flowing waterbody, after dilution of the
substance by the receiving waterbody:
  c=-
where
                                          (3-4)
    Qe
    Qt
          =  concentration  of substance  in  stream,
             (mass/volume).
          =  concentration  of substance  in effluent,
             (mass/volume).
          = effluent  flow rate, (volume/time).
          =  combined  effluent  and stream flow
             rate, (volume/time).
This equation predicts the concentration of substance
in the waterbody resulting from contaminant releases
from the subject site alone; it does  not  take  into
account additional  sources of  contamination
("background"   concentrations)  that may  also
contribute to  the total level  of contamination  in the
waterbody.
In cases where hazardous waste is introduced into a
stream  through  intermedia transfer  from air, soil,
ground  water,  or  nonpoint source, or where the
release  rate is known in terms of mass per unit time
rather than per unit effluent  volumes, in-stream
concentrations can  be  estimated  by  use  of the
following equation:
where

   Tr

   Qt
                                         (3-5)
                 intermedia  transfer rate,
                 (mass/time)
                 stream  flow  rate  after  intermedia
                 transfer    has    occurred,
                 (volume/time).
Assumptions implicit in these equations  are the
following:

*   Mixing of the hazardous substance in the water is
    instantaneous and complete.
                                                 53

-------
&& The hazardous material  is  refractive  (i.e., all
   decay  or removal processes are disregarded).

^ Stream flow and rate of contaminant release to
   the  stream  are constant  (i.e.,  steady-state
   conditions).

The  assumption  of complete mixing of a hazardous
substance in a flowing water body is not valid within a
mixing zone  downstream from the point or reach
where the substance is introduced. Under certain
conditions,  this mixing zone can  extend downstream
for a considerable distance,  and concentrations can
be considerably  higher within the  mixing zone  than
those estimated by the foregoing dilution equations.

The  length of the mixing  zone is  estimated  by the
following equation (adapted from  Fischer et al. 1979,
Liu 1977, Neely 1982):
  MZ =

where

   MZ
    w
     u
     si
    g
          0.4
       0.6
                                          (3-6)
 mixing zone length, (length units).
 width of waterbody, (length units).
 stream velocity, (length/time).
 stream depth, (length units).
 slope  of  the  stream  channel,
 (length/length).
acceleration due to gravity, (32 ft/sec2).
These equations  provide  in-stream  contaminant
concentrations resulting from  site releases  only.  If
total  in-stream  contaminant  concentrations  are
desired,  these  should be estimated by  adding
background  (i.e.,  upstream from the site) in-stream
contaminant concentrations to  those  estimated by
Equations 3-4 and 3-5.

If  the hazardous  substance  is introduced into a
flowing waterbody over a length of that body, rather
than from a  point source,  assume that the mixing
zone begins at the downstream end of the reach  over
which introduction takes place.  Neely (1982) presents
an  estimation procedure for hazardous  substance
concentration at exposure points within  a mixing zone
that incorporates  an expression  for dispersion.

The dilution equations  (3-4, 3-5) and the  procedure
presented by Neely (1982) assume that the introduced
hazardous substance is  conservative. Therefore,  they
predict an  estimated stream/river concentration  that
remains  constant from the downstream end of the
mixing zone throughout the remaining  length of the
stream,  or  decreases only  with further dilution
resulting  from additional stream flow from tributaries.
This  is useful  as a  basic model for the fate of
conservative  hazardous   substances;    for
nonconservative substances,  it provides a useful
worst-case estimate.  If the released  substance is
                                         found through this estimation procedure to be diluted
                                         to  concentrations below  a predetermined  level of
                                         concern,  and  no important exposure points  exist
                                         within the mixing zone, the fate of the substance in
                                         this medium  may need no  further analysis.  However,
                                         where  the  concentration  after dilution  of  a
                                         nonconservative  substance  is still above  a
                                         predetermined  critical level, it  may  be useful to
                                         estimate the  distance  downstream  where  the
                                         concentration will remain above this level, as well as
                                         the concentration of the  substance at  selected
                                         exposure points downstream.

                                         This type of estimation can be performed through use
                                         of  an  overall decay  coefficient, which  represents  a
                                         combination of all decay and loss rates affecting the
                                         removal of  a  substance  from a  waterbody. The
                                         concentration  of  a nonconservative  substance  at  a
                                         selected point downstream from the release  point and
                                         below the mixing zone (complete  mixing is assumed)
                                         can be estimated by the  following equation  (from
                                         Delos et al. 1984), which employs the  concept of an
                                         overall decay coefficient:
                                                       W(x) = W(O)e
                                                                                              (3-7)
                                                    where

                                                    W(x)

                                                    W(0)
                                            e
                                            K
                                            x
            concentration at downstream distance x,
            (mass/volume).
            concentration immediately below point of
            introduction,  (from  Equations  3-4,  3-
            5).
            2.71828.
            overall  decay coefficient, (time)'1.
            distance  downstream  from point  of
            introduction, (length).
            stream velocity, (length/time).
                                        The overall decay  coefficient can  also be  used to
                                        estimate the  distance  downstream  over which  a
                                        nonconservative  substance  remains  above  a
                                        predetermined critical concentration level W(CL). This
                                        is estimated  by substituting  W(CL) for W(x) in
                                        Equation 3-7, and  solving this equation  for  x,  as
                                        follows:
       u
 x= — —
       K

where
                                            u
                                            K
                                        W(CL)
                                                      /W(CL)\
                                                       	
                                                      V  W(O)  /
                                                                                               (3-8)
                                                                distance  downstream  from  point  of
                                                                introduction, (length).
                                                                stream velocity, (length/time).
                                                                overall decay coefficient,  (time)-1.
                                                                predetermined  critical  concentration
                                                                level, (massVolume).
                                                 54

-------
W(0)   =  concentration  immediately  below point  of
           introduction,  (from dilution Equations  3-
           4, 3-5).

This equation incorporates the following assumptions:

eses Mixing is  complete.

*   Conditions are steady state.

esx Longitudinal  dispersion  is  negligible;  the
    substance  transports downstream at stream
    velocity.

esx All  decay  and transfer  processes can be
    described  as  first-order coefficients  (i.e.,  decay
    rates  are  a  direct function of  hazardous
    substance concentration).

Values for K  can be derived  empirically  where
monitoring  data  are available, or can be  estimated
based on  decay rate  constants  available  for many
hazardous substances in the technical literature.

Concentration data from  immediately below the point
of substance release into  a  stream (after complete
mixing of waste stream into the waterbody), and from
at least one point downstream of the mixing zone are
required for the  empirical estimation of K.  Note  that
overall decay  coefficients are substance-  and site-
specific and  can  vary with  climatic and hydrologic
conditions. Care must  be taken  in calibrating  the
coefficient  empirically.  Data  covering seasonal
fluctuations must be used, and seasonal values for  K
corresponding to the various  observed  conditions, or
a worst-case K value (i.e., lowest reasonable value)
for the purpose of conservative estimation, should be
developed.

For  estimation  of K  through  the summation of
published  decay rate  constants, the most  important
removal process affecting the compound of concern
in  the receiving waterbody  must be known. For  this
information, see the discussion below (Section 3.4.2),
or see Callahan et al.  (1979),  or Mabey et al. (1992).
Additional  references that provide decay rate constant
values for a wide variety of compounds include:
Verschueren (1984) Dawson et al.  (1980)  USCG
(1974), and Schnoor et al. (1987).

Reliable values for K, which have  been developed for
a given waterbody and hazardous  substance  under
no-action  conditions (i.e.,  during  remedial
investigation), can be used to  estimate the fate of this
same substance  resulting from  the  release  rates
projected  after  implementation  of various remedial
action alternatives.

3.4.2 In-Depth  Analysis
When aquatic concentration estimates developed by
the above  simplified methods (or  methods covering
 estuaries  or  impoundments provided  by Mills et al.
 1982) indicate  that  these concentrations pose a
 potential human health  hazard at one or more
 exposure  points, more  accurate  estimates of short-
 term  and  long-term concentrations of the hazardous
 substance may  be required. A  large number of in-
 depth methods and computer models  exist to assess
 the fate  of substances in the aquatic  environment.
 Each of these models differs in the number and types
 of aquatic fate  processes that  it incorporates.  The
 most important of these  aquatic  processes  are
 described below, and information  is provided to allow
 identification  of  those processes  most likely  to be
 significant  at the site,   and for the  hazardous
 substances under analysis.
3.4.2.1  Intermedia Transfers
The major processes by which hazardous substances
can be  transferred from surface water to other
environmental media are as follows:

(1) Volatilization
Volatilization  of a substance from water depends on
the physicochemical properties of the  substance  and
characteristics of the  waterbody and  body  of air
involved.  Volatilization increases in importance for
substances  with higher  vapor  pressure,  and  for
waterbodies with higher surface  area-to-volume
ratios  and higher turbulence (Delos et  al. 1984).
Callahan et al.  (1979)  stress  the importance of
volatilization as a route of intermedia transfer for  129
priority  pollutants.  If volatilization is  considered  an
important  process for the substance  being  studied, or
if the importance of volatilization is unknown, the  rate
of volatilization  can be  estimated  by  the  method
provided by Mills  et al. (1982)  for quiescent
waterbodies or by Delos  et al.  (1984)  for turbulent
bodies. Lyman  et al.  (1982) provide  methods  for
estimating volatilization rates from water.

(2) Sedimentation
Hazardous  substances   released to  a surface
waterbody in  the solid,  particulate form will settle out
over  time and become  mixed into  the  bottom
sediment.  In  addition,  liquid  hazardous substances
with high affinities  for  adsorption to  suspended
particulates will settle out of surface waters with these
particulates. The rate of sedimentation  is  governed by
the  difference  between settling  velocity  and
resuspension  velocity. The former  increases with
mean  particle size and  density and  with water
temperature, and  can be estimated by the procedure
presented  by Delos  et  al.  (1984).  Resuspension
velocity is a function of bottom shear stress. Delos et
al. (1984) provide a procedure to estimate this rate.
Where sedimentation is considered to be  an important
process, use  a surface  water fate  model  that has the
capability  of accounting for bed-water  exchange  and
sediment load transport.
                                                 55

-------
(3)  Sorption
Substances dissolved in surface waters can sorb onto
solids suspended in the water or onto bed sediments.
This process, in effect, transfers the substances from
the  water to the sediment medium, and proceeds until
an equilibrium  point is reached.  This equilibrium  point
(and the resulting water  and sediment concentrations
of the substance)  is determined  by  the  soil-water
partition coefficient (a parameter that is a function of
sediment type, water pH,  cation exchange  capacity,
and  organic  content  of  sediment)  and the
physicochemical  properties of  the  hazardous
substance.  In  general,  metals and hydrophobic,
nonpolar organic compounds have a high tendency to
sorb onto entrained or bottom sediment. See  Lyman
et al.  (1982)  for  methods  of  estimating sediment
adsorption of waterborne contaminants.

3.4.2.2 Intramedia Transformation Processes
The following  is a  brief description of the important
intramedia transformation  processes that may be
significant for  the  surface  water  fate of hazardous
substances. Rate-controlling factors  are stated for
each. Callahan et al. (1979), Mabey et al. (1982),
Verschueren (1984), and Sax (1984) provide rate
constants for these  processes for  numerous
compounds.

(1)  Photolysis
Chemical transformation due to  photolysis utilizes
energy from sunlight, and for some chemicals, can
occur by  several processes.  Direct photolysis  rates
are a function of photon availability, light absorption
coefficients for the chemical in  question,  and a
reaction  yield  constant  (i.e., the  efficiency  of
substance transformation with the  available solar
energy).  Indirect photolysis occurs through the  action
of intermediate substances naturally occurring  in the
medium. These intermediates absorb light energy  by
various  processes  and  in  this energized  state,  react
with the hazardous  substance. Indirect photolysis is a
function of photon  availability, concentration and light
absorption coefficient of the  intermediate, and a rate
constant  for the  reaction between the  energized
intermediate and the hazardous material.

(2)  Oxidation
Oxidation  is the reaction of substances  with oxidant
species.  Oxidation rates  are a  function of the
concentrations  of the substance in question,
concentration of the oxidant,  and  a rate constant for
reaction between them.

(3)  Hydrolysis
Hydrolysis is  the  nucleophilic displacement  of an
electronegative substituent on a carbon  atom  by  an
hydroxyl  group. The nucleophilic  reactant can  be
either a water molecule or an hydroxyl ion. Hydrolysis
of most compounds is highly  dependent  on the pH of
the waterbody medium and can be promoted by both
acid  and base conditions. The rate of hydrolysis is  a
function  of the  concentration  of the  hazardous
substance and  the  rate constants  for the acid- and
base-promoted  processes at each pH value.

(4)  Biodegradation
Biodegradation is  the breakdown  of  substances
through  the  enzymatic  action  of biota present in the
water. Most  biodegradation is  carried out by microbial
biota.  It  depends  on the  metabolic  rates  and
characteristics and the population density of the biotic
agents,  which are in part functions of the availability
of other  nutrients, pH and temperature of the
medium,, and  sunlight availability,  among  other
factors.

3.4.2.3 Computer Models
Tables 3-5,  3-6,  and  3-7  summarize  the  features,
data  requirements,  resource requirements,  and
references or contacts  for  selected computer-based
models  appropriate  to  the  in-depth  analysis of the
aquatic  fate of hazardous  releases from  Superfund
sites. Additional  details for  certain  of the models
addressed in the tables  are provided below:

Exposure Analysis Modeling System (EXAMS-II)
(Burns et al.,  1982) is a steady-state and  dynamic
model designed for rapid evaluation of the behavior or
synthetic  organic  chemicals  in lakes,  rivers,  and
estuaries. EXAMS-II is an  interactive  program that
allows the user to specify and store the  properties of
chemicals and ecosystems,  modify  the characteristics
of either  via  simple  English-like commands,  and
conduct rapid,  efficient evaluations of the probable
fate of chemicals. EXAMS-II  simulates  a  toxic
chemical  and  its  transformation  products using
second-order  kinetics for  all significant  organic
chemical  reactions.  EXAMS-II, however, does not
simulate the solids with which the chemical interacts.
The concentration  of  solids  must be specified for
each  compartment;  the model accounts for sorbed
chemical  transport  based on solids concentrations
and  specified  transport fields. Benthic  exchange
includes  pore-water advection,   pore-water
diffusion, and solids  mixing.  The latter describes a net
steady-state exchange  associated  with  solids that  is
proportional  to pore water diffusion.

A data  set of average  or typical  values for
waterbody-specific  data   is  presently  being
developed by Battelle  Northwest Laboratories,  under
contract to EPA. This data file will  contain parameter
values for a number of major U.S. river  systems,
lakes, and reservoirs, and will be  integrated with the
EXAMS program. These values will be accessible for
fate  modeling  of the  waterbodies  included (GSC
1982).

MINTEQA1  (Felmy  et  al., 1984; Brown and Allison,
1987)  is  a  geochemical model that is capable of
                                                  56

-------
calculating  equilibrium  aqueous  speciation,
adsorption,  gas  phase  partitioning,  solid  phase
saturation  states,  and precipitation-dissolution of 11
metals (arsenic,  cadmium, chromium,  copper, lead,
mercury,  nickel, selenium, silver, thallium,  and zinc).
MINTEQA1  contains an  extensive  thermodynamic
data base and contains six  different algorithms for
calculating adsorption.  Proper application  of
MINTEQA1  requires applicable expertise, because
kinetic limitations  at particular sites  may  prevent
certain reactions even though  they  might  be
thermodynamically possible.

Hydrological  Simulation Program  - FORTRAN
(HSPF) (Johanson et al., 1984;  Donigian et al., 1984)
is a comprehensive package  for simulation  of
watershed hydrology and  water quality for both
conventional and  toxic organic pollutants. HSPF
incorporates  the watershed-scale ARM (Agricultural
Runoff Model) and NPS (Non-Point Source)  models
into a  basin-scale  analysis  framework that  includes
pollutant  transport and transformation  in stream
channels.

The model uses  information  such as the time history
of rainfall, temperature,  and  solar radiation; land
surface characteristics such as  land  use patterns and
soil properties;  and  land management practices  to
simulate  the  processes that  occur in  a watershed.
The result of this simulation is  a time  history of the
quantity  and quality of  runoff from  an urban  or
agricultural watershed. Flow  rate, sediment load, and
nutrient and pesticide concentrations  are predicted.
The program  takes these   results,  along with
information  about the  stream network  and point
source discharges,  and simulates instream processes
to produce a time history of water quantity and quality
at any point in a watershed - the  inflow to  a lake,
for example.  HSPF includes an  internal  data base
management system to process the large amounts of
simulation input and output.

Water Analysis  Simulation  Program  (WASP4)
(Ambrose et al.,  1986, 1987) is a  generalized
modeling framework  for contaminant fate and
transport  in surface water.  Based  on the flexible
compartment  modeling  approach, WASP  can  be
applied in one, two,  or three  dimensions. WASP is
designed  to  permit easy  substitution  of user-written
routines  into  the program structure.  Problems that
have been studied using WASP include  biochemical
oxygen  demand, dissolved  oxygen  dynamics,
nutrients  and eutrophication, bacterial contamination,
and toxic chemical movement.

A variety of water quality problems can  be addressed
with the  selection  of appropriate kinetic subroutines
that may be either selected from a library or written
by the user. Toxics WASP (TOX14; Ambrose et al.,
1987)  combines  a kinetic  structure adapted from
EXAMS with the WASP transport structure and simple
sediment balance  algorithms to predict dissolved  and
sorbed  chemical concentrations in the  bed and
overlying waters.

Eutrophication WASP (EUTR04; Ambrose  et  al.,
1987) combines a kinetic structure adapted from the
Potomac  Eutrophication Model with the  WASP
transport  structure.  EUTR04  predicts  dissolved
oxygen,  carbonaceous biochemical oxygen demand,
phytoplankton, carbon, and chlorophyll a, ammonia,
nitrate, organic nitrogen,  and  orthophosphate in  the
bed and overlying waters.

SARAH  (Ambrose and Vandergrift,  1986) is  a
steady-state mixing  zone  model  for  back-
calculating  acceptable concentrations  of  hazardous
wastes discharged to  land disposal  or waste water
treatment  facilities.  For steady or  batch  waste
streams,  SARAH  considers the  following
concentration  reductions:  dilution and loss  during
treatment,  initial Gaussian mixing at the edge of  a
stream,  lateral and longitudinal diffusion in the mixing
zone,  sorption,  volatilization, hydrolysis,  and
bioaccumulation  in fish.  The user  must specify,
appropriate in-stream criteria for protection  of  the
aquatic community, and humans through consumption
of fish  and water.  The  benthic community  is  not
presently  considered.  Treatment loss  is handled
empirically.  The  human  exposure  pathways
considered  include ingestion of treated drinking water
and consumption of contaminated fish.
3.4.2.4 Short- and Long-Term Concentration
Calculations
Long-term  average  ambient water  concentrations
should be calculated using (1) the average  release
rate (from Chapter 3) projected for the time period of
interest, and (2) the annual average stream flow rate
as input to the above estimation procedures.

Short-term concentration levels are obtained through
use of the short-term release rate developed during
contaminant release  analysis and  the  lowest
reasonable  24-hour flow rate,  or the 7-day, 10-
year  (7-Q-10) low flow rate  for the  period  of
record,  as  presented in the  above  estimation
procedures.

Table  3-6 indicates several aquatic fate models
capable  of  estimating  both  short-  and long-term
ambient water concentrations that  are appropriate to
in-depth  analysis of the  aquatic fate of contaminants
released from Superfund sites.
3.5  Quantitative Analysis of Ground-
Water Fate

To  model the migration of contaminants in  ground
water the following factors should be estimated:
                                                57

-------
            Table 3-5.     Resource Requirements and Information Sources: Surface Water Fate Models
                              Model
                                                                     Description
                                             Resource Requirements, comments
                                            References, sources of documentation,
                                                         software
            Water Quality Assessment Methodology
            (WQAM)
             Simplified Lake/Stream Analysis (SLSA)
             Michigan River Model (MICHRIV)
             Chemical Transport and Analysis Program
             (CTAP)
01
09
             Exposure Analysis Modeling System
             (EXAMS-II)
• Steady-state, 1 -dimensional model
*  Requires only desk top calculations
xx Provides canonical information
xx Models lakes, rivers,  and estuaries

xx Steady-state, 1 -dimensional model
x  Solution either by desk top calculations or
   simple  FORTRAN program
•  Suitable for simplified lake and  river
   systems
xx Steady-state, 1  -dimensional model
xx Computer program written in FORTRAN
x  Similar to SLSA, but can model more
   than one reach
xx Intended for metals
xx Models rivers and streams

xx Steady-state, d-dimensional
   compartmental  model
xx FORTRAN IV program suitable for
   numerous computers
x  Similar to SLSA except  more
   sophisticated; each CTAP compartment
   is equivalent to one SLSA "lake"
xx Models streams, stratified rivers, lakes,
   estuaries, and coastal embayments
xx Steady-state, 3-dimensional
   compartmental  model
• Complex computer program
xx Contains comprehensive second-order
   decay kinetics for organics; most models
   only have first-order  kinetics
xx Models organic chemicals
xx Suitable for freshwater, non-tidal aquatic
   systems
x  Easy to set up and use
xx No computer programming needed;
   requires only hand calculator
xx Recommended if time, costs, or
   information are  restrictive
xx Easy to set up and use
• Computer programming not  necessary; if
   used, only 280 bytes are required;
   suitable  for microcomputers
*«rWell documented and suggested for use
   before use of a more sophisticated model
xx May be  used with hand  calculator
xx Easy to set up and use
xx Requires minimal computer programming
xx Requires extensive data input
xx FORTRAN program - suitable for IBM
   360/370, UNIVAC 108, CDC 6600
   mainframe computers
xx Microcomputer version available requiring
   32 K bytes storage
xx One of the better documented models,
   which may make it more desirable than
   other complex models
xx Requires extensive data input
xx Has been incorporated into EPA-OTS
   GEMS system (see Section 4.1)
*«rWell documented and recommended for
   use over most other models
• Available on magnetic tape for installation
   on mainframe or small computers (e.g.,
   PDP-11 or HP 3000); batch version
   requires 64 K bytes memory at a
   minimum, more  for complex modeling
ji^Also available in interactive version,
   requiring 164 K bytes memory plus 2 K
   bytes for each chemical and  2.5 K bytes
   for each environment
   An estimated 350 hours required for
   installation and setup, assuming all data
   are readily available
Reference: Mills et al. 1982
Documentation:
  ORD Publications
  USEPA, Cincinnati,  Ohio 45268
  (513) 684-7562
Reference: HydroQual  1982
Documentation:
  William  Gulledge
  2581 M Street, N.W.
  Washington,  DC. 20037
  (202) 887-1183

Reference: Delos  et al. 1984
Technical Assistance  Available from:
  Bill Richardson
  USEPA
  Environmental Research Laboratory -
  Duluth
  Large Lakes  Research Station
Reference: HydroQual  1982
Documentation:
  William  Gulledge
  Chemical Manufacturers Association
  2581 M Street, N.W.
  Washington,  D.C. 20037
  (202) 887-1183
Reference: Burns et al. 1982
Documentation:
  ORD Publications, Center for
  Environmental Research Information
  USEPA
  Cincinnati, Ohio 45268
  (513) 684-7562
  Center for Water Quality Modeling
  Environmental Research Laboratory
  USEPA
  Athens, Ga. 30613
  (404) 546-3585
                                                                                                                                                                   (Continued)

-------
         Table 3-5.
(Continued)

    Model
                                                                  Description
 Resource Requirements,  comments
   References, sources of documentation,
                software
          Metals Exposure Analysis Modeling System
          (MINTEQA1)
          Hydrological  Simulation  Program
          FORTRAN (HSPF)
          Transient One-Dimensional Degradation
          and Migration Model (TODAM)
01
CD
          Channel Transport Model (CHNTRN)
          Finite Element Transport Model (FETRA)
                               Steady-state, 3-dimensional
                               compartmental model
                               Complex computer program
                               Designed for modeling of metal loadings
                               Suitable for freshwater, non-tidal aquatic
                               systems

                               Time-varying, 1  -dimensional model
                               Designed for year-round simulation
                               Models organic pollutants
                               Second-order decay mechanisms
                               Models non-tidal rivers, streams, and
                               mixed lakes
                               Time-varying, 1 -dimensional model
                               Second-order  decay mechanisms
                               Models river and estuarine systems
                               Requires exterior hydrodynamic model
                               (e.g., EXPLORE) to provide channel and
                               flow velocities to TODAM
                            •f  Time-varying, 1 -dimensional model
                            *  Models organic pollutants
                            *  Second-order decay mechanisms
                            •s  Models rivers, lakes, estuaries, and
                               coastal waters
                            •  Can be coupled with a hydrodynamic
                               model, CHNHYD, to estimate  flow
                               dynamics where such data are not
                               available
                            *  Time-varying, P-dimensional model
                               (longitudinal and lateral)
                            *  Second-order decay mechanisms for
                               organic pollutants
                            *s  Models rivers, estuaries, coastal systems,
                               and completely mixed lakes
                            *  Can be coupled with EXPLORE
                               hydrodynamic model to generate flow
                               velocities where these are  unknown
Complex metal dynamics requiring
extensive data input
Can be used with mainframe or small
(e.g., PCP 11/70 or HP 3000) computers
ineractive format
Contains data base with thermodynamic
properties of 7 metals
Requires extensive data input
Most suitable to minicomputers (e.g., HP
3000,  PRIME. HARRIS) as model
utilizes direct access input-output, which
can be costly on mainframe computers
Requires 250 K bytes of overlay-type
storage
Has been used  on IBM 370 series
computers
Requires extensive data input
Complex FORTRAN program, written  in
the preprocessor language FLECS or  in
FORTRAN  IV
Applicable to VAX or POP 11/70
computers (batch mode)
TODAM has been applied; however,
documentation is currently under review;
release date unknown
Requires extensive data input, and
extensive setup  time
Has not been field-tested, and
documentation is currently under review
FORTRAN  IV program language
Applicable to IBM 3933 computer, and
others
Input data requirements are extensive
Computer program written in FORTRAN
IV
Can be used on IBM. VAX, or CDC-
7600 computers
Has been field-validated
Setup and execution time requirements
are extensive
Further information:
 Yasuo Onishi
 Battelle Pacific Northwest Laboratories
 Richland, WA 99352
 (509) 376-8302
Reference: Johanson et al. 1984
Software:
  Center for Water Quality Modeling
  Environmental Research Laboratory
  USEPA
  Athens, GA 30613
  (404) 546-3585
Reference: Onishi et al. 1982
Further information:
  Yasuo Onishi
  Battelle Pacific  Northwest Laboratories
  Richland, WA 99352
  (509)  376-8302
Reference: Yeh 1982
Documentation:
  Dr. G. T. Yeh
  Environmental Sciences  Division
  Oak Ridge National Laboratory
  P.O. Box X
  Oak Ridge, TN 37830
  (615) 574-7285

Reference: Onishi 1981
Further information:
  Yasuo Onishi
  Battelle Pacific Northwest Laboratories
  Richland, WA 99352
  (509) 376-8302
                                                                                                                                                                (Continued)

-------
Table 3-5.     (Continued)

                  Model
                                                         Description
                                             Resource Requirements, comments
                                            References, sources of documentation,
                                                         software
 Sediment-Contaminant Transport
 (SERATRA)
xx Time-varying,  P-dimensional model
   (longitudinal-and vertical)
x* Complex sediment transport mechanisms
xx Second-order  decay mechanisms for
   organic  pollutants
•  Models rivers and  lakes
 Estuary and Stream Quality Model  (WASP4) xx Time-varying, 3-dimensional model
                                         xx Sophisticated second-order organic
                                            decay kinetics
                                         xx Models rivers, lakes, and estuaries
Surface Water Back Calculation Procedure
(SARAH)
xx Steady-state, 1  -dimensional analytical
   solution
xx FORTRAN  Code
xx Models contaminated leachate plume
   feeding the  downgradient surface
   waterbody (stream or river)
xx Monte Carlo simulated  generic
   environment
xx Degradation, dilution, sorption, and
   volatilization
xx Broaccumulation in fish
xx Requires extensive data input
xx Computer program written in FORTRAN
   preprocessor language FLECS < in
   batch   mode
xx Has been field-tested and is available
   for use
xx Requires an estimated  750 man-hours
   for setup, assuming all  required data are
   readily available
                                         xx Very data-intensive  model
                                         xx User must provide hydrodynamic flows
                                            between model compartments
                                         x  Applicable to IBM 370 or POP 11/70
                                            systems
                                         xx FORTRAN IV program  requires 64 K
                                            bytes memory
                                         xx Requires  150-300 man-hours for
                                            setup
xx Generic environment, minimal data input
•  FORTRAN model
Reference: Onishi and Wise  1982a, Gnishi
  and Wise 1982b
Documentation:
ORD Publications
  Center for Environmental Research
  Information
  USEPA
  Cincinnati, OH 45268
  (513) 684-7562
Technical Assistance:
  Robert Ambrose
  EPA Athens Environmental  Research Lab
  Center for Water Quality Modeling
  Athens, GA  30613
  (404) 546-3546
Documentation and Software:
  Dr. John Connolly
  Environmental Engineering and Science
  Manhattan College
  Bronx, N.Y. 10471
  (212) 920-0276 or:
  Dr. Parmely  H.  Prichard
  Environmental Research Laboratory
  Gulf Breeze, FL 32561
  (904) 932-5311

  Robert Ambrose
  Center for Water Quality Modeling
  USEPA
  Athens, GA 30613
  (404) 546-3546
Documentation: Jan.  14, 1986
  Federal  Register, Hazardous Waste
  Management System, Land  Disposal
  Restrictions,  Proposed Rule
Software.
  David Disney, Environmental Research
  Laboratory, Environmental Protection
  Agency,  College Station Road, Athens,
  GA 30613, (404) 546-5432, or (404)
  546-3123
Source: Versar 1983a.

-------
Table 3-6.    Features of Surface Water Fate Models
    MCHMV
    CT»f
    EXA
    MMTEQA1
    TODAM
    CMNTM
    FfTKA
    KMTRA
    WASP 4
Sourcti:  USEPA 1985h;   Dttot »l (I. 1984

-------
                             Table 3-7.    Data Requirements for Surface Water Models
en
IVJ
                                  CTAP
                                  EXAM I
                                  MMTEOA1
                                  TOOAM
                                  CNNTRN
                                  FITHA
                                  MRATMA
                                  WASP 4
                              Sourcti:  USEPA 1985H;  D*k>i •! al. 1984

-------
Direction - The direction  of contaminant migration is
important  in  predicting  the  potentially exposed
population.

Velocity  -  The migrating contaminant's velocity is
important in assessing when contamination will reach
the  exposed population  and  how long the
contamination will  be affecting that population.

Concentration  - Concentration  of  the contaminant in
ground  water  at  the exposure locations  is  used  to
calculate dose to  the population.  This factor is used
to convert the amount of water consumed each day
to the  mass of contaminant  received each day. The
mass  information  is  then used to predict health
effects associated with exposure  to the contaminant
(USEPA1985d).

Volume  -  The contaminated region's  volume  is
important in evaluating  the  extent  of the
contamination, which is essential  to estimating costs
of remedial measures  and  viability of  specific
alternative remedial  measures for the particular site. It
is also  useful for determining  how  long  a  remedial
measure will have to be taken.

The  following  ground-water  discussions are  divided
into three sections:

1. The  minimum technical foundation that is needed
   in order for the  analyst to apply  and interpret the
   equations  and  models for ground water. This
   discussion  is  meant  to  support the  hydrologist
   familiar with water  supply  calculations,  providing
   an  introduction  to   contaminant hydrology.
   Readers needing a more complete introduction to
   hydrology  may  wish  to  read EPA's Handbook
   titled  "Groundwater"  (EPA/625/6-87/016).

2. Equations that can  predict  average contaminant
   velocity and  mass flux for  dilute solute and
   concentrated  contaminant  plumes.  Knowing  the
   travel time and  the degradation  half-life, one can
   predict contaminant attenuation.  A  nomograph is
   provided for predicting dilution and  contaminated
   front velocity of dilute  solute  plumes,  as are
   equations that are useful in assessing the extent
   of contamination. The narrative contains guidance
   for  interpreting  available  monitoring data from
   existing wells and from monitoring wells. All of the
   equations  apply to homogeneous  and  isotropic
   media;  fractured rock  flow and karstic terrain flow
   are not addressed.

3. Computer models  that predict dilution,
   attenuation, and contaminated  front  velocity  of
   dilute solute  plumes  only.  All  of  the computer
   models assume homogeneous and isotropic
   media.  Computer models that predict organic fluid
   migration are  not discussed, nor are models that
   describe karstic terrain flow. The state of the art
    for these models is  not well-developed, and thus
    they are considered beyond the scope  of this
    report. The analyst wishing to model organic fluid
    migration  in  porous  media should  use the
    equations in Section  3.5.2.

3.5.7    Discussion of Ground- Water Modeling

3.5.1.1 The Contamination Cycle
The  two  primary types of ground-water
contamination  at uncontrolled hazardous waste  sites
involve  leaching of solid  contaminants and percolation
of liquid contaminations to the  underlying  aquifer.
Solid  material itself does  not generally contaminate
ground  water directly,  because  it does  not move
through  the porous soil.  Thus, it will not migrate until
precipitation  (or ground water)  leaches (dissolves)
some of it and carries  it  down  to the water table.
Ground-water  contamination  by  this  route depends
on the precipitation rate  and the solubility of the solid
contaminant. A variation of this route  involves
dissolution  of the  solid contaminant  by a complex
leachate that contains organic constituents  as  well  as
water.  The  existence   of dissolved  organic
constituents  in the  leaching fluid causes  organic
contaminants  to  have a higher solubility.  The
importance  of this  phenomenon is greatest for
contaminants with a high octanol/water partition
coefficient (Enfield  1984, Jaw-Kwei  n.d.).

Liquids  do  not need infiltrating  precipitation to carry
them down to the water table;  they move on their own
with  help  from  gravity. Thus, ground-water
contamination by  liquids  is  not  dependent  on the
precipitation  rate or the  solubility of the contaminant.
The viscosity and density of a liquid affect  its  rate  of
migration. After the liquid has percolated through the
soil, some will  remain in the  interstitial pore spaces;
this  material will dissolve  into the  percolating
precipitation  and migrate downward as a function  of
its water solubility and the rainfall  rate. Another
source of contamination  by  liquid  material arises from
intentional injection into the aquifer itself  (deep-well
injection) or  "injection" into  the vadose (unsaturated)
zone (unlined lagoons).

Hazardous waste is  often  assumed to be primarily
solid waste;  however,  studies showing the relative
proportion of solid to  pourable  hazardous  RCRA
wastes  indicate  that  pourable  hazardous  waste
constitutes  60  to  95  percent of the  total (Skinner
1984).  The  equations  for modeling liquid  waste
migration pertain to a larger percentage of the waste
migration situations than the dilute solute transport
models  (computer models/nomograph).

Two other types of ground-water  contamination may
also occur.  These are contamination  by gaseous
contaminants  and contamination  by intermedia
transfers. Gases constitute  a relatively small  source
of ground-water contamination, since  they are more
                                                 63

-------
 likely to contaminate air than ground water. The main
 mechanism for gases  contaminating ground  water is
 equilibration of gases  leaking from buried containers
 or injected into the ground,  with percolating rainwater
 causing subsequent downward migration  and mixing
 of this contaminated water with ground  water.
 Intermedia contamination  of ground water can come
 from  either air or surface water.  Contamination from
 air  can result from two  mechanisms: rain-out  and
 wash-out.  Rain-out  occurs when  airborne  con-
 taminated  particulates form  condensation nuclei  for
 the formation  of rain  drops. Wash-out occurs when
 falling rain captures  gaseous  or  particulate  con-
 taminants  as  it  falls to earth. The concentrations  of
 contaminants  entering ground water as  a  result  of
 gaseous  contamination or  intermedia  transfers  are
 generally  very small, and these are not considered to
 be  significiant  sources  of ground-water  contam-
 ination in  most cases.

 A third  source of contamination  that  may  be
 significant at  some sites is  through  ground-
 water/surface-water system interconnections.  That
 is,  contaminated surface  water  may recharge  a
 ground-water  system.  This  occurs  only  in  reaches
 where the surface-waterbody is  a "losing  stream"
 (i.e.,  one  that supplies water to the  ground-water
 system).  Frequently,  ground water feeds  surface
 water  (gaining reaches). For gaining  reaches,  the
 ground water,  if contaminated,  contaminates  the
 surface-waterbody into which  it discharges.

 One aspect of the contamination cycle  that should be
 considered is the  ratio of contaminant  to con-
taminated  ground water. A very small  quantity of
 concentrated  contaminant can contaminate  a large
volume of ground water to the ppm or ppb level.

 3.5.1.2 Ground-Water Flow Conditions
After precipitation infiltrates the surface  of the ground,
 it travels  vertically down  through the  vadose zone
 (unsaturated zone) where it meets the water table,
 and  it then flows approximately  horizontally. The
 horizontal  flow within the aquifer is saturated.

 (1)  Saturated  Zone
A simplified flow  equation  is used  to  describe  the
volumetric flow of water through  a porous  medium
 under  saturated  conditions.  The  volumetric  flow (or
discharge) is proportional to  the product of the driving
force,  the soil's  ability to  transmit water,  and  the
cross-sectional  area perpendicular to the flow
direction.  The  driving  force  is the  difference in  the
energy (hydraulic  head)  between  two  points in  the
aquifer divided  by the distance  between  the two
 points.  This driving  force  is called  the hydraulic
gradient.  A  soil's ability to  transmit water is
 represented by an  empirically determined coefficient
of hydraulic conductivity.  This  equation is called
 Darcy's law.  The properties of the liquid (water or
contaminant)  and the permeability of the  porous
 medium determine the hydraulic conductivity. The soil
 has an  intrinsic  property of permeability, which  is
 determined  by the  size,  orientation,  and  con-
 nectedness of the pore spaces.

 Soil permeability is  a function of soil  pore space,
 which is determined by soil  particle  size.  Small
 diameter clay soil particles cause clay soil to have low
 permeability, while larger diameter sandy soil particles
 result  in the high permeability of sandy  soils.  The
 permeability, and  therefore the hydraulic conductivity,
 of a homogeneous soil is  constant under conditions  of
 saturated flow.

 In cases where  the vadose zone is saturated and the
 flow direction is vertical, the change  in height  of the
 water  per unit  of vertical travel  distance is always
 one. Thus, the hydraulic gradient for vertical saturated
 flow is unity, and the  volumetric flows are proportional
 to the permeability alone.


 (2) Unsaturated Zone
 Darcy's  law governs flow  anywhere in  the  porous
 medium, including  the vadose, or unsaturated, zone.
 In the vadose zone, however, the pore spaces are not
 saturated with water or any other liquid. The hydraulic
 conductivity of any liquid through a porous medium  is
 partly dependent on the  amount of liquid in the pore
 spaces,  and  hydraulic conductivity  for unsaturated soil
 can  be  expressed as a fraction of the  hydraulic
 conductivity at saturation.

 When  the  pore  spaces are  entirely filled  with liquid
 (i.e., saturated), the hydraulic  conductivity for that
 medium  is at its maximum value. This is called the
 saturated hydraulic conductivity (or simply hydraulic
 conductivity), and it  is  essentially  constant for  a
 specific liquid saturating a specific soil medium.

 The unsaturated  hydraulic conductivity  at  residual
 moisture content is very  small. When the soil is very
 dry, most of the moisture  is tightly bound by capillary
 forces  in the void spaces, and the water will not flow
 easily.  Unsaturated  hydraulic  conductivity  increases,
 gradually at first and then  more rapidly, as the degree
 of saturation increases  from the residual moisture
 content to  the saturated  moisture  content. Since the
 hydraulic conductivity  is dependent upon  the moisture
 content,  the  specific  discharge through  the vadose
 zone  varies with the degree  of  saturation at any
 depth.

 The rate of infiltration  at the ground surface may be
 limited  by the capacity of the  soil  to accept water or
 by the delivery  rate of water at the  ground surface
 (e.g., the precipitation rate).  The infiltration rate into
 soil cannot exceed the value for that soil's saturated
 hydraulic conductivity. When  the hydraulic  loading to
the surface of the ground  is low, such as light rainfall
 alone,  the  flow of water  through the vadose zone  is
                                                  64

-------
unsaturated; however, when the hydraulic loading  is
large, such  as beneath a lagoon,  the flow of water
through  the vadose zone can be saturated. When the
hydraulic loading is small, it is limiting and the vertical
flow through the vadose zone is unsaturated. When
the hydraulic loading is  larger than the flow that can
move through the  soil with  saturated  flow,  the
permeability of the soil  is limiting  the flow,  and the
vertical flow through the  vadose  zone is saturated.


351.3 Multiphase Flow
The water solubility of any  particular chemical will
determine whether it will be transported as a solute,
as  a colloid, or as a separate,  concentrated phase.
Many  chemicals that  have  been identified  as
contaminants in ground water are sparingly soluble in
water. When introduced to the  ground-water system
as  liquids,   such  chemicals can flow  as  an
independent species through  the  porous medium.
When the immiscible contaminant comes into contact
with the water in the pore spaces of the vadose zone
or at the water table (phreatic surface), the liquids do
not  mix but  essentially remain  as two  separate
phases. Some of the  chemical  will  go into  solution
with the water, but since the solubility of the chemical
is very low, the bulk of the contaminant will remain as
a separate layer that could saturate the pore spaces it
is flowing through. Thus,  the migration  of two
immiscible liquids in porous media is  called two-
phase flow.

Complete descriptions  of two-phase flow  require an
additional equation for each separate phase  present
in the flow system. Several general rules that can be
applied  in  analyzing   ground-water contamination
problems  involving  immiscible chemicals,  are  as
follows:

(1) Floaters
The  specific gravity  of  an  immiscible  liquid
contaminant will determine whether water will displace
it or it will be displaced by water.  In downward flow,
water can displace the lighter, immiscible liquid so the
water is found below the immiscible liquid.  In
horizontal flow, the less dense, immiscible liquid  will
tend to  float  upward  until  the separate immiscible
phase floats  on  top of the water table. Thus,  the
immiscible liquids  that  have  a specific gravity of less
than one  are  sometimes referred to as "floaters." As
a general rule,  immiscible  hydrocarbons that  are
nonchlorinated are floaters (less dense than water).

(2) Sinkers
Immiscible  contaminants more dense than water,
whose  specific gravity values are greater than  one,
can displace water when flowing through the porous
medium.  Gravity will cause dense  immiscible liquids
to  sink as they flow horizontally through the porous
 medium.  Thus, the immiscible liquids more dense
than water are  often referred to as  "sinkers."
Generally, chlorinated hydrocarbons that contaminate
water are more dense than water.

Density and specific  gravity are intrinsic properties  of
a chemical, and values for  natural or manufactured
chemicals are usually published (Verschueren 1984;
Callahan et al. 1979).

(3)  Hydraulic Gradient for Immiscible Fluids
The hydraulic gradient, the difference in the hydraulic
heads at two points divided by the distance (along the
flow path) between the points, is the driving force for
ground-water movement  in  a porous  medium. With
regard to an immiscible separate phase, however, the
gradient that causes the immiscible liquid to flow  is
not necessarily the same as  that which influences the
ground water.  If contaminants  in an immiscible phase
that is more dense than water reach the bottom of the
aquifer,  that  separate  phase  may  alter its flow
direction to  conform to  the  shape  and slope  of the
aquitard  surface.  In some  cases,  the base  of the
aquifer may be sloped in a different direction from the
direction of flow determined by the hydraulic gradient.
This  possibility should be considered  when the
analyst tries to identify   the direction  of the
contaminant plume's  migration.

The assumption that the hydraulic gradient  of the
separate, immiscible  phase  approximates  that  of
ground water  is  quite reasonable for the less dense
immiscible  liquids.  Since  these contaminants float on
the water table, the hydraulic gradient of the phreatic
surface is probably also the gradient of the immiscible
phase.

(4)  Hydraulic Conductivity of Immiscible Fluids
If the  saturated  hydraulic  conductivity  of water
through a porous medium is  known, it is very  easy  to
modify  that  value to calculate the  hydraulic
conductivity of that  same porous  medium saturated
with a different liquid, such as a separate  layer of an
immiscible phase.

3.5.1.4 Contaminant Flow and Hydrodynamic
Dispersion
In  contaminant  transport,  contaminants  can  be
thought of as a  mass flowing  through a  cross-
sectional  area  of the  porous  medium  that  is
perpendicular  to  the flow  direction.  The  discussion
presented  here is for solute transport (mass  that  is
transferred with the  flowing  ground water), but basic
concepts  also  apply to the flow  of immiscible,
separate phases.

The movement of contaminants in  ground water can
be  described by  two principal mechanisms:  gross
fluid movement  (advective  flow) and dispersion.
Gross fluid movement can  be either ground-water
movement or organic fluid movement (the waste itself
moving as a concentrated liquid). Dispersion also can
be  described by  two  principal mechanisms: fluid
                                                   65

-------
mixing (mechanical  dispersion) and diffusion.  The
next section  addresses the  underlying mechanisms
for fluid mixing.

Fluid mixing is important for two reasons:  (1) precise
modeling of contaminant movement and (2) modeling
of dilution of the contaminant concentration  between
source and exposed population,

Dilution (mixing) in  ground  water  is different from
dilution in air  and in  surface water. In  both  air  and
surface water, dilution  is  a  major phenomenon.  In
ground  water, the magnitude of  dilution is much
smaller. Flow  in  both air  and surface  water can be
turbulent. Turbulent flow means that all the flow paths
are not essentially parallel to the  gross direction  of
motion;  some  flow paths  are at right  angles to the
bulk fluid  motion. The flow components that  are
perpendicular to the bulk  fluid  motion  cause  the
plume  to  spread  laterally. This reduces  the
concentration  in  the  plume,  while making the plume
contaminate a  larger volume  of air or surface water.

In ground water, turbulent  flow rarely exists. The slow
speed of ground water coupled with the straightening
effect of many soil particles keeps the flow  smooth
and  laminar.  In  an idealized conceptual model, the
interconnecting pore   spaces can  be thought of as
forming flow channels or tubes; any tendency for the
flow to eddy  is  resisted  by the sides of the flow
channel.  Since the interconnecting  pore spaces  do
not make a  continuous flow channel,  some lateral
mixing  will occur in real soil.

Dispersion  in air and  surface water is caused by the
eddy currents  (and diffusion). If the  flow is broken up
into two components,  longitudinal flow and eddy flow,
the gross motion is due to the longitudinal flow,  and
the  eddy  flow  is responsible for mixing. The
magnitude  of  the  eddy currents is the same in all
directions  (longitudinal,  transverse,  and  vertical).
Since the concentration gradients are weaker in the
longitudinal direction  than  they  are  in the transverse
and  vertical directions (for  continuous  steady state
sources), the  net effect of mixing  in the longitudinal
direction is  small compared to the effect of mixing  in
the directions perpendicular to the flow direction.
When  air  and  surface  water are modeled,  the
longitudinal mixing is  often neglected: lateral mixing  is
modeled as the principal mixing  phenomenon.

Dispersion  in  ground water  is  not  caused  by eddy
currents. Dispersion  (neglecting  diffusion   for  the
moment) is caused   by four principal  phenomena:
varying  pore  sizes, varying  path length, variation  in
velocity gradient  across pore space, and flow  splitting
around  soil particles with  mixing  within the pore
space.  The first  three phenomena contribute to
longitudinal dispersion; the last phenomenon causes
lateral  dispersion. In  ground water,  the magnitude  of
the mixing is much greater for longitudinal mixing than
for lateral  mixing.   Researchers  have  reported
longitudinal dispersivity values ranging from 2 to 25
times  higher than  transverse  dispersivity values
(Gelharetal. 1985).

In ground water, dilution occurs at a much slower rate
than it does  in  air or surface water. The overall
magnitude of mixing  is smaller, and the component of
mixing that is most important to dilution (lateral) is the
smaller  component  of  ground-water mixing.  For
short-term releases  (spills),  longitudinal  mixing  is
useful in diluting  plume concentrations. This  is
because  the plume can  effectively  mix with the
uncontaminated water in front  of and behind the  slug
of contamination, whereas continuous  sources make
the length of the plume so long that its  middle section
cannot effectively mix with clean water in front  or
behind it.

3.5.1.5 Transformation and Retardation
Movement of contaminants can  be modeled  by  fluid
movement, fluid  mixing,  and  diffusion;  however, for
more accurate modeling, chemical transformation and
retardation  should  also be considered. Some
contaminants are  subject to transformation  and
retardation  while  others are  not;  the relative
significance  of  transformation  and retardation  for
specific contaminants determines the  need to model
these  mechanisms.  Transformation is the  term  used
to describe loss of the contaminant from the plume.
The mass of the contaminant  is not lost;  rather, the
molecular structure  is changed  so that the toxicity
associated with  the  initial  molecular structure is no
longer present.  When the molecular structure  of
degradation  products  is more  toxic than the  original
contaminant,   degradation  is  not  considered
attenuation. Attenuation is used to describe chemical
structure  changes that reduce or eliminate the toxicity
of the  contaminant,  and to describe phenomena that
function as sinks for toxic contaminants. Phenomena
that  are reversible  are  not  sinks  for toxic
contaminants.

Chemical  interactions between contaminants and the
soil matrix that are  reversible  delay the migration  of
contaminants but do  not act as a sink. The effect  of
these  chemical interactions is modeled as retardation.
Retardation  is modeled using  a coefficient to scale
down   the velocity  of ground water  to the slower
effective velocity  of  the  contaminant  mass.
Attenuation reduces  population  risk; retardation delays
population risk.

Many  reversible  interactions  can  cause  retardation
phenomena;  however, only  two   retardation
mechanisms  apply to wide classes of  contaminants
and are well enough understood  to be  modeled on a
regular  basis.  Organic  retardation  and  cationic
retardation  are the most frequently  modeled
phenomena. Organic retardation refers to hydrophobic
contaminants sorbing onto organic material in the soil
                                                 66

-------
matrix. Cationic retardation  refers to positive charged
ions associating with the soil  matrix. This association
can be due to polar species in the ground water being
attracted to the  ionic double layer surrounding clay
particles in the soil, or it can be due to ionic bonding
with the soil matrix.
(I) Retardation of Organics
Organic  retardation,  which  refers to  hydrophobic
contaminants sorbing onto organic material in the soil
matrix, is estimated in ground water by the use of a
retardation  coefficient.  The velocity  of  each
compound in ground water is a  function  of  the
characteristics of the soil media and the compound's
octanol-water  partition  coefficient.  The  octanol-
water partition coefficient measures the compound's
degree of hydrophobicity. The parameter of the  soil
media that determines the presence of sorption sites
is the percent of organic carbon in the soil.
When the contaminant concentration in  the water is
high and the quantity of contaminant on the surface
of the soil organic carbon is  low, the net  transfer is
from the water to the  soil. Since the transfer is an
equilibrium  process,   it  reverses when the
concentration in the water is  low and the quantity of
contaminant on the surface of the soil organic carbon
is high.
If the leachate contains sufficient quantities of organic
material to affect the solubility of the contaminant, the
modeling  of  retardation is more  difficult.  The  toxic
constituent flow will still be retarded, but not as much.
Instead  of partitioning  between  the  water and  soil
organic carbon, the contaminant will partition between
the polar-organic  fluid  and the soil organic  carbon.
The toxic  contaminant will  spend a smaller fraction of
time on the solid soil particles and a larger fraction of
time in  the  fluid; this will increase its  migration
velocity.  Modeling  this phenomenon,  however, is
complex  and has already  been well documented
elsewhere.  The  analyst  interested  in  modeling
retardation  in complex  leachates  is  referred  to
Nkedi-Kizza  et al.  (1985),  Rao et al.  (1985), and
Woodburn et  al. (1986).
Once a contamination source stops contaminating the
ground water (either a one-time slug  or the end of a
long-term  loading),  the saturated  sorption  sites start
to lose  contaminants to  the clean ground  water that
flows  after  it. This phenomenon  causes the
development of a plume shape that has a long tail of
decreasing  contamination.  Since  the rate  of
desorption is  high  when  the degree  of saturation is
high, and  is lower as the quantity of contaminant  on
the sorption  sites diminishes, the desorption
phenomenon  can  provide a degrading influence  on
the ground water for a long time.
(2) Retardation of Cations
In  cationic retardation, positively  charged ions'
associate with  the soil matrix (clay particles). There is
a smaller effect for anion exchange. Anion exchange
is  due  to  positive charges  associated with hydrous
oxides.  Since soils  typically have  more  negatively
charged clay particles  than positively  charged
hydrous oxides, cations  flow with a  more  retarded
velocity  than  do  anions.  Contaminants  that are  not
charged  are   not  subject to  ionic retardation.
Contaminants  that are compounds or complexed ions
also are not retarded  by ionic retardation.


Cationic retardation is  reversible,  as  is  organic
retardation,  and it forms a  trail of low-level
contamination after the  source of  contamination
stops. Once a  source stops contaminating the ground
water, the  saturated  ion exchange  sites start to lose
contaminants to the clean  ground water. This
phenomenon  causes the development of  a  plume
shape  that  has   a  long  tail  of  decreasing
contamination.  Since  the rate of release is high when
the degree  of saturation is  high, and lower as  the
quantity of  contaminant on  the  ion  exchange sites
diminishes, the reversible ion exchange phenomenon
can  provide a degrading influence  on the ground
water for a  long time.

(3) Transformation/Attenuation
Transformation/attenuation is the  term used  to  model
sinks for  contaminants.  The  particular  type  of
chemical fate  modeled depends  on the contaminant
and the soil characteristics. The  following is a list of
different fate mechanisms:


x*  Hydrolysis
xx  Complexation-chelation
xx Acid/base  reactions
xx  Oxidation/reduction reactions
xx  Biodegradation
xx  Radioactive decay
xx Chemical precipitation
xx  Coagulation
•   Peptization  reactions.

Attenuation  is  modeled with the use  of a  "half-life"
parameter. Whether the degradation is due  to
hydrolysis  or biodegradation, the time  necessary  for
the concentration  to  drop  by  half is the measure of
degradability.


Appropriate  individual decay rates  or overall decay
coefficients  have  been  developed  for some
substances and are  available in the technical
literature. Sources for such data  include: Callahan et
al. (1979); Dawson et al. (1980); Mabey et al. (1982);
Sax (1984); USCG (1974); and Verschueren (1984).
Methods  of  estimating  decay coefficients are
presented by Lyman et al. (1982).
                                                 67

-------
3.5.1.6 Higher Velocity Transport
Some situations  can cause the  migration  velocity of
contaminants  to  be faster than the  ground-water
velocity. Macromolecules can themselves move faster
than  the ground  water, and any hydrophobic
contaminants that  are sorbed onto them will also
move faster. Until recently, hydrophobic contaminants
were  thought to flow with a  retarded velocity only
because  of  preferential  sorption  onto  stationary
organic  soil particles. In  such cases,  the time the
contaminant spends on the stationary soil  particles
lowers its average  velocity. Conversely,  the time  a
contaminant  spends on  a  "high  speed"
macromolecule raises  the average  velocity of the
contaminant.  Since  hydrophobic  contaminants sorb
onto both  stationary and higher velocity
macromolecules, both must be  considered  in order
for the  modeling  of  transport for  hydrophobic
contaminants to be complete.

Macromolecules  may be  found  in ground water in
concentrations ranging from 1  mg/l to  10 mg/l and are
large  enough  that  only the large pore spaces are
available for migration. This means that their average
velocity is the average velocity of  the  large pore
spaces,  and not the average velocity of all the pore
spaces.  The velocity of flow through each pore space
is  a function  of the size of the pore space, and the
larger pore spaces allow faster flow than do the small
spaces.  The velocity difference between the average
large pore space  and the  average pore space is
approximately one order of magnitude.

Macromolecules with large hydrophobic surface area
and small polar surface area will flow with a retarded
velocity  because  of reversible sorbtion onto soil
carbon.  These macromolecules will not cause  higher
speed transport.  Macromolecules with  large polar
surface  areas and  small  hydrophobic surface areas
will travel faster  than the  ground water. These
molecules can speed  up the migration  velocity of
hydrophobic contaminants.

Macromolecular transport  is not frequently modeled;
however,  when  such  modeling  is  necessary,  the
analyst  can  refer to  Enfield and  Bengtsson (n.d.) for
detailed guidance.
3.52 Ground- Water Modeling Equations and
Nomograph
This  section provides a  number of  hydrologic
modeling equations and  a  nomograph.  In no  cases
will all  equations  be necessary; depending on the
observed chemical contaminant,  a  discrete subset of
the equations will be useful  in assessing  the ground-
water  contamination problem  at  a specific
uncontrolled hazardous waste site.

Five discrete classes of contaminant are  discussed.
Each class is based  on a different technique for
calculating contaminant migration.  The five classes of
contaminant can have dramatically different calculated
velocities and  concentrations; use of the appropriate
analytical techniques for each class is thus necessary
for accuracy.

Estimating contaminant  velocity  is  based on
estimating water velocity. For those contaminants that
flow as water flows, contaminant velocity equals water
velocity  (vertical or  horizontal). For those  that flow at
rates different from water, the  estimated  water
velocity  must be adjusted to approximate that  of the
contaminant.
3.5.2.1 Calculating Ground-Water Velocity
Ground-water velocity can be determined for both
the saturated zone  and the  vadose  (unsaturated)
zone. Vadose zone velocity is discussed in the next
section; saturated  zone velocity  is discussed in this
section.

Ground-water  velocity in the  saturated zone is
calculated using Darcy's Law (Bouwer 1978):
v =  Ksi
where
    Ks

     i
                                 (3-9)
  Darcy velocity of water, also termed
   superficial  velocity,  or  specific
   discharge, (length/time).
  hydraulic conductivity  of  soil or aquifer
   material, (length/time).
  hydraulic gradient, (length/length).
However, v, the  Darcy  velocity,  is  not  the real
macroscopic velocity of the water, but the velocity as
if the water were  moving through the entire  cross-
sectional area normal to the flow, solids as well  as
pores (Bouwer  1978).  The  ground-water velocity  is
calculated from the  Darcy  velocity by  dividing  it  by
soil  porosity, or,  for  more precise modeling,  by
effective  porosity  (thus  taking into account  the fact
that the entire  cross-section of the  pore is not
flowing (i.e., due to boundary layer effects).  For clay
soils, the effective porosity also corrects for the effect
of electro-osmotic  counterflow and the  development
of electrokinetic streaming potentials  (Bouwer  1978).
The equation for  calculating  ground-water velocity
from Darcy  velocity using  effective  porosity  is  as
follows (Bouwer 1978):
Vpw = V/Pe

where
                               (3-10)
   VPW


     V
= ground water (pore water)  velocity,
   (length/time).
= Darcy velocity (superficial velocity,
   specific discharge), (length/time).

-------
    p     =  effective porosity,  (dimensionless
              fraction).

The  above terms  should  be determined for the  site
being studied. If this is not possible for all parameters,
then  literature  values can be used for the  few
parameters that are not available. Literature values for
saturated hydraulic  conductivity  are  presented in
Table 3-8  (Rawls et al. 1982) and Table  3-9
(Freeze and Cherry 1979).

The  hydraulic gradient (the change in  the elevation of
the water table  over distance  from the site) should
also be taken from  field  data  developed during  site
investigation.  Water levels  in  existing  nearby wells
can  also provide an  indication  of hydraulic gradient.
Table  3-10  provides values for saturated moisture
content, which is roughly equal to the  effective
porosity,  or Pe, for several soil types.

It  must  be  emphasized  that site-specific  data  are
highly  preferable to regional data, or data obtained
from  any of  the  above-referenced  tables.  If site-
specific information on effective porosity is  available,
it should be used; however, literature values for soils
with  the same hydraulic conductivity provide sufficient
accuracy. Effective porosity (P,)  can be approximated
by the difference between the moisture content at
saturation and at the wilting  point (-15 bar)*.  The
equation is as follows (Rawls 1986):
Pe = e8-9(-15)

where
(3-11)
This estimation  procedure addresses the fraction of
the pore spaces that is contributing to flow, but does
not address  the  effect  of electro-osmotic
counterflow and the development of electrokinetic
streaming  potentials.  For  clays,  this can  be  a
significant difference. Literature values listed in Table
3-10 should  be  used for clay  solids (these  values
incorporate the effects of the clays ionic double layer)
(Rawls  et al. 1982); either technique can be used for
sand or loam soil.

The above method for predicting the average velocity
of  ground  water is the  most  widely  accepted
approximation;  however, it is only an approximation
            and  further  refinement  of this  approach  would
            improve accuracy.  Corrections for the  path  length
            difference  between  the  straight line  distance  versus
            the  tortuous path through which ground water flows
            can improve the precision (Freeze  and Cherry 1979),
            although the literature does  not provide a  consistent
            correction  factor to  apply. To provide  a feel for the
            magnitude of this correction  the analyst can  review
            Das  (1983) which suggests a correction of  1.41. This
            value can  be  used to correct  the  velocity  or the
            distance (not both)  by  dividing the  number  by 1.4.
            However, the  analyst must interpret  the  results
            obtained through such  correction  with  care,  as the
            degree to which the factor cited in  Das applies to any
            given site is uncertain.
3.5.2.2 Calculating the Velocity of Infiltrating
Rainwater
This section discusses the calculation  of the velocity
of percolating  rainwater  flowing through the vadose
zone.  Darcy's law  can  be  used  to  calculate  the
unsaturated flow  velocity;  however,  the hydraulic
conductivity must be corrected to reflect the effect of
partially-filled  pore  spaces when  the  hydraulic
loading  is below that necessary to  support saturated
flow.

Interstitial  pore  water  velocity  for  unsaturated
transport through the vadose zone can be calculated
as follows (Enfield  et al. 1982):
                                                       'pw
                =  q/0
                                          (3-12)
    Pe    =   effective  porosity,   (fraction,    where
              dimensionless).
    9S    =   water content when the pores are fully
              saturated,  (fraction, dimensionless).
9(-15)    =   wilting point moisture content, (fraction,
              dimensionless).
                pw

                 q

                8
  'Wilting point is determined by drawing a suction of -15 bar to
  draw water out of the soil in a manner similar to the suction of a
  plant root. Bar is a measure of pressure (dynes/cm2).
          =   interstitial  ground water  (pore water)
              velocity, (length per unit time).
          =   average percolation or recharge  rate,
              (depth per unit time).
          =   volumetric  moisture  content  of  the
              unsaturated  zone, (decimal  fraction,
              representing  volume of  water  per
              volume of soil).

This equation applies to  steady-state conditions, or
those  that  can  be  assumed  to   be  steady.  For
unsteady hydraulic loading, the "q"  and "0"  will vary
with time and depth. Additionally, the distribution of
"q" and "0"  will vary as the  moisture migrates down.
This makes  determination of the average  transport
velocity burdensome.  For situations where  steady-
state  conditions  cannot  be  assumed,  the  analyst
should use a computer model; for example, SESOIL
(one of EPA's GEMS  computer system) calculates
the time of travel for seasonally varying rainfall rates.

The volumetric water content (0)  in the  unsaturated
zone can be estimated using the following  equation
(Clapp and Hornberger 1978*:
                                                   69

-------
Table 3-8.     Representative Values of Saturated
              Hydraulic Conductivity
Hydraulic
conductivity
Soil texture Number of soils3 («,.; cm/sec)b
Sand
Loamy sand
Sandy loam
Loam
Silt loam
Sandy clay loam
Silt clay loam
Clay loam
Sandy clay
Salt clay
Clay
762
338
666
383
1,206
498
366
689
45
127
291
5.8 x
1.7 x
7.2 x
3.7 x
1.9 x
1.2 x
4.2 X
6.4 x
3.3 x
2.5 x
1.7x
10-3
10-3
10-4
10-4
10-4
10-4
10-5
10-5
10-5
10-5
10-5
Table 3-9.     Saturated Hydraulic Conductivity Ranges
              for Selected Rock and Soil Types
         Saturated Hydraulic  Conductivity (cm/sec)
aNumber of Individual soil samples included in data
compiled by Rawls et al. 1982.
bpredicted values based on compiled soil properties.
Source:  Adapted from Rawls et al. 1982.
Soils
Unweathered marine
clay
Glacial till
Silt, loess
Silty sand
Clean sand
Gravel
Rocks
Unfractured
metamorphic and
Igneous rock
Shale
Sandstone
Limestone and
dolomite
Fractured Igneous and
metamorphic rock
Permeable basalt
Karst limestone

5xlO-"
10-10
IO-J
10-5
10-4
10-1

10-2
5 x ID-12
10-8
5 x 10-8
10-6
10-5
ID-4

- io-J
ID-4
I0-3
- 10-1
i
- 102

- 10-8
IO-J
5 x IO-4
5 x 10-4
_ 10-2
1
1
                                                              Source: Adapted from Freeze and Cherry  1979.
Table 3-10.   Representative Values for Saturated Moisture Contents and Field Capacities of Various Soil Types
                                      Saturated moisture content (9s)a
                   Field capacity (cm3/cm3)b
Number of soils
Sand
Loamy sand
Sandy loam
Loam
Silt loam
Sandy clay
loam
Clay loam
Silty clay loam
Sandy clay
Silty clay
Clay
762
338
666
383
1,206
498
366
689
45
127
291
Mean
0.437
0.437
0.453
0.463
0.501
0.398
0.464
0.471
0.430
0.479
0.475
± 1 Standard deviation
0.347
0.368
0.351
0.375
0.420
0.332
0.409
0.418
0.370
0.425
0.427
- 0.500
- 0.506
- 0.555
0.551
0.582
- 0.464
_ 0.519
- 0.524
- 0.490
- 0.533
- 0.523
Mean
0.091
0.125
0.207
0.270
0.330
0.255
0.318
0.366
0.339
0.387
0.396
± 1 Standard deviation
0.018
0.060
0.126
0.195
0.258
0.186
0.250
0.304
0.245
0.332
0.326
- 0.164
-0.190
0.288
- 0.345
- 0.402
- 0.324
- 0.386
- 0.428
- 0.433
- 0.442
- 0.466
aFrom total soil porosrty measurements compiled by Rawls et al. (1982) from numerous sources.
bwater retained at -0.33 bar tension; values predicted  based on compiled soil property measurements

Source: Rawls et al. 1982.
                                                       70

-------
0 = (0s)*(q/Ks)l/<2b + 3)>

where
                                          (3-13)
     0


    0s


     q
          =  volumetric  water content  in  the
             unsaturated zone,  (volume/volume  or
             unitless).
          =  volumetric  water  content of soil under
             saturated conditions, (volume/volume
             or unitless).
          = percolation  rate (assumed to be equal
             to  the  unsaturated  hydraulic
             conductivity term  in original  Clapp and
             Hornberger equation),  (depth  per unit
             time).
    KS    = saturated  hydraulic conductivity,  (depth
             per unit time).
     b    =  soil-specific  exponential parameter,
             (unitless).

Representative  values  of  "b" and  the term
"  1/(2b+ 3)" are listed in Table 3-11.
Table 3-11. Representative Values of Hydraulic Para-
meters (Standard Deviation in Parentheses)
Soil texture No- of bb — 1— esC
soils3 2b + 3
Sand
Loamy sand
Sandy loam
Silt loam
Loam
Sandy clay
loam
Silt clay loam
Clay loam
Sandy clay
Silt clay
Clav
13
30
204
384
125
80

147
262
19
441
140
4.
4
4.
5
5
7

7
8
.05
.38
.90
.30
.39
.12

.75
.52
10.40
10.40
11
.40
(1.78)
(1.47)
(1.75)
(1.87)
(1.87))
(2.43)

(2.77)
(3.44)
(1.64)
(4.45)
(3.70)
0
0
0
0
0
0

0
0
.090
.085
.080
.074
.073
.058

.054
.050
0.042
0
0
.042
.039
0.395
0.410
0.435
0.485
0.451
0.420

0.477
0.476
0.426
0.492
0.482
(0.056)
(0.068)
(0.086)
(0.059)
(0.078)
(0.059)

(0.057)
(0.053)
(0.057)
(0.064)
(0.050)
aNumber of individual soil samples included in data compiled by
  Clapp and Hornberger (1978).
bEmpirical parameter relating soil matric potential and moisture
  content; shown to be strongly dependent on soil texture.
cWolumetric soil moisture content (volume of water per volume of
  soil).
Source: Adapted from Clapp and Hornberger 1978.
The  saturated volumetric water  content  (@s)>
saturated  hydraulic  conductivity  (Ks),  and  the
exponential  function  (b)  are  all  related  to  soil
properties.  The  most  reliable values  for  these
parameters  are  empirical values (if  available)
measured during site  investigation.  Where  empirical
values  are  unavailable,  values  in Tables 3-10
through 3-11 provide guides for the rough estimation
of 0S,  Ks,  and the term  1/2b + 3  . Representative
values from two different  sources are presented for
Ks (Tables 3-8 and 3-9)  and  0S  (Tables 3-10  and
3-11), in order to  demonstrate the variability  in
estimates for these values.

Note  that  the value  6 cannot exceed  0S, the
saturated soil  moisture  content. When  0 calculated
by Equation 3-13 equals or exceeds  0s, it must  be
assumed that saturated conditions  exist.  In  such
cases, use  Equations  3-9 and  3-10.

Similarly, the minimum value for 0 that is applicable
to Equation 3-13 is the field capacity of  the soil. This
value represents  the volumetric  moisture content
remaining  in  the  soil  following  complete  gravity
drainage and  is the moisture  content  below  which
downward  flow  of  water due to gravity through
unsaturated soil ceases. Field capacity is a function of
soil type; the most reliable values are those measured
empirically. Where measured values are  not available,
default  values can be taken  from  Table  3-10.
Wherever Equation 3-13 results  in a value for 0 that
is less than the specific retention of the  soil, it should
be assumed that no downward movement of moisture
(and  dissolved contaminant)  occurred  for the
associated  time increment, and  that Vpw is equal to
zero.

Note  that the percolation rate  (q) cannot exceed the
saturated hydraulic conductivity (Ks) for the site soil.
Whenever q > Ks (and therefore 0 as  calculated  by
Equation 3-13 > 0S)  for the duration  of  the  study
period, it must be assumed that saturated conditions
exist  and that saturated  flow prevails. Equations 3-9
and  3-10 in the preceding subsection provide  a
means of estimating saturated flow velocities.

The  following  equation  provides  an estimate  of the
term q (Enfield et al. 1982):
                                                               Pr-ET-
                                                                                                (3-14)
                                                      where
                                                         HL
                                                          Qr =
          =  hydraulic  loading from  manmade
             sources, (depth per unit time)
          = precipitation,  (depth  per  unit time)
          = evapotranspiration, (depth per unit
             time)
         runoff, (depth per unit  time).
                                                      Records  of  estimated  percolation  rates for the site
                                                      locality during the time period  in question  (or annual
                                                      average percolation rate estimates) are often available
                                                      from local climate or soil authorities, including regional
                                                      U.S.  Geological Survey  (USGS)  and  U.S. Soil
                                                      Conservation Service offices.

                                                      An  estimation  procedure can  be  used to  evaluate
                                                      percolation rates (q) at sites where the sources  listed
                                                      above cannot provide them directly.  This  estimation
                                                      procedure requires data for precipitation, evaporation,
                                                      and runoff rates. In addition  to the above two sources,
                                                   71

-------
  Table 3-12.  Suggested Value for Cet Relating Evaporation from a US Class A Pan to Evapotranspiration from 8 to 15-cm
             Tall, Well-Watered Grass Turf
                              Pan surrounded by a short green crop
                          Pan surrounded by a dry surface ground
Upwind fetch of
crop (m from
Wind pan)

Light < 170 km/day



Moderate 170-425 km/day



Strong 425-700 km/day



Very strong > 700 km/day


0
10
100
rtioo
0
10
100
1000
0
10
100
1000
0
10
100
1000
r-vvciayc i cv
20-40
0.55
0.65
0.7
0.7
0.5
0.6
0.65
0.7
0.45
0.55
0.6
0.65
0.4
0.45
0.5
0.55
jiui iai iciau v<
%*
40-70
0.65
0.75
0.8
0.85
0.6
0.7
0.75
0.8
0.5
0.6
0.65
0.7
0.45
0.55
0.6
0.6
	 y' Upwind fetch of
dry fallow (m
>70 from pan)
0.75
0.85
0.85
0.85
0.65
0.75
0.8
0.8
0.6
0.65
0.7
0.75
0.5
0.6
0.65
0.65
0
10
100
1000
0
10
100
1000
0
10
100
IOOO
0
10
100
1000
rvvciayc icyiuii
relative humidity,
20-40 40-70
0.7
0.6
0.55
0.5
0.65
0.55
0.5
0.45
0.6
0.5
0.45
0.4
0.5
0.45
0.4
0.3
0.8
0.7
0.65
0.6
0.75
0.65
0.6
0.55
0.65
0.55
0.5
0.45
0.6
0.5
0.45
0.4
ai
%*
>70
0.85
0.8
0.75
0.7
0.8
0.7
0.65
0.6
0.7
0.65
0.6
0.55
0.65
0.55
0.5
0.45
  'Mean of maximum and minimum relative humidities.
  Source: Jensen 1973, as presented by Enfield et al. 1982.

the National Weather Service, Forest Service offices,
National  Oceanic  and Atmospheric  Administration
(NOAA) gauging  stations,  or other first order weather
stations (e.g.,  at local airports)  are  possible sources
for these three types of data.


The  average precipitation rate  per unit time (P,)  for
the study period  can be obtained from various local
weather authorities such as those listed above.


ET is  estimated by  using  measured Class A  pan
evaporation rates (a  measure  of local  evaporation
rates under standardized  conditions,  available from
the nearest NOAA gauging station) in the equation:
              C
               veg
             correction  factor  for converting
              evapotranspiration from  turf grass to
              evapotranspiration  from other
              vegetative  cover types, (unitless).
ET = EVAP x Cet x Cve

where
(3-15)
EVAP    =  region-specific  or  site-specific
              measured evaporation rates, (depth per
              unit time).
    Get    =  correction  factor  for  converting
              measured  pan  evaporation rates to
              evapotranspiration  rates  from  turf
              grass, (unitless).
Values  for  Cet  are taken  from Table  3-12,  which
requires  climatological and pan descriptive  in-
formation.

The  term Cveg  is available  mainly for agricultural
crops (Table  3-13), and varies  with  the thickness,
depth, and characteristics of vegetative cover. Typical
values are 0.87 for shorter broadleaf plants (alfalfa) to
0.6 for taller broadleaf plants  (potatoes,  sugar  beets)
and  0.6  for taller grains and grasses.  Where  crop-
specific  data  are  unavailable, a  conservative default
value for  this  term is the smallest  reasonable  value,
or 0.6.

Qr, or  the average runoff over the study period,  is
estimated by the method presented  in Section  2.4  of
this manual. A more reliable value for this term can
be obtained from  local  USGS gauging  stations. For
relatively  level  sites,  a reasonable  conservative
default  value for  the  purposes of this  estimation
procedure is  that  Qr  = 0,  where  site-specific data
are unavailable or  cannot be estimated.
                                                   72

-------
     Table 3-13.
        crop
     Crop Coefficients for Estimating
     Evapotranspiration
                       Period
                          Coefficient
                            (cveg)
     Alfalfa       April 1 - October 10         0.87
     Potatoes     May 10 - September 15      o.e5
     Small grains April  1 - July 20            o.e
     Sugar beets  April 10 - October 15         o.e

     Source: Jensen 1973, as presented by Enfield  et al.
     1982.


The above method for predicting the velocity  of water
migrating  through  the  vadose zone is the  best
approximation  available;  however, real  world  non-
homogeneities, such  as  root holes and macropores,
can  result in  faster velocities  than  predicted.  The
analyst  is  not expected  to  correct for  this,  yet  it  is
important  to  be  aware  of the  limitations of the
method.

3.5.2.3 Corrections for Viscosity and  Density
When the  movement of liquids other  than  water  is
calculated, the saturated and the  unsaturated
hydraulic conductivity must be  corrected  for the
density  and viscosity of the non-water  liquid.  The
equation for this correction is as follows:

Kc =  Kc*(density  of chemical/density of water)   (3-16)
      '(viscosity of water/viscosity  of  chemical)
where
    Kn
= hydraulic  conductivity of water (Darcy's
   coefficient), (saturated or unsaturated)
= hydraulic conductivity  of chemical,
   (saturated or unsaturated).
When the migration velocity through the vadose zone
is calculated,  density  and  viscosity  should  be
corrected with the above equation.  For horizontal flow
below the water table,  density and viscosity should be
factored  in when the hydraulic gradient is the slope of
the chemical plume. In many cases, one can assume
that the  thickness of the  concentrated chemical
plume  is relatively  constant. For such situations,  the
slope of the concentrated chemical is  zero  and  the
analyst  should  not  correct for the density. The  slope
(hydraulic gradient)  is that of water, and the Darcy
coefficient reflects the  density of water. However,  the
viscosity of the chemical  is the  viscosity of the
flowing  fluid  of concern,  and the analyst should
correct for the viscosity.

3.5.2.4  Retardation Effects
Hydrophobic or  cationic contaminants  that are
migrating as a dilute solute are subject to retardation
effects.  Concentrated  plumes are not  subject to this
phenomenon.  Contaminant  migration as a  dilute
solute in  complex leachates  containing  organic
constituents will  show some retardation, although not
as much  as in pure ground water.

When a  hydrophobic contaminant flowing in a dilute
plume flows past a soil particle that contains organic
carbon, the contaminant partitions between  the  polar
solvent (water) and the  solid  organic  carbon. When
the concentration in  the  water is  high  and  the
concentration on the  soil particle  low, the  net
migration is  from the water to  the soil. When the
reverse  occurs and the concentration in the water is
low and  the concentration on the soil particle is  high,
the net migration  is from the soil particle to the water.
When the  water  and  soil  concentrations  are in
equilibrium,  there  is no  net migration. However, the
flux from the soil to the water and the flux from the
water to  the soil are not zero; rather, they  are positive
fluxes that are equal and are in opposite directions.
When the  partitioning  is  between  concentrated
chemical and soil  particles, the contaminant does not
prefer the solid  "solvent"  effects of the  organic
carbon in the soil to the organic  liquid  solvent effects
of  the   concentrated   chemical plume.   Hence,
hydrophobic  contaminants partition out  of polar
solvents  (water) but not  out of hydrophobic solvents,
and thus,  retardation effects are modeled  for dilute
plumes only.

Retardation can  be modeled for complex leachates,
but the methods are not presented in this report. The
reader is referred to Nkedi-Kizza et al. 1985, Rao et
al. 1985, and Woodburn et al. 1986, for guidance on
performing these calculations.

The retardation  protocol is based on the  assumption
that adsorption of hydrophobic contaminants is due to
sorption  to organic carbon in the  soil.  Basing the
adsorption coefficients on soil organic carbon  rather
than total  mass eliminates  much, but not all, of the
variation  in sorption coefficients between different
soils.  The remaining variation may  be due to  other
characteristics such as  surface  area of soil particles
per mass of soil  (function of particle size). Numerous
studies  of  the  correlation  of Kd with  various soil
variables have found that the  organic  carbon  content
usually  gives  the most  significant correlation.
Furthermore,  this  correlation often extends  over a
wide  range of organic  carbon  content  —  from 0.1
percent to nearly 20 percent of the soil in some cases
(Lyman et al. 1982).

This  protocol estimates hydrophobic  retardation
based on  soil  organic  carbon,  but  it  should  not be
taken to  imply that hydrophobic  contaminants will not
adsorb  on  minerals free of organic  matter. Some
adsorption  will  always  take place, and it may  be
significant under certain conditions, such  as  clay soils
(high  surface area  per  mass of soil) with  very low
organic  carbon  content (no appreciable  sorption to
nonexistent organic carbon). Unfortunately,  methods
                                                   73

-------
for  estimating  adsorption coefficients under these
conditions are  not  currently  available (Lyman et al.
1982). The protocol discussed in this report relies on
the  percent of organic carbon content of the soil.

To  simplify  modeling,  equilibrium  conditions  are
modeled as the contaminant velocity  being a  fraction
of the ground-water velocity. If  the  analyst thinks  of
the  time an individual portion  of the contaminant mass
is in  the water as the  time  it  has ground-water
velocity, and  the time the contaminant is  on the  soil
particles as the time the contaminant does not have a
velocity,  the  contaminant velocity is related to  the
ground-water velocity  by the  ratio  of time  on  soil
particles to time in the water. The ratio of time in the
water to time on the soil particles is the same  ratio as
the  concentration ratio at equilibrium.

In complex leachates containing organics, the time a
hydrophobic contaminant spends on the solid carbon
is reduced because the  ratio  of the contaminant's
solubility in the fluid to its solubility on soil carbon  is
increased. The hydrophobic contaminant partitions
between the organics in the flowing fluid  and  the
organics that  are solid.

The same logic applies to cation retardation, and the
contaminant velocity for  cations is also  modeled as
fraction of ground-water velocity.

The equation used to  calculate  the  retardation is as
follows (Kent  et al. 1985):
Rd - 1

where
          (B* Kd)/pt
(3-17)
            R=
            pt=
            Kd=
                   retardation factor, (unitless).
                  bulk  density,  (g/ml).
                   total porosity, (unitless).
                   distribution factor  for sorption  on
                   aquifer  medium  (from sorption
                   isotherm column studies, or from
                   regression equation  based on the
                   octanol/water partition  coefficient,
                   (in ml/g).
The  use of the  retardation factor is described  in the
following equation (Kent et al. 1985):
where
         Rd  =
         V   =
         vpw

         vd   =
                                          (3-18
                 retardation factor, (unitless).
                 velocity  of ground  water, (same
                 units as Vc' length/time).
                 velocity  of  contaminant,  (same
                 units as V^' length/time).
                                                      The term Kd  is based on sorption  isotherm  column
                                                      studies. While this is the  more precise approach, the
                                                      analyst will typically  have to work with  estimated
                                                      parameters. For hydrophobic  contaminants, the term
                                                      Kd can be estimated from the term Koc (Lyman et al.
                                                      1982):


                                                      Knn  =  Kr|/f,
                                                              d"oc
                                                      where
                                                               Kd  =
                             partition  coefficient for  organic
                             carbon, (ml/g).
                             distribution factor for soil, (ml/g).
                             fraction of organic  carbon in  the
                             soil.
The term "fraction of organic carbon" (foc) is  precise
when taken from empirical measurements of the soil
in the  study area. For cases  where this  is  not
possible,  estimates  can  be made.  For the vadose
zone velocity, a  value of foc from Rawls (1986)
provides a good estimate.  Rawls'  work focused  on
soils near the surface,  the area of interest to
agriculture. For  saturated zone  velocity, the  analyst
has two choices. If the subsoil came from igneous or
metamorphic rock, the foc decreases with depth.  The
actual value may be quite low; however, the model to
predict retardation  is only useful down to 0.1 percent.
For this situation, the analyst should use 0.1 percent
for the foc.  If the subsoil came from sedimentary rock,
the foc distribution may  be similar to the distribution for
agricultural soils done  by Rawls.  The variation of foc
with depth may  be relatively constant.  The  carbon
was at the surface at one time, and  has been buried
over geological time. Hence, the analyst should use a
value of foc from the Rawls  (1986) distribution  for the
saturated  zone velocity  determination  (Trask  and
Patnode 1942).  Soil/water partition   coefficients  have
been developed for many contaminants of importance
(Callahan et al. 1979 and Mabey et al. 1982).

If Koc  is  not known,  it can be  estimated from
regression  equations  that relate  Koc  to  Kow
(octanol/water partition  coefficient). There  are six
regression equations  that  relate Koc to Kow.  The
equation that was  based on a chemical class  closest
to the   subject contaminant  should  be used. If the
contaminant does not fit into a specific class, the first
regression  equation  should  be used  because it was
based  on the  largest  sample.  The  regression
equations are  as follows (Lyman et al. 1982):
                                                      Log KOC = 0.544 log K™ + 1.377
                                                                                                (3-20)
            based  on  a  wide variety of contaminants,  mostly
            pesticides
                                                      or
                                                   74

-------
log Koc = 0.937 log Kow - 0.
006
(3-21)
based  on aromatics, polynuclear aromatics, triazines
and dinitroaniline herbicides

or

log Koc = 1.00 log K,,w - 0.21                (3-22)

based  on mostly aromatic or polynuclear aromatics

or

log Koc = 0.94 log Kow + 0.02               (3-23)

based  on s-triatines and dinitroaniline  herbicides
or
logKoc= 1.029 logKow -0.18
                 (3-24)
based  on  a  variety of insecticides, herbicides, and
fungicides
or
log Koc = 0.524 log Kow + 0.855
                 (3-25)
based  on substituted  phenylureas  and  alkyl-N-
phenylcarbamates.

The retardation effects are  computed  from  the
octanol/water partition coefficient (Kow), which  relates
the concentration in polar  solvent  (water)  to  the
concentration  in  hydrophobic  solvent  (octanol
simulating  the  soil  organic  carbon).  If  the
contaminated plume has  a large concentration of
organic chemicals dissolved in the ground water, the
actual  partitioning  will  be from  a solvent/organic
chemical system. This will  raise the concentration in
the fluid  and lower  the concentration  on the  soil
organic carbon. This shift in partitioning will  lower Rd,
(i.e., the contaminant will migrate at a speed closer to
that of ground water). Much of the  solubility of
extremely hydrophobic contaminants in  the water of
an  octanol/water partition  coefficient  test is  due to
dissolution in the octanol that is dissolved in  the water
rather than dissolution into water. This effect depends
on  the degree to which the water  is not pure water;
for most low-level  contamination  situations,  this
effect can be ignored. This  manual does not present
equations for calculating a numerical correction for
this effect.  The analyst should  be aware  of the
general influence of this effect,  but  not model the
precise numerical  difference. For  dilute  plumes, the
analyst should model full retardation; for concentrated
plumes, the analyst should model no retardation.

3.5.2.5 Contaminant Velocity
The velocity of concern is the actual contaminant
velocity.  The  determination  of ground-water velocity
discussed earlier is done  to provide a foundation for
calculating  the  contaminant  velocity.  The  particular
method used  for  determination of  the  contaminant
velocity is  dependent  on the  type  of ground-water
transport the chemical undergoes. Thus, the first step
in  calculating  the  velocity is  classifying the subject
contaminants as to migration class.

Once the molecular  identity  of the  contaminant is
known,  three  determining parameters can  be  taken
from literature:

1.  Physical  state  at  room temperature  (i.e., is it  a
    solid or a liquid?)

2.   Hydrophobicity  (i.e.,  is it  hydrophilic or
    hydrophobic?)

3.  Density  (i.e., is  it  less dense than water?,  Is its
    density near that  of water?, Or is it more dense
    than water?)

The five migration classes are as follows:
Migration
class #
A)

B)

C)

D)

E)

Vadose zone transport
Solid/carried by
precipitation
Hydrophilic liquid/
waste percolation
Hydrophobic liquid/
waste percolation
Hydrophobic liquid/
waste percolation
Hydrophobic liquid/
waste percolation
Saturated zone
transport
Solute transport

Solute transport

Low density/
floater transport
Medium density/
buoyant transport
High density/
sinker transport
                             Although the specific chemical will migrate  according
                             to the above classes, it is important to note that the
                             concentrated  plumes will  also  have a  dilute  plume
                             near  them.  For  mass flux considerations,  the
                             concentrated plume will dominate.

                             (1) Migration Class #A: Solid Material
                             Solid  material  will  dissolve  into  percolating
                             precipitation  and  migrate  as a  solute. Precipitation
                             provides  the hydraulic loading that drives the rate  of
                             release.  The plume exists  as a single plume (for
                             single chemical  contaminant)  that has  a single
                             average velocity. Unretarded contaminants move with
                             the  ground  water,  and  hence, the ground-water
                             velocity  is  the  contaminant  velocity.  Retarded
                             contaminants move with a velocity that is slower than
                             ground-water  velocity,  and  therefore  the
                             contaminant  velocity is based  on  the ground-water
                             velocity adjusted for retardation.  Typically, the velocity
                             is a fraction  of  the ground-water velocity.

                             (2) Migration Class 49: Hydrophilic Liquids
                             Liquids will directly percolate into the soil (i.e., without
                             waiting  for precipitation  to  cause leaching). The
                                                   75

-------
hydraulic loading  is  due to  the combination of
chemicals'  hydraulic loading  and  that due to
precipitation. The velocity of transport through the
vadose  zone must be  calculated with corrections for
the density  and viscosity  of the  contaminant.  The
plume exists  as a single plume  (for a single chemical
contaminant)  that can be considered to have a single
average velocity. Unretarded contaminants move  with
the ground  water,  and  hence,  the ground-water
velocity  is the contaminant velocity. This  is only exact
after the plume  has  mixed with the  ground-water to
the point that its density and viscosity  are similar to
those of water.  When the plume first  reaches the
water table,  it has not mixed with  very  much water,
and  its  density and viscosity differences  suggest
calculating a contaminant velocity that is  different
from the  ground-water velocity. Since the velocity
difference varies gradually from the source to the
point downgradient where  it is  well  mixed,  this
calculation is complex. Therefore, the analyst should
calculate as  if the ground-water velocity represented
the contaminant  velocity for the length  of the plume.
The analyst  should be  aware  of the limitations of this
method.  Retarded  (cationic) contaminants move  with
a  velocity that is slower than  ground-water velocity.
In this  case, contaminant  velocity is based on the
ground-water velocity adjusted for  retardation, and  is
a  fraction of the ground-water velocity.

(3) Migration Class #C: Hydrophobia Liquids Low
Density
Once hydrophobic  liquids  reach the water table,  they
form  two distinct plumes (for a single  chemical
contaminant), with  each  having  its own  average
velocity. The concentrated plume will  float on the
surface of the  water table and move  in the same
direction as  the ground-water flow.  Its  velocity  is a
function of the contaminant's  viscosity. If mounding is
significant, the density must also be  considered. The
dilute  plume is formed  by  small amounts of the
chemical dissolving  in  water as  limited  by the
hydrophobic  chemical's solubility.  This  plume will be
found below  the concentrated plume, with  the highest
concentration near the concentrated plume. From the
point where  the  contamination  leaves  the
concentrated plume to form  the  dilute plume, the
dilute plume  will move with the ground-water flow (at
a  retarded  velocity).  The concentrated  plume will
have  a  single average velocity, and it will  start at the
location  of the  source. The dilute  plume will have a
single  average  velocity, but its starting  point can be
from the  location of the  source, or it can form  from
the concentrated plume anywhere along the  length of
the concentrated plume.

Retarded contaminants in the dilute plume move with
a  velocity that  is slower  than  ground-water velocity.
Thus,  contaminant  velocity,  based  on the ground-
water velocity adjusted for retardation,  is typically a
fraction  of the  ground-water  velocity.   Contaminants
in  the concentrated plume do  not  move with  the
ground-water velocity;  their  velocity must be
determined  by  considering the  effect of  the
hydrophobic contaminant's  viscosity.  The
concentrated  plume  does  not  exhibit  retardation
effects.  If mounding is significant, the  analyst  also
must factor in the  density.

(4) Migration Class #D: Hydrophobic Liquids/Medium
Density
This  class of compounds  migrates similarly to Class
#3, except that the  concentrated plume  will not  float
or sink, but will have more  or less neutral buoyancy. It
will move in the  direction of ground-water flow, but
its migration velocity will be a function  of its viscosity.
Again, the dilute plume will surround the  concentrated
plume, forming  a transition zone  between  the
uncontaminated water  and  the  concentrated  plume
body. From  the point where  the contaminant  leaves
the concentrated  plume to form the dilute  plume, the
dilute plume will move  with the ground-water flow (at
a  retarded velocity).  The concentrated plume  will
have  a single average velocity, and it will start at the
location  of the   source,  or  it can form from  the
concentrated plume anywhere along the  length of the
concentrated plume.

Retarded contaminants in  the dilute plume move  with
a  velocity that is  slower than ground-water velocity.
Thus, the contaminant  velocity,  based on  the
ground-water velocity  adjusted for retardation,  is  a
fraction of the ground-water velocity.  Contaminants
in  the concentrated plume do  not  move with  the
ground-water  velocity;   their  velocity must be
determined  by  considering the  effect  of  the
hydrophobic contaminant's  viscosity.  The
concentrated  plume  does  not flow with  retardation
effect.

(5) Migration Class #E: Hydrophobic Liquids/High
Density
As with low and  medium density hydrophobic% once
a  high density  plume reaches the water table,  it forms
two   distinct plumes  (for a single  chemical
contaminant)  with  each  having  its  own average
velocity.  The  concentrated plume will  sink  to the
bottom of the  aquifer. Its velocity is a  function of the
contaminant's  viscosity. If mounding  on  the aquitard
is  significant,  the density  must  also  be considered.
The  dilute plume will be above  the concentrated
plume, with the highest concentration  near  the
concentrated  plume and  the lowest concentration  at
the farthest distances  from  the concentrated   plume.
The  concentrated plume will have a  single average
velocity and will  start at the location  of the  source.
The  dilute plume  will have a single average velocity,
but its starting point can be from the  location of the
source, or it can form from the concentrated plume
anywhere along the length of the concentrated plume.
                                                  76

-------
Retarded contaminants in the dilute plume move with
a velocity that is slower than  ground-water velocity.
Thus,  the contaminant  velocity, based  on the
ground-water  velocity  adjusted  for retardation,  is a
fraction  of the  ground-water velocity. Contaminants
in  the  concentrated plume do not move with the
ground-water velocity;  their velocity  must be
determined  by  considering the  effect  of the
hydrophobic  contaminant's viscosity. If the sinker
mounds above the  aquitard significantly, the density
should  be taken  into consideration. The concentrated
plume does not flow with retardation effects.

3.5.2.6 Nomograph Technique
The  following nomograph  is based on  a solution  to
solute transport in an aquifer from  a point source that
extends  throughout the thickness of the  aquifer.
Contaminant  transport from  the source  includes
advective flow with the ground water and longitudinal
and  transverse  dispersion  (see  Wilson and  Miller
1978). The nomograph  is  taken  from  Kent et al.
(1985);  the  analyst  is  referred to  this  source
document for further  discussion  of the use of the
nomograph and its limitations.

The  nomograph,  which is  a  one-dimension model
(results  restricted  to a  line, dispersion  is two-
dimensional),  is intended as a  rapid means to obtain
an approximate  solution. Scale factors  are  used  to
translate Wilson and Miller (1978) to nomograph form.
Dilution/dispersive mixing and retardation parameters
are included in the solution.

Three scale  factors that must  be calculated before
using the nomograph are:

  X  =—                                (3-26)
      RdDx
               Dy
                                          (3-27)
(3-28)
Two of the three ratios are computed directly, and the
third is found using the nomograph (Figure 3-8). The
procedure for calculating the scaling factors and using
the nomograph is presented as follows:

(1) Scale Factor Development
This nomograph models  the  same variety  of
conditions that  the  Wilson and Miller model  (from
which it was derived) does,  yet it does it with only
one graph.  This was achieved by scaling the
parameters to make them dimensionless.  Distance X
is  made dimensionless  by dividing  by the distance
scaling factor  (XD,  the  characteristic  dispersion
length). The  mass  flux (Q *  C,)   is  made
                                                     dimensionless by  dividing  by the mass flux scaling
                                                     factor  (QD). And time  (T) is  made  dimensionless by
                                                     dividing  by the  time scaling factor  (TD). Obtain XD
                                                     using the following:
                                                            Dx
                                                                                               (3-29)
                                                     where  variables are  defined as  in  Figure 3-8,
                                                     Definition of Terms.
                                                     Calculate TD using
                                                                                               (3-30)
                   V*


           where

              V
              Rd
              Td,Dx  =


           and where
                                                                     (Ks*i)/Pe.
                                                                     1 + p* Kd/Pt.
                                                                     defined in  Figure 3-8,  Definition of
                                                                     Terms,
                                                                       =  foe * Koc.
                                                     Calculate QD using:

                                                     QD = Pe * m * (Dx * Dy)1/2
                                                    (3-31)
where variables  are defined  as in Figure  3-8,
Definition of Terms.

(2) Application of Scale Factors
Use the three  scale  factors and the nomograph
(Figure 3-8) to calculate  the concentration at time T
and distance X.

(a)  Find T/Td curve desired.

(b)  Find  X/Xd  on the x-axis.

(c)  Plot the point of intersection  of the T/Td curve
    and X/Xd.

(d)  Use this point and the point on  the Q * C0/Qd line
    to draw a  straight line. Where  this line intersects
    the  concentration line,  the  concentration  at
    distance X and time T is indicated.

3.5.2.7 Extent of Plume
As discussed earlier, a large volume of contaminated
ground  water can  result from  a  small volume of
chemical release.  For example,  a lO-gallon  spill of
solvent can contaminate  a billion  gallons  of ground
water to  10  ppb.  Similarly, a  5000-gallon tanker
truck  can contaminate 500 billion  gallons  of ground
water to  10 ppb. The  analyst must be aware of the
relationship between volume of contaminant  released
and volume of contaminated  ground water.  The
                                                  77

-------
Figure 3-8.    Nomograph for solutions of time, distance, and concentration for any point along the principal direction of ground-
              water flow.
                                                     Nomograph for
                                                   Plume Center-Line
                                                     Concentration
          io-6H


          1 fl-




          IC-3-1
                                                                              QC0
                                                                                   H
                                                                              (Ib/ft3)
                                                                                J«H
           103-


           104-


           105
•10-"

•io-1
                                                                                         mg/l
 102





 104


 10s


 10'


-107





•10'
                                                                                      •'
                                   100
                                               1,000      10,000
100,000
                                          X
                                          XD
                                                                                                  MO"
                                                                                                       C
                                                                                                     (mg/l)
                                                                                                   "   i
                                                                                                  -10
                                                                                                  F-102
                                                                                                  HO3
                                                                                                  =-104
Source: Kent et al. (1985)

-------
Figure 3-8. (Continued)
                                                        Definition of  Terms
               Primary  Variables
                C   —    concentration of leachate at a specific time and distance.
                X    =    distance from source where concentration of leachate is computed.
                         distance is measured in  direction of ground-water flow (perpendicular to gradient).
                Y    =    transverse distance  measured from the  centerline of ground-water flow
                         (assumed  to be zero in the nomograph).
                t     =    sample time from  beginning of leachate source flow.
         Units
        (M/L3)
            (L)

            (L)
            (T)
               Aquifer Parameters :
                m   =   effective aquifer thickness or zone of mixing.
                Pe  =   effective porosity of aquifer or zone of mixing.
                v    =   velocity of ground-water flow within voids, estimated  directly from:
            (L)
(Dimensionless)
                                                                    Ki
                          where
                                               coefficient of permeability or hydraulic
                                               conductivity of aquifer or zone of mixing.
                                               gradient of ground-water flow.
(Dimensionless)
                Transport Parameters:
                     =   longitudinal dispersion coefficient (mixing rate) with respect to
                         distance in x direction and time, estimated directly or from:
         (L2/T)
                          where
                                               longitudinal dispersivity
                                               molecular diffusion coefficient, which is assumed
                                               to be negligible for velocities typical of permeable
                                               aquifers.  D* may be the dominant process in
                                               aquitards where ax V would be negible.
           2
         (L2/T)
                                                                 79

-------
Figure 3-8.    (Continued)
                                                       Definition of Terms
               Dy   =   Transverse dispersion coefficient (mixing  rate) with respect to
                        distance in the y direction and time, estimated directly or from:
                                                            Dy=ayv+D*
                          where
                                                                                                       Units
                                              transverse dispersivity,
            (I)
                          or estimated as:
                                              Dx divided by a ratio, which commonly ranges between
                                              5 and 10 for medium to coarse sand aquifers.
                        Retardation factor estimated directly or from:
(Dimensionless)
                          where
                                                             0b(Kd)         v
                                                     R,= l+	(or)R,= —
                                                       d       p         d  vd
                                   Pb     =   bulk density of aquifer medium.                             (M/L3)
                                   Pt     =   total porosity.                                    (Dimensionless)
                                   Kd     =   distribution factor for sorption on aquifer
                                              medium (from sorption isotherm column studies)              (L3/M)
                                   v      =   velocity of ground water.                                    (UT)
                                   vd     =   observed velocity of leachate for a given concen-
                                              tration and chemical species.                                (UT)
                                   Y      =   coefficient for radioactive or biological decay. For
                                              no decay, the value of y is one. (Assumed to be
                                              one in the nomograph.) Calculated from:            (Dimensionless)
                                                            4Dx       4Dxlog(2)
                                                       ~  +  2           2t
                                                                        vtl/2
                                   where
                                                        X= decay constant =
                                                                            log (2)
                                                                                                         (I/I)
                                   t1/2    = Half-life:   time when half of the original mass remains.
            (T)
                                                                80

-------
Figure 3-8. (Continued)


                                                     Definition of Terms


                                                                                                    Units



               Source Rate of Leachate:



               QC0 =    Mass flow fate:                                                                (M/T)



                         where



               Q   =    Volume flow rate estimated directly or from:                                       (L3/T)
                                                             Q=Aq
                         where



                         A   =   area of source.                                                       (L2)

                         q    =   recharge rate.                                                       (L/T)



                         C0  =   Initial concentration                                                  (M/L3)



               Intermediate Variables (used for nomograph only):



               Xd  =   A characteristic dispersion length or scale factor given by:                             (L)

               TD=     A characteristic dispersion time or scale factor given by:                               (T)
                                                                R,D

                                                           T  = —
               Q    = A characteristic dilution-dispersion flow given by:                                     (L/T)
                                                       Q  =P m*/D D
                                                        0>   e   >  x  y
                                                              81

-------
equation is a  simple  mass  balance  equation and  is
expressed as follows:
For liquid contaminants:
        = Vgw*Cgw
(3-32)
where
     V.,   =  volume of liquid chemical released.
     Vgw  =  volume of contaminated ground water.
     C1   =  average  concentration of chemical
             contaminant in the released liquid.
      Cgw =  average  concentration  of contaminant
             in ground water.

Both  volumes and  concentrations should  be  in the
same units.
For solid contaminants:
MC*CC  =  Cgw*Vgw
where
(3-33)
     Mc   =  mass of solid waste, (in milligrams).
     Cc   =  concentration  expressed  as  mass
             fraction,  fraction  of  contaminant in
             waste, (dimensionless).
     Cgw  =  concentration of contaminant in ground
             water, (mg/liter).
     Vgw  =  volume of contaminated ground  water,
             (liters).

To convert the quantity of contaminated ground water
to a volume  of  contaminated  soil,  the following
equation is used:
(Vgw*0.13368)/Pt   =  Vc

where
(3-34)
     Vgw  =  volume  of contaminated ground water,
             (in gallons).
     pt   = total porosity,  (dimensionless fraction).
     Vc   = volume of contaminated soil,  (in cubic
             feet).

Or alternatively:

VgW/Pt = Vc                               (3-35)

where both volumes are in the same units.
3.5.2.8    Use  of Monitoring  Data
The analyst should take  care  when using monitoring
data to  assess  the depth of contamination in order to
calculate  volume or  mass  in  the plume. The
difference between monitoring  and pumping wells  will
affect the interpretation of the concentrations found in
the wells. Monitoring wells are the more desirable,  but
since most existing  wells  will be  pumping  wells,
monitoring wells will typically have to be installed. The
cost associated with  drilling monitoring wells most
likely will  cause  the analyst to rely  on  existing
pumping wells.

Monitoring  wells extract a small quantity of water (a
sample); this minimizes  the well's influence on  the
flow of the  ground-water. They do  not induce a large
vertical  component in the  ground-water flow,  and
thus they sample a horizontal slice of the aquifer. The
concentration in a sample removed from a monitoring
well represents a  concentration at the depth  of  the
well screen. Thus,  monitoring wells at various depths
can be used to  assess the depth of contamination.

Pumping wells draw large quantities of water from an
aquifer (a pumping well provides water). This causes
a cone of depression  to form on the  water table and
influences the flow direction above and  beneath  the
well screen. Pumping  wells induce  vertical  flow in  the
aquifer near the well.  This vertical movement causes
the concentration  in  the  well  to  reflect  the average
concentration for a  depth range that  is substantially
greater than the length of the well screen. Water will
be drawn from above  and below the well screen. The
well water  does  not  reflect the  concentration of a
particular  depth,  but rather  reflects  an average
concentration from a  range of depths.  This makes an
assessment of  the depth  of  contamination difficult.
However, it makes assessing the mass in the plume
easier since the well draws a sample  that represents
the concentrations at a wide range  of depths near the
well screen depth.

3.5.2.9 VMS Model
In addition  to the nomographic technique, the Office
of Solid  Waste (OSW)  has  developed  a  simplified
model for its delisting program that relates  leachate
concentration to receptor well  concentration  500 feet
downgradient  from  the edge of a landfill. The
approach is called the  VMS  model (Vertical and
Horizontal  Spread  model). The  only reduction  in
concentration provided by the VMS model is that due
to vertical  and  horizontal dispersion  (OSW  plans  to
add  hydrolysis  and biodegration  for  organics). The
approach involves back  calculating  from a  health-
based  ground-water  concentration at the exposed
population location  to an acceptable leachate
concentration at the site.  Wastes with  leachate above
this  concentration  must  be  managed as  hazardous
wastes.  Those with leachate below this concentration
can be managed   in  a municipal  landfill  or
                                                  82

-------
nonhazardous  industrial landfill  (i.e., outside the
hazardous waste system).

The  only data the VMS  model  requires  are the
leachate concentration  and the annual volume of
wastes  disposed of (constituent  concentrations of
toxicants are also  required in  order  to  ensure that
they are present  in sufficient  mass  to  sustain
leaching). The  model calculates a different dilution
factor depending on the  annual  volume  of waste
disposed of.  For  Superfund  purposes, the  total
volume  of waste at the site would be used as the
"Annual Volume of  Waste"  term. A small volume of
waste  can rely on greater dilution,  while  a  large
volume  of waste  is assigned a smaller  dilution
potential. All other input  parameters are  fixed at
reasonable  worst-case  values.  By  fixing  the
environmental  parameters, the model  assumes  a
generic  environment that is consistent with OSW's
requirements. For CERCLA purposes, the model is
considered to be useful as a  simplified analytical
procedure,  and use of the VMS model for site-
specific,  in-depth analysis is not recommended.

When  using  this  model,  one  should keep its
limitations in mind. The  VMS model simulates soluble
toxic constituents dissolving  into percolating
precipitation and moving with  the ground water.  It
does not address solvent transport of  organics  (two-
phase flow)  or  the  percolation  of  organic fluids into
the ground.

Critics  of the  VMS model  have pointed  out  two
weaknesses  of the  approach. The  first point is that
the model upon which the VMS model was based (the
Domenico and Palciaukas model) does not relate
leachate concentration to  exposed  population well
concentration. This  model relates the concentration in
ground water immediately below the hazardous waste
sites to  the exposed  population  well  concentration.
When leachate  enters ground water, it will be mixed
with  ground  water.  This contaminates ground water
and  at  the  same time  dilutes the concentration of
leachate. It is wrong to use the  C0  term in the model
as leachate  concentration,  because it represents the
concentration in ground water  at  the vertical  point
where  leachate enters.  This concentration  must be
measured on a site-specific basis to  make  the use
of the model consistent with the boundary conditions
used in  the derivation  of the model.  The model is
derived from the following assumptions:

1. Steady-state  concentrations are achieved under
   the  conditions  that the concentration  C0 in
   ground water is  maintained  on  a vertical  plane of
   finite size.

2. No longitudinal dispersion occurs; dispersion only
   in the y and z directions is assumed.
3. Recharge  or  dilution  mechanisms,  other than
    ground-water  flow  and the  above-mentioned
    dispersion, are  ignored.

4. The contaminant  velocity in ground  water is
    known.

The second  weakness  is the  method  used to
determine the cross-sectional area of the  plume at
the edge of the landfill. The depth of the plume is
determined  by the  horizontal velocity of ground water
and  the  vertical velocity  of the  contaminant.  The
model  presumes  that the vertical  velocity  of the
contaminant in the vadose  zone  is also the  vertical
velocity of the contaminant  in the  saturated zone. In
the vadose  zone,  the contaminants  are under the
influence  of gravity; in  the saturated zone, the vertical
velocity is much smaller  because  the effect of gravity
is canceled by the  buoyancy forces. The VMS model
assumes  that the velocities are the same.

These  two weaknesses were present in  the VMS
model at the  time  this  document was written:
subsequent revisions may address  these problems.

3.53   In-Depth Methods and Models
Several references  are available that provide detailed
derivations  and outline the application  of more
sophisticated  equations for  the  analysis of
contaminant migration  in the saturated  and
unsaturated zones. The analyst  is  referred to the
following  documents:  USEPA 1985J;  Van Genuchten
and Alves (1982) Walton (1984), and Javendel et al.
(1984), USEPA (1986a),  Geotrans (1986), and  van
derHeijde (1985 and 1987).

Tables 3-14,  3-15, and 3-16 provide  information
regarding several  modeling  procedures  for  the in-
depth  assessment of  the ground-water fate of
hazardous substances. Note that  in order to  provide
the analyst  with an indication of the large number of
computer models that  could be applied to analysis of
contaminant  fate in  ground water,  Table 3-15
(Features  of Unsaturated  Zone  and Ground-Water
Fate Models) provides  data for 24 models in addition
to the  11 for which  more  detailed  information is
provided in  Tables  3-14  and 3-16. Two  of  the
models addressed  in these tables  are part of  GEMS:
SESOIL  and AT123D.  The latter is  described in
greater detail below, because it is  more versatile  and
is applicable to a wide  range  of  fate  analysis
situations.  Additionally, following that  discussion
further  detail for certain  of the models addressed in
Tables  3-14,  3-15, and  3-16 is  also provided.

AT123D  (Analytical  Transient  1-,  2-,   or 3-
Dimensional  Simulation  Model)  is  capable  of
simulating  the transport  and fate  of  hazardous
material under 300 different user-selected  situations
(Yeh 1981).  One of eight  source  configurations  can
be selected: a point source; line  sources aligned in
                                                 83

-------
             Table 3-14.   Resource Requirements and Information Sources :  Unsaturated Zone and Ground-Water Fate Models
                               Model
               Description
     Resource requirements, comments
                                                                                                                                            References, sources of documentation,
                                                                                                                                                          software
             Unsaturated zone
                  Seasonal Soil  Compartment Model
                  (SESOIL)
CD
.b.
• Long-term fate  simulations
x*s Accounts for numerous hydrologic,
   meteorologic characteristics of site
xx Accounts for numerous transfer,
   transformation processes: adsorption
   volatilization, degradation, brodegradation
• Models  organics,  inorganics
xx Produces contaminant concentration
   distribution in Unsaturated zone, quality of
   ground-water runoff
&  Handles up to three layers of soil types,
   permeabilities
xx integrated into GEMS (see Section 3.1)
• Versatile, easy to use
xx FORTRAN program language;  has been
   Implemented on IBM 370, VAX 11/780
Documentation: Bomazountas and Wagner
1981

Contact for access to GEMS system:
Mr. Loren Hall
US. EPA, Exposure  Evaluation Division
Washington, D.C.
(202)  382-3931
             PRZM  (Pestcide  Root Zone Model)
• One-dimensional
xx Organic substances
xx Degradation  is simulated
xx Provides pollutant velocity, distnbution,
   and concentration data
xx Accommodates various  release  rates,
   schedules
xx PC  Based Model
xx Requires 256 K RAM minimum, 640K
   preferred,
x  Intel 8089 or 80287  math coprocessor
xx Has been field-verified with pesticides
xx FORTRAN program  language
Reference: Carsel et al. 1984

Information:
David Disney
USEPA Environmental Research Laboratory
Athens, Ga. 30613
(909) 546-3132
             PESTAN
xx One-dimensional
xx Organic substances
xx Degradation  is simulated
xx Provides pollutant velocity, distribution,
   and concentration data
xx Accommodates various  release  rates,
   schedules
xx Considered a screening model
xx Rapid evaluations
xx Inexpensive, easy to use; requires only
   hand-held  calculator
xx Has been field-verified with pesticides
                                                                                                                                          Reference: Enfield et al.  1982
                                                                                                                                                                      (Continued)

-------
             Table 3-14.   (Continued)
                               Model
                                                          Description
                                                                                                       Resource requirements, comments
                                                                                        References, sources of documentation,
                                                                                                     software
                  Hydrologic evaluation of landfill
                  performance (HELP) (as modified by
                  Anderson-Nichols)
                                               One-dimensional
                                               Models leaching from landfills to
                                               unsaturated soil beneath landfill
                                               Has four options to handle modeling the
                                               solubilization of toxic constituents
                                               Models organics/inorganics
                                               Uses rainfall and waste solubility to
                                               model leachate concentrations leaving
                                               landfill
                                           Four options allow modeling with available
                                           data
Information:
Brian Bicknell
Anderson-Nichols
Palo Alto, Calif. 94303
(415) 493-1864
oo
Ol
Saturated zone
     Random Walk Solute Transport Model
     (RWSTM)
     (a.k.a. TRANS)
     (requires PLASM for flow modeling)
xx One- or two-dimensional
xx Time-vanant release rates
xx Accommodates well-injected  release
xx Incorporates dispersion, retardation
xx Handles nonconservative  pollutants
xx Accounts for well pumping
xx Provides contaminant concentration at
   user-selected points
                                                                                                  xx Requires mathematical programming,
                                                                                                     hydrogeological knowledge on part of
                                                                                                     user
                                                                                                  xx Has been field-validated
                                                                                                                                            Documentation: Prickett et al.  1981
                  Coupled Fluid,  Energy and Solute
                  Transport (CFEST) Combined with
                  UNSAT-ID
                                          xx Three-dimensional
                                          xx Accommodates heterogeneous,
                                              anisotropic, multilayered soil
                                              configurations
                                          x   Handles saline aquifers as well as fresh
                                              water
                                          xx Transport mechanisms of dispersion,
                                              advection simulated
                                          xx Sorption, degradation  mechanisms not
                                              incorporated
                                          xx Time-variant release and flow rates
                                          xx Combination covers unsaturated and
                                              saturated zones
                                            < Has been applied for arsenic and
                                             organic wastes
Documentation: Gupta et al. 1987
                                                                                                                                                                          (Continued)

-------
             Table 3-14. (Continued)
                               Model
               Description
     Resource requirements, comments
                                                                                                                                             References, sources of documentation,
                                                                                                                                                           software
                  Sandra Waste Isolation Flow and
                  Transport Model (SWIFT and SWIFT
CD
              Leachate Plume Migration
              Model (LPMM)
    Three-dimensional
    Transport processes of advection,
    dispersion simulated
    Sorption, degradation processes
    accounted for
    Appropriate for waste-infection,
    waste-isolation modeling
    Code was based on SWIP Model
•    Has been field-verified
•    Has associated user's guide in self-
     teaching format
•    FORTRAN program; has been
     implemented on various CDC systems
     including CDC 7600
•    1986 version has been released
XMS Continuous source model
xx Dispersion is simulated
xx Degradation processes accounted for
xx Has been field verified
x  A simplistic model; results may not be as
   sophisticated as necessary for Level III
   work
   Can be used in nomographic, hand-held
   calculator, or computer form
   Relatively easy to use
Documentation: Reeves  and Cranwell
1981; Finley and Reeves 1968
Software:
National Energy Software Center
Argonne National  Laboratories
Argonne, III. 60439
Information:
Intera Environmental Consultants,  Inc.
11999 Katy Freeway, Suite  610
Houston, Tex. 77079
                                                                                                                                           References: Kent et al. 1982
              Analytical Transient One-, Two-, and
              Three-Dimensional Simulation Model
              (AT123D)
                                                        See Section 4.4.2 of text
                                         •  FORTRAN program applicable to wide
                                             range of computers
                                         •  May require extensive setup time
                                         •  Available through GEMS (see Section
                                             4.1)
                                                                                                                                           Documentation: Yeh 1981
                                                                                                                                                                       (Continued)

-------
Table 3-14. (Continued)
                  Model
                                                   Description
     Resource requirements, comments
   References, sources of documentation,
                software
 Unsaturated and Saturated Zones

     Finite-Element Model of Waste
     (FEMWASTE)  and
     Finite Element Model of Water Now
     (FEMWATER)
                                     Two-dimensional
                                    Aeflnterzone transfer is modeled
                                    ^Incorporates convectron, dispersion
                                    £gSimulates degradation of nonconservative
                                       substances
                                    && Absorption is accounted for
                                    «KCapable of modeling layered,
                                       heterogeneous soil zones
                                    .e^f FEMWATER is a model for ground-
                                       water flow, while FEMWASTE simulates
                                       the transport/fate of contaminants
.efts'Has been Implemented on IBM 360
,«•,«• May require background in hydrogeology,
   differential equations, programming
g& Field-verified
Documentation: Yeh and Ward 1981

Information;
Dr. George T. Yeh
Oak Ridge National Laboratory
Environmental Science Division
P.O. Boxx
Oak Ridge, Tenn. 37830
(615)  574-7285
Solute Transport and Dispersion Model  &&  Two-dimensional                       e& Field Verified
                                    .efts' Conservative substances (no decay       .efts' Relatively inexpensive, easy to use
                                       simulation)
                                    .efts' Heterogeneous soil conditions accounted
                                       for
                                    .efts' Pumping or recharging well effects
                                       modeled
                                    •  Thickness of saturated zone may vary
                                                                                                                           Documentation:
                                                                                                                           Kowikow and Bredehaeft 1974
Sources: USEPA 1982b; Brown et al. 1983; Kufs et al. 1983; Versar 1983.

-------
              Table 3-15.   Features of Unsaturated Zone and Ground-Water Fate Models
\
SOIL/ROCK
CHARACTERISTICS
\
FLUID
CONDITIONS
\ FLOW \
\CONDITIONS \

AT1IJO
LPMM
OME-D
ptwna (to)
R«**«.
•LM
VHS
MW-Analyl.
•ta*tr Belt
Klnztlkach
WlHon M
Belult
FtiMrait
F*mw«l«r/F«m




•






























itcrenenve CIIAIIUK ^
Ttk 1M1 1
K*M tl M. 1MB
«WI tttniMMtll MM Ahw* IBM
C*«*H tl •. tMI
JntmM tl H. IM4
ItaycMm tl H. 1MT
OMIKMC* mt PikjIiMkjt tMI
urn «tf HtlkJt m* Bikilittai tMt
•Mr mn4 VtmUfl tMT
KMutlkM* 1MB
WHIM tMt
B*yn tMI
MMIHwi tMI
T»* tMT. Tth, »M Wtrt tMI
NwwM M M. 1MB
Onto unt Btfrt IBM
Caratl at «. IBM
BWMMUMOT mt WtfMf IBM
mi atBMMtn tBTB
V*M IBM
•wtflhlM IBM
Ol»U tl Bl tMT
Ttft n« MM tMB
KlBP tMT
KwilWw unt BtOMtn 1BTB
Bswlwil Mitf RwiNiMV IMfl
Traqr IBM
PrkkMI tl M. tMt
MMtra tMt
Rttvw tl M. tMB
Trmto IBM
HwBhswn M •! 1W7
Mill tl tl. tMT
EnlltM tl al. tMI
••• taklt S-14
00
do
                •NON-AQUEOUS PHASE  LIQUIDS  1) FOR UNSATURATED ZONE ONLY.

-------
                              Table 3-15. (Continued)
OO
CD
                                 \
GEOMETRY
\
TRANSPORT AND
FATE PROCESSES
\
                                                                                                                          REFERENCE CITATION

ATI MO
LPMM
ONE-D
Plwiw (SO)
R««q
SLN
VH»
RW-AMlyt.
BVAVOT J>tl
KlKMltlMlt
wftlMfi 3S
••tut*
F*«tr*n
F<«w«t«f>FMN»Mt«
Peril*
oat/osi
PRZM
• MOIL
•UMATRA-1
•UTRA
TRIPM
CF68T
FEWA'FEMA
HSTID
MOC(U»QS S-0)
MOC OCNSC
HOC NRC
RW8TM
•WENT
•WIFT •
TRACR M>
TRAFRAP/WT
• IOPLUME-H
PE3TAN
HELP

•

•
•

•
•

•
•
•
•



















•






•


•


•
•






•
•
•












•

•
•
• ^

•












•




•


•

•
•
•
•



•
•


^








•
•


•
•
•




•
•



4
•
•




•



\



•

•
•




•











•











^



•



























•
•
•
•
" \

•
•
•

•
•
•
•
•
•
•

•
•

•

•
•
•

•
•
•
•
•
•



•
•
•

^ \































•



' \































•
•
•

•" i









•






•
•

•



•
•

•





•
•
•
' \
















•




























*






•
•














•



•
•
•
•

•

•
•
•
•
•
•
•
•
•



















* \

















•

















' \



































^
•















•
•

















^














•




•

•

•




•
•

•




ft* 1M1 1
K*M M *. IMS
•Ml OWWOMM *Ht AtVM 1MI
C»4tH M M. IMS
JcvMltfcl M •!. 1M4
Hur«K*rn M d. 1M7
Dcmmlo* tut PctotaMlIM 1M1
VMI 4>f tMMto m4 •ftalvMWi 1MC
•MC *n4 VwraV 1M7
KNn»»*Mi 1M«
W««m IMf
•*l|Ni 1MI
MMftnMS • ••*
Ted 1MT. Vik, mt W«N 1M1
Rim»lMt •! tf. 1M»
Dwte ant •** 1M»
C«M4 M M. 1M4
••IMMHHM0 M|4 WMMf 1M4
•M owNwMwi irr»
V«M 1M4
QufMhtan IMS
Oupu M ri. 1MT
T««l (Ml NMI IMf
KI|V 1MT
KMliMr «M >l«l»lltl im
•wrt*r« *M Kwilnw IMS
Tragy IMS
PiMwn •( «L 1M1
Inln* 1MI
RWVM M •!. ISM
Travta 1M4
Hufikom M d. 1MT
RNd M d. 1MT
EnlMd M «. IMS
8M trill* 1-14
                                           PHASE LIOMM   1, FOR UNSATUMnGO ZONE ONLK.

-------
Table 3-15. (Continued)
K        BOUNDARY/SOURCE  \
         CHARACTERISTICS    \
CAPAilLITIES
A
    Modal
                                                                                 REFERENCE CITATION
AT123D
LPMM
ONE-D
Plum. (3D)
R»tq
SLM
VMS
RW-Analyt.
•aavar Sett
Klntalbaeh
Walton 3$
Soluta
Famtran
Fam watar/Famwaata
Porllo
GS2/OS3
PRZM
SESOIL
SUMATRA-1
SUTRA
TRIPM
CFEST
FEWA/FEMA
HST1D
MOC(USGS 2-D)
MOC DENSE
HOC NRC
RWSTM
SWENT
SWIFT II
TRACR 3D
TRAFRAP/WT
•IOPLUME-II
PEST AN
HELP




















•



•
•
•
•




•
•
•
•











•
•






•



•
•
•
•




•


•



•
•

•
•















•
•
•
•




•





































•
•
•
•
•


•
•

•
•
•
•


•
•
•

•



•
•
•
•




•


•

•

•

•
•
•

•
•
•
•


•
•
•

•



•
•
•
•




•














•
•


•
•
•

•












•
•














•























































•


•


•




•
•
•
•

•
•
•

•
•

•
•

•

•
•

•
•


•
•
•
•
•

•

•

•
•





























































































Y«a 1M1 1
Kant at al. IMS
van QamioMan and Alvaa 1M2
Catfafl M •!. tSS2
JavanM at al. ISM
HuyaKorn at •!. 1M7
DcMMnfco mtt PrietankM 1SS2
nn «w H^fd* •>* SrlnhriMn 1Mt
SMT mnt V«nul|i 1W7
Kfei»lbMh ISM
WiHon IMS
toll* IMS
Mcrthwi IMS
Y«* 1M7, V«k, and Ward 1M1
Runekal at al. IMS
Davfc an* 8*g«l IMS
Caracl M al. 1M4
SwuMuntu ami Wagnw 1M4
van QanueMan 1*T»
VOM 1M4
Ouragklan 1M3
Oupta al al. 1M7
Va* tnt Hutf IMS
KlM> »M7 •
Konfcow and Badahecti 107»
Sanfetd and Konlkow IMS
Tracy 1M2
Prtekan « al. 1M1
Mara 1M3
Rwvaa •( al. 1MC
Trarb 1M4
Huyakem a( al. 1M7
Mai at al. 1M7
EnftaM M al. 1M2
Saa tabto 3-14
    •NON-AQUEOUS PHASE LIQUIDS  1) FOR UNSATURATED ZONE ONLY.
                                                        90

-------
              Table 3-16.   Data Requirements for Unsaturated Zone and Ground-Water Fate Models
CD
                 HWSTM PL»SM
                 MMMSTC • FCMMTtft
                SOURCES:  USEPA 1982b;  Brown It ll. 1913;  Kufe el il. 19*3;  Versir 1983.

-------
one  of three different ways with  respect to ground-
water flow; area sources, also aligned in one of three
different configurations; or a volume source  (existing
plume).  Release types can  be  instantaneous,
longer-term but finite,  or constant. Aquitard locations
can be specified below or on both sides of the aquifer
in  any configuration; or the aquifer can  be treated as
infinite in  all  directions.  Advection and dispersion
transports are simulated.  Losses  resulting from
volatilization, degradation, and  adsorption  are
modeled. The  model predicts contaminant movement
in  one, two, or three dimensions (Yeh 1981).

Use  of AT123D  requires the following  information:
dispersion  coefficients  in  horizontal,  vertical,   and
longitudinal  direction: geometry  of  the aquifer,
especially   regarding  configuration of  aquitards;  soil
properties,  including  bulk  density, effective  porosity,
hydraulic conductivity (permeability); source type;  and
release duration  and strength,  soil-waste  stream
partition  coefficient,  hydraulic  gradients,  and  an
overall decay  constant  (or  soil  half-life  figures)  for
the substance studied  (Yeh 1981).

The  model determines contaminant concentration at
any  point, at a downstream and  lateral distance  and
depth specified by the  user, as  a  function of time
from the beginning of source release.

AT123D can be accessed through the  GEMS system
(see Section 3.1). It is written in FORTRAN  and  can
be installed on a wide  range of computer types.

In  addition, the  Office  of Solid Waste (OSW)  has
developed a national model that uses the Monte Carlo
simulation for relisting  hazardous wastes on  a generic
basis.  This FORTRAN  computer  model  is  a three-
dimensional  advective-dispersive transport  model.
The  model  currently  considers the mechanisms of
hydrolysis,  dispersion, and rainfall recharge  into the
ground-water  plume.  OSW is  using  the model to
back-calculate from a  health-based  standard at the
exposed  population  well  to an acceptable  on-site
leachate  concentration.  If a treated waste  produces
leachate with a contaminant concentration below the
acceptable concentration,  then it is  considered
protective  of the public health.

The  model currently uses the HELP  model to provide
leachate  release  rates. Leachate  strength
(concentration) is  provided  by  the  Toxicity
Characteristic Leaching  Procedure (TCLP). OSW
plans to  add the geochemical  model MINTEQ to
handle  metal  speciation.  Biodegradation  processes
are being evaluated for incorporation  into the model.

Since EPA's  model is a national  model that uses a
generic environment, the  data  requirements  are
minimal.  The model approximates  an  average
environment by making multiple runs (typically several
thousand  runs for each chemical constituent) with
varying  environmental  data. By applying  this
approach, called a Monte  Carlo simulation, one can
model the dilution potential of all possible  sites as  a
cumulative frequency  distribution  versus expected
concentration  at  an  exposed  population  well.  The
extent to which a particular CERCLA site matches the
OSW model  depends on  the closeness  of  site
characteristics and  the model assumptions.  If  a
particular  CERCLA site has adequate  hydrogeologic
data and satisfies the model assumptions,  the model
can  be  used  for  site-specific analyses. Before  final
assessment of the desired level of cleanup, however,
application of the model on  a  site-specific basis  will
typically be required.  Generic modeling  is appropriate
for OSW's purposes,  but may suggest cleanup levels
beyond  those  necessary at  a  particular  site.
Preliminary  work  or screening-level  efforts  at
CERCLA sites  where adequate,  good  quality
hydrogeologic  data do  not exist can benefit from the
model's  minimal  data  requirements  for site-specific
environmental parameters.

The  model  is being updated  to  incorporate  flow
through  fractured  media and the unsaturated zone.
The  data base  for  MINTEQ  is being  enlarged  to
handle additional  metals,  and  more  data  are being
collected to validate the model results.

Since OSW's  model  uses a Monte Carlo simulated
environment, it should be applied with this limitation in
mind. Other limitations in the use of this model derive
from two sources: (1) limitations in the scope of the
model, and  (2) specific modeling  choices made so
that  the  model would  support OSW's  requirements.
The  model's scope is limited by the leachate release
algorithm HELP,  which  models  soluble toxic
constituents dissolving  into  percolating rainwater and
moving  with that  water.  It  does not address
percolation  of organic fluids  into  the  ground  or
associated leaching by  concentrated organics.

Additionally, the TCLP does  not fully  predict leachate
concentrations due to leaching with water containing
dissolved  solvents. It does assume the presence of
acetic acid in leach  water,  thereby  providing some
measure  of hydrophobic solubility. Although HELP
can model a variety of landfill cover situations, OSW's
requirements were such that it modeled a landfill with
a failed  liner but  an intact  (aged) cover.  The
permeability of the hypothetical  cover was chosen at
1 x  10-6  cm/sec to represent an aged (deteriorated)
cover with an  initial permeability  of 1 x 10-7 cm/sec.
OSW states that it found the range of permeabilities
for aged clay actually to be  between 1.4  x 10-6 and
43  x 10-6 (USEPA 1986).  For  CERCLA  sites,
selection of a permeability within that range may be
more appropriate. Also, many  CERCLA sites do not
have a cover, or the cover may be breached. In either
case,  the mass flux leaving the site will  be
considerably  larger.  Even if the site has an intact
cover,  one may wish  to  predict long-term potential
                                                  92

-------
releases and  also to consider  the eventual
subsidence  and breaching  that may  occur  in  the
future.

Pesticide Root  Zone  Model (PRZM) (Carsel  et  al.,
1984) simulates the  vertical movement of pesticides
in  unsaturated soil,  both  within and  below the plant
root  zone,  and  extending to the water table  using
generally available  input data that are  reasonable in
spatial  and  temporal  requirements.  The  model
consists  of hydrology  and  chemical transport
components that  simulate runoff,  erosion,  plant
uptake,  leaching,   decay, foliar  wash off,  and
volatilization  (implicitly) of a  pesticide. Predictions  can
be made daily, monthly, or annually.

3.5.4 Short-  and Long-Term Concentration
Calculations
Long-term  average   ground-water concentrations of
contaminants at exposure points are  a function of the
concentration profile  over the  time  period of study,
which are,  in  turn,  a function   of hydrologic
fluctuations,  release rate fluctuations, and  the
effectiveness of  remedial actions.  Average
concentration values are  obtained from steady-state
methods. Several  of the in-depth analysis  models
tabulated in Section  3.5.3 accept time-weighted
input data,   and  provide  long-term average
concentrations,  as well as the concentration profile as
a function of time.

Short-term concentrations  at  exposure  points  are
obtained  by  examining  the ground-water
concentration profile  at the  selected exposure point
over time, and  identifiying of the period of maximum
concentration.

3.6  Biotic  Pathways

3.6. J Estimation Procedures
After the fate  of a  contaminant in  air,  water,  and
ground water has been estimated, one can assess its
fate in  biotic  populations.  Using  the ambient
concentration  data  developed  for  each of these
media, a determination  is  made whether any biotic
populations that can  potentially serve  as pathways for
human exposure to  hazardous materials  (i.e., vector
organisms) are within  zones of  elevated hazardous
material  concentrations. Such vector  populations may
include  agricultural  crops; agricultural livestock; fish,
shellfish,  or  crustaceans  that   are  important
commercial or sport species; and  game  populations in
hunting areas.

In  assessing  the   biological fate  of  hazardous
materials, the following processes,  which determine
the rate  of introduction of hazardous material  to  and
the  final concentration of hazardous material  within
vector organisms, should  be considered:
^ The  concentration  of  hazardous material  in
    environmental  media  containing  or  supporting
    vector organisms.

A* The  metabolic rate of the  vector  organisms.
    Metabolic  rates  are functions of  several
    environmental parameters  including  temperature
    and  the  availability  of sunlight, oxygen, nutrients,
    and  water or other factors.

MS* Substance  bioavailability:  the affinity  of  each
    hazardous  substance for partitioning  into the
    organic phase or its availability for other forms  of
    uptake.  The  bioavailability  of each  substance
    differs, as does that of various chemical species
    of an individual substance: the octanol/water
    partition coefficient is an  indication  of this
    parameter.  Bioavailability  of  a given substance
    can  vary with  environmental conditions.  Factors
    that  influence the  physiochemical speciation  of
    substances, and  thus their bioavailability, include
    salinity,  pH, Eh, organic carbon  concentration,
    and  temperature.

** Characteristics  of species  metabolic  processes.
    These characteristics differ among species and
    include feeding habits and  ability  to metabolically
    degrade,  store,  and eliminate  the   substance.
    Bioconcentration  factors  (or  BCFs, the  ratios  of
    organism tissue  concentration to ambient
    environmental  concentration)  for  many  species
    and  hazardous substances have been empirically
    determined and are  discussed below.

Consider the  following transport  mechanisms  in
assessing the  distribution  of hazardous  substances
within  the  biologic medium and  identifying the
potential points of human exposure:

*   Transport and distribution  of vector organisms  as
    a result of human commercial or sport  activity.

MS* Migration of  organisms,  or movement  of  these
    organisms  with  advective  flow of environmental
    substrate media.

x* Movement  of contaminants through the  food
    chain. This  mechanism often results  in very high
    concentrations  of  hazardous  materials in the
    tissue of higher trophic level organisms within and
    without contaminated areas.

General theoretical  relationships  between the above
factors and  concentrations of  hazardous  substances
at human exposure points are not available. This is
because such  relationships  are highly  specific  to
individual  ecologies,   biotic  species,   hazardous
substances, and human activities  associated with
involved biotic species.
                                                  93

-------
For  this  reason,  the  assessment  of biotic
concentrations of  hazardous  substances  at human
exposure  points is  limited  to  the qualitative
identification of  major pathways,  and  the  rough
quantification of exposure  levels wherever some
means of relating  ambient  soil,  water, or  air
concentrations to  edible tissue concentrations  are
available.

The available  methods of  estimating tissue
concentrations in aquatic animals, terrestrial animals,
and  terrestrial plants are discussed in  the  following
sections.

3.6.1.1  Aquatic Animals
Because  aquatic animals  are immersed in the
contaminated water medium  to  which  they are
exposed,  it  is  commonly  assumed  that tissue
contaminant  concentrations  are a  function  of
contaminant  equilibrium partitioning between water
and  organic tissue, and  are therefore directly related
to contaminant ambient water  concentrations. The
bioconcentration  factor  (BCF)  represents the ratio of
aquatic  animal  tissue concentration to water
concentration.  This ratio  is  highly  contaminant-
specific and is also dependent on the aquatic species
and on site parameters.

The  most  reliable source of aquatic  animal  BCF
values  is monitoring data for the site. Wherever water
concentrations and biotic tissue concentrations  have
been surveyed simultaneously, a site-specific BCF
can  be calculated for the species and  substance
involved (assuming water  column  concentration
values  represent  relatively steady concentrations over
at least the previous several  weeks,  and not short-
term high  or low concentrations). This  BCF can  be
used to  project  changes in tissue  concentrations
resulting  from projected changes in  ambient water
concentrations of the  involved hazardous substance.

In cases where site monitoring data are insufficient for
development of a BCF, one can use the BCF values
reported in technical  literature. A substantial amount
of research is available  regarding the bioconcentration
of hazardous substances,  especially in aquatic
organisms (see USEPA Office of Water Regulations
and  Standards:  Ambient  Water  Quality Criteria
documents, for a review of research current to 1980;
or Verschueren  1984;  Dawson,  English,  and  Petty
1980; Mabey at al. 1982; and Callahan et al. 1979 for
BCF factors). Exercise care  to  match  contaminants,
species, and site conditions  (e.g.,  temperature, pH,
water salinity) for  which  reported BCF values  were
measured  with conditions at the site. BCF values for
different species  or contaminants or those measured
under dissimilar conditions may not be applicable.

A third alternative for derivation of  BCF values is to
calculate these  values  based  on  the structure or
physiochemical  properties  of  the hazardous
substance.  See  Lyman  et al. (1982),  Kenaga  and
Goring (1978), and Veith et al. (1980) for instructions
on BCF estimation procedures.

3.6.1.2 Terrestrial Animals
Little data are available allowing the quantification  of
contaminant concentrations in  edible terrestrial animal
tissue  based  on  ambient  environmental
concentrations. Kenaga (1980) compiled and studied
data comparing dietary  concentrations of several
organic compounds with  the  concentration  of these
compounds in the fat of beef cattle. He found that the
fat/diet  BCFs for these compounds  correlate
reasonably well  with  the  water  solubility  (negative
correlation)  and  octanol-water partition  coefficient
(positive correlation)  of these  compounds.  BCFs
could only be predicted within three to four orders  of
magnitude,  however.  This method of  tissue
concentration estimation  must be  considered
semiquantitative at best.

Human  exposure  to contaminants through  the
terrestrial animal pathway can be  reliably determined
only through identification  of potential  vector
organisms and  exposure points,  and through  a
sampling and analysis program for determining tissue
concentrations at these exposure points.

3.6.1.3 Terrestrial Plants
Plant adsorption  of environmental contaminants  has
been studied by various researchers,  and some  data
are available regarding the uptake of pesticides  and
other contaminants by edible crops. These data cover
specific crop uptake  of specific  contaminants  (see
CDHS  1985 for  a review of pesticide  research),
however, and  no  relationships  allowing  reliable
extrapolation  of soil/plant tissue concentration ratios
are presently identified. Where plant/soil BCF data are
available in  the  technical  literature for the specific
plant species, contaminant, soil type, and tissue  type
of concern  in a Superfund exposure assessment,
these BCF data  can be used for a semiquantitative
estimation of edible tissue concentrations.

As  is  the case  with  terrestrial  animals, the most
reliable technique  for assessing  contaminant
concentrations at points of human exposure to plant
tissue  is the  identification of potential  vector
organisms and exposure points, and the surveying  of
tissue contaminant concentration in these organisms.
                                                 94

-------
                                              Chapter 4
                                    Uncertainty in the Analysis
This chapter  provides  a  brief introduction  to the
evaluation of uncertainties  inherent in  the  exposure
assessment process. When  applying  the  exposure
assessment  tools outlined in  the preceding sections,
uncertainty  may be  a  factor  at  each step.  Such
uncertainty  can  involve  variations  in the values of
variables used as  input  to a given model, the
accuracy with which the model itself represents actual
environmental  processes, and the manner in which
the exposure scenario is developed. Each of these
categories of potential uncertainty is discussed  below.
Once the exposure  assessment  is completed, its
results must be reviewed  and  evaluated to identify the
degree  of uncertainty involved. This  factor  should
then be considered  when using the  assessment
results for remedial decisionmaking.

The  following discussions focus on the uncertainties
of assessing the average  daily exposures to toxic
chemicals; uncertainties  related to the'  human  health
response to  these exposures  are not discussed. The
information  provided here does not constitute  a
comprehensive  treatment of  uncertainties  in the
exposure assessment  process. It is intended to  make
the analyst  aware  of  the categories of uncertainties
that  may be involved in exposure  assessments. In-
depth  guidance for  the execution of uncertainty
analyses is provided in various references   in the
literature. Specifically,  the analyst may wish to  review
the following  sources  of information concerning
various  aspects  of  uncertainty analysis pertinent to
the exposure assessment process:

 - Cohen (1950)
 - Eisenhart (1968)
 - Henrion  and Morgan  (1984)
 - Hoffman  et  al. (1984)
 - Kleijnen  (1974)
 - Morgan et al. (1984)
 - Rubinstein  (1981)
 - USEPA (1987e)

4.1 Sources of Uncertainty

4.1.1 Input Variable Uncertainty
Most  of the analytical procedures presented  in this
manual are  quantitative  in  nature,  and their   results
may be  highly dependent upon the accuracy  of the
input  variables used. For  example,  hydraulic
conductivity  and other parameters that  determine the
velocity of ground water and the contaminants that it
may carry can  vary  significantly over relatively short
distances, thereby affecting  one's  ability to  estimate
average contaminant velocities  with  confidence.
Similarly, the  presence of hydrogeologic hetero-
geneities  can  affect the  speed  with  which
contaminants arrive at a given  well from their point of
release and also their direction  of travel. Often, the
presence of such heterogeneities may be unknown.
Thus,  the  accuracy with  which  values for  such
parameters can be quantified is critical  to the degree
of confidence  that  the  decisionmaker  has in the
assessment  results.

Most scientific computation  involves a limited number
of input variables, which are tied together by  a  model
to provide the desired output. The  input parameters
can be broadly classified into the following categories:
constants, state variables, and natural variables.

A constant  has a single value irrespective of the
nature  of  other variables.  In some cases, the
variability of a parameter may  be so small that it can
be considered constant. In other cases,  even  if the
value  varies, its effect on  the  final answer  may  be
minimal. The results  are not sensitive to variation in
that parameter's value.

A state variable is one that has a fixed value, but that
value  is  not known  accurately.  The errors  in  such
variables  are  due  to  limitations  in  experimental
techniques. A relevant example  is the octanol/water
partition coefficient. While this has a single value for a
given  system,  some  degree of  uncertainty is
introduced  through  experimental  errors.  In  some
instances the values  of  state  variables  are estimated
rather than measured; therefore, the  uncertainties  for
such values  are even higher.

A natural  variable is one  that can  exhibit  different
values. An example is soil porosity, which can  exhibit
different values within a range because the soil  matrix
varies with location,  and because a  given area may
include many soil types.
                                                  95

-------
If  the  actual  values  for such variables are  not
accurately  known  for the location in question,  the
estimated exposure may be significantly in error. This
problem is  illustrated  by a study where the values of
parameters needed  to  calculate the velocity of a
solute in ground water were varied randomly,  using
Monte  Carlo simulation  techniques  (Mercer,  Silka,
and  Faust  1985). This analysis determined that  the
velocity estimates may vary  over  four  orders of
magnitude.

The  selection of accurate input  parameters is
essential to estimate the contaminant velocity and
other components of the exposure  assessment.
Often,  however,  the analyst  will not be able  to
determine the value of such parameters with absolute
certainty. It  is important that one be aware of the type
and degree of uncertainties involved at each stage of
the analysis,  and  interpret the  results obtained
accordingly.

The  different  values  of input  parameters that  are
measured  many  times can  be  expressed  as a
parameter  distribution.  A  parameter distribution
typically appears as a bell-shaped  curve.  The mode,
or the most likely value, is represented by the peak of
the bell-shaped  curve.  The  tails  to either side
represent the  relative frequency of times  when  the
measured values are greater or less  than the mode.
For a  parameter  that  varies considerably,  the bell-
shaped  curve will be wide (standard deviation is
large).  For  those that do not vary  appreciably,  it will
be narrow (standard deviation is small).

Input parameter distributions can be used to generate
the output  parameter distribution.  The  shape of the
parameter  distribution conveys the degree  of
uncertainty  of the  parameter (input  or output). This is
the most rigorous way to  define  the uncertainty  of the
predicted  output  parameter;  however,  it is  used
infrequently in  the environmental field  due to the lack
of input parameter distributions  upon which to base
the predicted output  parameter distribution. This
subject will  be discussed  further in the section on the
Monte Carlo technique.

 In the  environmental field, the methods used for
discussing the  degree of uncertainty are  often
qualitative  rather than  quantitative.  Qualitative
methods involve  discussing whether the data  are
thought  to  be  representative or not.  Some exposure
modeling is done  based on  literature  values  rather
than  measured values. In such  cases the degree of
certainty may  be expressed as  whether the estimate
was  based  on literature values  or  measured values,
not on  how well  defined the distribution of  the
parameter is. Some exposure estimates are based on
estimated  parameters;  the  qualitative  statement  that
the exposure  was based on  estimated  parameters
defines the  certainty in a qualitative  manner.
4.2  Modeling Uncertainty

4.2.1 Model Simplification
The  degree to which a specific contaminant transport
and  fate  model  accurately  represents  the  actual
conditions that are  present in the environment
constitutes a large source of potential  uncertainty.
The  analyst must  choose the model that  addresses
the appropriate aspects of interest.

Models are typically simplifications of the complexities
of reality.  There is some accuracy lost when making
these simplifications. While such loss may be small in
some cases,  in others  it  may be  unacceptably  large.
Two  assumptions  that illustrate  this idea are  the
assumptions of homogeneous soils and isotropic soils
for ground-water  models.  In  most  cases,  these
assumptions do not materially change the answer.  If
the soil under the  site  has layer cake stratigraphy, the
assumption of homogeneity is invalid. Typically,  most
cases will  be in-between  the  two extremes  of
homogeneous   soils and  completely  non-
homogeneous soils. The analyst will have to decide if
the assumptions are valid for each case.

In some cases the simplification of the real world into
an  actual model  is  acceptable and,  although
producing  uncertainty,  it is a necessary  evil. There is
a point at  which the level of the discrepancy between
the model and the real world constitutes an error in
the  use  of the  model and not  an  acceptable
simplification that  is necessary to  model  a
complicated  real world. At this point, the deviation is
an error and not an uncertain prediction.

4.2.2 Averaging Hydraulic Conductivities
An example of this would be  the  modeling of ground
water flow by averaging the hydraulic  conductivities
across all  aquifer materials. For contaminant transport
modeling,  this would constitute an error;  however, for
modeling well production, this is  an  accepted
practice.  Ground water modeling  with  numbers has
been occurring for the last 100 years. For the first 90
years of this period, most of the modeling was for
water supply; contaminant migration was  not
modeled.  The practice of averaging the hydraulic
conductivities across the cross-sectional area  of the
aquifer produced  answers  that had high  certainties
when predicting the volume of water that  could be
produced  by  a well during a period of time.  Some
modelers  applied  this  technique  to the  problem  of
modeling contaminant migration and produced
erroneous results. Although they were accustomed to
this practice, it was not acceptable in this case.

Modeling contaminant  migration requires that areas of
different hydraulic  conductivity be treated  separately
(sometimes it is not possible to differentiate the  areas
and  the  model  results must  be  viewed as  less
certain). For example,  if the site overlies  a sand  layer
and  a  clay layer, the analyst should model the two
                                                 96

-------
layers separately. The result of the separate modeling
will show that the time of arrival  in the sand is much
sooner than in the clay layer. Effectively the majority
of the  contaminant  mass would  migrate through the
sand layer and hardly  any would use the clay layer for
migration. Assuming  an  average  hydraulic
conductivity would predict a time  delay  between
release and arrival that is 100 to  1000 times too long.
Such  uncertainties,  however, constitute an  error  of
approach, and are not unresolvable uncertainties.

4.2.3 Dispersion  Simulation
Different  ground-water models simulate dispersion  in
different  ways.  The  degree to  which a  particular
model  accurately models  the  dispersion  at a given
site affects the accuracy of using that model for that
site. Ground  water dispersion modeling is a  young
field and  the state of the art is rapidly advancing. The
analyst should  become familiar  with the  dispersion
simulation technique for each model he/she uses.

Also, some ground  water models presume an aquifer
of infinite depth,  while  some  model  a finite aquifer
depth.  Contaminants dispersing  in an aquifer of finite
depth will effectively reflect off the lower aquitard and
cause  the resulting  downstream  concentrations to be
larger.  Use of a model  appropriate to the constraints
of the  site  is necessary for accurate  modeling  of the
drop-off  in  contaminant  concentration with  travel
distance.  Additionally,  some  models will simulate
lateral constraints  of the aquifer to model  this
limitation on  the  reduction in  downstream
concentrations.

Dispersion modeling  in air and  surface water has
been performed for  a  much longer time, and as such,
the methods for modeling  dispersion have coalesced
into  a  consistent approach. However, limitations on
the extent of dispersion  for air modeling can vary. For
example, a valley model will simulate the constraint of
lateral  dispersion by  the valley  walls. A model that
handles  inversions  will  simulate  the build-up  of
contaminant  concentration due  to  limited vertical
mixing.  Surface water models may vary on the
approach  they take to modeling initial mixing.  Some
surface water models use  compartments to manage
the modeling task.  If the modeler uses  a  small
number of large compartments,  small scale effects
may not be accurately modeled and the results will be
less certain.

4.2.4 Numerical Models and Analytical Models
Different types of models provide varying accuracy in
different situations. Two types   of models  are
numerical  (finite-element)  and  analytical  models.
Neither is best in all  cases, but  one  is usually better
in a given situation.  The numerical models are
typically  more difficult to use,  and thus ease of use
may enter into the decision of model selection.
Analytical  models  often  involve  mathematical
simplifications. These simplifications  are made  in
order to  find a closed-form solution.  In  most cases
the accuracy lost is  negligible; however,  in extreme
cases the inaccuracy  will be large.

Typically, analytical  models  require  less computer
time than do numerical models. If the  grid is large, a
numerical model requires a substantial  amount  of
computer time  for  each  run.  Numerical models
typically  require  more  input data.  Different program
needs cause different questions  to  be raised.  A
preliminary  scoping  problem will  rarely require  a
numerical model;  conversely,  a problem that requires
maximum defensibility will  suggest  that the additional
data and operational  burdens of a numerical model
are justified  in light  of the greater certainty  of the
output.

In cases where the question involves simulating what
will happen  in typical  generic  situations  across the
country,  an  analytical  model will  give a  better  picture
than a numerical model. Numerical models address
site-specific conditions  better  than do analytical
models:  they do  not  necessarily  model a typical
situation  with any increased accuracy.

4.2.5 Chemical Degradation Simulation
Some  models do  not  describe all  of the processes
that may potentially occur.  For example, degradation
is not accounted  for  in  some models. If the con-
taminant is extremely  refractory (i.e., does not
degrade), this limitation will not  materially affect the
answer.   If  the  contaminant  degrades  quickly,
however, this limitation  will cause the model results  to
be in  substantial error. Some models  simulate the
effect on the  reaction  rate  kinetics  of two con-
centrations while  some use only one  concentration.
The simpler approach of 1st order  reaction kinetics is
acceptable if the other concentration  does  not vary
appreciably, and is  less  accurate if both the con-
centrations vary substantially.  The  analyst must rely
on his/her judgment to ensure that  the  uncertainty is
minimized.

4.2.6 Model Operational  Parameters
Certain modeling  parameters  specified by  the analyst
can have a profound effect  on the accuracy and
viability of the  output.  An example is the  parameter
"time step." Time  step is used  on iterative models.
Models may either calculate an answer explicitly  or
they may determine  their  solution  with  a  successive
iteration  approach. For iterative  models, the analyst
will have to  make many model runs, and not stop until
he/she has a good run. The challenge of choosing an
appropriate  time step is that both  too large and too
small  time  steps  cause  inaccuracies.  The analyst
must find the optimum size for the  time step.  A time
step that is too small  causes  numerical  error
                                                  97

-------
propagation  (see below),  while one that is too large
causes a less accurate calculation of each step.

Numerical error propagation in  iterative models can
cause  inaccurate answers. If the analyst uses too
many  iterations,  the  truncation error of digital
representation  of numbers can build upon the
successive  iterations and produce  output that  is
totally  erroneous,  The degree of  error that  can be
present can  make the output totally meaningless. For
example,  the  estimated  output concentrations can
include concentrations that are greater than  a million
PPM  or  concentrations  that  are  negative.  Clearly,
concentrations in these ranges signify  a bad  run. The
analyst must also watch  out  for iteration errors that
produce  errors that are  less obvious and, hence,
there  is the possibility that the analyst will not be
aware  of  their occurrence. Conversely, the analyst
may choose too few iterations and the  resulting time
step between each iteration then becomes too  large.
In this case, the model  will  inaccurately calculate
each step. The analyst must become familiar with the
models he/she is using so as to stay in  the safe area
between the two extremes. Knowing the precise limits
is difficult, but staying between them is important.

4.2.7 Source Shape
The degree to which the  shape of the source  is
modeled  can effect uncertainty. For example,  if the
analyst uses a point source model to model an area
source, the nearby concentrations  will  be less
accurate than they would be  if the analyst  used an
area source  model. Line sources and volume sources
can  provide the same  problem. At large  distances
from the source, the effect of the shape of the source
is less important, and may often be neglected. Some
sources are best  modeled  as a vertical line source
and some are best modeled as a  horizontal line
source; hence, orientation is a factor as  well as
shape. It is a matter of fit  between  the model and the
actual  site  rather  than   choosing the  best source
shape for all cases.

4.2.8 Steady State Modeling
Use of a  steady-state  model to  model  a true
steady-state  scenario provides accurate  results. Use
of a steady-state model  to model a  truly  dynamic
scenario  can produce inaccurate  answers.  In most
cases,  the analyst will have to  make a judgment as  to
whether the actual scenario  is  close enough to steady
state  to  justify  using  a  steady-state model.  The
analyst must match the model to the  question  being
asked,  and to the details  of the specific site, in order
to minimize the uncertainty of the output.

4.2.9 Number of Dimensions  Addressed by the
Model
Choice of  a one-, two- or three-dimensional
model can  affect  the  uncertainty  of the  results.
Neither is  best in  all cases  and, typically, one  is
preferred  in a given  site-specific  scenario.  The
three-dimensional model  generally  has less
uncertainty than the  one- -or two-dimensional
models, but, this is not always the case. For example,
when  modeling the migration  of contaminants  in
ground water  through  a lO-foot thick  aquifer, a
two-dimensional  model  will  produce more  certain
results than  the  blind  application  of a three-
dimensional  model.  It is  not just a trade-off between
difficulty of the model and  quality of the output, but a
matching situation as well.

4.3 Scenario Uncertainty

The  analyst needs to  be aware of uncertainties that
result  from  using conservative  assumptions  when
data  are lacking. While  it is  traditional in exposure
assessment to make conservative assumptions  in  the
absence  of data, such assumptions  must  be
reasonable  and  the assessment  results must  be
interpreted with  caution. Use  of reasonably
conservative assumptions at each step may produce
cumulative  assessment results  that are  overly
conservative and thus unreasonable.

In addition, conceptual errors may result in the use of
assumptions that  affect the selection of the modeling
technique  applied to the exposure assessment.  For
example,  using a three-dimensional  model  in
situations where the aquifer  thickness is  not "large"
in relation to the  areal extent  of contamination would
not be  appropriate.  Thus, the concepts upon  which
the exposure  scenario  is based  must be carefully
considered to make sure that they  adequately reflect
the situation  under evaluation.

Quantitative  descriptions  of scenario  uncertainty  are
often  impractical,  and  qualitative descriptions  of  the
level of uncertainty are more common  for  the young,
and  developing,  field of exposure  assessment. Any
exposure prediction  has  cases of overstatement and
understatement of  risk. Where possible, the
understatements and overstatements of risk are
minimized.  Where this  is not possible, the  analyst
attempts  to  balance them  so as  to produce a
prediction that is most realistic.

4.4  Approaches   for  Dealing  with
Uncertainty

4.4. J Sensitivity Appraisals
Variation in  the  values  of input parameters  causes
variation in  the  values of the output parameters. The
ratio  of the  input parameter variation to  the  output
parameter variation will be different for parameters in
different parts  of the equation.  Sensitivity appraisals
involve assessing which  parameters have the highest
ratios and which  have the lowest. The accuracy of
parameters that have  the  largest  effect on the
accuracy of the output  parameters should be  high,
while parameters that have only  a small effect on  the
accuracy of the output parameters  can be estimated
                                                 98

-------
or determined by  less  accurate and less costly
methods.

Sensitivity appraisals can  be quantitative  or
qualitative. A quantitative  sensitivity appraisal involves
plotting the output parameter as a function of variation
of a single  input  parameter, while holding  all of the
other input parameters constant. As one can imagine,
there may  be a  different functional  relationship
between the output parameter and the varying  input
parameter  for each combination  of fixed  input
parameters. For complex models the approach can
become  overwhelming. Typically,  the analyst will  be
able to  interpret the equation and set up  the  fixed
input variables so as to  minimize  the  number of
functional relationships  produced.  However,  it  may
still be burdensome, and it may produce results that
are more precise than necessary.

In the environmental modeling field,  the  qualitative
approach  has  strong advantages  over  a  quantitative
approach.  The  qualitative  approach  involves
inspecting  the  model's  equations, and ascertaining
which input variables are the most sensitive.  This is
usually done  by  visual inspection,  with an
understanding  of the mathematical relationships in the
equation. For  example,  if one input  parameter
multiplies all the other terms, the analyst can expect
the input parameter to have a sensitivity ratio of one.
If the input parameter is the  exponent of  the  other
terms, the analyst can expect this parameter to  have
a very high sensitivity ratio. If the input parameter is
part of a separate term that is added to the rest of the
equation,  and  it is multiplied  by a constant of low
value, the input parameter can  be assumed  to have a
low sensitivity  ratio. A qualitative appraisal  is usually
the most efficient technique for determining  the  input
parameter accuracy needs.

4.4.2 Monte-Carlo Simulations
The Monte-Carlo  technique  involves running a
model  a  large number  of times with  varying  input
parameters. The values  for the input parameters are
chosen  from  the  parameter  distributions,  with  its
relative  frequency  of a  particular value being  used
being equal to the relative frequency in the parameter
distribution.  This is based on the assumption that the
input variables vary independently from each  other.
This  technique  generates  an  output  parameter
distribution,  which provides a mode and a  statement
of the uncertainty  associated with the prediction.

One difficulty with this technique is the  assumption of
independent variation. The input variables are chosen
as  if there were no relationship among them. If the
variables are  truly independent, the results are
accurate. Typically, however, the variables are related
to each  other and are, thus, dependent  variables. For
example, if the two input variables are  hydraulic
conductivity and hydraulic gradient, the analyst could
assume  that they  are  either independent  or
dependent. The  analyst could  assume  independency
because the two variables  represent different factors
that do  not have a  direct functional relationship
between them. But, if the  analyst looked  at enough
sets of data,  the sites with high conductivity  would
have  more  gradients  that  are flat;  conversely, the
sites with low  conductivity would have more gradients
that are  steep. Thus,  the two variables would  exhibit
covariation and  cannot  be  considered strictly
independent. This weakens the validity of  using the
Monte-Carlo  approach.

While  it is   possible to use  input parameter
distributions  to  generate model  output distributions
using  Monte-Carlo  simulations,  it is usually not
possible  to get the input parameter  distributions. The
input  parameter distribution shows the variation  of
parameter values. It must be based on  a  large
number of observations (actual measurements). The
environmental  field is young and  growing.  As such,
most  sampling (to date) falls  short of providing the
mass  of data  necessary to generate an  input
parameter distribution. Faced with  this dilemma, some
analysts have  fallen back  on assuming  such
distribution. Since they do  not  have a way to  gauge
the distribution, a uniform distribution from  the  lowest
to highest possible value is  assumed. This distribution
states that there is an  even probability  that the value
could be any value  between  the  lowest and the
highest value of the range.

Assuming an  input parameter  distribution  does not
help to  reduce uncertainty,  however, as the certainty
of the output is then  a function  of  the  assumed
certainty of the input parameter. For example,  if you
assume that the input parameters are  very precise,
then the  certainty of the output is  high.  Conversely, if
you  assume the parameters  may have  an  equal
probability to be any  value  across the range  of
possible values, the certainty of the output will be low.
Using a Monte-Carlo  approach with assumed  input
parameter distributions  that  are uniform  only indicates
how accurate  the model is at  predicting the  output
parameter when  you  have no idea what  the  input
parameters are, since models predict output based on
the relationship to the  input parameters. Thus, using
the Monte-Carlo  technique  to assess the certainty of
a  model's predictions  cannot be done with assumed
input parameter distributions.

4.4.3 Using Monitoring Data to Calibrate fhe
Model
One of the best ways to reduce the uncertainty  of the
predicted parameter  is to use monitoring  data  to
calibrate the  model. If you  have measured
contaminant concentrations that are comparable  to
modeled contaminant concentrations, the analyst can
correct  for over-  or  under- predictions.  If for
example, the measured values are always 90%  of the
predicted values, the  analyst  can multiply all  of the
output values by  90%.
                                                  99

-------
The difficulty of this technique is that the values must
be comparable.  In  many  cases  the model is being
used  to  predict future  events. Current contaminant
concentrations can be determined  more accurately by
monitoring, thus the need for modeling is reduced.

In air and surface  water modeling  the  difference
between  current and future events is  much  smaller
than for ground water modeling. Air and surface water
move more quickly than does ground water.  Hence,
calibration is  a more useful  technique in  air and
surface  water  modeling than  in ground water
modeling.

If the ground  water  model  predicts  a  certain
contaminant concentration  1  mile  from  the  source
after  20  years,   and  monitoring shows  no
contamination  at 1 mile from  the  source, this cannot
be used  to calibrate the model. The plume may not
have  reached  the point  1 mile away, as of yet. In  20
years, monitoring  may very  well  show  the same
contaminant concentration that was predicted by the
model. Care  should be taken to  ensure that the
monitoring data  used  to  calibrate the model  are
comparable in  time and space.


4.5  Level of Uncertainty Appropriate for
Exposure Modeling

There is  no one level of certainty that is appropriate
for all situations.  Each  program has different needs,
and various parts of a program have diverse needs. A
screening level study has less need  for accuracy than
a court case  that will require a  substantial  sum of
money from a PRP. The level of defensibility  required
will vary from one situation to another

EPA  program offices have developed a multi-tiered
approach. A desk top model  may be sufficient  for a
first-tier  analysis,  an analytical  model may  be
sufficient for a second-tier analysis, and a numerical
model may be required for a third-tier  analysis. For
example,  the method of screening sites for inclusion
on the National Priorities List should be less rigorous
than the  method of supporting a decision  on  various
site  clean-up  options.  Data  requirements will also
vary.

Although it would be nice to have  maximum accuracy
in all  cases, it would also imply maximum  difficulty in
all cases. Clearly,  a balance must be found between
difficulty and accuracy of the prediction.


4.6 Risk Communication

Once the analyst  has completed  the modeling  task,
the results of  the task must be communicated to the
analyst's  supervisor. This information should  include
the predictions of exposure over  time, and it should
include some communication  regarding the  level of
uncertainty associated with the prediction. The level
of uncertainty  can be expressed in a quantitative or
qualitative form.  Further  guidance on risk
communication can be found in USEPA (1987e).

A quantitative appraisal of the uncertainty is the most
preferable way to express  the uncertainty. A
quantitative presentation  may be an output parameter
distribution which  tells the most probable  value
(mode)  and the  relative probability that  the value is
larger or smaller than the  mode. Or, the presentation
may consist of the predicted value and a standard
deviation. The  standard deviation provides the level of
precision or uncertainty. Another  approach  involves
providing the predicted value and the 95% confidence
limits. The 95%  confidence  limits express that 95%
of the  possible  values of the parameter will be
between the upper and  lower  confidence  limits, The
main  catch to  precise  numerical  expression   of the
uncertainty is the  lack of sufficient data upon which to
base the quantitative expression of the uncertainty. In
the future, it  may  be possible to  use  this precise
approach.

A qualitative appraisal of the uncertainty  is the most
viable  way to express  the level  of  uncertainty. A
qualitative  presentation  will  describe  the significant
factors  that determine the level of uncertainty. The
quality of the prediction  is a function of the quality of
the inputs to the prediction. Major inputs that affect
quality are: data  precision, model  sophistication, and
defensibility of  the scenario.

Expressing the  quality of the data  would  entail
describing the  sources of the data. For example: Did
the data come from literature values or were the data
taken from actual site measurements? Were the data
measured by the best  available techniques or were
they sampled  by  another  technique?  Were replicate
samples taken? Was the sampling protocol sufficient
to obtain  representative samples?  Are the costs of
the sampling program appropriate for  the use  of the
results, or could more expensive  data gathering
techniques be  used?

Expressing the quality of the model used would entail
a description  of  the type  of model. For example: Is
the  model a  desk-top  calculation,  an  analytical
model, or a numerical model? Has the model been in
use  for some time or  is  it new? Is the model a
standard model used by  the agency or is it new to the
agency? Have other people used  the  model? Does
the model address  all of the important facets  of the
situation, or does  it neglect some potentially important
factors?  Has the  model been  used in  court  cases
before? How  good  is  the  model relative to other
possible models?  Is it the best  available model  at this
point in time? Is the model the  most defensible  model
available? Were monitoring data used to calibrate the
model  predictions? How  comparable  were the
monitoring data to the model predictions?
                                                  100

-------
Expressing the quality of the scenario is more difficult.
Reasonableness of the scenario  is important. Use of
similar scenarios by the agency in the  past is useful
information. Questions to ask would include: Was the
scenario  used in court cases,  for rulemaking activity
that  has been  published in  the Federal Register,
and/or did it receive public comment? Was the public
comment favorable  or  did it bring  out potential
difficulties?  Does the  scenario  neglect  certain
exposure routes that have  been  neglected  by  the
agency in the past?

The  important aspect to consider is  how good  the
prediction is,  not how imperfect the model is.
Modeling is  a  young field that  is rapidly growing.
Uncertainties are  minimized  but  never eliminated.
Modeling produces  state-of-the-art estimates,  and
nothing more.
                                                  101

-------
Preceeding Page Blank
                                           Chapter 5
                                          References
Ambrose, R.B. and Vandergrift, S.B. 1986. SARAH, a
  surface  water assessment  model for back
  calculating reductions in abiotic hazardous  waste.
  Athens,  GA:  U.S.  Environmental Protection
  Agency.  EPA-600/3-86/058.

Ambrose,  R.B., Vandergrift,  S.B.  and Wool, T.A.
  1986. WASPS, A hydrodynamic and water  quality
  model-model theory, a user's manual, a pro-
  grammer's guide. Athens, GA:  U.S. Environmental
  Protection Agency.   EPA-600/3-86-034.

Ambrose,  R.B.,  Wool, T.A., Connolly,  J.P.  and
  Schantz,  R.  1987.  WASP4, a hydrodynamic  and
  water quality model-model  theory,  a  user's
  manual, a programmer's guide. Athens, GA: U.S.
  Environmental Protection Agency.

Bear,  J. and Verruijt, A. 1987. Modeling groundwater
  flow and  pollution. Dordrecht, The Netherlands: D.
  Reidel Publishing Company.

Beljin,  M.S. 1985. A  program package  of analytical
  models  for solute  transport in groundwater
  'SOLUTE'.  Indianapolis, IN: International  Ground
  Water Modeling  Center,  Holcomb Research
  Institute, Butler University.

Bonazountas,  M.  and Wagner, J. 1981. SESOIL, a
  seasonal soil compartment model. Cambridge,  MA:
  A.D.  Little Inc. for  U.S. Environmental  Protection
  Agency.  Contract No. 68-01-6271.

Bonazountas,  M.,  Fiksel, J.,  et al.  1982.
  Environmental mathematical pollutant fate  modeling
  handbook/catalogue (Draft). Washington, DC: U.S.
  Environmental Protection Agency,  Office  of Policy
  and Resource Management.  Contract  No.  68-
  01-5146.

Bouwer, H. 1978.  Groundwater hydrology.  McGraw-
  Hill Pub. Co., New York, NY.

Bowers, J.F., et al. 1979. Industrial  source complex
  (ISC) dispersion model user's guide, volumes I and
  II. Washington, DC: U.S. Environmental  Protection
  Agency.  EPA 45014-79-030.  As reviewed  in:
  Bonazountas  M,  Fiksel   J.,  et  al. 1982.  En-
  vironmental mathematical  pollutant  fate modeling
  handbook/catalogue (Draft). Washington, DC: U.S.
  Environmental Protection Agency, Office of Policy
  and  Resource Management.  Contract  No.  68-
  01-5146.

Brown, D.S. and  Allison, J.D. 1987. MINTEQA1,
  equilibrium metals  speciation  model:  a  user's
  manual.  Athens,  GA:  U.S.  Environmental
  Protection  Agency.

Brown, S.M., Boutwell, S.H.,  Roberts, B.R. 1983.
  Selection and use  of models for remedial action
  evaluation  at uncontrolled  hazardous waste sites.
  Palo Alto,  CA: Anderson-Nichols &  Co., Inc. Draft
  report.  EPA  Contract No. 68-03-3116. Work
  Assignment No. 5.

Burns, L.A., Cline,  D.M.,  Lassiter, R.R.  1982.
  Exposure analysis modeling system  (EXAMS) user
  manual  and  system documentation. Athens, GA:
  Environmental Research  Laboratory, Office of
  Research  and  Development,  U.S.  Environmental
  Protection  Agency.  EPA-600/3-82-023. As
  reviewed in:   Versar Inc.  1983.  Methodology for
  assessing  exposures to chemical substances via
  the  ingestion of drinking water.  Washington, DC:
  U.S. Environmental Protection Agency. Contract
  No.  68-01-6438.

Burt,  E.  1977.  VALLEY  model user's  guide.
  Washington,  DC: U.S. Environmental  Protection
  Agency. EPA-450/2-77-018.  As  reviewed in:
  U.S.  EPA.  1982. Environmental  Modeling
  Catalogue. Washington, DC:  U.S.  Environmental
  Protection Agency.  Information Clearing  House.
  PM-211A.

Busse, A.D.,  Zimmerman, J.R. 1976. User's guide for
  the  climatological  dispersion  model.  Research
  Triangle Park, NC:  U.S. Environmental Protection
  Agency.  EPA-R4-73-024. As  reviewed in:
  Bonazountas  M, Fiksel J., et al.  1982.
  Environmental mathematical pollutant fate modeling
  handbook/catalogue (draft). Washington, DC: U.S.
  Environmental Protection Agency, Office of Policy
  and  Resource Management.  Contract  No. 68-
  01-5146.
                                               103

-------
Callahan,  M., Slimak, M., Gabel, N.,  et al.  1979.
  Water-related environmental fate of 129  priority
  pollutants. Washington,  DC: U.S. Environmental
  Protection  Agency.  EPA-440/4-79-029a,b.

Carsel, RF., Smith,  C.N., Mulkey, L.A., Dean, J.D.
  and Jowise,  P.  1984.  User's manual  for the
  pesticide root zone  model (PRZM): release  1.
  Athens, Georgia:   Environmental Research
  Laboratory,  U.S. Environmental  Protection  Agency,
  EPA-600/3-84-  109.

Chico, T. and  Catalano, J.A. 1986. Addendum  to the
  user's guide for MPTER. Research Triangle Park,
  NC: U.S. Environmental Protection Agency.

Christiansen, J.H. 1976.  Design and application of the
  Texas  episodic  model.  Proceedings of the
  Conference  on Environmental Modeling and
  Simulation.  Washington,  DC: U.S. Environmental
  Protection  Agency.    EPA-600/9-76-016.  As
  reviewed in: U.S.  EPA.  1982. Environmental
  Modeling Catalogue.  Washington, DC: U.S.
  Environmental  Protection Agency.  Information
  Clearing House. PM-211 A.

Clapp, R.B. and Hornberger, G.M.  1978. Empirical
  equations for some soil  hydraulic properties.  Water
  Resources  Research.  14:601-604.

Codell, R.B.,  Key, K.T. and Whelan,  G. 1982. A
  collection of mathematical models for  dispersion in
  surface  water and ground water. Washington DC:
  U.S.  Nuclear  Regulatory Commission.
  NUREGICR-0868.

Cohen, A.C. 1950. Censored samples  from  normal
  distributions. Ann. Math.  Stat. Volume 21.

Cowherd,  C.,  Muleski, G.E., Englehart, P.J. and
  Gillette,  D.A. 1985. Rapid assessment of exposure
  to particulate emissions from surface contamination
  sites.  Kansas  City,  MO. Midwest Research
  Institute.   PB85-192219.

Craig, D.G., Turelle,  J.W.  1964. Guide for wind
  erosion  control  on cropland in the  Great  Plains
  states.  U.S. Department of  Agriculture. Soil
  Conservation Service. Washington, DC.


Cupitt, L.T. 1980.  Fate  of  toxic and  hazardous
  materials in  the air environment. Research  Triangle
  Park,  NC: U.S.  Environmental  Protection Agency.
  EPA-600/3-80-084.
Das, B.M.  1983.  Advanced  Soil Mechanics.  New
  York,  McGraw-Hill Book  Company.
Davis, L.A.  and  Segol, G. 1985.  Documentation and
  user's guide:  GS2  and GS3 - variably saturated
  flow  and  mass  transport  models. Washington,
   D.C.:  U.S.  Nuclear  Regulatory Commission.
   NUREG/CR-3901.

Dawson,  G.W.,  English,  C.J. and Petty, S.E.  1980.
   Physical  chemical properties of  hazardous  waste
   constituents. Athens, GA: U.S. Environmental
   Protection  Agency,   Environmental  Research
   Laboratory.

Delos,  C.G., Richardson, W.L., DePinto, J.V.,  et al.
   1984.  Technical  guidance  manual for performing
   wasteload allocations, book II:  streams and  rivers.
   U.S. Environmental  Protection Agency.  Office  of
   Water  Regulations  and Standards. Water Quality
   Analysis Branch. Washington, DC. (Draft Final.)

Domenico,  P.A.  and  Palciauskas,  V.V.  1982.
   Alternative boundaries in solid waste management.
   Ground water. Vol. 20, No. 3.

Donigian, A.S., Imhoff, J.C., Bicknell, B.R. and  Kittle,
   J.L.  1984. Application  guide for  hydrological
   simulation  program  -  Fortran  (HSPF).  U.S.
   Environmental  Protection Agency.  EPA-600/3-
   84-065.

Eisenhart, C. 1968.  Expression of the uncertainties of
   final  results. Science. Volume 160. June 1968.

Enfield, C.G. and  Bengtsson,  G. Not  dated. Macro
   molecular transport of hydrophobic contaminants in
   aqueous environments.  EPA/ORD, Ada, OK.

Enfield, C.G., Carsel, R.F., Cohen, S.Z.,  Phan, T. and
  Walters,   D.M.  1982.    Approximating  pollutant
  transport to  ground water.  Ground  Water  20(6)
   711-722.

Enfield, C.G. 1984. Chemical transport  facilitated by
   multiphase  flow systems. Robert  S.  Kerr
   Environmental  Research Laboratory.  U.S.
   Environmental  Protection Agency, Ada,  OK.
   Presented at seminar  on Degradation, Retention
   and  Dispersion  of Pollutants  in  Ground Water,
   Copenhagen,  Denmark, September 12-14.

Farino,  W.,  Spawn, P., Jasinski, M.  and Murphy, B.
   1983.  Evaluation and selection of models for
   estimating air emissions  from hazardous waste
  treatment, storage and disposal facilities. Revised
   Draft  Final Report.  GCA  Corporation.
   GCA/Tech nology  Division.    Bedford,
   Massachusetts,  prepared  for U.S. Environmental
   Protection Agency, Office of Solid  Waste.  Land
   Disposal  Branch.  Contract No. 68-02-3168.

Farmer, W.J.,  Yang, M.S., Letey, J., Spencer, W.F.
  and  Roulier, M.H.  1978.  Land  disposal of
   hexachlorobenzene  waste:  controlling  vapor
   movement in  soils. San Antonio,  TX. 4th Annual
  Symposium on Land Disposal.
                                               104

-------
Felmy,  A.R.,  Girvin,  D.C., Jenne,  E.A. 1984.
  MINTEQ  — a  computer program  for calculating
  aqueous geochemical equilibria. Athens,  GA:  U.S.
  Environmental  Protection Agency.  EPA-600/3-
  84-032.

Finley, N.C. and Reeves,  M.  1968. SWIFT  self-
  teaching  curriculum. Washington,  DC:  Nuclear
  Regulatory Commission.  NUREG/CR-1968,
  SAND  81-0410.  As reviewed in: Lo  TYR, Scott
  BH, Benjamin  RR.   1983. Remedial  action
  assessment  models  for hazardous waste sites.
  Review draft. Athens, GA:   U.S. Environmental
  Protection Agency. Contract  No. 68-03-3116.

Fisher, H.B., List,  E.F., Koh, R.C.Y., Imberger,  J.,
  Brooks, N.H.  1979.  Mixing in inland and  coastal
  waters. New York, NY: Academic Press.

Freeze,  R.A., Cherry, J.A. 1979.  Groundwater.
  Englewood  Cliffs, NJ: Prentice-Hall, Inc.

GSC.  1982. Graphical  exposure modeling  system
  (GEMS) user's guide.    General  Software
  Corporation. Washington, DC: Office of Pesticides
  and Toxic Substances,  U.S. Environmental
  Protection Agency. Contract  No. 68-01-6618.

Gelhar, L.W., Mantoglou, A., Welty, C. and Rehfeldt,
  K.R. 1985.  A review  of field-scale  physical solute
  transport  processes  in  saturated and  unsaturated
  porous  media.   Palo Alto, CA:  Electric Power
  Research Institute. EPRI EA-190.

Geotrans. 1986. Analytical models  for  evaluating
  leachate  migration  in groundwater systems.
  Washington, DC: U.S.  Environmental  Protection
  Agency, Office  of Solid Waste.

Gillett, J.W.  1980.  Terrestrial microcosm  technology
  in assessing  fate, transport  and effects of  toxic
  chemicals.  In  Dynamics, Exposure  and  Hazard
  Assessment of Toxic  Chemicals, R. Hauge, editor.
  Ann Arbor Science,  Ann Arbor,  Ml. As quoted in
  Trabalka and Garten (1982).

Gupta, S.K., Cole,  C.R.,  Kincaid, C.T., and Monti,
  A.M.  1987.  Coupled fluid,  energy,  and solute
  transport (CFEST) model:  formulation  and user's
  manual. Columbus, Ohio:  Office of Nuclear Waste
  Isolation, Battelle Memorial Institute.

Gureghian  A.B.  1983.  TRIPM: a two-dimensional
  finite-element  model for  the  simultaneous
  transport of water and reacting solutes through
  saturated  and  unsaturated  porous  media.
  Columbia, OH: Battelle Memorial  Institute, Office
  of Nuclear Waste Isolation. BMI/ONWI-465.
Haith, D.A.  1980. A mathematical  model  for
  estimating pesticide losses in  runoff. Journal of
  Environmental  Quality. 9(3):428-433.

Hanna, S.R.,  Hosker,  R.P., Jr.  1980. Atmospheric
  removal  processes for  toxic  chemicals. Silver
  Spring,  MD: National  Oceanic and Atmospheric
  Administration.  Technical Memorandum ERL
  ARL-102.

Haynes, W.A.  1966. Guide for wind erosion  control in
  the northeastern United States. U.S.  Department of
  Agriculture. Soil Conservation Service. Washington,
  DC.

Hendry,  D.G.,  Kenley,  R.A. 1979.  Atmospheric
  reaction products of organic  compounds.
  Washington, DC:  U.S.  Environmental Protection
  Agency.    EPA-56011 2-79-001.

Henrion, M. and Morgan,  M.G. 1984. A computer aid
  for  risk  and  other policy analysis. Pittsburgh,  PA:
  Carnegie-Mellon  University, Department  of
  Engineering and Public Policy.

Hilsmeier,  W.F.,  Gifford,  F.A.  1962. Graphs  for
  estimating atmospheric diffusion. Oak  Ridge, TN:
  Atomic  Energy  Commission.  ORO-545.  As
  reviewed in:  Turner  DB.  1970.  Workbook  of
  atmospheric dispersion estimates. Research
  Triangle  Park,  NC: U.S.  Environmental Protection
  Agency, Office of Air Programs. AP-26.

Hoffman,  F.O.,  Miller,  C.W., Little, C.A. 1984.
  Uncertainties  associated  with  predictions  derived
  from models and  parameters. In: Models and
  parameters  for  environmental  radiological
  assessments. C.W. Miller, Ed. Oak Ridge National
  Laboratory, Health and  Safety Research Division.

Hwang, S.T.  1982.  Toxic emissions  from land
  disposal  facilities. In Environmental  Progress. Vol.
  1, No. 1.

Huyakorn,  P.S., Ungs,  M.  J.,  Mulkey,  L.A. and
  Sudicky, S.A.  1987.  A three-dimensional
  analytical method for predicting  leachate migration.
  Ground water, Vol. 25, No. 5.

HydroQual,  Inc. 1982.  Application guide for CMA  -
  HydroQual  chemical fate models.  Prepared  for:
  Chemical Manufacturers  Association, Washington,
  DC. As reviewed in: Versar Inc.  1983. Methodology
  for  assessing  exposures to chemical  substances
  via the  ingestion  of drinking  water. Washington,
  DC: U.S.  Environmental  Protection  Agency.
  Contract No.  68-01-6271.

Intera. 1983.  SWENT: A  three-dimensional finite-
  difference code for the simulation of fluid, energy,
  and "solute  radionuclide transport.   Inter-a, Inc.,
                                               105

-------
  Columbus,  OH: Office of Nuclear Waste Isolation,
  Battelle Memorial Institute.  BMI/ONWI-457.

Javandel, I.,  Doughty, C.  and Tsang,  C.F.  1984.
  Groundwater transport:  handbook of mathematical
  models. AGU Water Resources  Monograph.
  10:240.

Jaw-Kwei, A.M. and Luthy,  R.G. not dated. Aromatic
  compound solubility  in  solvent/water mixtures.
  Pittsburgh,  PA: Carnegie-Mellon University.

Jenson, M.E.,  ed.  1973.  Consumptive  use of water
  and  irrigation water requirements.  New York, NY:
  American Society of Civil Engineers.  As presented
  by  Enfield et al.  1982.  Approximating pollutant
  transport to ground  water.  Ground Water 20(6)
  711-722.

Johnanson,  R.C.,  Imhoff, G.C.,  Davis,  H.H.  1984.
  Hydrocomp Inc.  Users  manual for  hydrological
  simulation program  -  FORTRAN (HSPF). Athens,
  GA:  Office  of Research and  Development,  U.S.
  Environmental Protection Agency.    EPA-600,9-
  80-015. As reviewed  in: Versar Inc.  1983.
  Methodology for  assessing  exposures  to chemical
  substances  via  the  ingestion  of  drinking  water.
  Washington,  DC: U.S.  Environmental Protection
  Agency. Contract No. 68-01-6271.

Kenaga,  E.E. and  Goring,  C.A.I.  1978.  Relationship
  between water  solubility, soil  sorption, octanol-
  water  partitioning, and concentration of chemicals
  in biota. In:  Eaton JG, Parrish  PR,  Hendricks AC,
  eds. Aquatic  toxicology. Proc. of  third annual
  symposium  on aquatic  toxicology,  New Orleans,
  LA,  17-18 Oct. ASTM special technical publication
  707.

Kenaga,  E.E.  1980. Correlation  of bioconcentration
  factors of chemicals  in aquatic and terrestrial
  organisms  with their physical  and chemical
  properties.  Environ.  Sci. Technol. 14(5):  553-556.

Kent,  D.C., Pettyjohn,  W.A. and Prickett, T.A. 1985.
  Analytical methods  for the  prediction  of leachate
  plume  migration.  Ground Water Monitoring Review.
  Spring 1985. Vol. 5 #2 pp.  46-59.

Kent,  D.C., Pettyjohn, W.A., Witz,  F. and Prickett, T.
   1982.  Prediction  of leachate plume migration and
  mixing in ground  water. Solid and Hazardous Water
  Research and Development Annual  Symposium
  proceedings. Columbus,  OH: National  Water Well
  Association. As   reviewed  in:  Versar Inc.  1983.
  Theoretical evaluation of sites located  in the zone
  of saturation.  Draft  final  report. Chicago,  IL: U.S.
  Environmental Protection Agency.   Contract  No.
  68-01-6438.
Kinzelbach,  W.  1986.  Groundwater modeling: an
  introduction with sample  programs  in  BASIC.
  Amsterdam, The Netherlands:  Elsevier Science
  Publishers.

Kipp,  K.L. Jr., 1987. HST3D: A  computer code for
  simulation of heat and solute  transport  in three-
  dimensional groundwater flow systems. Lakewood,
  CO: U.S. Geological  Survey, WRI 86-4095.

Kleijnen,  J.P.C.  1974.  Statistical techniques  in
  simulation. New York. NY: Marcel Dekker, Inc.

Konikow, L.F.,  and Bredehoeft, J.D.  1978. Computer
  model  of two-dimensional  transport  and
  dispersion in ground  water. Washington,  DC:  U.S.
  Geological Survey. Techniques  of Water  Resource
  Investigation,  Book 7, Chapter  2.  As  reviewed in:
  Versar  Inc.  1983. Theoretical  evaluation  of  sites
  located  in the zone of saturation. Draft final report.
  Chicago, IL:   U.S.  Environmental  Protection
  Agency. Contract No. 68-01-6438.

Kufs,  C.,  Repa,  E., Rogoshewski, P.,  et  al. 1983.
  Leachate plume  migration control  (unpublished
  draft).  Cincinnati,  OH: U.S. Environmental
  Protection Agency.  Contract No.  68-03-3113.

Liu,  H. 1977. Predicting  dispersion coefficient  of
  streams. J.  Environmental Engineering  Division.
  Proceedings of the American  Society of  Civil
  Engineers. Vol. 103.

Lyman,  W.J., Reehl, W.F.,  Rosenblatt,  D.H. 1982.
  Handbook  of chemical property  estimation
  methods. New  York. McGraw-Hill.

Mabey,  W.R., Smith, J.H., Podoll,  R.T., et  al. 1982.
  Aquatic  fate process data for organic priority
  pollutants.  Final  draft report.   Washington,  DC:
  Office of Water Regulations and Standards,  U.S.
  Environmental Protection Agency.

Mackay, D., Leinonen,  P.J. 1975. Rate of evaporation
  of  low-solubility  contaminants from  water  bodies
  to atmosphere.  In Environ. Sci.  Technol.   Vol.
  9(13).

Martinez,  M.J. 1985.  A finite element computer
  program for simulating  radionuclide transport
  through  porous media. Albuquerque,  NM: Sandia
  National Laboratory.  SAND 84-0747.

Mercer, J.W.,  Silka, L.R.  and  Faust,  C.R. 1985.
  Modeling ground-water flow  at Love Canal.
  Journal  of Environmental Engineering, August.

Miller,  F.J., Gardner, D.E.,  Graham, J.A., Lee,  R.E.
  Jr.,  Wilson,  W.E., Bachmann,   J.D.    1979.  Size
  consideration  for establishing a standard for
                                                106

-------
  inhalable particles.  In Journal of the Air Pollution
  Control Association.

Mills, W.B., Dean, J.D., Porcella,  D.B., et  al.  1982.
  Water quality assessment:  a  screening procedure
  for toxic and conventional  pollutants:  parts  1,  2,
  and 3. Athens, GA:  U.S. Environmental Protection
  Agency. Environmental  Research  Laboratory.
  Office of Research  and Development.   EPA-
  600/6-821004 a,b,c.

Mockus, J. 1972. Estimation of direct  runoff from
  storm rainfall.  In  national  engineering handbook.
  Section  4:  hydrology.  Washington, DC:  U.S.
  Department of Agriculture.  Soil Conservation
  Service.

Morgan, M.G., Henrion, M., Nair,  I., Mullin, T. and
  Wiecha, C. 1984. A  generic  "pre-HEED" on
  characterizing and dealing  with  uncertainty  in
  health  and environmental  risk assessment.
  Pittsburgh, PA:  Carnegie-Mellon  University,
  Department of Enginering and  Public Policy.

Neely, W.B.  1982. The definition and use  of mixing
  zones. Environ. Sci.  Technol. 16(9):520A-521A.

Nkedi-Kizza,  P.,  Rao,  P.S.C.  and  Hornsby,  A.G.
  1985. Influence of Organic cosolvents on sorption
  of hydrophobic organic  chemicals by soils.
  Environmental  Science  and Technology  Vol.  19,
  No. 10, p.  975-979

Onishi,  Y., Whelan,  G. and Skaggs,  R.L.  1982.
  Development  of  a multimedia radionuclide
  exposure  assessment methodology  for  low-level
  waste management.  Athens,  GA: Office  of
  Research  and Development,  U.S. Environmental
  Protection  Agency.  As  reviewed in: Versar Inc.
  1983.  Methodology for  assessing  exposures  to
  chemical substances  via the  ingestion of drinking
  water.  Washington, DC: U.S.  Environmental
  Protection Agency.  Contract  No. 68-01-6271

Onishi, Y. and Wise, S.E. 1982a. Mathematical model,
  SERATRA, for sediment-contaminant transport  in
  rivers and  its application to pesticide  transport  in
  Four Mile and  Wolf Creeks in Iowa.  Athens, GA:
  Office  of  Research and Development,  U.S.
  Environmental  Protection Agency.    EPA-600-
  3182-045.  As reviewed  in:  Versar  Inc.  1983.
  Methodology for assessing  exposures  to chemical
  substances via the ingestion of drinking water.
  Washington, DC:  U.S.  Environmental Protection
  Agency.  Contract No.  68-01-6271.

Onishi,  Y.  and Wise,  S.E.  1982b.  User's manual for
  the  instream  sediment-contaminant transport
  model SERATRA. Athens, GA: Office of  Research
  and  Development,  U.S.  Environmental Protection
  Agency.    EPA-60013-82-055.  As reviewed in:
  Versar  Inc. 1983.  Methodology  for assessing
  exposures to chemical substances via the ingestion
  of drinking  water.  Washington, DC:  U.S.
  Environmental Protection Agency.  Contract No.
  68-01-6438.

Onishi, Y.  1981.  Sediment-contaminant transport
  model.  Journal of the  Hydraulics Division, ASCE.
  107(HY9):1089  -  1107.  Proc.  Paper 16505. As
  reviewed in: Versar Inc.  1983. Methodology for
  assessing  exposures to  chemical substances via
  the ingestion  of drinking water.  Washington, DC:
  U.S. Environmental Protection  Agency.  Contract
  No.  68-01-6271.

Pasquill, F. 1961. The estimation of the dispersion of
  windborne  materials. Meteorol. Mag. 90,  1063,
  33-49.  As reviewed  in:  Turner  DB.     1970.
  Workbook of atmospheric dispersion estimates.
  Research  Triangle  Park, NC: U.S.  Environmental
  Protection Agency,  Office  of Air Programs. AP-
  26.

Perry, R.  and Chilton, C.  1973. Chemical engineers
  handbook. 5th edition.  New York: McGraw-Hill.

Pierce, D. and Turner, D.B. 1980. User's guide for
  MPTER.  Research Triangle  Park, NC:  U.S.
  Environmental Protection Agency.   EPA-600:8-
  80-016. NTIS No.  PB-80-176361,

Prickett,  T.A.,  Naymik, T.G.  and Lonnquist.  C.G.
  1981. A "random-walk" solute transport model for
  selected  groundwater  quality   evaluations.
  Champaign, IL:  Illinois  Department  of Energy and
  Natural  Resources.    ISWS/BUL-65181.  As
  reviewed in:  Versar  Inc.  1983.  Theoretical
  evaluation  of  sites  located   in  the zone of
  saturation.  Draft final  report.  Chicago  IL:  U.S.
  Environmental Protection Agency.  Contract No.
  68-01-6438.

Rao, P.S.C., Hornsby, A.G., Kilcrease, P.P.,  Nkedi-
  Kizza,  P.  1985.  Sorption and  transport of
  hydrophobic organic chemicals in  aqueous and
  mixed  solvent  systems:  model  development and
  preliminary  evaluation.  Journal   of  Environmental
  Quality, Vol.  14, No.  3, July-September.  Rawls
  WJ,  Brakensiek DL, Saxton  KE. 1982. Estimation
  of Soil Water Properties.  Transactions of the
  American Society  of  Agricultural Engineers.
  25(5):1316-1320 and 1328.

Rawls,  W.J.  1986. Computer printouts  from the soils
  data base, dated  August  28,  1986, from  W.J.
  Rawls,  Beltsville Agricultural Research  Center,
  Beltsville, Maryland.

Reeves, M. and Cranwell, R.M.  1981. User's manual
  for  the Sandia Waste-Isolation Flow  Transport
  model  (SWIFT).  Washington,  DC: Nuclear
                                               107

-------
  Regulatory  Commission.  NUREG/CR-2324,
  SAND81-2516.  As  reviewed  in: Lo  TYR,  Scott
  BH, Benjamin  RR.   1983.  Remedial  action
  assessment models for hazardous  waste sites.
  Review draft. Athens, GA:  U.S. Environmental
  Protection Agency.  Contract No. 68-03-3116.

Reeves, M., Ward, D.S.,  Johns, N.J. and Cranwell,
  R.M.  1986. The Sandia waste-isolation flow and
  transport model for fractured media; release  4.84:
  theory  and  implementation. Washington, D.C: U.S.
  Nuclear Regulatory Commission. NUREGICR-
  3328.

Rifai,  H.S.,  Bedient,  P.B., Borden,  R.C.  and
  Haasbeek,  J.F.  1987. BIOPLUME  II;  computer
  model  of two-dimensional contaminant transport
  under the  influence  of  oxygen limited
  biodegradation  in  ground water. Houston,  TX:
  National  Center for  Ground water Research, Rice
  University.

Rubinstein, R.Y. 1981.  Simulation and the Monte
  Carlo  method. New  York, NY: Wiley Science
  Publishers.

Runchal,  A., Sagar, B., Baca. R.G. and Kline,  N.W.
  1985.  PORFLO -  a  continuum model for  fluid
  flow, heat transfer,  and mass  transport in  porous
  media.    Richland,  WA: Rockwell  Hanford
  Operations.  RHO-BW-CR-150P.

Russell, R.S., Bartlett, B.O.  and Bruce, R.S. 1971.
  The significance  of long-lived  nuclides  after  a
  nuclear war.  In D.W.  Bensen  and A.M. Sparrow,
  eds. Survival  of food  crops and livestock in the
  event of nuclear war.  Oak Ridge, TN: US  Atomic
  Energy  Commission. As quoted in  Trabalka and
  Garten  (1982).

Salame,  M.  (no  date) Permeability-structure
  relationships of high polymers.  Obtained by private
  communication. Monsanto Co., Bloomfield, CT.

Salame, M. 1961. The prediction  of liquid permeation
  in polyethylene and  related polymers.  SPE Trans.
  1(4): 153.

Salame,  M. 1973. Transport properties  of nitrile
  polymers. J. Polymer Sci.  41:1-15.

Salame, M. 1985. Private communication. Monsanto
  Co., Bloomfield, CT.

Sanford,  W.E.  and  Konikow, L.F.  1985. A  two-
  constituent solute transport model for groundwater
  having variable  density. Reston,  VA:  U.S.
  Geological Survey.  WRI 85-4279.
Sax, N.I.  1984.  Dangerous properties of industrial
   materials,  6th  edition.    New York,  NY: Van
   Nostrand Reinhold.

Schnoor, et al.  1987.  Processes, coefficients,  and
   models for simulating toxic  organics and  heavy
   metals in  surface waters. Athens,  GA:  U.S.
   Environmental Protection Agency. EPA-600/3-
   87/015.

Schwab,  G.O.,  Frevert,  R.K.,  Edminster,  T.W.  and
   Barnes,  K.K. 1966. Soil  and water conservation
   engineering. 2nd  edition.  New York:  John Wiley
   and Sons.

Seely, D., Turina,  P., Pangaro,  N., et  al.  1983.
   Development of protocols  for ambient air sampling
   and monitoring at hazardous waste  facilities:
   methods summary report.  Draft  Report.  GCA
   Corporation.  GCA/Technology Division.  New
   Bedford, MA. Prepared  for U.S.  Environmental
   Protection Agency,  Office of Solid Waste,  Land
   Disposal Branch.  Contract No.  68-02-3168.

Sehmel, G.A.  1980.  Particle resuspension:  a review.
   In Environment International. Vol.  4.  Pergamon
   Press, Ltd.

Shen, T. 1981. Estimating hazardous air  emissions
   from  disposal  sites.  Pollution Engineering  13(8):
   31-34.

Skidmore, E.L., Woodruff, N.P. 1968. Wind  erosion
   forces in the United States  and their  use in
   predicting soil loss. Agriculture Handbook No. 346.
   Washington, DC:  U.S. Department of Agriculture,
   Agricultural  Research Service.

Skinner. 1984. Banning wastes from  land disposal.
   Paper presented by EPA Office of Solid Waste and
   Emergency Response  at  the First Public Briefing
   on  the  1984 Amendments to the Resource
   Conservation and  Recovery Act.  December 11,
   1984, Washington, DC.

Smith, W.J.,  Whicker,  F.W.,  Meyer, H.R.  1982.
   Review  and  categorization of saltation, suspension
   and resuspension models.  Nuclear Safety. 23(6).

Steingiser,  S., Nemphos, S.P., Salame,  M.  1978.
   Barrier polymers.  In:   Kirk-Othmer  encyclopedia
   of chemical  technology,  3rd edition.  John Wiley
   and Sons, New York, NY.

Texas Air Control Board.  1980. User's guide  to the
   Texas  climatological Model. Austin, TX:  Texas  Air
   Control  Board. As  reviewed  in:  U.S.  EPA. 1982.
   Environmental  Modeling  Catalogue.  Washington,
   DC: U.S.  Environmental  Protection  Agency.
   Information  Clearing House. PM-211A.
                                               108

-------
Thibodeaux,  L.J., Hwang,  ST. 1982. Landfarming  of
  petroleum wastes  -  modeling the air emission
  problem. In Environmental Progress. Vol. 1, No. 1.

Thibodeaux,  L.J.  1981.  Estimating the air emissions
  of chemicals from hazardous waste  landfills.  In
  Journal of Hazardous  Materials. Vol. 4.

Trabelka, J.R., Garten,  C.T., Jr. 1982.  Development
  of predictive models for xenobiotic bioaccumulation
  in terrestrial ecosystems. Oak Ridge,  TN:   Oak
  Ridge National Laboratory, Environmental Sciences
  Division,  (NTIS  DE83 003171).  ORNL-5869.

Turner, D.B.  1970.  Workbook  of atmospheric
  dispersion estimates.  Research Triangle Park, NC:
  U.S.  Environmental Protection  Agency,  Office  of
  Air Programs.  AP-26.

Tracy, J.V. 1982. User's guide  and documentation for
  adsorption and  decay modifications to the USGS
  solute transport  model.  Washington,  D.C.:   U.S.
  Nuclear  Regulatory Commission. NUREGiCR-
  2502.

Trask,  P.O., Patnode,  H.W. 1942.  Source  beds  of
  petroleum. Tulsa, OK: The American Association  of
  Petroelum Geologists.

Travis,  B. 1984. TRACR3D:  A model of flow and
  transport in  porous/ fractured media. Los Alamos,
  NM:  Los  Alamos National  Laboratory. LA-0667-
  MS.

Turner,  D.B., and  Novak, J.H.  1978. User's guide for
  RAM.  Publication No.  EPA-60018-78-016 Vols
  a and b. Research Triangle  Park, NC:  U.S.
  Environmental Protection Agency.

USCG.  1974.  Hazardous chemical data.  Washington
  DC: United  States Coast Guard CG-446-2.

USDA.  1974.  Department  of Agriculture.  Universal
  soil  loss  equation. Agronomy technical  note no.
  32.  Portland, Oregon.  U.S. Soil  Conservation
  Service. West Technical  Service Center.

USDC.  1961.  Rainfall frequency atlas of the United
  States. Washington, DC:  U.S. Department  of
  Commerce. Weather Bureau.  Technical Paper
  Number 40.

USDC.  1968.  Climatic atlas  of the United States.
  Washington, DC: U.S.  Department of  Commerce.
  Environmental   Sciences  Services Administration.
  Environmental Data Service.

USEPA. 1977a. Guidelines for  air quality  maintenance
  planning  and  analyses, Vol. 10   (Revised).
   Procedures  for evaluating air quality impact of new
  stationary sources.  Research Triangle  Park, NC:
  Office of Air Quality Planning and Standards, U.S.
  Environmental Protection Agency.

USEPA.  1977b.  User's  manual  for single  source
  (CRSTER)  model.  Research Triangle  Park, NC:
  Office of Air Quality Planning and Standards, U.S.
  Environmental Protection  Agency. EPA  Publication
  No.  EPA-45012-77-013.

USEPA.  1979. Environmental modeling  catalogue.
  Washington, DC:  U.S. Environmental  Protection
  Agency. Contract  No. 68-01-4723.

USEPA.  1980a.  Hazardous waste  background
  document  for the control  of  hazardous waste
  leachate. Cincinnati, OH:  Municipal  Environmental
  Research Laboratory, Solid and  Hazardous Waste
  Research  Division, Briefing  document for  Review
  Committee  Meeting held at the U.S.  Environmental
  Protection  Agency, Washington,  DC. January 23,
  1980.

USEPA. 1980b. Land disposal  of  hexachlorobenzene
  wastes: controlling  vapor  movement in soil. U.S.
  Environmental Protection Agency.  Municipal
  Environmental  Research Laboratory. Office of
  Research   and  Development.    Cincinnati, OH.
  EPA-60012-80-119.

USEPA.  1982a. Environmental modeling  catalogue.
  Washington, DC:  U.S. Environmental  Protection
  Agency. Information Clearing House. PM-211  A.

USEPA.  1982b.  Establishment  of guidelines for
  modeling  groundwater contamination  from
  hazardous waste facilities. Preliminary groundwater
  modeling  profile (discussion draft). Washington,
  D.C.: U.S. Environmental Protection Agency, Office
  of Solid Waste.

USEPA. 1983a. Compilation of air  pollutant emission
  factors: AP-42.   Research  Triangle  Park, NC:
  U.S. Environmental Protection  Agency. Office of
  Air,  Noise  and Radiation.  Office of Air  Quality
  Planning and Standards.

USEPA.  1983b. Methods for assessing exposure to
  windblown  particulates.   Washington, DC:  U.S.
  Environmental  Protection  Agency. Office of Health
  and Environmental Assessment.  Office  of
  Research  and Development.   EPA-60014-83-
  007.

USEPA.  1983c. Technical  assistance document  for
  sampling and analysis of  toxic organic compounds
  in ambient air.  Research  Triangle Park, NC: U.S.
  Environmental   Protection Agency,  Environmental
  Monitoring  Systems  Laboratory. EPA-60014-83-
  027.
                                                109

-------
USEPA.  1985a. Guidance on remedial investigations
  under  CERCLA.    Washington,  D.C.:  U.S.
  Environmental  Protection  Agency,  Office of
  Emergency and Remedial  Response.

USEPA.  1985b. Guidance on feasibility studies  under
  CERCLA. Washington, D.C.: U.S. Environmental
  Protection Agency,  Office  of Emergency and
  Remedial  Response.

USEPA.  1985c.  Superfund  public health  evaluation
  manual. Draft. Prepared by ICF, Inc. for the  Policy
  Analysis Staff, Office  of Emergency and Remedial
  Response. Washington, D.C.: U.S. Environmental
  Protection Agency.  October 1, 1985.

USEPA.  1985d. Methods for assessing exposure to
  chemical substances.   Versar  Inc.  Washington,
  D.C.: U.S. Environmental Protection Agency,  Office
  of Toxic Substances.  EPA 56015-85-001.

USEPA.  1985e. Methods for assessing exposure to
  chemical substances  in the  ambient  environment.
  Versar Inc. Washington, D.C.: U.S. Environmental
  Protection Agency, Office of Toxic  Substances.
  EPA 56015-85-002.

USEPA.  1985f. Methods for assessing exposure from
  disposal  of  chemical substances.  Versar Inc.
  Washington,  D.C.:  U.S.  Environmental  Protection
  Agency, Office  of  Toxic Substances. EPA 560/5-
  85-003.

USEPA.  1985g. Methods for enumerating and
  characterizing populations exposed  to chemical
  substances. Versar Inc.  Washington, D.C.:  U.S.
  Environmental Protection Agency,  Off ice of Toxic
  Substances.  EPA  560/5-85-004.

USEPA.  1985h. Methodology for assessing exposures
  to chemical substances via  the ingestion of
  drinking water. Versar Inc. Washington,  D.C.: U.S.
  Environmental Protection Agency,  Off ice of Toxic
  Substances.   EPA 560/5-85-005.

USEPA.  19851.  Methods  for assessing environmental
  pathways  of food contamination.  Versar Inc.
  Washington,  D.C.:  U.S.  Environmental  Protection
  Agency, Office  of  Toxic Substances. EPA 560/5-
  85-008.

USEPA.  1985J.  Modeling remedial actions at
  uncontrolled hazardous waste sites. Cincinnati OH:
  Hazardous Waste  Engineering  Research
  Laboratory, Gffice  of Research  and Development,
  U.S.  Environmental  Protection  Agency.
  EPA/540/2-85/001.

USEPA.  1986a. Mathematical model  selection criteria
  for performing exposure assessments: groundwater
  contaminants from hazardous  waste facilities.
  Washington,  D.C.:  Office  of Health  and
  Environmental  Assessment, U.S. Environmental
  Protection Agency.

USEPA. 1986b.  Guideline on air quality  models.
  Research Triangle Park, N.C.: Office of Air Quality
  Planning and  Standards.  U.S.  Environmental
  Protection Agency.  EPA-450/2-78-027R.

USEPA. 1986c. Guidelines for exposure assessment.
  Washington, D.C.:   Office of Research  and
  Development,  U.S.  Environmental   Protection
  Agency.  51 FR  34042.

USEPA. 1986d. Office of Solid  Waste.  Hazardous
  waste management  system;  land  disposed
  restrictions;  proposed rule.  Federal Register Vol.
  51, No. 9. January 14, 1986.

USEPA. 1986e. Development of  advisory levels  for
  polychlorinated  biphenyls  (PCB)  cleanup.
  Washington,  DC:  U.S.  Environmental  Protection
  Agency, Exposure Assessment Group, Office of
  Research and  Development.  (NTIS  #PB86-
  232774) May 1986.  EPA/600/6-86-002.

USEPA. 1987a. Data quality objectives for  remedial
  response activities.  Washington,  D.C.: Office of
  Emergency and Remedial Response and  Office of
  Waste Programs  Enforcement,  Office  of Solid
  Waste  and  Remedial  Response,  U.S.
  Environmental Protection Agency.

USEPA. 1987b.  Compendium of superfund  field
  operating methods.  Washington,  D.C.: Office of
  Solid Waste and Emergency  Response,  U.S.
  Environmental Protection  Agency.  EPA-540/P-
  87/001.  OSWER  Directive 9355.0-14.

USEPA. 1987c.  Proposed criteria for selection of
  groundwater  exposure  assessment  models.
  Washington,  D.C.:  Office  of Health  and
  Environmental  Assessment. U.S. Environmental
  Protection Agency.  OHEA-E-219.

USEPA. 1987d. Mathematical model selection criteria
  for performing exposure  assessments:  surface
  water models.  Washington,  D.C.: Office of Health
  and  Environmental  Assessment.  U.S.
  Environmental  Protection Agency.  OHEA-E-245.

USEPA.  1987e.  Risk  Communication, Risk
  Assessment,  Management Communication, Guide
  to Selected Sources. USEPA Office of  Information
  Resource  Management, Office  of Toxic
  Substances,  March  1987.  EPA-IMSD/87/002.

van der Heijde, Paul K.M., Bachmat, Y., Bredehoeft,
  J., Andrews, B.,  Holtz, D.,  Sebastian, S.  1985.
  Groundwater management:  The use of numerical
  models.  Washington,  DC: American Geophysical
                                              110

-------
  Union,  Water  Resources  Monograph #5, 2nd
  Edition.

van der Heijde, P.K.M. and Srinivasan,  P.  1986.  A
  "random-walk"  model  for solute transport in
  groundwater 'RANDOM  WAL/RWH'. Indianapolis,
  IN:  International  Ground Water  Modeling  Center,
  Holcomb Research Institute,  Butler University.

van der Heijde, Paul K.M., Beljin, M.S. 1987. Model
  assessment for  delineating  well  head protection
  areas. Washington,  DC: Office of Groundwater
  Protection,  U.S.  Environmental Protection  Agency.
  International  Groundwater  Modeling  Center,
  Holcomb Research  Institute.   Butler  University,
  Indianapolis,  Indiana.

van Genuchten,  M.T.  1978. Mass transport in
  saturated-unsaturated media:   one-dimen-
  sional solutions. Princeton, NJ: Department of Civil
  Engineering. Princeton  University. 78/WR-11.

van  Genuchten,  M.T.  and Alves, W.J.  1982.
  Analytical   solutions  of  the one-dimentional
  convective-dispersive  solute transport equation.
  Riverside, CA: U.S. Salinity Laboratory, Agricultural
  Research Service, U.S. Department of Agriculture.
  Technical Bulletin 1661.

Veith,  G.D.,  DeFoe,  D.L., Bergstedt,  B.V.  1980.
  Measuring   and  estimating  the bioconcentration
  factor of chemicals in fish. J. Fish Res Board Can.
  36:  1040-1048.

Versar.  1983.  Theoretical evaluation of sites located
  in the zone  of saturation. Draft final  report. Versar,
  Inc.  Chicago IL: U.S. Environmental  Protection
  Agency. Contract No. 68-01-6438.

Verschueren,  K.  1984. Handbook of environmental
  data on organic chemicals.  New  York: Van
  Nostrand/Reinhold Press.

Voss,  C.I. 1984.  SUTRA: A finite element simulation
  model  for  saturated-unsaturated  fluid  density-
  dependent ground water flow with energy  transport
  or chemical  reactive single species solute
  transport.    Reston,  VA: U.S.  Geological  Survey,
  Water Resources Investigation.  84-4369.

Walton, W.C.  1984. Handbook of analytical ground
  water models.    International Ground  Water
  Modeling Center, Holcomb Research  Institute,
  Butler University, Indianapolis, Indiana.

Walton, W.C.  1985. Thirty-five  BASIC  groundwater
  programs  for desktop microcomputers
  'WALTON84-35BASIC'.  Indianapolis,  IN:
   International  Ground Water Modeling  Center,
   Holcomb Research Institute, Butler University.
Williams,  J.R.  1975.  Sediment-yield  prediction with
  the universal  equation using  runoff energy  factor.
  In Present  and prospective  technology for
  predicting  sediment yields  and  sources.  U.S.
  Department of Agriculture.  ARS-S-40.

Wilson,  J.L.,  Miller,  P. J.  1978.  Two-dimensional
  plume  in uniform ground-water flow. Journal of
  the Hydraulics Division,  ASCE 104(4): 503-514.

Wischmeier,  W.H., Smith,  D.D. 1978. Predicting
  rainfall erosion losses - a guide to conservation
  planning. Washington,  DC:  U.S. Department of
  Agriculture.  Agriculture Handbook No. 537.

Wischmeier, W.H.  1972.  Estimating  the cover  and
  management factor on undisturbed areas. U.S.
  Department of Agriculture.  Oxford,  MS:
  Proceedings of the  USDA Sediment  Yield
  Workshop.

Woodburn, K.B.,  Rao, P.S.C., Fukui, M.,  Nkedi-
  Kizza, P.  1986.  Solvophobic  approach for
  predicting sorption  of  hydrophobic organic
  chemicals on synthetic  sorbents and soils.  J. of
  Contam.  Hydrology.  1:  227-241.

Yeh, G.T.  and Huff,  D.D.  1985. FEMA:  A finite
  element model  of material transport through
  aquifers. Oak Ridge,  TN:  Oak Ridge  National
  Laboratory,   ORNL-6063.

Yeh, G.T. 1981. AT123D. Analytical  transient  one-,
  two-,  and  three-dimensional  simulation  of  waste
  transport in the  aquifer system. Oak Ridge,  TN:
  Oak Ridge National Laboratory,  Environmental
  Sciences Division  Publication No.  1439.  ORNL-
Yeh,  G.T.  1982. CHNTRN: a  chemical transport
   model for simulating  sediment  and  chemical
   distribution  in a stream/river network. Washington,
   DC: Office of Pesticides and  Toxic Substances,
   U.S.  Environmental Protection Agency. Contract
   No. W-7405-eng-26.  As reviewed in:  Versar
   1983.  Methodology for assessing  exposures to
   chemical  substances via the  ingestion  of drinking
   water. Washington,  DC:   U.S.  Environmental
   Protection  Agency.  Contract  No. 68-01-6271.

Yeh, G.T. 1987. FEMWATER: A finite element model
   of  water flow through  saturated-unsaturated
   porous media - first revision.  Oak Ridge, TN:  Oak
   Ridge Natational Laboratory, ORNL-5567/RI.

Yeh,  G.T.,  Ward, D.S.  1981. FEMWASTE: A finite-
   element  model  of waste  transport through
   saturated-unsaturated porous  media.  Oak Ridge
   National  Laboratory,  Environmental  Services
                                                111

-------
Division:  Publication No.  1462,  ORNL-5602.  137
p.  As reviewed in: Versar Inc.  1983. Theoretical
evaluation  of sites  located  in the  zone  of
saturation.  Draft final report. Chicago, IL:  U.S.
Environmental Protection Agency.  Contract  No.
68-01-6438.
                                             112

-------
                                         Appendix A
    Analysis of Exposed Human Populations  and Exposure Calculation and Integration

                                     Table of Contents

 Chapter                                                                             Page
1      QUANTITATIVE ANALYSIS OF EXPOSED POPULATIONS  	    114

      1.1    Introduction   	   114
      1.2   Exposed Populations Screening   	     114
      1.3   Quantitative  Exposed Populations Analysis   	    116
      1.4   Identification and Enumeration of Exposed Human Populations  	    118
           1.4.1    Populations Exposed Through Air  	    118
           1.4.2   Populations Exposed Through Surface Water or Ground Water  	   119
           1.4.3   Populations Exposed Through Food   	    119
           1.4.4   Populations Exposed Through Soil   	    120
      1.5   Population Characterization  	     120
      1.6   Activity Analysis   	     120

2     EXPOSURE CALCULATION AND INTEGRATION  	    121

      2.1    Inhalation Exposure  	     122
      2.2   Dermal Exposure   	     123
      2.3   Ingestion Exposure  	     128
           2.3.1    Food/Soil   	     128
           2.3.2   Water 	     128
      2.4   Exposure Integration  	     129

3     APPENDIX A REFERENCES 	     131
                                              113

-------
                                            Chapter  1
                         Quantitative Analysis  of Exposed Populations
1.1  Introduction

    The results  of contaminant release  and fate
analyses provide the  basis  for assessing exposed
populations.  This  assessment compares environ-
mental contamination data with populations data to
determine  the  likelihood of  human  contact with
contaminants of  concern. This  chapter details
methods  useful in evaluating  the following
components of exposed populations analysis:

   1.   Identification  and  enumeration of exposed
       populations;

   2.   Characterization of exposed populations; and

   3.   Analysis  of activities that bring  populations
       into contact with contaminants.

Each of these components is detailed in the following
subsections.

As with  other  evaluations,  exposed populations
analysis begins  with a screening assessment,  which
identifies exposure  pathways that are incomplete,  i.e.,
those situations where contaminants are released  and
migrate from a site, but  do not  contact human
populations and are not likely to do so in the future.
Such situations  require  no further analysis.  At the
same time, exposed populations screening also  points
out those  exposure pathways that are  complete  and
that  will require quantitative analysis to estimate the
extent of human exposure.

Data needed  to quantify potentially  exposed
populations are  readily  available.  In essence,  all
quantitative exposed populations  evaluations can  be
considered in-depth analyses.  For  each population
segment identified in  this  portion  of  the  exposure
assessment process, exposures  are quantified  and
integrated  as described in  Chapter 2 of this Appendix.
 1.2  Exposed  Populations Screening
 Exposed populations screening  is primarily qualitative.
 This evaluation draws on the results  of  contaminant
 fate  analysis  (presented in  Chapter 3) to determine
the likelihood and  extent of human population contact
with contaminants.

Exposed populations screening  is guided by  the
decision network  provided in  Figure A-l.  The
following numbered paragraphs  each  refer  to
particular numbered boxes in the  figure.

1. Human  exposure through inhalation should  be
evaluated for contaminants that have  migrated or may
migrate from the site into air. The assessment should
consider  both contaminated  dust  and volatile
compounds. For  screening  purposes,  comparing
contaminant concentration isopleths with maps of the
local area will  identify the  potential  for such  human
population  inhalation  exposure.  The user should
realize,  however,  that exposure can occur  in
recreational  areas  as well  as  in residential,
commercial, or  industrial areas,  and  should interpret
local area maps accordingly.

2. In  cases where  surface waterbodies  have been
contaminated by  toxics migrating from  a site,  the
water's potential commercial use  as a fish or shellfish
source  should be  evaluated.  If the waters are
commercially  fished,  fishermen may be  exposed
through  dermal contact with contaminated  water,
although  such  exposure will  generally   be
overshadowed by other exposure mechanisms.

3. In  cases where recreationally  or  commercially
caught fish/shellfish  are taken  from  contaminated
waters,  persons consuming the  catch  may  be
exposed. For  chemicals  that tend to bioaccumulate,
consumers may be  exposed  to  contaminant
concentrations  in  fish/shellfish tissue that are  many
times greater than those present in the water column
or sediments. When performing  exposed  populations
screening, the  analyst need  only determine whether
waters identified in the environmental fate analysis as
having received contaminants from the hazardous
waste site are used commercially or recreationally.

4. Individuals who swim in contaminated  waters can
experience dermal exposure to toxics over their entire
body.  In addition,  significant  quantities  of
contaminated water may be  ingested inadvertently
while swimming, and  swimmers will be  exposed to
                                                114

-------
Figure A-1.  Exposed populations decision network
                                                                             Environmental Fate Analysis

                                                                                       +
4, + 4r + +
Have Toxics Migrated
into Air?

Have Toxics Migrated
into Surface
Water?

Have Toxics Migrated
into Ground
Water?
I Have Toxics Migrated 1 1 Is Site Accessible 1
into Ott-site Soils? i— 1 1 to Public? i— 1
fH 1 L3
Are Persons
Potentially Exposed
Via Inhalation?

Is Contaminated
Surface Water
Fished
Commercially?
Is Contaminated
Surface Water
Used
Recreationally?

Is Contaminated
Surface Water
A Drinking
Water Source? Hf
                                                                                 Is Contaminated
                                                                                 Ground-water
                                                                                   A Drinking
                                                                                 Water Source?
  Are Persons
   Potentially
  Exposed Via
Dermal Contact?
Are Workers
Potentially
Exposed Via
Dermal
Contact? r—
[2

^V
Are Persons
Potentially Exposed
Via Consumption of
Contaminated
Fish/Shellfish?
[T
W

Are Persons
Potentially Exposed
Via Dermal Contact,
Ingestion, or
Inhalation
While Swimming?
[T

Are Persons
Potentially
Exposed Via
Ingestion?

A
Are Persons Potentially
Exposed Via
Inhalation During
Showering/Bathing?
r-ft-
Are Persons
Potentially
Exposed Via
Ingestion of
Contaminated Soil
or Home Grown Food?

Are Persons
Potentially
Exposed Via Derma
Contact or Soil
Ingestion On-site?
A.

Are Persons
Potentially
Exposed Via
Inhalation
On-site?
A
                                                                               Go on to Integrated Exposure
                                                                               Analysis for Each Potentially
                                                                                   Exposed Population

-------
volatile contaminants in  the water through inhalation.
Other  screening should evaluate the  existing or
potential  degree to which the local  population uses
contaminated water-bodies  (fresh  or marine)  for
swimming.

5. If contaminated ground water or surface  water is
a source of potable water, the population  served may
experience considerable  ingestion exposure. Similarly,
the population may also  be exposed to toxics through
both dermal  absorption   and  inhalation  (of volatiles)
while showering  or bathing. When  undertaking a
screening  analysis,  it is  only  necessary to determine
which  residences  or commercial/institutional  estab-
lishments  are likely  to obtain  their potable water from
contaminated water sources.

6. If contaminants  migrate to off-site  soils,  persons
contacting such soil may be exposed.  Individuals who
grow their own fruit or vegetables  at  home may
experience additional  exposure from  ingesting food
grown  in contaminated  soils,  as do those consuming
contaminated commercially-grown foods. Similarly,
livestock  that  have  grazed  on  contaminated
vegetation may  constitute  a source of ingestion
exposure  for consumers.  Screening  analysis should
strive to correlate  areas of  human   habitation with
areas  of contaminated soil, as  defined  in  the
environmental fate analysis.

7. Similarly, if direct access  to the site  is possible,
children may be  attracted  to the location  and may
directly contact hazardous materials or contaminated
soil.  Such activity may  result in inhalation or dermal
exposure,  as well  as  intentional  or  inadvertent
ingestion  of contaminated soil.  For screening
purposes,  the proximity of residential areas to the site
should indicate  the potential for direct access  by
children.

1.3  Quantitative Exposed  Populations
Analysis

Quantitative analyses of potentially  exposed human
populations comprises three distinct  steps, which  are
illustrated in  Figure  A-2.  First,  the  results of
environmental fate analysis are compared with data
identifying  and  enumerating  nearby  human
populations to provide  boundaries and quantify  the
population(s) potentially or actually coming  into
contact with  contaminated  air, water, and  soil.
Populations  consuming contaminated  food (home
grown  vegetables,  fish) can  similarly be  identified
once the areal extent of  contamination  is known.

Population  characterization, the second step, involves
identifying those groups  within the exposed population
that, because of the specific  health effects  of some
pollutants  or  factors related  to the  population itself,
would  experience  a higher risk than  would  the
average population  as  a result  of a given  level  of
exposure. Indeed,  the health effects  of the
contaminants under  evaluation  will often dictate the
need  for population  characterization. For example, if
mutagenic  or  teratogenic substances  are  involved,
women  of  childbearing age  should be considered a
high-risk group. In addition, factors relating to the
exposed population may cause  certain  groups to
constitute high-risk  subpopulations.  These include:

•   Persons with a  genetic  predisposition to certain
    health effects;

•   Persons whose health or resistance to disease is
    impaired by  behavioral factors such as smoking,
    use of alcohol or drugs, etc;

•   Infants, children, and  the elderly,  who are  more
    susceptible  to  health  impacts from  a  given
    exposure than are persons of other ages;

•   Persons who are already  suffering from disease
    and  may be  more  susceptible  to  further
    impairment as  a result of a given  level of
    exposure than are healthy persons;

•   Persons  who are  exposed to naturally  high
    background levels of contaminants (e.g., selenium
    or arsenic)  and may be at greater risk to  small
    incremental  increases of  hazardous  substances
    than  are persons who are  not  exposed  to  such
    background levels; and

•   Nutritionally  deficient  populations  who  may  be
    less  resistant to exposure  than those  with
    adequate diet.

While most studies  will  consider  only the exposed
population  as a  whole and not as  separate  discrete
subpopulations,  in  certain  cases, such  detailed
population  analysis  may  be  warranted for  in-depth
studies.

Age and  sex influence the  average inhalation  rate, the
rate of food and water intake, the body area subject
to dermal exposure, and the  types of food consumed,
all of which can affect the level of exposure actually
experienced. Some quantitative  assessments may
require  further  characterization  of populations to
determine age-  and sex-specific  exposure  factors.

The third step  is  activity  analysis. Once population
identification and characterization  have  answered the
question "Who  may be  exposed?",  the question
"How and  to what level are component portions of
this population exposed?" may next be asked in  order
to refine the  evaluation.  This refinement involves
determining  the  exposed  population's activities.
Comprehensive analysis can  encompass the range of
indoor, outdoor,  and in-car activities.  For Superfund
Feasibility  Studies,  however, average values for
activity-related  considerations usually suffice.
                                                  116

-------
Figure A-2. Quantitative exposed population analysis
Site Sampling
Data
1
Identification
and
Enumeration
of ^
Exposed ^N
Population
Characterization
of
Exposed '
Population
Activity
Analysis
r
\
+ ^
JL^^"""1^
r ~W
Inhalation of Inhalation o
Contaminants Contaminant
On-Site Off-Site
^
r ^ r
Identify Populations Comb
with Possible Access to Site with
(Workers, Children Fat
L ^
V.
^
*
9
»

Environmental
Fate Analysis
4-
I
1 Ingestion

-i- T ^
1 Water 1 PICA 1 Food 1

ne Census Combine Population Combine Population Data Use Recre
Air Data Data w'tn Ground-or wjtn Environmental to Identify
e Results Surface Water pate Results and Food of Exposed S
Fate Results and Production Statistics
•« » ^ r ^ r ^
*

Determine Determine
Site-Specific Age/Sex National Age/Sex
Distribution from Distribution from
Census of Population Census of Population
± ±

Identify/Quantify
Exposure-Related
Activities

Site Sampling
Data
+ +


4
4, 4,
Swimming 1 I Bathing 1 Direct Contact 1

ation Data Combine Population Identify Populations
Population Data with Ground-or Njtn Possible Access to Site
wimmers, etc Surface Water (Workers, Children)
Fate Results and



-------
The  activity  analysis can also help to identify high-
risk  groups.  For example, those  groups that may
experience a significantly higher frequency or duration
of exposure as compared with the general population
can also be considered high-risk groups.

1.4  Identification and  Enumeration  of
Exposed Human Populations

The major population data base that can be accessed
to determine  the  size,  distribution,  and demographic
characteristics of  a  geographically defined population
is the Census of Population.

The  data collected in  the Census are  organized
according to  geographic areas. Within these areas,
data  are further  broken  down into Census-defined
statistical areas  and  government  units.  Population
data  are available within  Standard  Metropolitan
Statistical Areas  (SMSAs)  down  to the  level of the
"block"  and  in  non-SMSAs to the level  of  the
Enumeration District (ED).

These data  are especially  useful in quantifying  and
characterizing populations exposed as a result of their
presence in  a specific  locale  (e.g.,  those  exposed to
toxics in ambient air or soil). An  isopleth  map of
varying  concentrations around a source  can be
overlaid with Census maps. Such maps are available
for areas within SMSAs and can be purchased from
the Bureau of the Census.  Also,  Census Tracts
(Series PHC80-2)  contains detailed  characteristics
of the population (e.g., age,  sex,  race, education)
within  each tract, a division of an  SMSA containing
4,000 residents  each.  Census  Tracts  is currently
available on  microfiche  by  SMSA  and on  computer
tape.

Many Super-fund  sites are not within SMSAs. Census
data for non-SMSA  areas are not available on  maps,
but can be transcribed from Census  publications.

The  most useful Census publications for  this type of
data  are Number of inhabitants  (Series  PC80-1-A)
and   General  Population  Characteristics  (Series
PC80-1-B).  Each series is currently available  and
consists of a separate volume for each state, together
with  a national summary  volume.  Number of
inhabitants  provides only population counts, with no
demographic data. It provides data down  to the level
of county subdivision and incorporated town. General
Population  Characteristics provides  population counts
by  age, sex,  and other  demographic data,  and
contains data down  to the level of small towns  (1,000
or more inhabitants).

All printed Census  information   is available for
purchase through  the Government Printing  Office
(GPO);  all series issued on  microfiche,  maps,
computer tapes,  and  technical  documentation  are
available directly  from the Customer Services Branch
at the  Bureau of the  Census,  Department  of
Commerce, Washington, D.C., and can be ordered  by
calling  (202) 763-4100. Alternatively, it may be more
convenient to contact one  of  the  Census Bureau
regional offices. Cities where such offices are located
and phone numbers for the public information service
within each regional office are listed in  Table A-l.
       Table A-l.
                  Regional Census Bureau
                  Offices
       Atlanta. Ga.
       Boston,Mass.
       Charlotte, N.C.
       Chicago, IL.
       Dallas,Tex.
       Denver.Colo.
       Detroit, Mich.
       Kansas City, Kans.
       Los Angeles, Calif.
       New York, N.Y.
       Philadelphia, Pa.

       Seattle, Wash.
(404)  881-2274

(617)223-0226

(704)371-6144

(312)353-0980

(214)767-0625

(303) 234-5825

(313)226-4675

(913)236-3731

(213)209-6612

(212)264-4730

(215)597-8313

(206) 442-7080
7.4.7  Populations Exposed through Air
A  convenient  means  of accessing quantitative
population data for a specific area impacted by  air
contaminants is to directly  link environmental fate and
exposed populations analysis through use of  an
integrated  computer-based  fate  model,  and
population  data retrieval  program called ATM-
SECPOP. Developed by  the EPA Office  of Toxic
Substances, Exposure  Evaluation Division  (OTS-
EED),  this  model primarily  analyzes point  source
emissions, but can  also be  adapted to  area  or line
source analyses.  ATM-SECPOP integrates the
output of a concentration prediction model  (ATM)
(Patterson et al. 1982); a  population distribution data
base  (the  proprietary 1980 Census Master Area
Reference  File (MARF)),  which is accessed via a
population distribution model  called  SECPOP; and
graphic  and mapping  information  displays. This
integration  affords  a rapid  and efficient  means  of
generating and  presenting exposure data  relating  to
the airborne release of chemical  substances. The
graphic display functions can be used  to illustrate the
relationship  of variables such as the distribution  of
exposure or concentration  versus distance  for any or
all directions around  a facility. Graphic displays may
be  in  the form of bar charts,  scatter plots, rose
diagrams, or maps. Because of the proprietary nature
of the  data  contained in MARF, ATM-SECPOP's use
is  restricted to personnel and contractors of EPA,
Office of Toxic  Substances (EPA-OTS). Special
arrangements can  be made for others to use  the
data.  Inquiries  should be directed  to the  Modeling
                                                 118

-------
Section of the  Exposure  Assessment Branch of
EPA-OTS  in Washington, D.C. A detailed discussion
of ATM is presented in Chapter 3 of this manual.

Where sites  are  accessible, the  possibility  that
children may  enter  and  explore or play  on the site
should be evaluated. On-site,  children may
experience  inhalation exposure to contaminated dust,
volatiles, or both. In some  cases, the  site  boundary
may adjoin residential properties, and the area of
contamination may actually  include such residences.
Accurate  estimation of the potentially  exposed
population in  such  a case  is difficult; it can  be
assumed that  each  household with  children  in the
immediate vicinity of the site has  one child  who  may
find the site inviting.  This  should provide  an upper
bound estimate on the actual number of children  who
may enter the  site.  The  Bureau of the Census (1986)
reports that in   1984,  50.1 percent of  all  U.S.
households included children. This percentage can be
applied to  the total  number of local households to
enumerate those in  the area with  children.  The
analyst  must decide which  households  are close
enough to  the site to be considered.

Similarly, workers conducting activities at the site  may
also experience inhalation exposure. Local authorities
(e.g.,   Zoning   Board) may  be able  to supply
information on the  likelihood  of on-site work-related
activities that can be used to estimate the number of
workers who may  become exposed. Remediation
workers are not  included in this  estimated exposed
population.
7.4.2 Populations Exposed through Surface Water
or Ground Water
Environmental  fate analysis results can  be used to
identify  geographically-defined  sources  of
recreational (aquatic)  dermal exposure, such as river
reaches downstream of an uncontrolled hazardous
site. The exposed population comprises swimmers in
those  specific  contaminated  waters. The local
government agency concerned with recreation should
be  able to  provide  estimates of the  populations
swimming in local waters; this will usually  be the
state,  city, or  county Department  of Parks  or
Recreation. Alternatively, one  can use  the following
national average value from the Bureau of Outdoor
Recreation (USDOI  1973): 34  percent  of the total
population  swims   outdoors  in natural surface
waterbodies (including oceans, lakes, creeks, and
rivers).

All  persons  served  by a water supply  system that
draws water from a  contaminated  water  source must
be  considered  as  potentially  exposed through
ingestion and  dermal exposure while  bathing.
Information concerning  local surface  drinking water
sources and populations served can be obtained from
the  local  Department of Public Works,  Planning
Department, or Health Department. These sources
should  be able to provide information on  public
departments or private  drinking  water treatment
companies  that use ground water as their raw water
supply, and also  on  the number of households
drawing water from private  wells.
1.4.3   Populations Exposed through Food
Exposure to  contaminated  food will  usually  be
associated with fruit and  vegetables grown  in home
gardens or with  game residing  in or using
contaminated areas. In order to identify  the number of
persons  consuming contaminated home grown  fruit
and vegetables,  first consult General Population
Characteristics,  Series  PC80-1-B  to learn the total
number of  households  in  a given  geographic area.
Then the data  presented in Table A-2,  which provide
estimates of the percent of households in urban  and
rural areas that have fruit and vegetable gardens  and
the average number of persons  per household,  can
be applied to the local population  data to estimate the
number of  persons likely  to consume contaminated
home grown produce.

The USDA Food  Consumption of Households report
series can  be consulted to estimate the local
population using a given food item for  urban,  rural
non-farm, and  rural farm locales. These  reports
present  seasonal food  use survey  data on  the
following bases:  Northeast (USDA 1983a),  North
Central  (USDA 1983b),  South (USDA 1983c),  and
West (USDA 19834). More aggregated  data are also
provided  for the entire United  States in a companion
report  (USDA  1983e).  The percent of households
using  a given food item  can be obtained from these
reports.  The product of  this value  and  the total
resident population of an  area is an estimate of the
local exposed  population.  Similar national level data
are also provided on the basis  of age and sex in Food
and Nutrient Intakes of  Individuals in  1 Day in the
United States  (USDA 1980). In  addition, the  U.S.
Food  and  Drug Administration (FDA)  can  be
contacted for data concerning  daily intakes of various
food  items.  Such  data  have  been  compiled  for the
FDA Total Diet Study (Pennington  1983).
   Table A-2.   U.S. Home Fruit and Vegetable Garden
              Use, 1977
                 Percent of
                 households  Household   Percent of
                   with     size (no. of   total U.S.
Urbanization
Urban
Rural non-farm
Rural farm
gardens
43
41
84
persons)
3.17
3.44
3.86
population
32
9
3
    Source: USEPA 1980.
                                                 119

-------
Monitoring data may indicate  whether fish  and game
are contaminated  in the  subject area.  One can
estimate the fishing  population by contacting the local
agency  responsible  for  issuing  fishing licenses;  this
may be the state fish and  game  commission or the
state department of  natural resources. Since there are
2.69 persons in the average household (Bureau of the
Census 1986)  one  can  estimate the actual exposed
population  by  multiplying  2.69 by  the  number of
licensed hunters or fishermen  in  the area.

7.4.4 Populations Exposed through Soil
Exposure to contaminated soil constitutes a potential
exposure route  for workers  or  children  playing
outdoors. Neighborhood  children playing  at the  site
can be exposed to  high levels of contaminants. Soil-
related  exposure in such  cases  would be through
direct  dermal  contact  with the contaminated  soil.
Another potentially significant,  but infrequently
encountered, exposure  mechanism  involves  children
who eat dirt; this eating behavior, known as pica, may
lead to their  actually  ingesting contaminated soil.
Hand-to-mouth contact during  normal  play  is  a
more common means of ingesting  soil, however. For
any site located near residential areas, the degree of
accessibility  to children should be considered.  Bureau
of the  Census data can  be  used  as described in
Section A-1.4.1  to  estimate the  number of  local
children who may have access to the site.

In addition,  workers conducting activities  at the  site
(other  than remediation) may have direct  dermal
contact with  contaminated  soils.  Section  A-1.4.1
provides general guidance to  identify and  enumerate
exposed worker populations.

1.5 Population  Characterization

After exposed  populations  have been identified  and
enumerated, they can be characterized  by age  and
sex factors.  The  physiological  parameters that
determine the  dose received  per  a  given level of
exposure (e.g., breathing rate, skin surface area,  and
ingestion rate)  are  often age-  or  sex-specific. Also,
from a toxicity  standpoint, subpopulations defined by
age or sex,  such  as  the  elderly  or women  of
childbearing age,  may be especially susceptible to a
chemical substance. Superfund  studies will generally
use average values, but  by  characterizing exposed
populations, one can determine  exposure distributions
within  the population at large and delineate  specific
high-risk  subpopulations.

The Census  Publication series General  Population
Characteristics  (PC80-1-B)  cites  figures   for the
age and sex structure of the  population residing in a
specific area. Separate volumes  for each state
contain age and sex  breakdowns  at the level of
county subdivisions  and  small towns. If more detail is
required, the Census Bureau microfiches  containing
this  information  at  the Census  tract level (only
available by SMSAs).

In the case of exposure resulting  from ingestion of
food,  the food consumption surveys of the USDA
(1983a-e) record  age and sex data for the sampled
population. These data are contained in five separate
regional reports;  the appropriate one  should be
consulted.

In lieu  of obtaining  site-specific data,  one can  use
the population  characteristics of the U.S. as a whole,
provided  in the  yearly Statistical Abstract of the
United  States (for  example,  see  Bureau  of the
Census 1986),  to  approximate  the population
distribution in the area of concern.

1.6 Activity Analysis

Activities engaged  in by  members of a given
population or  subpopulation can  dramatically affect
the level of  human exposure to  environmental
contaminants.  For example,  persons whose  lifestyle
or employment involves frequent  strenuous  activity
will inhale larger volumes of air per unit time than will
those living a  less strenuous life, and will experience
a higher level of exposure to  airborne contaminants.

Activity  analysis allows   refinement of  certain
parameters  used  in the  calculation of  exposure,
including:

x* Inhalation rate;

x   Frequency of exposure; and

xx Duration of exposure.

The  procedure  for  integrating  activity-related
inhalation, frequency, and  duration  data into the
exposure assessment  process is detailed in the
following chapter.
                                                  120

-------
                                             Chapter 2
                             Exposure Calculation and Integration
This chapter provides  guidance  for calculating  and
integrating exposures  to  all  populations affected by
the various exposure routes  associated with a given
uncontrolled hazardous waste site.  Specifically,
guidance is provided to estimate exposure from:


1.  Inhalation
   a.   Ambient air
   b.   Indoor air (contaminants  released during
       showering)

2. Dermal contact
   a.   Water (swimming)
   b.   Soil

3.  Ingestion
   a.   Food
   b.   Water
   c.   Soil
This analysis is based  on the  results of all previous
analyses, and  is the  final  stage  of  the  exposure
assessment. This  guidance is complete; no additional
documentation is required to finish the analysis.

Integrated exposure  analysis is  conducted for  only
those  contaminants  having  complete exposure
pathways (i.e., those contaminants  that are released
and migrate  from  the site and that do contact human
populations).  Therefore, no  screening  evaluation  is
included  in the exposure  integration  process. While
calculating the  exposure incurred  is traditionally the
final step in  the  quantitative exposure  assessment
process,  it can also be  viewed as a component of the
human health risk assessment.  Therefore,  the
material detailed in this chapter is  also discussed  in
the Superfund Public Health Evaluation Manual
(USEPA1985).

Exposure is defined  as  the  amount of pollutant
contacting   body  boundaries (skin,  lungs,  or
gastrointestinal tract). Exposure calculation considers
how often  populations come  into  contact with
contaminants in  specific environmental media, the
mode  of  such  contact,  and  the  amount of
contaminated medium  that contacts the  internal  or
external  body surface  during each  exposure  event.
The goal of this analysis is to quantify the amount of
contaminant contacted within a given time interval.

Short-term and long-term  exposures  are calculated
in the same manner. First, for each exposure scenario
under consideration,  an exposure per event is
estimated. This exposure value quantifies the  amount
of contaminant  contacted during each exposure
event,  with  "event"   being  defined differently
depending on the  nature of the scenario  under
consideration  (e.g.,  each day spent swimming in  a
contaminated  river  is a single  swimming  exposure
event, each day's  inhalation of contaminated air is an
inhalation exposure  event).  Event-based  exposure
estimates take into account the  concentration of
contaminant  in the medium through  which exposure
occurs,  the  rate  of  contact with such  media
(inhalation rate, ingestion rate, etc.), and the duration
of each event.

The  analyst  can convert event-based exposure
values to final exposure values by multiplying the
exposure  per  event by the  frequency  of exposure
events over  the timeframe being considered.  Short-
term exposure is  based on the  number of exposure
events that  occur during  the short-term timeframe
(10 to  90 days), while  long-term  exposures are
based on  the  number of events that occur within an
assumed  70-year lifetime. The 70-year  assumed
average  lifetime  is  traditionally used  in  exposure
assessments,  and it provides a conservative  upper
bound  of lifetime  exposure. Certain  exposure
scenarios, however, may  only  apply to  short-term
exposure. Whenever practical,  the analyst  should
strive to determine the timeframe over which a  given
exposure  pathway would be expected to affect the
exposed  population. Once  determined, the timeframe
will  indicate  whether that  pathway  should  be
evaluated  on a short- or a long-term basis.

Exposure  estimates  are expressed in  terms of  mass
of contaminant/unit of body mass/day by  dividing daily
exposure  by the  value for total body  mass of an
average  individual in the exposed  population. For
Superfund studies, an average adult body mass of 70
kg will  usually be adequate for this  conversion. In
cases where exposure to specified subpopulations
must  be evaluated,  values for  other than  average
                                                 121

-------
adults  may  be required. Consult  Anderson  et al.
(1984) to obtain  alternative  body  mass  values.
Similarly,  average values for activity-related
parameters (e.g.,  inhalation  rate) generally will be
adequate for Superfund site evaluations.  For special
situations  and  detailed exposure  analysis,  analysts
can refer to the discussion of activity data in  Freed et
al. (1985). An exposure factors handbook is  currently
under development  (USEPA 1987), and  the analyst
performing exposure assessments after publication of
this  manual  should consult that document for the
most up-to-date  exposure  factors.

The following sections  address the  exposure
calculation  process  specific  to  each  exposure
mechanism.  Data  management  forms  designed to
organize and  tabulate  the data in  the exposure
calculation process  are presented in  Appendix C.

2.1  Inhalation Exposure

Inhalation exposure per event is estimated based on
the  hours  per event, the inhalation  rate of the
exposed  individual  during  the  event,  and  the
concentration of contaminant in the  air breathed. The
formula for calculating event-based exposure  is the
following:
    IEX = t xIxC(X)xF-BW-2.56xlO\.i,  .
          e                             lifetime
where

  IEX    =  estimated  inhalation   exposure,
             (mg/kg/day).
     te    =  duration  of  an  exposure event,
             (hours/event).
     I    =  average  inhalation rate  of exposed
             persons, (m3/hr).
 C(X)    =  contaminant  air  concentration
             throughout the  exposure  period,
             (milligrams/m3).
   BW    =  average adult body weight, (70 kg).
     F    =  frequency of  exposure  event,
             (number/lifetime).

Short-term exposure is  calculated using the  short-
term  contaminant  air concentration,  and  long-term
exposure  is  based on the long-term concentration.

Inhalation exposures  are  keyed to geographic
locations  delineated  during  the  Environmental Fate
Analysis. Ambient concentration is generally assumed
to be  homogeneous throughout a limited area or
sector (within an  isopleth); however, this assumption
is not always well-founded.  Numerous  studies have
shown that there can be marked differences in  indoor
and  outdoor  concentrations  of pollutants  (Budiansky
 1980,  Moschandreas  et  al.  1978) or among
microenvironments in the  same area (Ott  1981). To
account  for  these differences when  calculating
exposure, several  investigations have coined the term
"microenvironment,"   which  refers  to a  type of
physical setting  where  concentrations  of  pollutants
can be expected to be similar. For Superfund studies,
it is usually unnecessary to disaggregate analysis on
a microenvironment basis. Instead, it can generally be
assumed that contaminants have been  present  long
enough for  indoor-to-outdoor  concentrations to
have reached equilibrium.

To calculate exposure duration, the  analyst considers
the amount of time exposed persons actually spend in
the contaminated  area. For example, if a site is in a
residential  area,  one  can conservatively  estimate
exposure by  assuming that all  residents  spend  the
entire 24-hour day within the contaminated  zone.  If a
site is located in an  industrialized  area,  it may be
more appropriate to  base  duration on  an 8-hour
workday,  if it  can  be reasonably  assumed  that
workers  do  not also  live in  the immediate
industrialized  area. Such factors must  be  evaluated
on a  case-by-case basis.  For inhalation  exposure,
frequency is assumed to be daily.

For a general application, use an average adult  value
for inhalation rate. An  example of  an  adult average
derived from experimental results (USEPA 1981) is an
inhalation  rate of  1 m3/hour. This value can be used
to conservatively estimate exposure  regardless of
microenvironments or activity.

Generating time-weighted  average inhalation  rates
provides  a more  precise estimate  of inhalation  rate.
This  calculation  is  based on  microenvironment-
related data and activity stress levels/ventilation rates
associated with  the  individual microenvironment. If
this level of detail is  warranted,  the inhalation  rates
presented in  Table A-3 can  be used.  Freed et al.
(1985) cite  directions  for  developing  time-weighted
average inhalation rates.

To calculate ambient inhalation exposure, one should
obtain contaminant air concentration values from the
results  of the environmental fate  analysis. In one
case, however, concentration values will have  to be
calculated in the exposure  integration  stage  of the
exposure assessment.  Persons showering  or bathing
in potable water contaminated with toxics may be
exposed  through  inhalation  if the  contaminants  are
volatile. This is especially true of showering, since the
high  turbulence, combined  with  the  elevated
temperature  of the  shower  water, can  produce a
significant release of volatile components.

Various approaches are  available  to  estimate
contaminant  concentrations  indoors. These
approaches depend on a number of factors, including
the room air volume,  air exchange and  mixing factors,
contaminant concentration  in the water, the amount of
water used, and the manner in which  a contaminant
                                                 122

-------
  Table A-3.   Summary of Human inhalation Rates for
             Men, Women and Children by Activity Level
             (m3/hour)a

               Resting"    Light0   Moderated  Heavy6
Adult male
Adult female
Average adult'
Child, age 6
Child, age 10
0.6
0.6
0.6
0.4
0.4
1.3
1.3
1.3
1.4
1.7
2.6
2.4
2.6
2.1
3.3
7.1
4.9
6.0
2.4
4.2
  aValues of inhalation rates for males, females, and children
   presented in this table represent the midpoint of ranges of
   values reported for each activity level in Anderson et al. (1984)
  "includes watching television, reading, and sleeping.
  Includes most domestic work, attending to personal needs and
   care, hobbies, and conducting minor indoor repairs and home
   improvements.
  dlncludes heavy indoor cleanup and performance of major
   indoor repairs and alterations and climbing stairs.
  Includes vigorous physical exercise and climbing stairs carrying
   a load.
  'Derived by taking the mean of the adult male and adult female
   values for each activity level. A representative  24-hour
   breathing rate for an average adult is 1.1 m3/hour. This value
   is based on the assumption that the average adult spends 93.2
   percent of the time at the light/resting level of activity, 5.8
   percent at a moderate level of activity, and 0.9 percent at a
   heavy level of activity.  Values for the  percent of time spent at
   each activity level are from Freed et al.  (1985).


is released  into  room air  (instantaneously,  con-
tinuously,  time-dependent). If showering/bathing
exposure estimation  is required  for  a Superfund
exposure assessment,  the analyst is referred to
Versar (1984) for a detailed discussion  of techniques
to estimate indoor air contaminant concentration.  For
both showers and  baths, the analyst  should assume a
continuous contaminant release during the  bathing/
showering period. Values for the other variable factors
mentioned  above can be  obtained  from  Versar
(1985).

To evaluate inhalation exposure to  contaminants
volatilizing  from potable water  while showering,  the
analyst should  again assume frequency to be daily.
Each shower is assumed to last  15 minutes.

Inhalation exposure  to  swimmers can  be based on
monitored  or  estimated ambient air concentrations
above a  contaminated water body.  To estimate
concentrations, calculate the  rate of volatilization of
the contaminant from the  water body  and use  this
value as  the  input  to  a "box  model"  air migration
model. The dynamic release  rate can  be  calculated
using Equations 2-10, 2-15, 2-16,  and 2-17.  The
recommended  air model  is BOXMOD (in EPA's
GEMS system,  see Chapter 3).

2.2 Dermal Exposure

Dermal exposure is  determined by the  concentration
of hazardous  substance in a contaminated  medium
that is contacted, the extent of contact (i.e., the body
surface  area contacted),  and  the  duration  of such
contact.  For exposure to contaminated water, dermal
exposure per event is calculated as follows:
                                                         DEX=tPxAVxCxPCxFxl liter/1000 cm£
                                                                         4_days_

                                                                           lifetime
                                                                                                   (A-2)
where

 DEX

     te
   AV
     C

   PC

     F

   BW
           =  estimated  dermal   exposure,
              (mg/kg/day).
           =  duration of exposure, (hours/event).
           =  skin surface area available for contact,
              (cm2).
           =  contaminant concentration  in water,
              (mg/liter).
           =  dermal  permeability  constant for the
              subject contaminant, (cm/hr).
           =  frequency  of exposure events  per
              lifetime.
           =  average adult body  weight, (70 kg).
The term 1  liter/1,000 cm3 is a volumetric conversion
constant for water.

When  possible,  it is important to consider the degree
to which a given contaminant is actually able to enter
the body. Some compounds  will not readily  penetrate
the skin, while others  may do so at a rapid  rate. The
above equation can  only be used in cases where
dermal permeability constants for the contaminant(s)
of  concern are known. Table A-4 lists dermal
permeability  constants for selected compounds.  For
many  compounds, however,  dermal permeability
constants will not  be available. In such  cases,  the
analyst must assume that  contaminants are  carried
through the  skin  as  a  solute in water  which  is
absorbed (rather than being preferentially  absorbed
independently of the water),  and that the  contaminant
concentration in  the water being absorbed is equal to
the ambient  concentration. Thus, the permeation rate
of water across  the skin  boundary is  assumed to  be
the factor controlling the contaminant absorption rate.
Short-term dermal  exposure  per event is calculated
using  the  short-term  contaminant concentrations  in
water  or  soil, and  long-term exposure  is based  on
the long-term contaminant concentrations.

The local recreation department  may have  detailed
data quantifying  the duration  and frequency  of water
use for  swimming.  When such locale-specific data
are  not available,  the following  national  average
figures, based on  data from the Bureau of Outdoor
Recreation (USDOI  1973), can be applied:

*   Frequency of exposure =  7 days/year.
                                                   123

-------
Table A-4.   Permeability Constants for Various Compounds
                               Permeability
Compound
SURFACTANTS
Decanoic acid
Dodecanoic acid
Tetradecanoic acid
Hexadecanoic acid
Octadecanoic acid
Sodium dodecyl sulfate
Sodium dodecyl isothionate
Sodium p-1 -dodecyl
benzenesulphonate
Sodium laurate
IONS
Aluminum
Potassium
Bromide
Palmitate
Laurate
DRUGS
Methotrexate
Benzoyl peroxide
Estradiol
Amphetamine
Ouabain
Burimamide
Metramide
Cimetidine
PHENOLS
Resorcinol
p-Nitrophenol
n-Nitrophenol
Phenol
Methylhydroxybenzoate
n-Cresol
o-Cresol
p-Cresol
beta-Naphthol
o-Chlorophenol
p-Ethylphenol
3,4-Xylenol
p-Bromophenol
p-Chlorophenol
Thymol
Chlorocresol

constant3 (cm//hr)

1 .OOE-03
2.00E-03
6.00E-04
1 .20E-05
6.00E-06
2.00E-03
5.40E-05
6.00E-06
1 .OOE-03

7.20E-06
6.70E-05
1 .80E-05
4.20E-05
3.00E-03

6.00E-10
5.10E-07
3.90E-03
1 .40E-05
3.90E-06
1 .70E-07
1.10E-07
3.30E-07

2.40E-03
5.58E-02
5.58E-02
8.22E-02
9.12E-02
1.52E-01
1.57E-01
1.75E-01
2.79E-01
3.31 E-01
3.49E-01
3.60E-01
3.60E-01
3.60E-01
5.28E-01
5.50E-01

Reference

Howes 1975
Howes 1975
Howes 1975
Howes 1975
Howes 1975
Howes 1975
Howes 1975
Howes 1975
Tregear 1966

Tregear 1966
Tregear 1966
Tregear 1966
Tregear 1966
Tregear 1966

McCullough et al. 1976
Nachtet at. 1981
Galey et al. 1976
Galey et al. 1976
Sutton 1973
Sutton 1973
Sutton 1973
Sutton 1973

Roberts et al. 1977
Roberts et al. 1977
Roberts et al. 1977
Roberts et al. 1977
Roberts et al. 1977
Roberts et al. 1977
Roberts et al. 1977
Roberts et at. 1977
Roberts et at. 1977
Roberts et al. 1977
Roberts et at. 1977
Roberts et at. 1 977
Roberts et at. 1977
Roberts et al. 1977
Roberts et al. 1977
Roberts et al. 1977
(Continued)
                              124

-------
Table A4.   (Continued)
Compound
PHENOLS (Continued)
Chloroxylenol
1 ,4,6-Trichlorophenol
2,4-Dichlorophenol
STEROIDS
Progesterone
Pregnenolone
Hydroxypregnenolone
Hydroxyprogesterone
Cortexone
Testosterone
Cortexolone
Corticosterone
Cortisone
Hydrocortisone
Aldosterone
Estrone
Estradiol
Estriol
Dihydroepiandrosteroneb
Dihydrotestosteroneb
ALCOHOLS
Methanol
Ethanol
Propanol
Butanol
Pentanol
Hexanol
Heptanol
Octanol
Nonanol
Decanol
GLYCOL ETHERS
2-Methoxyethanol
2-Ethoxyethanol
2-Ethoxyethanol acetate
2-n-Butoxyethanol
1 -Methoxypropan-2-OI
2-(2-Methoxyethoxy)ethanol
2-(2-Ethoxyethoxy)ethanol
2-(2-n-Butoxyethoxy)ethanol

Permeability
constant3 (cm//hr)

5.90E-01
5.94E-01
6.01 E-01

1 .50E-03
1 .50E-03
6.00E-04
6.00E-04
4.50E-04
4.00E-04
7.50E-05
6.00E-05
1 .OOE-05
3.00E-06
3.00E-06
3.60E-03
3.00E-04
4.00E-05
1 .70E-04
3.90E-04

5.00E-04
8.00E-04
1 .20E-03
2.50E-03
6.00E-03
1 .30E-02
3.20E-02
520E-02
6.00E-02
8.00E-02

2.89E-03
8.42E-04
8.07E-04
2.14E-04
1 .25E-03
2.06E-04
1 .32E-04
3.60E-05

Reference

Roberts et al. 1977
Roberts et al. 1977
Roberts et al. 1977

Scheuplein et al. 1969
Scheuplein et al. 1969
Scheuplein et al. 1969
Scheuplein et al. 1969
Scheupliln et al. 1969
Scheuplein et al. 1969
Scheuplein et al. 1969
Scheuplein et al. 1969
Scheuplein et al. 1969
Scheuplein et al. 1969
Scheuplein et al. 1969
Scheuplein et al. 1969
Scheuplein et al. 1969
Scheuplein et al. 1969
Schaefer et al. 1 982
Schaefer et al. 1982

Scheuplein and Blank 1971
Scheuplein and Blank 1971
Scheuplein and Blank 1971
Scheuplein and Blank 1971
Scheuplein and Blank 1971
Scheuplein and Blank 1971
Scheuplein and Blank 1971
Scheuplein and Blank 1971
Scheuplein and Blank 1971
Scheuplein and Blank 1971

Dugard et al. 1984
Dugard et al. 1984
Dugard et al. 1984
Dugard et al. 1984
Dugard et al. 1984
Dugard et al. 1984
Dugard et al. 1984
Dugard et al. 1984
(Continued)
                               125

-------
Table A4.    (Continued)
Compound
PESTICIDES0
Azodrin
Ethion
Guthion
Malathion
Parathion
Baygon
Carbaryl
Aldrin
Dreldrin
Lindane
24-D
Diquat
OTHER
Water
Ethylbenzene
Styrene
Toluene
Anilrne"

N-nitrosodiethanolamine
Ethyl ether
2-Butanone
1-Butanol
2-Ethoxyethanol
2,3-Butanediol
Benzeneb

Desoximetasoneb'd
Linoleic acidb'd
Dithranolb
Theophyllineb
Caffeineb
8-Methoxypsoralenb.d
p-Butoxyphenylacethydroxamic
acidbd
Triacetoxyanthraceneb
Heparinb'd
Carbon disulfideb

Permeability
constanta (cm//hr)

9.80E-04
2.20E-04
1 .06E-03
5.50E-04
650E-04
1 .31 E-03
4.90E-03
5.20E-04
5.10E-04
6.20E-04
3.90E-04
2.00E-05

8.00E-04
1 .OOE-03
6.00E-04
9.00E-04
2.00E-02

5.50E-05
1.70E + 01
5.00E + 00
4.00E + 00
3.00E-01
5.00E-02
4.10E-01

3.40E-05
1 .60E-05
2.10E-04
2.50E-05
3.30E-04
9.90E-04
9.80E-05
5.80E-05
8.20E-05
5.50E-02

Reference

Feldman and Maibach 1974
Feldman and Maibach 1974
Feldman and Maibach 1974
Feldman and Maibach 1974
Feldman and Maibach 1974
Feldman and Maibach 1974
Feldman and Maibach 1974
Feldman and Maibach 1974
Feldman and Maibach 1974
Feldman and Maibach 1974
Feldman and Maibach 1974
Feldman and Maibach 1974

Blank et al. 1984
Dutiewicz and Tyras 1967
Dutiewicz and Tyras 1968
Dutiewicz and Tyras 1968
Baranowska-Dutkiewicz
1982
Bronaugh et al. 1981
Scheuplein and Blank 1971
Scheuplein and Blank 1971
Scheuplein and Blank 1971
Scheuplein and Blank 1971
Scheuplein and Blank 1971
Baranowska-Dutkiewicz
1982
Schaefer et al. 1982
Schaefer et al. 1982
Schaefer et al. 1982
Schaefer et al. 1982
Schaefer et al. 1982
Schaefer et al. 1982
Schaefer et al. 1982
Schaefer et al. 1982
Schaefer et al. 1982
Baranowska-Dutkrewicz
1982
  a  Permeability constants are for contaminants as a dilute solution  in water, except
    as noted.
  b  Calculated permeability constant, subject to error.
  c  Permeability constants are for contaminants in acetone. These values should not
    be  used for  dermal exposure due  to contact with  contaminated water. These
    values should be used for dermal exposure to  pure wastes.
  d  Permeability constants are for contaminants in gel. These values should not be
    used for dermal exposure due to contact with  contaminated water. These values
    may be used for dermal exposure to pure wastes.
                                     126

-------
ss   Duration of exposure  = 2.6 hours/day.

Dermal  absorption of waterborne contaminants may
be  a significant  exposure  route.  The  factors that
influence dermal absorption of chemicals are: (1) the
nature  of  the  compound  (molecular weight,
lipophilicity), (2) the presence  of  other compounds
that might facilitate passage of a chemical though the
skin  (e.g.,  chelating  or complexing  agents), and (3)
the permeability  of  the  skin. Generally only  lipid-
soluble, non-ionized compounds are absorbed
significantly  through the  skin. Also,  the skin is
normally  permeable only  to compounds  whose
molecular weights are less than 500  Daltons. The
permeability of the skin  to  larger  molecular weight
compounds and to less  lipophilic compounds can  be
increased  when corrosive agents such as  acids are
present  or when there  are skin abrasions.  For
waterborne chemicals, exposure through the skin is
almost directly proportional to concentration.

Brown,  Bishop, and  Rowan  (1984)  recently reported
that when compared with  ingestion, dermal  absorption
of volatile  organic contaminants  in drinking water
accounted for approximately 29 to  91  percent of the
total dose incurred, with the average being about 64
percent.  The  dermal  exposure  route  becomes
especially pertinent when organic  contaminants are
present  in very dilute aqueous solution, as may often
be the case at Super-fund sites. In certain cases, then,
dermal exposure to contaminants contained  in ground
or surface  water may actually overshadow ingestion
exposure.

When persons become  exposed to contaminants in
drinking water, the dermal  exposure associated with
bathing  or showering should also be considered. One
can  use the  same  approach  to  assess
bathing/showering  as  was  used for swimming.
Generally,  an  average  frequency of one bath or
shower  per day can be assumed, and each  event can
be estimated to last 15 minutes.

For  swimming or bathing exposure, the  surface area
available for dermal exposure is assumed to equal the
total amount of human  skin surface  area. Average
availability values are given below for adults and
children. If the exposed population is not  separated by
age groups, both availability  values should  be used to
represent a general range of exposure for the total
swimming  or  bathing population.  Both  availability
figures  cited below are taken from Anderson  et al.
(1984):

• Average adult  (male  and female, 20-30 yrs) =
     18,150 cm2.

•  Average  child  (male  and female,  3-12  yrs) =
    9,400 cm2.
Direct dermal  contact with contaminants present in
soil is calculated as follows:
   DEX=ClXAVxDAxF
                   t  days
                   lifetime
                                           (A-3)
where

  DEX


   AV

   DA
     F

   BW
            dermal exposure, (mg/kg/day).
            weight fraction  of  chemical  substance
             in soil, (unitless).
            skin  surface area available for contact,
             (cm*).
            dust adherence,  (mg/cm ).
            frequency  of exposure  events  per
             lifetime.
             average adult body weight, (70 kg).
Values for contaminant  weight  fraction in the
contaminated  soil will  be available  from the site
survey. Skin  surface  availability depends on the
nature of activity being  conducted, and can vary for a
given activity depending on the season  of the year.
Anderson  et al. (1984) provide  data on  skin surface
areas  of different  parts of the  body  for adults and
children. Based on a projection  of the  type of activity
at the site and the  age of the exposed  population
(i.e., workers or children), the data in Anderson et  al.
can  be used  to develop skin surface estimates  for
use in estimating direct dermal exposure.
Data on dust adherence to skin (DA) are limited,
although the following experimental values for (soil-
related) dust adherence were reported by the Toxic
Substance Control Commission of the State  of
Michigan (Harger 1979):

es   Commercial  potting soil adheres to hands at 1.45
    mg/cm2.

*   Dust of the clay mineral kaolin adheres to  hands
    at 2.77 mg/cm2.

The degree to  which  these values represent dust
adherence at  any given site  is uncertain,  as  such
adherence will depend on a  variety  of  site-specific
factors. Therefore, instead of selecting  one of the
above  values to  estimate direct dermal exposure, it  is
suggested that  the  analyst use  both values  and
generate  an exposure  range.  The lifetime  frequency
of direct dermal  exposure will  also vary considerably
and will depend  on the nature of the site, its ease  of
access, and a variety  of other factors.  Thus, contact
frequency should  be   estimated  on  a case-by-case
basis,  based on knowledge of  the site and  its
environs.
                                                  127

-------
 Note that  this approach is  conservative  in that it
 assumes that all of the contaminant  adsorbed to the
 soil (dust) particles is available for absorption through
 the  skin. In fact, only  a  percentage  of the  total
 adsorbed contaminant mass may actually be available
 for such absorption, as some  percentage may remain
 bound to the soil particle.

 The  site  survey  will provide values for  the
 contaminant weight  fraction in the contaminated soil.
 Skin surface availability depends on the  nature of the
 activity being  conducted, and can vary for a given
 activity depending on  the  season of  the year.
 Anderson  et al. (1984) provide data  on  skin surface
 areas of different parts  of the body for adults  and
 children.  Based on a projection of the type of activity
 at the site  and the  age of the exposed  population
 (i.e., workers  or children),  one can  use the data in
 Anderson  et al. to develop  skin surface  estimates for
 use in estimating direct dermal exposure.

 The  lifetime frequency of direct dermal exposure will
 vary considerably and will  depend on the nature of
 the  site,  its  ease  of access,  and  other factors.
 Contact frequency should  be  estimated on a case-
 by-case basis, based on knowledge  of  the site  and
 its environs.
2.3  Ingestion  Exposure

2.3. J  Food/Soil
Food ingestion  exposure is estimated  as the  product
of contaminant  concentration in  the  food consumed
and  the  amount  of food  consumed per  day.
Frequency is daily for foods that  are a regular part of
the diet. For recreationally caught fish, frequency can
be estimated based on the seasonal nature of fishing
involved.

USDA source materials listed in  Section A-1.4.3 are
also  useful in quantifying the amount of contaminated
food  ingested. The Food  Consumption of Households
report series  provides data quantifying the amount of
various food categories  consumed  by  households on
a seasonal basis. Similar data are presented  in food
and  Nutrient Intakes of Individuals in 1 Day in the
United States. The first source can be used to derive
estimates  of the amount of various foods consumed
by the  overall exposed population by  applying
seasonal  percentage use values to local population
census data.  The second source  is  used  in
subpopulation analyses  by applying  sex-  and  age-
specific consumption values to census  data  for the
exposed population.

Consumption of fish caught in  contaminated  waters
may  be an  important  ingestion  route, since  certain
contaminants of concern tend to  biomagnify in the
food  chain.  This phenomenon  results in  tissue
concentrations  of  contaminants  in  predator fish
exhibiting  levels that  greatly  exceed the  ambient
concentration in  the waterbody. An average daily fish
ingestion  rate for the U.S.  population  has been
estimated  as  6.5  grams  per day  (USEPA 1980b).
Persons for whom fish constitutes  a major portion of
the overall diet  may consume up  to 124  grams per
day  (USDA 1980).  A West Coast  study of
consumption of fish caught in contaminated waters by
sport fishermen (Puffer et al. 1981) reports a median
fish ingestion rate  of 37 grams/day.  This  report  also
lists a maximum rate of 225 grams/day.

Ingestion  exposure estimates are calculated in the
same manner, regardless  of the type of food
ingested.  Multiplication  of the contaminant
concentration in  the ingested food by the  amount of
contaminated food  ingested  per  day yields exposure
per day.

Children may ingest soil during play both inadvertently
and  intentionally (pica   behavior).  In  those
assessments where the exposed populations analysis
has found that children may have access to areas of
contaminated  soil, this  exposure route  should  be
evaluated.  Data  quantifying the amount of  soil
ingested   by  children  are conflicting  and  vary
considerably. For  example,  Calabrese  et  al.  (1987)
report that estimates range from a low of  10 mg/day
(for 2-year-old children) to a high  of 10,000 mg/day
(for  1.5- to 3.5-year-old  children).  Within  this
range, reasonable typical values can be identified  and
associated with  various age groups, if desired.  For
studies warranting  such  detail, the daily soil ingestion
rate values presented  in Table A-5 can be used. For
studies  that do not require such detail, one can  use
an overall average  soil ingestion value of 100 mg/day.

       Table A-5.   Typical Daily Soil Ingestion
                  Rates for Children by Age
                  GrouP    Soil ingestion range
       	Age	(mg/day)	
       0-9 months
       9-18 months
       1.5 - 3.5 years
       3.5 - 5 years
       5-18 years
0
50

200

50

10
       Source: Calabrese et al. 1987.
2.3.2 Water
Event-based  water Ingestion  exposure equals  the
daily total amount of contaminant ingested from either
surface  or  ground  waters affected  by  the Superfund
site.  This exposure is  determined by the contaminant
concentration in  the water and  the amount of water
ingested per day. On average, an adult ingestion
coefficient of 2.0 liters per day  (USEPA 1980b) can
be used for  Superfund  site  analyses.  Frequency of
drinking  water exposure is daily.
                                                 128

-------
When  contaminated  surface waters  are  used
recreationally,  it  may  be appropriate  to  estimate
exposure that results  from  inadvertently ingesting
contaminated water while swimming. For this analysis,
the same  values  for event frequency and duration
previously  presented in Section  A-2.2  should be
used.  In  addition,  to estimate  the  amount of
contaminated water ingested  per event, an assumed
value of 50  ml per hour can be used.


2.4  Exposure Integration

The  final step in  the exposure assessment  process
for uncontrolled hazardous  waste sites  is the
integration of  all exposures experienced by individual
exposed populations.  This simply  involves organizing
the results of the previous  analyses  to  total all
exposures to a given  hazardous  substance
experienced by each  population  segment.  Because
different chemicals  exhibit  different toxicological
properties, exposures  to each contaminant of concern
are considered separately.  Note that in some cases,
individual populations may be exposed  to a given
chemical in a  particular medium through more than
one  exposure  scenario.  For  example, persons who
swim in contaminated waters  may obtain their
drinking water  from  the  same  contaminated
waterbody. In such  cases, the dermal exposure
experienced while swimming can  be added  to that
experienced during bathing or showering  to  generate
an overall dermal  exposure value for that population
segment. The  data  management forms  supplied  in
Appendix C are designed to help organize the results
of exposure calculation and integration.
                                                129

-------

-------
                                            Chapter 3
                                    Appendix A  References
Anderson  E,  Browne  N,  Duletsky  S,  et  al.  1984.
  Development of statistical distribution or  ranges of
  standard factors  used  in exposure assessments.
  Revised draft final  report. Washington, DC: Office
  of Health and Environmental Assessment, U.S.
  Environmental Protection Agency. Contract  No.
  68-02-3510.

Baranowska-Dutkiewicz  B.  1982. Skin absorption of
  aniline from  aqueous solutions in  man. Toxicology
  Letters 10:367-72.

Blank IH,  Moloney  J,  Ernslie AG,  et al. 1984.  The
  diffusion of water across the stratum corneum as a
  function of its water  content.  The Journal of
  Investigative  Dermatology.  82:188-194.

Bronaugh  R, Condon E, Scheuplein RJ. 1981.  The
  effect of cosmetic  vehicles  on  the penetration of
  n-nitrosodiethanolamine through  excised human
  skin.  Journal of  Investigative Dermatology  76:94-
  6.

Brown HS, Bishop DR, Rowan CA. 1984. The  role of
  skin  absorption as  a  route of exposure for volatile
  organic compounds  (VOCs)  in  drinking  water.
  Amer. J. of Public Health. 74 (5).

Budiansky S.  1980. Indoor air pollution. Environ.  Sci.
  Technol. 14(g): 1023-1027.

Bureau of the Census. 1986. Statistical abstract of the
  United States: 1986  (106th edition). Washington,
  DC:  U.S.  Department of  Commerce, U.S.
  Government Printing Office.

Calabrese EJ,  Kostecki PT, Gilbert CE. 1987. How
  much soil do  children  eat?  An emerging
  consideration  for environmental  health  risk
  assessment. Paper  submitted to  Comments in
  Toxicology.

Dugard PH, Walker M, Mawdsley SJ, Scott  RC. 1984.
  Absorption  of some glycol ethers through  human
  skin   in vitro. Environment Health  Perspectives
  57: 193-97.
Dutkiewicz  T,  Tyras H.  1967. A  study of skin
  absorption of ethylbenzene in  man.  British Journal
  Ind.  Med. 24:330-32.

Dutkiewicz  T,  Tyras H. 1968.  Skin  absorption of
  toluene, styrene, and xylene  by man. Department
  of Toxicological Chemistry and Industrial
  Toxicology, Medical Academy,  Lodz, Poland.

Feldman RJ,  Maibach HI. 1974.  Percutaneous
  penetration of some  pesticides and  herbicides in
  man. Toxicology and  Applied Pharmacology
  28: 126-32.

Freed JR, Chambers T, Christie WN, Carpenter CE.
  1985. Methods for assessing  exposure to chemical
  substances:  volume  2  - methods  for assessing
  exposure to  chemical  substances  in the ambient
  environment. Washington, DC: U.S. Environmental
  Protection Agency, Office of Toxic Substances,
  Exposure Evaluation Division.  EPA 560/5-83-
  015.

Galey WR, Londdale HK, Nacht S. 1976. The in vitro
  permeability of skin and buccal mucosa to selected
  drugs  and tritiated water. J. Invest. Dermatol.
  67:713-717.

Harger JRE. 1979. A model for  the determination of
  an action level for removal of curene contaminated
  soil. Memorandum to P.S. Cole, Executive Director.
  Lansing, Ml: Toxic Substance  Control Commission.

Howes D.  1975.  The percutaneous  absorption of
  some anionic surfactants. J.  Soc. Cosmet. Chem.
  26:47-63.

McCullough JL,  Snyder DS, Weinstein GD, Friedland
  A,  Stein  B. 1976.  Factors affecting  human
  percutaneous penetration of  methotrexate and its
  analogues in vitro. J. Invest. Dermatol. 66:103-
  107.

Moschandreas DJ, Stark JWC, McFadden JF, Morse
  SS.  1978. Indoor pollution in  the residential
  environment  - vols.  I and II. Washington, DC:
                                                131

-------
  Office of Air Quality  Planning and Standards, U.S.
  Environmental Protection Agency.

Nacht  S,  Yeung D,  Beasley JN Jr., Anfo  MD,
  Maibach HI. 1981. Benzoyl  peroxide: percutaneous
  penetration  and  metabolic  disposition.  Journal
  American Academy  Dermatology 4:31-7.

Ott  WR.  1981.  Exposure  estimates based on
  computer-generated  activity  patterns.  Paper
  presented  at the 74th annual meeting  of the Air
  Pollution Control  Association.  Philadelphia, PA.
  Paper No. 81-57-6.

Patterson MR, Sworski TJ,  Sjoreen AL, et al.  1982.
  User's manual for UTM-TOX, a unified transport
  model. Draft report.  Oak Ridge,  TN: Oak  Ridge
  National  Laboratory.  ORNL-TM-8182.  IEG-
  AD-89-F-1-3999-0.

Pennington,  JAT. 1983. Revision  of the total diet
  study. In J.  Amer. Dietetic Assoc. 82 (2).

Puffer  H.,  Azen SP, Young DR, et  al.  1981.
  Consumption rates of potentially hazardous marine
  fish caught in the metropolitan Los Angeles area.
  California  Department of  Fish and  Game.  EPA
  Grant No. R 807 120010.

Roberts MS, Anderson RA, Swarbrick J. 1977.
  Permeability of human epidermis  to  phenolic
  compounds.  Journal Pharm.  and  Pharmacol.
  29:677-83.

Schaefer  H,  Zesch A, Stuttgen G.  1982. Skin
  permeability. New York: Springer-Verlag.

Scheuplein RJ,  Blank  IH,  Brauner GJ, MacFarlane
  DJ.  1969.  Percutaneous absorption of steroids.
  Journal  of  Investigative Dermatology  54(1):63-70.

Scheuplein RJ, Blank  IH.  1971. Permeability of the
  skin.  Physiological Reviews 51(4):702-47.

Sutton  TJ. 1973. Dermal  toxicity and penetration
  studies following topical  application  of  three
  histamine  H2-receptor  antagonists  with  a
  comparison  with   an   H1-receptor  antagonist.
  Toxicology  and Applied  Pharmacology  50(3):459-
  65.

Tregear RT.   1966. Physical functions  of  skin.  New
  York: Academic Press.

USDA.  1980.  Food and nutrient intakes  of individuals
  in 1   day  in  the  United  States, spring 1977,
  nationwide food  consumption survey 1977-78,
  preliminary  report  no. 2. Washington,  DC: Science
  and Education Administration.
USDA. 1983a. Food consumption of households in
  the Northeast,  seasons and year 1965-66,  report
  no. 13. Washington, DC:  Agricultural Research
  Service. August 1972.

USDA. 1983b. Food consumption of households in
  the North Central region, seasons and year 1965-
  66, report  no. 14.  Washington, DC:  Agricultural
  Research Service. September 1972.

USDA. 1983c. Food consumption of households in
  the South,  seasons  and  year 1965-66, report no.
  15. Washington, DC: Agricultural Research Service.
  January 1973.

USDA. 1983d. Food consumption of households in
  the West,  seasons  and year 1965-66, report no.
  16. Washington, DC: Agricultural Research Service.
  January  1973.

USDA. 1983e. Food consumption of households in
  the United States,  seasons and year  1965-66.
  Washington, DC: Agricultural  Research  Service.
  March 1972.

USDOI. 1973.  Outdoor recreation: a  legacy for
  America. Washington, DC: U.S. Department of
  Interior.

USEPA. 1980a. Dietary consumption  distributions of
  selected food groups for the U.S.   population.
  Washington, DC: U.S.  Environmental Protection
  Agency.  Office  of Pesticides  and Toxic
  Substances, Office of Testing and  Evaluation. EPA
  560/11-80-012.

USEPA. 1980b.  Water quality criteria   documents.
  Federal  Register, Vol.\45 No. 231, November 28,
  1980.

USEPA. 1981. The exposure assessment group's
  handbook  for  performing exposure assessments
  (draft report). Washington, DC:  U.S. Environmental
  Protection Agency.

USEPA. 1985. Superfund  public health evaluation
  manual. Draft. Washington,  DC:  ICF, Inc. Prepared
  for the Policy Analysis Staff, Office of Emergency
  and Remedial  Response,  U.S.  Environmental
  Protection Agency. October 1, 1985.

USEPA.  1987. Exposure factors  handbook.
  Washington, DC:  Exposure Evaluation  Division,
  U.S. Office of Toxic  Substances,  U.S.
  Environmental Protection Agency. Contract  No.
  68-02-4254, Task No.  83.

Versar. 1984. Methods for  estimating concentrations
  of chemicals in indoor air. Draft  final report. Versar
  Inc. Washington, DC:  Prepared for the  Exposure
  Assessment Branch,  Exposure  Evaluation  Division,
                                               132

-------
  Office  of Toxic Substances,  U.S.  Environmental
  Protection Agency.

Versar.  1985.  Exposure  assessment  for
  perchloroethylene. Revised draft report. Versar Inc.
  Exposure  Assessment  Branch, Exposure
  Evaluation Division,  Office  of Toxic Substances,
  U.S. Environmental Protection Agency.
                                               133

-------

-------
                                 Appendix B

Possible Exposure Assessment Data Requirements for Uncoltrolled Hazardous Waste
                       Sites and Index to Variable Terms
                                     135

-------
Table B-1.    Possible  Data Requirements for  Estimation of Contaminant  Release and Transport and Exposed  Populations
     Type of Analysis
       Type of Site
                                                          Area of Concern
                                                                                      Area Subclass
                                                                                                                  Parameter
Contaminant release
Contaminated surface soil   Particulate  release
(includes spills and leaks)
                                                                               Wind erosion
                                                                                Unpaved  roads
• Soil erodibility index3
• Soil ridge roughness
  factor3
• Field length along
  prevailing wind direction
^Vegetative cover factor
jteConcentrationsof
  contaminantsb
^Volume of contaminated
  regionb

.««Silt content
j«Mean speed of vehicles
  traversing contaminated
  area"
j«Mean weight of vehicles
  traversing contaminated
  area"
j«Mean number of wheels
  of vehicles traversing
  contaminated  aread
                                                     Volatilization
                                                                                Excavation and transfer of
                                                                                soil
                                                                                Short-term release
                                                                                Long-term release
                                                     Runoff to surface water
                                                                               .e&tfSilt content0     p
                                                                               •Mean wind speed
                                                                               • Drop height
                                                                               • Material moisture content
                                                                               • Dumping device capacity

                                                                               • Vapor concentration of
                                                                                 contaminants in soil pore
                                                                                 spaces'

                                                                               • Depth from soil surface to
                                                                                 bottom of contaminated
                                                                                 regionb
                                                                               AS Area of contamination11
                                                                               • Depth of "dry"
                                                                                 (uncontaminated) zone at
                                                                                 sampling timeb
                                                                               • Concentrations of
                                                                                 contaminants in soil and
                                                                                 in liquid phaseb
                                                                               ..a? Soil porosityb'c
                                                                               j«Absolute temperatureb'e
                                                                               • Time measured from
                                                                                 sampling time

                                                                               • Soil erodibility factor9
                                                                               • Slope - length factor
                                                                               • Vegetative cover factord
                                                                                 Erosion control practice
                                                                                 factor"
                                                                               W\rea of contamination
                                                                               jsjsSoil bulk density0
                                                                               .ujsTotal areal concentrations
                                                                                 of contaminants
                                                      Release to ground water
                                                                                - See Chapter 3.6 of
                                                                                  Manual
                                                                                                                         (Continued)
                                                                136

-------
Table B-1.    (Continued)
     Type of Analysis
   Type of Site      Area of Concern
                                                                      Area Subclass
                                                                                                   Parameter
                           Landfill
                                             Volatilization
                                     NO internal gas generation
                                                                With internal gas
                                                                generation
                                             j«Area of contamination
                                             • Soil porosity0
                                             ^Effective depth of soil
                                               cover
                                             • Mole fractions of
                                               contaminants in waste
                                             ^Absolute ambient
                                               temperature6
                                             ^Absolute ambient
                                               pressure6'11
                                             ^Soil bulk density0'1
                                                                 contaminants in  soilb
                                                                j»Volume of contaminated
                                                                 regionb

                                                                j»Vapor concentration of
                                                                 contaminants in soil pore
                                                                 spaces'
                                                                • Area of contamination
                           Lagoon
                                              Release to ground
                                              water

                                              Volatilization
Contaminant fate
Contaminated
surface soil,
landfill, lagoon
Migration into
ground water

Atmospheric fate
                                                                - See Chapter 3.6 of
                                                                  Manual
                                                                jsd-iquid-phase
                                                                 concentrations of
                                                                 contaminants
                                                                j»Area of contamination
                                                                ^Absolute ambient
                                                                 temperature6
                                                                j«Volume of contaminated
                                                                 regronb

                                                                - See Chapter 3.6 of
                                                                  Manual
 • Distance from site to
   selected exposure point
 j»Mean wind speed6
 • Relative annual frequency
   of wind flow towards point
   x6
 • Relative annual frequency
   of stability class for wind
   flow towards point x6
 .^Stability classes
   (A = unstable, F = stable);
   according to Pasquill
   classification system6
 • Vegetative cover factord
	(Continued)
                                                        137

-------
Table B-1.    (Continued)
     Type of Analysis
Type of Site
                                                   Area of Concern
                                                                              Area Subclass
                                                                                                           Parameter
                                              Surface water fate
                                              Ground water fate
                                                                        Saturated zone
                                                                        Unsaturated zone
 Exposed populations
                                   All
                                              General
                                              Contaminated surface
                                              water
                                              Contaminated ground  water
                                                                     ^Combined effluent and
                                                                      stream flow data
                                                                     Intermedia substance
                                                                      transfer  rate'
                                                                     j«Widtri of water body'
                                                                     •  Stream velocity'
                                                                     j-asStream depth'
                                                                     .^Slope of stream channel'

                                                                     .easSoil hydraulic
                                                                      conductivity1*
                                                                     ^(Hydraulic gradient1
                                                                     ..^Effective soil porosity"1

                                                                     j««Average percolation or
                                                                      recharge  ratem
                                                                     •  Volumetric water content
                                                                      of soil in unsaturated
                                                                      zone'
                                                                     .^Hydraulic loading from
                                                                       manmade  sources''"
                                                                     •  Precipitation rate"'0
                                                                     j«Evapotranspiration rate
                                                                     •  Runoff rate''"
                                                                     j«Average depth of
                                                                      contaminated  area"
                                                                     ^Evaporation rate0

                                                                     •  Location of population
                                                                     .^Number of persons
                                                                     jieAge/sex distribution

                                                                     •  Recreation patterns
                                                                       (fishing, hunting,
                                                                       swimming)
                                                                     ^(Commercial fisheries
                                                                       present
                                                                     ^Drinking water intake
                                                                       locations and populations
                                                                       served

                                                                     ^Drinking water intake
                                                                       locations and populations
                                                                       served
                                                                                                                          f,n
 b  Some values can be obtained from existing literature.
 c  For calculation of long-term release ( > 70 years).
 d  Can be obtained from Soil Conservation Service (SCS) "Soils 5 File"  data base.
 e  Estimated indirectly from site survey information.
 f  Can be estimated based on existing meteorological station data.
   Can be calculated.
 9  Can be obtained from SCS office or from existing literature.
   Necessary only if diffusion coefficients for toxic components are not available from existing literature.
 '  Can be measured as an alternative to  measuring soil porosity.
 J  Can be obtained from USGS data.
 l<  Can be calculated or estimated from Table in  Manual.
   Can be obtained from USGS or local university geology/hydrogeology departments.
 m  Can be calculated via  equation in  manual, or can be obtained from  USGS, USDA, NOAA, or U.S.  Forest Service.
 "  Needed to calculate average  percolation/recharge rate when not measured at site.
 0  Available  from  local or National Weather Service.
                                                             138

-------
Table B-2.    Index to Variable Terms
Term
Used
E
1'
K'
C'
L'
V
EVT

k
s

SP
W
w
D p
H
D;
A



csi
p.

dsc
IVIj
T

MW,
MWa
2V,,
2V2
Pa
D'
MW
8


Definition
Potential annual wind erosion soil loss
Soil erodibility index
Soil ridge roughness factor
Climatic factor
Field length along the prevailing wind direction
Vegetative cover factor
Emission factor for vehicular traffic

Particle size multiplier
Sift content

Mean vehicle speed
Mean vehicle weight
Mean number of wheels
Number of days with at least 0.254 mm (0.01 in)
of precipitation per year
Emission rate of toxic component i
Diffusion coefficient of component i
Contaminated area



Saturation vapor concentration of component i
Soil porosity

Effective depth of soil cover
Mole fraction of toxic component i in the waste
Temperature

Molecular weight of contaminant i
Molecular weight of air
Molecular diffusion volumes of toxic contaminant
\/t) and air (V2)
Absolute pressure
Known diffusion coefficient of a compound with
molecular weight and molecular diffusion volume
close to that of the unknown (D,)
Molecular weight of the selected compound
corresponding to D'
Soil bulk density


Units
(mass/area/time)
(dimensionless)
(dimensionless)
(dimensionless)
(feet)
(dimensionless)
(kg/vehicle kilometer
traveled; Ib/vehicle mile
traveled)
(dimensionless)
(%)

(kph; mph)
(Mg; tons)
(dimensionless)
(dimensionless)
(g/sec)
(cm2/sec)
(cm2; areas; ha)
(100 in2)



(g/cm3)
(dimensionless)

(cm)
(g/g)
(K,C)

(g/mole)
(g/mole)

atm

(g/mole)
(g/cm3)


Equation(s) in
which term is used
2-1
2-1
2-1
2-1
2-1
2-1
2-2

2-2
2-2

2-2
2-2
2-2
2-2
2-3; 2-8; 2-9;
2-11; 2-15
2-3; 2-4; 2-5;
2-12
2-3; 2-8; 2-9;
2-11; 2-15; 2-
1 9; 2-21 ; 2-24;
2-25; 2-26;
2-30; 2-32;
2-33; 2-37
2-3; 2-7
2-3; 2-6; 2-12;
3-17; 3-34;
3-35
2-3
2-3
2-4; 2-7;2-10;
2-13; 2-16;
2-17
2-4; 2-5; 2-7;
2-10; 2-17
2-4
2-4
2-4
2-5
2-5
2-6; 2-25;
2-26; 2-27;
3-17

Source
calculated
site data and
literature
site data and
literature
literature
site data and
literature
site data
calculated

see text
site data, SCS
Soils 5 Fife
site data
site data
site data
Figure 2-3
calculated
calculated
site data



calculated
site data; SCS
Soils 5 File

site data
site data
site data

literature
see text
literature and
calculated
site data
see text
literature
site data; SCS
Soils 5 File

(Continued)
                                                   139

-------
Table B-2.    (Continued)
Term
Used
P
P

R


Q*
Vy
kiG

MWH20
k|G,H20

cs

CB
t
d
D


Hi'
HI
h

td

K,
kiL
MW02
K|_, 02

EM
vc
CC

E


Y(S)E
a

vr

KP

L

S

C


Definition
Particle density
Vapor pressure of the chemical

Gas constant


Vapor concentration of compound i
Mean landfill gas velocity in the soil pore spaces
Gas-phase mass transfer coefficient of
chemical i
Molecular weight of water
Gas phase mass transfer coefficient for water
vapor at 25°C
The liquid-phase concentration of component i

Bulk contaminant concentration in soil
Time measured from sampling time
Depth of dry zone at sampling time
Related to the amount of contaminant i that goes
from liquid to gas phase, and then from gas phase
to diffusion in air
Henry's Law constant in concentration form
Henry's Law constant
Depth from soil surface to the bottom of the
contaminated region
The time at which all contaminant has volatized
from the soil
Overall mass transfer coefficient
Liquid phase mass transfer coefficient
Molecular weight of oxygen
Liquid phase mass transfer coefficient for oxygen
at 25°C
Average release of contaminant i
Volume of contaminated region
Concentration of contaminant i in soil

Total release rate of contaminant i obtained by
summing all above-listed releases of the
contaminant at the site
Sediment yield in tons per event
Conversion constant

Volume of runoff
Peak flow rate
The soil-erodibility factor. Obtained from the
local Soil Conservation Service Office
The slope-length factor

The slope-steepness factor

The cover factor


Units
(g/cm3)
(mm Hg)

(62.3 mm Hg-liter/
k-mol; 8.2 x 10'5
atm-m3/-mol-k)
(g/cm3)
(cm/sec)
(cm/s)

(g/mole)


(g/cm3)

(g/cm3)
(seconds)
(cm)
(cm2/sec)


(dimensionless)
(atm-m3/mol)
(cm)

(sec)

(cm/sec)
(cm/sec)
(g/mole)


(mass/time)
(cm3)
(g/cm3, kg/ha, Ib/acre)

(g/sec)


(metric tons)


(acre-feet, m3)
(ft3/sec, m3/sec)
(commonly expressed in
tons per acre per R unit)
(dimensionless)

(dimensionless)

(dimensionless)


Equation(s) in which
term is used
2-6
2-7; 2-37

2-7; 2-13; 2-16


2-8; 2-9
2-8
2-9; 2-10; 2-16

2-10
2-10

2-11; 2-14;
2-15; 2-19; 2-34
2-11; 2-14; 2-19
2-11
2-11; 2-14; 2-19
2-11; 2-12;
2-14; 2-19

2-12; 2-13
2-13; 2-16
2-14

2-14

2-15; 2-16
2-16; 2-17
2-17
2-17

2-18; 2-19; 2-29
2-18; 3-34; 3-35
2-18; 2-25;
2-28; A-3
2-18


2-20; 2-27
2-20; 2-21;
2-23; 2-24
2-20; 2-21
2-20; 2-24
2-20; 2-30

2-20; 2-30

2-20; 2-30

2-20; 2-30


Source
see text
literature or
estimated (see text)
see text


site data
see text
calculated

see text
calculated

site data

site data
site data
site data
calculated


calculated
literature
site data

calculated

calculated
calculated
see text
literature

calculated
site data
site data

calculated


calculated
see text

calculated
calculated
site data, literature

see Figures 2-4
through 2-6
see Figures 2-4
through 2-6
see text and Table
2-4
(Continued)
                                                    140

-------
Table B-2.    (Continued)
Term
Used
P
Qr

Rt



sw

CN
Tr



Ss
DS
ec

Kd

PXi
PQi
B

N




Y(S)A
Rr
sd
Dd

LC
U
C0
QJ
b
Ks


K
c
Kw

DC
Dw
Uc
Uw

Definition
The erosion control practice factor
Depth of runoff

The total storm rainfall



Water retention factor

The SCS Runoff Curve Number
Storm duration



Sorbed substance quantity
Dissolved substance quantity
Available water capacity of the top cm
of soil
Sorption partition coefficient

Sorbed substance loss per event
Dissolved substance loss per event
Dissolved or sorbed loss per storm
event
Number of "average" storm events in
70 years



Annual soil loss in runoff
Rainfall and runoff factor
Sediment delivery ratio
Overland distance between site and
receiving waterbody
Contaminant Loading rate

Solubility of solid chemical
Volume loading rate
Soil specific exponential function
Soil hydraulic conductivity

Hydraulic gradient
Hydraulic conductivity of liquid
contaminant in site soil
Hydraulic conductivity of water in site
soil (same as K,)
Density of liquid contaminant
Density of water
Dynamic viscosity of liquid contaminant
Dynamic viscosity of water

Units
(dimensionless)
(in, cm)

(in, cm)



(in, cm)

(dimensionless)
(hour)



(kg, lb)
(kg, lb)
(dimensionless)

(cm3/g)

(kg, lb)
(kg, lb)
(kg, lb)

(dimensionless)




(tons)
(dimensionless)
(dimensionless)
(Ft)

(mass/time)

(mass/volume)
(volume/time)
(dimensionless)
(length/time)

(dimensionless)
(length/time)

(length/time)

(mass/volume)
(ma&volume)
[mass/(length x time)]
[mass/(length x time)]

Equation(s) in
which term is used
2-20; 2-30
2-21; 2-22;
2-24; 2-28
2-22; 2-24;
2-28


2-22; 2-23;
2-24
2-23
2-24



2-25; 2-27
2-28; 2-28
2-25; 2-28

2-25; 2-26;
3-17; 3-19
2-27
2-28
2-29

2-29




2-30
2-30
2-30; 2-31
2-31

2-32; 2-34;
2-37
2-32
2-33; 2-34
2-33; 3-13
2-33; 3-9;3-13

2-33; 3-9
2-35; 3-16

2-35; 3-16

2-35
2-35
2-35
2-35

Source
see text
calculated

National
Climatological Data
Center, Asheville,
NC; USDC (1961)
calculated

Table 2-6
National
Climatological Data
Center, Asheville,
NC; USDC (1961)
calculated
calculated
calculated
(see text)
literature

calculated
calculated
calculated
(see text)
National
Climatological Data
Center, Asheville,
NC; USDC (1961)
(see text)
calculated

calculated
site data

calculated

literature
calculated
Table 3-1 1
site data,
Table 3-8; 3-9
site data
calculated

site data

literature
literature
literature
literature
(Continued)
                                               141

-------
Table B-2.   (Continued)
Term
Used
PS


Ap


SH
*

d|
X

C(X)

Q

ay

°z

li

IT
C(X)

W(X)
CA(X)

*A

A,B,C,
D.E.F

C(CL)

Y(X)


C

ce

Qe

Ot
T,

MZ

W
u
dn

Definition
Permeation rate


Permeation constant for polymer


2nd permeation constant for polymer
Permachor value for polymer-
permanent pair
Thickness of liner
Distance from site to selected exposure
point
Atmospheric concentration of substance
at distance X from the site
Release rate of substance from site to
atmosphere
Atmospheric dispersion coefficient in
the lateral (crosswind) direction
Atmospheric dispersion coefficient in
the vertical direction
Mean wind speed
Tho \/£)liic> rii — "^ 141ft
1 1 lc ValUU (Jl O . I *r I o
Average atmospheric concentration of
substance at point X over long term
Relative annual frequency of wind flow
Atmospheric concentration at point X
during stability Class A
Relative annual frequency of stability
Class A for wind flow towards point X
Stability classes (A = unstable,
F = stable) according to Pasquill
classification system
Predetermined critical atmospheric
concentration level
Perpendicular distance from point X on
plume centerline to the C(CL) isopleth
boundary
Concentration of substance in stream
water
Concentration of substance in effluent

Effluent flow rate

Combined effluent and stream flow rate
Intermedia substance transfer rate

Length of mixing zone downstream of
effluent release to stream
Width of water body
Stream velocity
Stream depth

Units
[g-mil/
(100 in2 x day x
cmHg)]
[g-mil/
(100 in2 x day x
cmHg)]
(cc/cal)
(cal/cc)

(mils)
(length)

(mass/volume)

(mass/time)

(length)

(length)

(length/time)
, , . .
(Gtmensioniess)
(mass/volume)

(dimensionless)
(mass/volume)

(dimensionless)

(dimensionless)


(mass/volume)

(length)


(mass/volume)

(mass/volume)

(volume/time)

(volume/time)
(mass/time)

(length)

(length)
(length/time)
(length units)

Equation(s) in
which term is used
2-36; 2-37


2-36


2-36
2-36

2-37
3-1; 3-2; 3-3;
3-7; 3-8
3-1; 3-3, A-1

3-1

3-1; 3-3

3-1

3-1
3-1
I
3-2; A-1

3-2
3-2

3-2

3-1; 3-2


3-3

3-3


3-4; 3-5; A-2

3-4

3-4

3-4; 3-5
3-5

3-6

3-6
3-6; 3-7; 3-8
3-6

Source
calculated


Table 2-7


Table 2-7
Table 2-7; 2-9;
2-10
site data
site data

calculated

calculated

Figure 3-5

Figure 3-6

site data


calculated

site data
calculated

site data

site data


air quality criteria

calculated


calculated

contaminant
release analysis
contaminant
release analysis
site data
site data,
calculated
calculated

site data
site data
site data
(Continued)
                                            142

-------
Table B-2.    (Continued)
Term
Used
s.
g
W(X)

W(CL)

W(0)


K
e
v
• p w
q

V


P e

es

^(-15
e

HL

Pr
ET
Qr
EVAP
Get


vea
vi-y

Rd

vd
Koc



foe
KQW


xd
Td
Qd

Definition
Slope of stream bed
Gravitational acceleration constant
Water concentration of substance at
downstream distance X
Predetermined critical water
concentration level
Water concentration of substance
immediately below point of introduction
to stream
Overall aquatic decay coefficient
Exponential function
Interstitial pore-water velocity or
ground-water velocity
Average percolation or recharge rate

Darcy velocity


Soil Effective Porosity

Saturated Water Content soil (equal to
P.)
Wilting Point Moisture Content
Volumetric water content of soil

Hydraulic loading from manmade
sources
Precipitation rate
Evapotranspiration rate
Runoff rate
Evaporation rate
Correction factor for converting pan
evaporation rate to evapotranspiration
rate for turf grass
Correction factor for converting turf
grass evapotranspiration to that for
other vegetative cover
Retardation factor

Retarded velocity of hydrophobic
Partition coefficient for organic carbon



Fraction of organic carbon in soil
Octanol/water partition coefficient


Nomograph factor
Nomograph factor
Nomograph factor

Units
(dimensionless)
(32 ft/sec2)
(mass/volume)

(mass/volume)

(mass/volume)


(time-l)

(length&me)

(depth/time)

(length/time)


(dimensionless)

(dimensionless)

(dimensionless)
(dimensionless)

(depth/time)

(depth/time)
(depth/time)
(depth/time)
(depth/time)
(dimensionless)


(dimensionless)


(dimensionless)

(length/time)
(ml/g)



(dimensionless)
(ml/g)


(dimensionless)
(dimensionless)
(dimensionless)

Equation(s) in
which term is used
3-6
3-6
3-7

3-8

3-7; 3-8


3-7; 3-8
2-36; 3-7
3-10; 3-12;
3-18
2-32; 3-12;
3-13; 3-14
3-9; 3-10;
3-26; 3-27;
3-29; 3-30
3-10; 3-11;
3-28; 3-31
3-11; 3-13

3 11
3-12; 3-13

3-14

3-14
3 14; 3-15
3 14
3 15
3-15


3-15


3-17; 3-18;
3-27; 3-30
3-18
3-19; 3-20;
3-21; 3-22;
3-23; 3-24;
3-25
3 19
3-20; 3-21;
3-22; 3-23;
3-24; 3-25
3-26; 3-29
3-27
3-28; 3-31

Source
site data
	
calculated

water quality
criteria
calculated


literature, estimated
	
calculated

site data,
calculated
calculated


site data

site data, literature

site data, literature
site data,
calculated
site data,
calculated
site data
site data, calculated
site data, calculated
site data
Table 3-12


Table 3-13, see
text

calculated

calculated
calculated



site data, literature
literature


calculated
calculated
calculated
(Continued)
                                                143

-------
Table B-2
Term
Used
Dx


Dy
m
ax
ay
Y
A
T,/2
v,2

1
v
vgw
cgw

Mc
cc

IEX
te
1
Bw
F
DEX
AV
PC

DA
C
(Continued)
Definition
Longitudinal Dispersion Coefficient


Transverse Dispersion Coefficient
Aquifer thickness
Longitudinal dispersivity
Transverse dispersivity
Coefficient for decay
Decay constant
Half-life
Volume of liquid chemical released
Average concentration of chemical
contaminant in released liquid
Volume of contaminated ground water

Average concentration of contaminant
in ground water
Mass of solid waste
Concentration Expressed as Mass
Fraction
Inhalation exposure
Duration of exposure event
Average Inhalation rate
Average adult body weight
Frequency of exposure event
Dermal exposure
Skin surface area available
Dermal permeability constant for
subject contaminant
Dust adherence
Contaminant concentration
Units
(lengthytime)


(Iength2/time)
(length)
(length)
(length)
(dimensionless)
(1/time)
(time)
(lengths)
(mass/lengths)

(lengths)

(mass/length3)

(mg)
(dimensionless)

(mg/kg/day)
(hours/event)
(m3/hr)
(70kg)
(number/lifetime)
(mg/kg/day)
(cm2)
(cm/hr)

(mg/cm2)

Equation(s) in
which term is used
'6-2V. '6-21;
3-28: 3-29;
3-31
3-28; 3-31
(Fig. 3-8)
(Fig. 3-8)
(Fig. 3-8)
(Fig. 3-8)
(Fig. 3-8)
(Fig. 3-8)
3-32
3-32

3-32; 3-33;
3-34; 3-35
3-32; 3-33

3-33
3-33

A-1
A-1; A-2
A-1
A-1; A-2; A-3
A-1; A-2; A-3
A-2; A-3
A-2; A-3
A-2

A-3

Source
calculated


calculated
site data
literature
literature
literature
calculated
calculated
site data
site data

site data

site data

site data
site data

calculated
estimated
Table A-3
Eq A-1
estimated
calculated
estimated
Table A-4

See text

144

-------
                                          Appendix C
                                   Data Management Forms
This appendix presents  master  copies of data
management forms designed for use when applying
the various analyses  described in  this  manual.  The
forms  are  intended  to provide easy,  consistent
organization  of the  results  of  each  analysis
component  in the  human exposure  assessment
process (qualitative analysis, quantitative contaminant
release  analysis,  etc.) for  ready use in subsequent
analytical components.  In addition,  these  forms will
also organize exposure  assessment output in a form
most useful for  conducting  a risk  assessment
(executed following and based  on the results of the
exposure assessment) as well as the development  of
a  site  Endangerment Assessment  for  enforcement
purposes.
These  forms are included  as master copies,  that
should  be photocopied for  use in a given  site
investigation. In many  cases, a number of copies of
certain  forms will be required to tabulate all results of
the exposure assessments. For example, Form No. 7:
Exposure Integration  requires that the exposed
population segment be logged into the upper  left
corner  of the form, and exposure  information for that
population segment be  entered into  the  remaining
columns for each chemical to which the population is
exposed. If four distinct exposed population segments
are affected at  the site, four copies of the form will be
required.
                                                145

-------
         Form 1: Qualitative Exposure Analysis                                                                                 Site Name:.
                                                                                                                                    Date:.
                                                                                                                                  Analyst:,
                                       On-site Release              Release                                                            Potentially Exposed
                  Chemical                  Source           Likelihood Magnitude*      Release Mechanism       Receiving Medium       Population Segment     Exposure Mechanism
CD

-------
Form 1. Qualitative Exposure Analysis (Continued)                                                                   Site Name:
                                                                                                                            Date:.
                                                                                                                         Analyst:.
                              On-site Release              Release                                                             Potentially Exposed
         Chemical                  Source           Likelihood/Magnitude*      Release Mechanism       Receiving Medium        Population Segment      Exposure Mechanism
 6.
 7.
 10.
"Code each source as to: (1)  Likelihood of release and (2) Potential magnitude of release. Use H, M, L (high, medium, low) designation and provide a letter code for likelihood and
magnitude, each  separated by a "/".

-------
Form 2.  Quantitative Contaminant Release Data                                                      Site Name:=
                                                                                                          Date:.
                                                                                                       Analyst:.
 1.
 2.
 3.
 4.
 5
 7.
 8.
 9.
 10-
                                                                                                  Frequency of Short-
                             On-site Release                                Short-term Release       term Release Rate       Long-term Release
        Chemical                  Source            Receiving Medium           Rate (units)                (units)                Rate (units)

-------
Form 3.  Quantitative Environmental Fate Data
                                                  Site Name:_
                                                       Date:_
                                                     Analyst,
             Chemical
Affected Medium
Short-term Environmental
   Concentration (units)
Long-term Environmental
  Concentration (units)
 1.
 7.

-------
                  form 4: Quantitative Exposed Populations Data
                                                                                Site Name:_
                                                                                     Date;.
                                                                                   Analyst:.
                          Chemical
Affected Medium
                                                                                 Exposure Mechanism
     Population Segment          Number of Persons Potentially
(• denotes sensitive population)               Exposed
                   i.
                   2.
en
o
                   3.
                   4.
                   5.

-------
                Form 4: Quantitative Exposed Populations Data (Continued)                                                        Site Name:_
                                                                                                                                      Date:_
                                                                                                                                    Analyst:,
                                                                                                                Population Segment          Number of Persons Potentially
                        Chemical                 Affected Medium               Exposure Mechanism          (* denotes sensitive population)               Exposed
                 6.
cr

                 8.
                 10.

-------
            Form 5: Short-terrm Exposure Calculation                                                                         Site Name:_
                                                                                                                                 Date:.
                                                                                                                               Analyst:.
                   12                     3                      4567                     8

                                                                                                                                   Short-term Daily       Exposed
                                                    Short-term Exposure      Events per Time        Body Mass        Time Period     Exposure (mg/kg/day)        Segment
                Chemical       Exposure Mechanism     per Event (units)            Period               (kg)              (days)           [3x4E5x6]        (number of persons)
ui
ro
             5.

-------
           Form 5: Short-terrm Exposure Calculation (Continued)                                                             Site Name:.
                                                                                                                                Date:_
                                                                                                                              Analyst:_
                  12                    3                      4567                     8

                                                                                                                                  Short-term Daily      Exposed Population
                                                   Short-term Exposure      Events per Time        Body Mass        Time Period     Exposure (mg/kg/day)        Segment
                Chemical      Exposure Mechanism     per Event (units)            Period               (kg)              (days)           [3x4E5x6]       (number of persons)

            6.	    	   	    	
01
CO
            §>_
            9.

-------
           Form 6:  Long-terrm Exposure Calculation                                                                          Site Name:.
                                                                                                                                  Date:.
                                                                                                                               Analyst:.
                  12                     3                      4                   56                  7                     8

                                                                                                                  Time Period       Long-term Daily       Exposed Population
                                                    Long-term Exposure      Events per Time        Body Mass       (2.56 x 104     Exposure (mg/kg/day)         Segment
                Chemical      Exposure Mechanism     per Event (units)            Period               (kg)              days)            [3 x 4E5 x 6]       (number of persons)

            1.                                     	    	    	
01

-------
          Form 6:  Long-terrm Exposure Calculation (Continued)                                                             Site Name:.
                                                                                                                               Date:.
                                                                                                                             Analyst:.
                 12                    3                      4567                     8

                                                                                                                Time Period       Long-term Daily       Exposed Population
                                                  Long-term Exposure      Events per Time        Body Mass        (2.56 x io4     Exposure (mg/kg/day)         Segment
               Chemical      Exposure Mechanism     per Event (units)            Period               (kg)              days)            [3x4E5x6]       (number of persons)

           6.	   	    	
en
01
           10.

-------
                                Form 7:  Exposure Integration (Continued)
                                                                Site Name:.
                                                                     Dale:.
                                                                   Analyst:.
                                                                                                                  Exposure
                                    Population
                                     Segment
Chemical
Exposure Mechanisrr
Short-term
Long-term
Number exposed
                                                   6.
                                                   7.
en
-vl
                                                   9
                                                   10.

-------

-------