EPA/540/1-881001OSWER Directive 9285.5-1

April 1988

Superfund Exposure AssessmentManual

U.S. Environmental Protection AgencyOffice of Remedial Response

Washington, DC 20460

Notice

This report was prepared under contract to an agency of the United StatesGovernment. 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 anythird 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 suchthird party would not infringe on privately owned rights.

ii

Table of Contents

Chapter Page

List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1.1 Purpose . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1.2 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1.3

U s e o f t h e M a n u a l 1.4 . . . . . . . . . . . . . . . . . . . . . . . . .1.5 Timeframe of Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1.6 Analysis of Exposure Associated with Remedial Actions . . . . . . . . . . . . . . . . . .1.7 Organization of the Manual . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 CONTAMINANT RELEASE ANALYSIS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.12.2

2.3

2.4

2.5

2.6

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Contaminant Release Screening . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2.2.1 Contaminants in Soil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2.2.2 Contaminants Above-Ground

Quantitative Analysis of Atmospheric Contamination . . . . . . . . . . . .2.3.1 Fugitive Dust Emission Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.3.1.1 Beginning Quantitative Analysis . . . . . . . . . . . . . . . . . . . . . . .2.3.1.2 In-Depth Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.3.2 Volatilization Emission Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2.3.2.1 Beginning Quantitative Analysis . . . . . . . . . . . . . . . . . . . . . . .2.3.2.2 In-Depth Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.3.3 Long-Term and Short-Term Release CalculationQuantitative Analysis of Surface Water Contamination . . . . . . . . . . . . . .2.4.1 Beginning Quantitative Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.4.1.l Dissolved and Sorbed Contaminant Migration . . . . . . . . . . . . .2.4.2 In-Depth Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2.4.3 Long-Term and Short-Term Release Calculation . . . . . . . . . . . . . . . .

Quantitative Analysis of Ground-Water Contamination . . . . . . . . . . . . . . . . . .25.1 Beginning Quantitative Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

251.1 Leachate Release Rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2.5.2 In-Depth Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2.5.3 Long-Term and Short-Term Release Calculation . . . . . . . . . . . . . . . . .

Soil Contamination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2.6.1 Beginning Quantitative Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2.6.2 In-Depth Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3 CONTAMINANT FATE ANALYSIS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353.2 Contaminant Fate Screening . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

iii

viVIIIixxixii

1

7

788

10101010141414212222232325272929293131313131

Chapter

Table of Contents (Continued)

Page

3.2.1 Atmospheric Fate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3.2.2 Surface Water Fate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3.2.3 Soil and Ground-Water Fate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3.2.4 Biotic Fate

3.3 Quantitative Analysis of Atmospheric Fate . . . . . . . . . . . . . . . . . . 3.3.1 Screening Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3.3.2 In-Depth Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3.3.2.1 Intermedia Transfer3.3.2.2 lntramedia Transformation Processes . . . . . . . . . . . . . . . 3.3.2.3 The Effects of Terrain

3.3.3 Computer Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.4 Short- and Long-Term Concentration Calculations

3.4 Surface Water Fate Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.1 Beginning Quantitative Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3.4.2 In-Depth Analysis.

3.4.2.1 lntermedia Transformation Processes . . . . . . . . . . . . . . . . . . . . . . 3.4.2.2 lntramedia Transformation Processes . . . . . . . . . . . . . . . . . . .3.4.2.3 Computer Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3.4.2.4 Short- and Long-Term Concentration Calculations

3.5 Quantitative Analysis of Ground-Water Fate . . . . . . . . . . . . . . . . . . . . . . . .3.5.1 Discussion of Ground Water Modeling . . . . . . . . . . . . . . . . . . . . . . . . .

3.5.1.1 The Contamination Cycle351.2 Ground Water Flow Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5.1.3 Multiphase Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3.5.1.4 Contaminant Flow and Hydrodynamic Dispersion . . . . . . . . . .3.5.1.5 Transformation and Retardation . . . . . . . . . . . . . . . . . . . . . . .3.5.1.6 Higher Velocity Transport . . . . . . . . . . . . . . . . . . . . . . . . . . .

3.5.2 Ground-Water Modeling Equations and Nomograph3.5.2.1 Calculating Ground Water Velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5.2.2 Calculating the Velocity of Infiltrating Rainwater . . . . . . . . . . . .3.5.2.3 Corrections for Viscosity and Density . . . . . . . . . . . . . . . . . .3.5.2.4 Retardation Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3.5.2.5 Contaminant Velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3.5.2.6 Nomograph Technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3.5.2.7 Extent of Plume . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3.5.2.8 Use of Monitoring Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3.5.2.9 VHS Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3.5.3 In-Depth Methods and Models3.5.4 Short- and Long-Term Concentration Calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3.6 Biotic Pathways . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3.6.1 Estimation Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3.6.1.1 Aquatic Animals. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3.6.1.2 Terrestrial Animals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3.6.1.3 Terrestrial Plants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

4 UNCERTAINTY IN THE ANALYSIS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

4.1 Sources of Uncertainty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4.1.1 Input Variable Uncertainty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

4.2 Modeling Uncertainty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4.2.1 Model Simplification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4.2.2 Averaging Hydraulic Conductivities . . . . . . . . . . . . . . . . . . . . . . . . . . . .4.2.3 Dispersion Simulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4.2.4 Numerical Models and Analytical Models . . . . . . . . . . . . . . . . . . . . . . . .4.2.5 Chemical Degradation Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4.2.6 Model Operational Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

iv

3638404042424646474848485353555556565757636364656566686868697373757777828283939393949494

95

959596969697979797

Table of Contents (Continued)

Chapter Page

4.2.7 Source Shape . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 984.2.8 Steady State Modeling. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 984.2.9 Number of Dimensions Addressed by the Model . . . . . . . . . . . . . . . . . . . 98

4.3 Scenario Uncertainty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4.4 Approaches for Dealing with Uncertainty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

98

4.4.1 Sensitivity Appraisals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 984.4.2 Monte-Carlo Simulations . . . . . . . . . . . . . . . . . . . . . . . . . .4.4.3 Using Monitoring Data to Calibrate the Model

99. . . . . . . . . . . . . . . . . . . . . 99

4.5 Level of Uncertainty Appropriate for Exposure Modeling . . . . . . . . . . . . . . . . . . . 1004.6 Risk Communication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100

5 REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103

APPENDIX A Analysis of Exposed Human Populations andExposure Calculation and Integration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113

APPENDIX 13 Possible Exposure Assessment Data Requirements for UncontrolledHazardous Waste Sites and Index to Variable Terms . . . . . . . . . . . . . . . . . 135

APPENDIX C Data Management Forms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145

Number

List of TablesPage

l-l2-12-2

2-32-42-52-62-72-82-92-103-13-23-33-43-53-63-73-83-93-10

3-11

3-12

3-133-14

3-153-16

Technical Resource Contacts for Superfund Exposure Assessments . . . . . . . . . . . . . 5Potential Contaminant Release Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8Environmental Variables and Model Parameters for

the Wind Erosion Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Diffusion Coefficients of Selected Organic Compounds . . . . . . . . . . . . . . . . . . . . .“C” Values for Permanent Pasture, Rangeland, and Idle Land . . . . . . . . . . . . . . . .“C” Values for Woodland . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Runoff Curve Numbers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Parameter Values for Permeation Equation (at 25°C) . . . . . . . . . . . . . . . . . . . . . . .Polymer Categorization for Permeation of Water . . . . . . . . . . . . . . . . . . . . . . . . . .Permachor Values of Some Organic Liquids in Polyethylene and PVC . . . . . . . . . . .Water Permachor Value for Dry Polymers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Key to Stability Categories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Resource Requirements and Information Sources: Atmospheric Fate Models. . . . . .Features of Atmospheric Fate Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Data Requirements for Atmospheric Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Resource Requirements and Information Sources: Surface Water Fate Models . . . .Feature of Surface Water Fate Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Data Requirements for Surface Water Models . . . . . . . . . . . . . . . . . . . . . . . . . . . .Representative Values of Saturated Hydraulic Conductivity . . . . . . . . . . . . . . . . . . .Saturated Hydraulic Conductivity Ranges for Selected Rock and Soil Types . . . . . .Representative Values for Saturated Moisture

Contents and Field Capacities of Various Soil Types . . . . . . . . . . . . . . . . . . . .Representative Values of Hydraulic Parameters

(Standard Deviation in Parentheses) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Suggested Value for Cet Relating Evaporation from a US Class A Panto Evapotranspiration from 8 to 15-cm Tall, Well-watered Grass Turf . . . . . . . . . .Crop Coefficients for Estimating Evapotranspiration . . . . . . . . . . . . . . . . . . . . . . . .Resource Requirements and Information Sources:

Unsaturated Zone and Ground-Water Fate Models . . . . . . . . . . . . . . . . . . . .Features of Unsaturated Zone and Ground-Water Fate Models . . . . . . . . . . . . . . .Data Requirements for Unsaturated Zone and Ground-water Models . . . . . . . . . . .

131826262732323233454951525861627070

70

71

7273

848891

vi

List of Tables (Continued)

Number Page

A-1A-2A-3

A-4A-5B-1

B-2

Regional Census Bureau Offices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118U.S. Home Fruit and Vegetable Garden Use, 1977 . . . . . . . . . . . . . . . . . . . . . . . . 119Summary of Human Inhalation Rates for Men, Women, and Children

by Activity Level (m3/hour) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123Permeability Constants for Various Compounds . . . . . . . . . . . . . . . . . . . . . . . . . . . 124Typical Daily Soil Ingestion Rates for Children by Age Group . . . . . . . . . . . . . . . . . 129Possible Data Requirements for Estimation of

Contaminant Release and Transport and Exposed Populations . . . . . . . . . . . . 136Index to Variable Terms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139

vii

Number

List of FiguresPage

1-12-12-22-3

2-4

2-52-63-13-23-3

3-43-5

3-6

3-73-8

A-1A-2

Overview of the Integrated Exposure Assessment Process . . . . . . . . . . . . . . . . . . . . 2Contaminant Release Decision Network: Contaminants in Soil . . . . . . . . . . . . . . . . . 9Contaminant Release Decision Network: Contaminants Above-Ground . . . . . . . . . 11Mean Number of Days Per Year with > 0.01 Inches of Precipitation

(i.e., “wet days”) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15Slope Effect Chart Applicable to Areas A-l in Washington,

Oregon, and Idaho, and all of A-3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24Soil Moisture-Soil Temperature Regimes of the Western United States 24Slope Effect Chart for Areas Where Figure 2-5 is Not Applicable . . . . . . . . . . . . . 24Environmental Fate Screening Assessment Decision Network: Atmosphere . . . . . . 37Environmental Fate Screening Assessment Decision Network: Surface Water . . . . . 39Environmental Fate Screening Assessment Decision Network:

Soils and Ground-water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41Environmental Fate Screening Assessment Decision Network: Food Chain . . . . . . . 42Horizontal Dispersion Coefficient as a Function of Downwind Distance

from the Source . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43Vertical Dispersion Coefficient as a Function of Downwind Distance

from the Source . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44Area Within lsopleths for a Ground-Level Source . . . . . . . . . . . . . . . . . . . . . . . . . 47Nomograph for Solutions of Time, Distance, and Concentration

for Any Point Along the Principal Direction of Ground-water Flow . . . . . . . . . . . 78Exposed Populations Decision Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115Quantitative Exposed Population Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117

Viii

Foreword

The Super-fund Exposure Assessment Manual presents an integrated methodto help Remedial Project Managers and their contractors define the three majorcomponents involved in assessing human population exposure to contaminantsreleased from uncontrolled hazardous waste sites:

1. Analysis of toxic contaminant releases;2. Determination of the environmental fate of such contaminants; and3. Evaluation of the nature and magnitude of exposure to toxic

contaminants.

This report provides guidance for the development of exposure assessmentsusing monitoring data (which may provide the most dependable basis forevaluating some existing exposure levels), as well as modeling techniques topredict exposure over time.

Executive Summary

The analytical process outlined in the SuperfundExposure Assessment Manual provides a frameworkfor the assessment of exposure to contaminants at ormigrating from uncontrolled hazardous waste sites.The application of both monitoring and modelingprocedures to the exposure assessment process isoutlined. This process considers all contaminantreleases and exposure routes and assures that anadequate level of analytical detail is applied to supportthe human health risk assessment process.

The analytical process is structured in five segments:

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

2 . E v a l u a t i o n o f t h e t r a n s p o r t a n denvironmental fate of the contaminantsreleased;

3 . I d e n t i f i c a t i o n , e n u m e r a t i o n , a n dcharacterization of potentially exposedpopulations;

4. Integrated exposure analysis; and

5. Uncertainty analysis.

The Superfund Exposure Assessment Manualsupports the development of exposure assessmentsthat are consistent from site to site, and provides ameans of documenting that each site receivesadequate evaluation. The procedures presentedreflect current (at the time of publication) state-of-the-art methods for conduct ing an exposureassessment. However, it is important for the analystto recognize that exposure assessment is adeveloping science. Although the overall protocol forconducting exposure assessments at Superfund siteswill not change significantly over time and the basicparameters needed as input to the analysis are notlikely to change, alternative analytical methods maybe developed for many parts of the assessment. Themethods presented in this manual can serve as abenchmark against which such new methods can becompared.

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. Additionalguidance was provided by Peter Tong and Mary-Virginia Wandless of the Toxics IntegrationBranch.

Assistance was also provided by the following people:

Bob Ambrose ORD (Office of Research and Development)Doug Ammon Clean Sites, Inc. (formerly USEPA)Brint Bixler CH2M Hill (formerly USEPA)Robert Carsel ORD (Office of Research and Development)Richard Daley OWPE (Office of Waste Programs Enforcement)Carl Enfield ORD (Office of Research and Development)Tom Evans ORD (Exposure Assessment Group)Kevin Garrahan ORD (Exposure Assessment Group)Mark Garrison USEPA Region IIISteve Golian OERR (Office of Emergency and Remedial Response)Karen Hammerstrom OTS (Office of Toxic Substance)Seong T. Hwang ORD (Office of Research and Development)Joe Keeley Oregon Graduate CenterAshok Kumar University of ToledoSteve Ostrodka EPA Region VZubair Saleem OSW (Office of Solid Waste)Paul Schumann OSW (Office of Solid Waste)James Spatarella Versar, lnc. (formerly USEPA)Richard L. Stanford Roy F. Weston, Inc. (formerly USEPA)Sherry Sterling OWPE (Office of Waste Programs Enforcement)David Tetta EPA Region XLouis J. Thibodeaux University of ArkansasJawed Touma OAQPS (Office of Air Quality Planning and Standards)Georgia Valaoras OWPE (Office of Waste Programs Enforcement)Paul K.M. van der Heijde Holcomb Research InstituteLarry Zaragoza 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. LeeSchultz, Walter A. Palmer, Mark L. Mercer, Ruth A. Dickinson, Gary Whitmyre, Alan F. Gleit, GinaH. Dixon, and Van Kozak (currently Texas Department of Agriculture).

xii

Chapter 1lntroduction

1.1 Purpose

The Superfund Exposure Assessment Manualprovides Remedial Project Managers (RPMs) with theg u i d a n c e n e c e s s a r y t o c o n d u c t e x p o s u r eassessments that meet the needs of the Super-fundhuman health risk evaluation process. Specifically,the manual:

1. Provides an overall description of the integrated

health impacts resulting from the uncontrolled site.The risk assessment is based on the results of a siteexposure assessment, which evaluates:

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

2. The environmental transport and transformation ofcontaminants following release; and

exposure assessment process as it is applied touncontrolled hazardous waste sites; and

3. Implications of the resulting contact with exposedpopulations.

2. Serves as a source of reference concerning theuse of estimation procedures and computermode l ing techn iques fo r the ana lys is o funcontrolled sites.

This manual provides guidance for the assessment ofhuman population health risk only. Guidance forecological r isk assessment wi l l be providedseparately.

Sect ion 110 of SARA mandates that healthassessments be conducted by the Agency for ToxicSubstances and Disease Registry for all sites on theNational Priorities List. These health assessments canbe based on the results of site-specific exposureassessments. The exposure assessment, therefore, isan analytical tool that is used to comply with themandates of CERCLA.

1.2 Background 1.3 Scope

The Comprehensive Environmental Response,Compensation, and Liability Act of 1980 (CERCLA -42 USC 9601 et. seq.), as amended by theSuperfund Amendments and Reauthorization Act of1986 (SARA), was enacted to provide the FederalGovernment with the authority to respond to releasesor threatened releases of hazardous substances,pollutants, or contaminants into the environment. Asprescribed in the revised National Contingency Plan(see 47 FR 137, July 16, 1982), all sites designatedfor in-depth evaluation are included on the NationalPriorities List. These sites are evaluated for remedialaction through the application of a RemedialInvestigation, which defines the nature and extent ofcontamination, and a Feasibility Study, in whichpotential remedial alternatives are developed andanalyzed. Guidance for conducting these two majorcomponents of the remedial response process isprovided in USEPA (1985a and 1985b, respective-ly - currently under revision). As discussed in thatguidance, a part of the Feasibility Study is thedevelopment of a risk assessment that projects those

This manual provides guidance for the use (but notthe acquisition) of field data in the exposureassessment process. It does not serve as an all-encompassing guide to the use of computer modelsin the site remediation process, or direct the analysisof health risks that result from predicted exposure.This manual is intended to be used in conjunctionwith other related guidance, such as that for theacquisition of field data. As detailed in USEPA(1987a), field sampling Data Quality Objectives(DQOs) establish a phased sampling strategydesigned to guide the efficient acquisition of field datafor si te-specif ic exposure and publ ic healthassessments, and provide sampling plan guidanceaddressing the location of sampling points. Fieldoperating procedures for obtaining and handlingsamples have also been developed (USEPA 1987b).Other references, (USEPA 1986a, 1986b, 1987c, and1987d), address the utility, applications, andlimitations of computer models for predictingcontaminant concentrations and transport throughvarious environmental media. The process for

1

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

When conducting a comprehensive risk assessment,the analyst will need to refer to all of the above-citedguidance. While none of these guidance manualsstands alone, taken as a whole, they provide anoverall, integrated approach to analysis of sitecontamination and health risk.

1.4 Use of the Manual

This manual is used to apply state-of-the-artexposure assessment procedures to the uniqueanalytical needs of uncontrolled hazardous wastesi tes. The ul t imate goal of human exposureassessment at Super-fund sites is the determination ofthe type and magnitude of potential exposure tocontaminants present at and migrating from the site.To achieve this goal, many sites may require a mix ofqualitative and quantitative exposure analysis. Thelatter may range from simple analytical techniques(e.g., contaminant release or dispersion estimationequations) to more complicated computer modelingapproaches.

The general procedure for conducting an integratedexposure analysis is illustrated in Figure l-l. Thisprocedure is based on EPA’s published Guidelines forExposure Assessment (USEPA 1986c) and otherrelated guidance (USEPA 1985d-i) and is anadaptation of that process to the analytical problemsposed by abandoned hazardous waste sites. Aspreviously mentioned, target chemicals are selectedas part of the human health risk assessment process(see USEPA 1985c). Once these chemicals arechosen, the exposure assessment proceeds throughthe following stages:

1. Contaminant Release AnalysisEach on-site release point is identified for everytarget chemical, and the level of release (massloading) to each environmental medium isdetermined. Determination of contaminantre lease may be made e i the r by d i rec tmeasurement (monitoring) of such releases orby estimation. Although difficult to achieve for allmedia, monitored release values provide a moresound basis for projection of contaminantmigration later in the exposure assessmentprocess than do modeled estimates. When it isnot possible to obtain measured release rates,estimates can be based on measurements ofcontaminant concentrations in pertinent sourcemedia (e.g., estimates of contaminant release toground water based on measured concentrationsin contaminated soil). The results of the

Figure 1-1. Overview of the integrated exposureassessment process.

contaminant release analysis provide the basisfor evaluating the potential for contaminanttransport, transformation, and environmentalfate.

2. Contaminant Transport and Fate AnalysisTh is ana lys is descr ibes the ex ten t andmagnitude of environmental contamination (i.e.,con taminant concent ra t ions in spec i f i cenvironmental media). When possible, directmeasurement of contaminant concentrations ispreferred, and collection of samples during siteevaluat ion wi l l provide a clear basis fordetermining exposure potent ial for someexposure routes. However, the human healthrisk assessment process also requires projectionof potential exposure over a lifetime (see Section1-5), which can only be accomplished usingestimation procedures.

2

3. Exposed Populations AnalysisThe results of contaminant transport and fateanalysis allow the analyst to evaluate populationscontacting chemicals emanating from the site.Analysis of exposed populations involves theidentification, enumeration, and characterizationof those population segments likely to beexposed. The goal of this analysis is not only todelineate those populations coming into contactwith contaminants emanating from the site, butalso to determine how and with what frequencyand duration such contact occurs.

4. Integrated Exposure AnalysisIn this step, the individual chemical-specificexposure estimates for each exposure route(i.e., inhalation, ingestion, and dermal contact)are developed. For each exposed population, allexposures to each hazardous substance areidentified. In cases in which a population groupexperiences more than one exposure by a givenroute, exposures are summed to develop acumulat ive exposure value for the routeinvolved. For example, persons who reside inthe vicinity of a Superfund site may experiencedermal exposure to a given contaminant directlyon site as well as directly through basementseepage, and exposures via both of these routesshould be summed for exposure integrationpurposes.

5. Uncertainty AnalysisThe exposure assessment concludes with ananalysis of uncertainty. In this analysis each stepin the assessment is reviewed to identify anyuncertainties involved and to evaluate theirseparate and cumulative impact on assessmentresults. Uncertainties may result from the use ofdefault values for analytical input parameters,from the use of simplified estimation proceduresas opposed to more rigorous computer analysisor monitoring-based analysis, from an inabilityto define exposed populations with confidence,etc. The uncertainty analysis provides necessaryinput for remedial decisionmakers who mustevaluate the results of the exposure assessmentwith regard to their implications for potential risksassociated with the uncontrolled site andappropriate remedial alternative selection

This manual is intended to be used in conjunctionwith various other guidance to conduct Superfund siteRemedial Investigations and Feasibility Studies. Theuse of this manual is particularly linked to the PublicHealth Evaluation Manual. The two are intended to beused as two parts of the same process: the analysisof health impacts resulting from uncontrolledhazardous waste sites. In conducting a Superfundevaluation of exposure and public health impact, theanalyst initially applies the indicator chemical selectionprocess” outlined in the Superfund Public Health

Evaluation Manual to select the chemicals on whichthe site analyses will focus. Once the chemicals havebeen selected, the analytical framework of theSuperfund Exposure Assessment Manual is applied.Following completion of the exposure assessment,the analyst returns to the Superfund Public HealthEvaluation Manual for guidance in determining thedegree of human health risk for each exposedpopulation.

The user of this manual should understand that theseanalytical procedures are intended to be appliedsite-specifically. No two sites will be exactly alike interms of the extent and complexity of contamination,of contaminant migration, or of potentially exposedpopulations. Therefore, the specific analyticalprocedures to be applied in all Superfund exposureassessments cannot be fixed in general. Instead theapproach and methods applied to conducting anexposure assessment must be tailored to addressex is t ing s i te cond i t ions. In some s i tuat ionscontaminant releases or exposure routes may beadequately addressed by applying only screeningprocedures . In o ther cases more comp lex ,quantitative evaluation will be necessary.

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

1. Number and identity of chemicals present;

2. Availability of appropriate standards and/or toxicitydata;

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

4. Necessity for precision of the results; and

5. Quality and quantity of available monitoring data.

Simplified analyses may be used in the followinginstances: only a small number of chemicals must beevaluated; environmental standards or criteria forchemicals under study are available; a small numberof exposure pathways are present; release andtransport processes are relatively simple: or there is al imited need for detai l and precision in theassessment results (e.g., screening studies).Conversely, sites that have many contaminants forwhich no environmental standards or criteria areavailable, that exhibit multiple exposure pathways,that have complex contaminant release and transportprocesses in effect, or that require analytical resultsin great detail and precision will require more

?? Selection of indicator chemicals will be required only at thosesites where the number of contaminants present is too large toindividually evaluate exposure to each.

3

complex, quantitative analytical methods. Most sites o b t a i n a n d r e v i e w t h e o r i g i n a l s o u r c ewill fall somewhere between these two extremes. documentation cited for analytical components.

Procedures presented in this manual for conductingquantitative analyses include both simplified “desktop” approaches for developing order-of-magnitudeestimates and more resource-intensive, in-depthapproaches. Computer modeling and site monitoringare included. Generally, it is appropriate to applysimplified analysis to all pertinent exposure routes atthe beginning of quantitative evaluations so that thosecausing greatest concern can be identified forsubsequent in-depth analysis.

4. Results obtained through application of thesetools must be interpreted based on conditions atthe site being analyzed. These tools are providedto aid the analyst in making decisions, not tomake decisions for the analyst. When possible,models used in analyzing a given site should beverified with field monitoring data that test andvalidate model predictions at that site.

It is important to understand that analysis of exposureand resultant health impact is often a complexprocess in which selection and application of themost appropriate analytical tools, as well as theinsightful interpretation of their results, can be critical.The U.S. EPA encourages ongoing communicationbetween site analysts and experts in various exposureand health impact assessment fields. Thus, whenquestions arise regarding the utility of a particularmodel or mathematical solution, it is recommendedthat the analyst review the pertinent sectionsdescribed in this manual or contact the ToxicsIntegration Branch of the Hazardous Site EvaluationDivision of the Office of Emergency and RemedialResponse (FTS 475-9486). In addition, Table l-llists specific EPA contacts who can provide insightinto particular site assessment problems.

5 . T h e a p p r o a c h t o c o n d u c t i n g e x p o s u r eassessments ou t l ined in th is manua l i sconservative as are human health risk studies.However, the analyst needs to be sensitive to andto compensate, at least qualitatively, for theadd i t i ve e f fec t o f mu l t ip le conserva t ionassumptions. The degree of conservatism shouldnot be so extreme that the conclusions drawnfrom the analysis are unrealistic.

1.5 Timeframe of Analysis

In developing this manual, an attempt was made tocompile analytical methods appropriate for assessingexposure to chemicals migrating from uncontrolledhazardous waste facilities. There are limitations to theapplication of these analytical tools and to theinterpretation of the results obtained, including:

Quanti tat ive exposure assessments generateestimates of the long-term (chronic daily intake) andshort-term (subchronic daily intake) exposure tocontaminants. The output of each analyt icalcomponent (contaminant release, environmental fate,etc.) must be expressed in the same long-term andshort-term form. Long-term releases are defined asthe release rates of each contaminant migrating fromthe site averaged over an assumed 70-year humanlifetime. Short-term contaminant releases aredefined (USEPA 1985c) as those that occur over ashort period (usually 10 to 90 days) during the firstyear following site investigation.

1.6 Analysis of Exposure Associated withRemedial Actions

1. While some of these tools have been developedspecifically for application to Superfund sites,others were originally developed for differentpurposes and had to be adapted or directlyapplied to evaluation of conditions present atuncontrolled hazardous waste sites. The analystmust be careful in interpreting the resultsobtained from application of these tools and mustconsider their inherent uncertainties.

2. This manual assumes that the analyst has astrong technical background in engineering or thesciences. This background is essential to ensurethat analyses are carried out in a technicallysound fashion and that interpretations of theresults obtained are realistic.

3. It was not possible to include discussion of alltechnical limitations and caveats pertaining toeach analytical tool or procedure reviewed in thismanual. It may be beneficial for the analyst to

The analytical tools presented in this SuperfundExposure Assessment Manual are those appropriatefor analyzing exposure associated with the baselinecondition (i.e., the uncontrol led si te pr ior toimplementation of any remedial action). It should benoted, however, that waste treatment processes usedas part of a remediation strategy can themselvescontribute significant releases of contaminants to theenvironment. Stripping volatiles from wastewaters, forexample, generally involves artificial acceleration ofthe natural volatilization process, resulting in forcedtransfer of the volatile contaminants from wastewaterto air. Thus, analysts must evaluate the engineeringdesign of each remedial alternative to determine thelevel of contaminant release associated with itsimplementation. The user of this manual should referto Farino et al. (1983) for a discussion of methods toestimate wastewater treatment air emissions. Whenincinerating toxic wastes other than those containingPCBs, Destruction and Removal Efficiency (DRE)

4

Table 1-1. Technical Resource Contacts for Superfund Exposure Assessments

CommercialOffice 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 ResearchLaboratory; 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. lnternational 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 forOwners and Operators of Permitted Hazardous WasteFacilities; Subpart 0 - incinerators). For incinerationof wastes contaminated with PCBs, the analyst canrefer to 40 CFR 761.70 (Polychlorinated Biphenyls(PCBs) Manufacturing, Processing, Distribution inCommerce, and Use Prohibitions - Incineration).

Well engineered remedial alternatives planned foruncontrolled hazardous waste sites are not expectedin themselves to cause additional releases of toxiccontaminants to ground-water systems. Even if anunexpected spill of toxics occurs when remedialaction is taken, contaminant release should be slowenough to allow spilled substances to be isolatedprior to their reaching the saturated zone. Short-term release of contaminants to air may occur whileexcavating contaminated soil and loading it forremoval from the site. In such situations, the analystshould refer to USEPA (1983a), for release equationsfor 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 thismanual may be useful. For example, reductions incon taminan t re leases can be es t imated byrecalculating releases using altered (from the baselinecase) site-specific input variables based on theremedial action under consideration. Alternatively, onecan obtain a rougher approximation by applying theexpected remedial action percent control (based onengineering experience) to the source releaseestimates calculated for the baseline case. Inaddition, the analyst should refer to USEPA (1985j)

for a detailed discussion of both simplified analyticalmethods and numerical modeling approaches that canbe used to estimate remedial effectiveness.

1.7 Organization of the Manual

The following chapters of this manual detail methodsfor evaluating exposure to chemicals migrating fromSuperfund sites. The body of the manual providesguidance for the qualitative and quantitative evaluationof contaminant release, migration, and fate in theenvironment, along with that for evaluating uncertaintyin the analysis. Procedures for conducting exposedpopulations analysis and for developing an integratedexposure analysis are provided in appendices to thisreport.

5

Chapter 2Contaminant Release Analysis

2.1 Introduction

This chapter provides guidance for the analysis ofcontaminant releases from uncontrolled hazardouswaste sites. The goal of this analysis is to determinecontaminant release rates to specific environmentalmedia over time. The following sections address therelease of contaminants to air, surface water, andground water from wastes placed both above-ground and below-ground. In particular, guidance isprovided for the evaluation of the following categoriesof 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 twostages - sc reen ing o f con taminan t re leasemechanisms 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 eachrelease, and broadly defines the possible extent ofthe release. The following section is designed toestablish a consistent basis for the qualitativescreening of contaminant release from site to site.

Once the potential sources of on-site contaminantrelease have been screened, those requiring furtherevaluation are quantitatively analyzed. This mayinvolve the application of a simplified “desk-top”estimation approach, or a more in-depth methodsuch as computerized modeling or additional sitemonitoring. The goal of this analysis is to generate

release rate estimates (mass per unit time) for eachsource of contaminant release. Release rate valuesare necessary as input for subsequent environmentalfate analysis (see Chapter 3). Individual on-sitereleases of each contaminant are summed togenerate an overall, medium-specific release rate foreach chemical migrating from the site. Short-term(worst-case) release rates are developed, as arelong-term rates (averaged over 70 years).

The simplified estimation procedures that follow allowthe analyst to make release approximations based onchemical- and site-specific factors. However. thesecalculations do not take into account the full range ofvariables that affect on-site contaminant release.These approaches (with one exception) assumesteady state conditions. They do not directly addressthe reduction in contaminants present (due to releaselosses), or the associated reduction in release loadingover t ime corresponding with the decreasingcontaminant reservoir.*

When possible, monitoring should be used to quantifyrates of contaminant release. In some cases,however, this may not be feasible because methodsto directly measure releases from certain settings arestill being developed. Moreover, it may not always bepossible to monitor contaminant releases under theconditions of concern (e.g., dust releases under highwind conditions, surface water runoff releases duringstorm events, etc.). It may often be necessary toestimate release rates in the exposure assessmentprocess. All of the release rate estimation procedurespresented here, however, require some monitoredvalues as input. (Examples are measured contaminantconcentrations in soil, soil characteristics.) Theanalyst should be aware of the need to develop amonitored data base that is adequate to support theneeds of the contaminant release analysis portion ofthe exposure assessment.

In general, the procedures to estimate the rate ofcontaminant release are complete. When analyzing

* Estimation of the variation in the level of release over time iscalculated separately. See Long-Term and Short-TermRelease calculation subsections in this chapter.

7

wind erosion releases, however, the analyst shouldconsult other published guidance that addresses theapplication of the wind erosion equation in variousregions of the country. Depending on the location of aparticular site, one of the following three manuals willbe 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 sitedetermines whether contaminant release* occurs byany or all of the mechanisms summarized in Table2-1. In contaminant release screening, the likelihoodof release from each source, the nature of thecontaminants involved, and the probable magnitude oftheir release (relative to other on-site sources) areconsidered.

Figures 2-1 and 2-2 present the decision networksthat guide contaminant release screening analysis.Figure 2-1 deals with contaminants in or under thesoil and Figure 2-2 addresses above-groundwastes. Any release mechanisms evident at the sitewill require a further screening evaluation todetermine the likely environmental fate of thecontaminants involved (see Chapter 3).

2.2.1 Contaminants in Soil (see Figure 2-1)The following numbered paragraphs help to interpretand apply the steps of the contaminant releasedecision network presented in Figure 2-1. Eachparagraph refers to a particular numbered box in thefigure.

1. Most uncontrolled hazardous waste sites willexhibit some degree of surface or subsurface soilcontamination. This contamination may be the

result of intentional waste disposal underground(landfilling) or in surface soils (surface applicationor landfarming), or i t may be caused byunintentional waste releases from spills or leaks.

2. Landfilled wastes may become mobile if they arenot contained in impervious containers, or if thecontainers are leaking. Release of such wastesmay contaminate subsoils, ground water (throughpercolation), or air (through volatilization).

3. Landfilled wastes will be covered with soil;however, soil cover will not necessarily isolatewastes from the environment. If the cover can bepenetrated by rainwater or run-on, wastes canbe leached from the landfill cells and contaminatesubsoils, ultimately reaching ground water.Similarly, the soil cover may not be deep enoughto prevent the migration of volatile contaminantsinto the atmosphere. Estimations are that 60percent of hazardous waste is in liquid (sludge)form (USEPA 1980a). Infiltrating rainwater canincrease the migration rate of liquid or semiliquidmaterials by increasing the hydraulic headaffecting them, as well as by the leaching of toxiccomponents. Such factors as erosion or extremedrying (and cracking) can reduce the ability of asoil or clay cover to maintain the isolation ofwastes. Also, contaminated soil may cover thewaste cells themselves. When evaluating thepotential for landfill releases, current conditions,along with the long-term integrity of the landfilland its soil cover, should be evaluated. If thelandfill soil cover does not assure long-term

* For the purposes of this manual, contaminant “release” isdefined as any process that results in migration of contaminantsacross the site boundary. Within this context, volatilization,generation of surface runoff, or leachate, are considered to berelease mechanisms. Contaminant transport equates with thoseprocesses that carry released contaminants to points distantfrom the site.

Table 2-1. Potential Contaminant Release1 MechanismsMedia directly affected

Process (media indirectly affected) Timeframe

Volatilization Air Chronic

Overland flow2 Soils, surface water (ground water) Chronic, episodic

Direct discharges Soils, surface water (ground water) Chronic, episodic

Leachate generation4 Soils, ground water Chronic

Fugitive dust generation5 Air Chronic, episodic

Generation of surface runoff Soils, surface water (ground water) Chronic, episodic

Combustion3 Air Episodic

1See Section 2.2 for a definition of contaminant “release” as used in this manual.2lmpoundment overflow/failure, drum leakage, etc.3lncludes on-site treatment releases (e.g., wastewater/runoff treatment, incineration).4Buried wastes, wastes stored above ground (leaks), land application, lagoons.5contaminated soils, particulate wastes.

8

isolation of the wastes, one should evaluate theleachability and volatilization potential of thelandfilled wastes.

4. At some hazardous wastes sites, toxic materialsmay have been purposefully incorporated intosur face so i l s to p romote the i r m ic rob ia ldestruction. In such cases, toxic components maystill remain in the soil. At most sites, however,surface soils have become contaminated becauseof hazardous material spills or leaks duringmanufacturing, processing, storage, or transferoperations. In these situations, the potential forrelease of contaminants in surface soils throughfour mechanisms should be evaluated. Thesemechan isms a re : (1 ) re lease o f vo la t i l ecomponents to the atmosphere (via evaporation);(2) release of toxic particulate matter (via winderosion); (3) surface runoff -related releases; and(4) percolation of contaminants or leachate toground water.

5. The percolation of contaminated runoff maycontaminate surface soils and underlying groundwater. Surface water systems may be similarlydegraded by contaminated runoff inflow. Runoffmay also serve as a source of volatilizationre lease to a i r , a l though re leases f romcontaminated soi ls would be expected toconstitute a greater threat than that fromcontaminated runoff. Hydrophobic wastes maycontaminate surface waterbodies by adsorbingonto soil material that can be eroded from the siteand enter a waterbody in surface runoff. In awaterbody, sediment transport is much slowerthan is water movement, and contaminatedsediments may remain in the vicinity of thecontamination source for a long time.

6. Under high wind conditions, wind erosion maycarry solid particulate wastes or soil particles withsorbed hydrophobic toxic materials from the site.

7. If the site is accessible, direct contact withcontaminants may occur. Also children mayingest some contaminated soil during play. Suchingestion 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-GroundThe following numbered paragraphs help to interpretand apply the steps of the contaminant releasedecision network presented in Figure 2-2. Eachparagraph refers to a particular numbered box in thefigure.

1 . Wastes can be s to red above-ground inlagoons/ponds, in containers (drums, tanks), or inpiles. Unless containers effectively isolate wastes

2.

from the environment, above-ground storagec a n r e s u l t i n t h e d i r e c t i n t r o d u c t i o n o fcontaminants into air, soils, surface water, orground water.

Lagoons may introduce hazardous materials tothe environment through a number of pathways.Erosion or overflow resulting from heavy rainfallcan breach the lagoon and result in the outflow ofliquid wastes that contaminate surface soils,ground water, and surface waterbodies. Inaddition, unlined lagoons may introduce toxicsdirectly into ground water via percolation throughthe lagoon bottom. Also, because lagoons areuncovered, the release of volatile compounds tothe atmosphere is a common problem.

3. Wastes stored above-ground in containers maynot be effectively isolated from the environment.Over time, container corrosion and leakage occur.Leaked wastes will contaminate soils in thestorage area; may percolate to ground water; ormay contaminate surface runoff, which, in turn,can extend the area of soil contamination or canenter local surface waterbodies. Leaked materialsmay also evaporate into the atmosphere.

4. If the site is accessible to the public, directcontact with contaminants may occur. Also,children may ingest contaminated soils, eitherinadvertently or as a result of pica behavior.

2.3 Quantitative Analysis of Atmos-pheric Contamination2.3.1 Fugitive Dust Emission AnalysisEmissions of contaminated fugitive dusts (airbornewastes or contaminated soil particles) originating atuncontrolled hazardous waste sites can result from acombination of such factors as (1) wind erosion ofwastes and contaminated soils, and (2) vehiclestraveling over contaminated, unpaved roads.

Methods for analyzing such contaminant releases arepresented below.

2.3.1.1 Beginning Quantitative AnalysisThe following procedures are useful in estimating totalfugitive dust releases likely to result from the twofactors cited above. Once total suspendible dustgeneration levels have been calculated using theseequations, one can project the amounts of hazardoussubstances expected to enter the atmosphere infugitive dust using either of the following approaches:

?? Multiply the amount of dust generated by theweight percent of the toxic substance in soil orwaste. This approach does not take into accountfactors relating to such aspects as particle size oradsorption potential, which can affect the amount

10

of contaminant actually entering the atmosphere provides a conservative estimate of contaminatedas dust. fugitive dust release.

? Multiply the estimates for total dust generation bypercentages (by weight) of the substances ofconcern in actual fugitive dust samples obtainedwith on-site air monitoring. This approach takesinto account those chemical-specific and site-s p e c i f i c f a c t o r s t h a t a f f e c t r e l e a s e o fcontaminated dust in the field.

A series of publ icat ions issued by the U.S.Department of Agriculture provides directions forapplying this equation to a site-specific situation.Craig and Turel le (1964) present est imat ionprocedures for the Great Plains; Haynes (1966)addresses the Northeast; and Skidmore and Woodruff(1968) offer procedures for the entire nation.

(1) Wind Erosion Analysis*Wind e ros ion o f agr i cu l tu ra l so i l s and , byextrapolation, other disturbed soils, depends upon avariety of factors. These include surface roughnessand cloddiness; surface soil moisture content, kind,amount (and orientation, if applicable) of vegetativecover: wind velocity; and the amount of soil surface(length) exposed to the eroding wind force. The U.S.Soil Conservation Service (SCS) has developed amethod to estimate wind erosion based on a series ofgraphs relating variables presented below. Thegraphical method for calculating wind erosion basedon the functional relationship of these variables is notpresented in this manual; instead, the analyst isdirected to the Skidmore and Woodruff (1968) sourcedocument.

Although it is strongly recommended that site-specific soils data be obtained for each site underevaluation, it is not necessary to do so in order toobtain parameter values for use with the wind erosionequation (or other fugitive dust generation equations).Instead, when necessary, soils data can be obtainedfrom the local Soil Conservation Service office. SCShas on record a range of pertinent soils data for sitesacross the country where soil surveys have beenconducted. In addition, SCS maintains an extensivecomputerized soil properties data base called theSoils 5 File. This data base lists estimated soils data,based on surveys of surrounding soils properties, forareas where surveys have not been conducted todate. These data are readily available from local SCSofficials. Users of this manual should consult SCS toobtain more detailed information regarding the natureand accessibility of information contained in the soilsurveys and the Soils 5 File.

E = f(I’, C’, K’, L’, V)

where

(2 -1 )

E = potential annual average wind erosion soilloss, (mass/area/time).

I ' =soil erodibility 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 arelease rate in units of mass per time.

Table 2-2 identifies the factors that determine thevalues of the five variables used in the SCS equation.Note that the vegetative cover factor (V) specificallyapplies to crop residues, and care must be takenwhen extrapolating to the cover conditions present atuncontrolled waste sites. For Remedial Investigationand Feasibility Study estimation purposes, one canuse a “zero pounds per acre” vegetative cover value.This assumes a worst-case situation (from avegetation-related wind attenuation perspective) and

The SCS wind erosion equation is one of a number ofapproaches for estimating particulate release fromabandoned hazardous materials facilities. One suchsource (Cowherd et al. 1985) is specifically designedto guide rapid (less than 24 hours) evaluation of thepotential degree of particulate emission fromuncontrolled hazardous waste sites. This methoduses an emission factor approach to estimate releaseand procedures adapted f rom computer izeddispersion models for approximating concentrationisopleths. Concentration estimates and Bureau of theCensus data are used to identify the exposedpopulation and estimate the level of exposure. Thisapproach includes the three key components ofexposure analysis: re lease ra te es t imat ion ,contaminant migration analysis, and populationexposure determination. However, Cowherd et al.(1985) caution that their method is designed foremergency eva lua t ions o r as a p re l im inaryassessment tool, which may then be used inunder tak ing a more de ta i led inves t iga t ion .Nevertheless, the degree of accuracy attained usingthis method is consistent with simplified quantitativeestimation procedures. This approach provides theanalyst with estimates of short-term (worst-case,24-hour) release and exposure estimates, as well aslong-term (average annual) estimates.*

* Applied to nonadhering, noncompacted contaminated soil orwaste materials.

The SCS wind erosion equation is designed toprovide annual erosion losses only, and cannot be

12

Table 2-2. Environmental Variables and Model Parameters for the Wind Erosion EquationEquivalent SCS wind erosion equation primary wind erosion variables Parameters

Soil erodibility index, I (function of soil particle size distribution; readfrom a table) Soil and knoll erodibility, I’ (equal to I x I,)

Knoll erodibility, Is (function of knoll slope steepness; read from agraph)

Surface crust stability, Fs Disregarded-crust is transientSoil ridge roughness, Kr, (function of height, width, and spacing of Soil ridge roughness factor, K’ (estimated by comparison to a set of

clods and furrows) standard photographs Included in SCS wind erosion equationusers’ manuals)

Annual average wind velocity, v (read from map)Surface soil moisture, M (estimated using Thornthwalte’s (1931) Local wind erosion climatic factor, C’ (may be calculated but commonly

precipitation-evaporation index) read from maps of C’)Distance across field, Df (field width in direction of primary erosive

wind) Field length, L’ (the difference between Df and Db)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 Equivalent vegetative cover, V (the product of R’, S, and Ko) - can

effectiveness of residues from different crops) often be assumed = 0 for abandoned waste sites (see text)Orientation of vegetative cover, K0 (factor relating erosion reduction to

standing vs. fallen crop residues)

Source: Smith et al. 1982.

reliably altered to generate short-term estimates.** Inaddition, it cannot be used with data delineatingclimatic extremes for a given location, but must bebased on average annual climatic data. Instead, forexposure assessment purposes the short-termrelease, estimated using the wind erosion equation, isassumed to equal the average release over the firstyear following site investigation.

The user of this manual should review Cowherd et al.(1985) and compare that method with the SCS winderosion procedure before selecting an analyticalapproach for estimation of particulate contaminantrelease and related exposure. The analyst can alsorefer to USEPA (1983b), Farino et al. (1983) Sehmel(1980), and Smith et al. (1982) for a review of otherpossible approaches.

As noted in USEPA (1983b), the SCS wind equationcomputes the total wind erosion soil loss resultingfrom the combination of surface creep, saltation, andsuspension. Although appropriate for studies ofagricultural soil loss, in exposure evaluations theanalyst is generally concerned only with that fractionof the total soil loss that consists of particles ofsuspendible, wind transportable, and inhalable size.When the wind erosion equation is used to estimatecontaminated fugitive dust exposure situations, the

* Note: EPA (1985c) defines short-term concentrations toequate with a 10- to 90-day period. Thus, the 24-hourmaximum exposure may not adequately represent subchronicexposures.** 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 erosionequation must be adjusted (reduced) to reflect onlythat portion of the total soil loss that is suspendibleand transportable over significant distances by wind.

Considerable discussion of the cut-off point forsuspendible soil particle size exists in the literature(Sehmel 1980, Smith et al. 1982, and USEPA 1983a.b). As a group, particles < 100 µm aerodynamicequivalent diameter include those that can besuspended by and transported in the wind and thosethat can be inhaled (see Miller et al. 1979 and USEPA1986d for a discussion of the extent to which variousparticle sizes penetrate the human respiratorysystem). Particles in the 30 to 100 pm diameter rangewill often settle within a few hundred feet of thesource (USEPA 1983a), while those particles < 30pm in diameter can be transported considerabledistances 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-µm size range, theparticle size distribution of the parent soil determinesthe size distribution of suspended particles (Smith etal. 1982). Therefore, that proportion which is < 10 µmin diameter can be determined based on the soil sizedistribution of the parent soil. It can be assumed thatthis proportion of the total soil loss, as calculated viathe SCS wind erosion equation, is lost to suspensionand is available for inhalation.

(2) Unpaved Roads AnalysisThe following equation (USEPA 1983a) can be usedto estimate fugitive dust releases associated withvehicles traveling on contaminated unpaved roads.

13

where

EVT = emission factor for vehicular traffic, (lb/vehiclemile traveled: kg/vehicle kilometer traveled)

k = 0.45 = particle size multiplier for particles< 10 pm (i.e., particles that may remainsuspended once they become airborne andwhich can be inhaled into the respiratorysystem).*

s = silt content (of road surface material),(percent).”

Sp = mean vehicle speed, (mph; kph).w = mean vehicle weight, (tons; Mg).w = mean number of wheels.Dp = number of days with at least 0.254 mm (0.01

in) of precipitation per year (see Figure 2-3).

The emission factor (EVT) can be multiplied by a“vehicle kilometers traveled per time” value togenerate a “dust release per time value.” Short-term (maximum release) estimates can be made byusing a reduced value of “Dp” in the equation toreflect assumed drought conditions at the site. Figure2-3 reflects the range of average “Dp” values forlocations in the U.S. Consultation with the localNational Weather Service office may provide locale-specific insight into what “Dp” values should be usedto represent dry years at the site. Long-term(average) releases can be estimated by using theannual average value for “Dp.” USEPA (1983a)states that this equation is valid for situations thatcomply with the following source conditions:

? Road surface silt content = 4.3 - 20 percent;?? Mean vehicle weight = 3-157 tons (2.7-142

Mg);?? Mean vehicle speed = 13-40 mph (21-64

kph); and?? Mean number of wheels = 4-13.

For an overview of the utility and limitationsassociated with the application of emission factors toparticulate release estimation problems, the user of

*See EPA (1983a) for “k” values used when release ofspecific particle size groups other than < 10 pm is desired.**Soil silt content can be estimated from SCS Soils 5 File databy subtracting the “percent clay” value from the “percentmaterial passing No. 200 sieve” value. (Personalcommunication between Lee Schultz (Versar Inc.), and KeithYoung (U.S. Department of Agriculture, Soil ConservationService), Washington, D.C., May 1, 1984.)

this manual can refer to USEPA (1983a, b), Farino etal. (1983) Sehmel (1980), and Smith et al. (1982).

2.3.1.2 In-Depth AnalysisFor contaminated fugitive dust emissions, in-depthanalysis will consist of monitoring and modelingactivities. Generally, air sampling will be conducteddownwind and upwind of the uncontrolled hazardouswaste site. The difference in particulate loadingobtained at the two (or more) sampling locations willquantify the particulate mass loading attributable tothe site alone (assuming that air sampling stationscan be sited to eliminate interference from othersources). Using these data, either simple dispersionequations or computerized air dispersion modeling*can be used to back-calculate the emission level ata “virtual point source.” The use of dispersionmodeling to back calculate emission levels, however,is often quite unreliable because of the difficulty inob ta in ing accura te amb ien t mon i to r ing andmeteorological input data.

The virtual point source is a hypothetical sourcelocated upwind of the subject site that has ahypothetical release rate which would result in thecontaminant concentrat ions observed at theuncontrolled hazardous waste site (area source). Thevirtual point source release rate can then be used insubsequent contaminant transport analysis for thesubject site. The user of this manual should refer toUSEPA (1983c) and Seely et al. (1983) for a detailedpresentation of ambient air sampling strategies andprocedures appropriate for abandoned hazardouswaste facilities.

2.3.2 Volatilization Emission AnalysisVolatil ization of contaminants at uncontrolledhazardous waste sites can occur at the followingsources:

(1) Covered landfills - without internal gasgeneration;

(2) Covered landfills - with internal gasgeneration;

(3) Spills, leaks, landfarms - concentratedwastes on the surface or adhered to soilparticles below the surface; and

(4) Lagoons - wastes dissolved in or mixedwith water.

In the baseline situation, one or more of thesesources will contribute to the overall air loadingoriginating at the site, and will need to be controlledthrough remedial action.

2.3.2.1 Beginning Quantitative AnalysisThis section presents simplified analytical proceduresfor estimating releases from the above sourcecategories. Because the chemical properties of agiven substance largely determine the volatilizationrate, the equations presented require input of

14

quantified property values. These data are availablefor many chemicals that may be present atuncontrolled hazardous waste sites, and are found invarious chemical reference texts. In cases wherechemical data are missing, the analyst must estimatethe property values. This section provides equationsfor estimating certain requisite chemical properties.Comprehensive guidance for chemical propertyestimation is provided in reference materials such asLyman et al. (1982). Readily accessible computerizedsystems are available to predict a range of pertinentchemical properties. The computerized GraphicExposure Model ing System (GEMS), and i tssubsystem CHEMEST, is an example. The EPAOffice of Toxic Substances in Washington, D.C. hasdeveloped and is managing this system. Essentially acomputerized version of Lyman et al. (1982), it canbe rapidly accessed to estimate the chemicalcharacteristics necessary for volatilization estimation.

The user of this manual can refer to Farino et al.(1983) for a detailed review and evaluation of existingequat ions fo r es t imat ing vo la t i l i za t ion f romuncontrolled hazardous waste sites. This reportpresents a survey of available air release models forvolatile substances and a critical analysis of theapplications and limitations of each.

(1) Landfills Without Internal Gas GenerationEquation 2-3 can be used to estimate volatilereleases from covered landfills containing toxicmaterials alone, or toxic materials segregated fromother landfilled nonhazardous wastes. Equations 2-4through 2-7 are used to calculate certain inputvariables that are required to apply Equation 2-3.Farmer et al. (1978) developed an equation toestimate the effectiveness of various landfill covertypes and depths in controlling volatile releases. Thisequation, based on Fick’s First Law of steady statediffusion, assumes that diffusion into the atmosphereoccurs at a plane surface where concentrationsremain constant. It ignores biodegradation, transportin water, adsorption, and production of landfill gas.Diffusion of the toxic vapor through the soil cover isthe controlling factor. It also assumes that there is asufficient mass of toxicant in the landfill so thatdepletion of the contaminant will not reduce theemission rate.

Equation 2-3, simplified by Farmer et al. (USEPA1980b), incorporates a number of assumptions (seeFarino et al. 1983 for a complete discussion), such ascomple te ly d ry so i l (wors t case) and zero

*Although computerized dispersion modeling can be used toobtain contaminant release rates, it is primarily a tool fordetermtning contaminant atmospheric fate. Thus, refer toChapter 3, Environmental Fate Analysis, for detaileddiscussions of air dispersion models applicable to uncontrolledhazardous waste facilities.

concentration of volatilizing material at the soilsurface. Shen (1981) converted Farmer’s simplifiedequation for calculating the vapor flux rate to a formthat provides a toxic vapor emission rate bymultiplying the basic equation by the exposedcontaminated surface area. In addition, Shen modifiedthe equation to allow calculation of the volatilizationrate of a specific component of the overall toxicmixture by multiplying by the weight fraction of thecomponent in the mixture. However, as pointed outby Farino et al. (1983), a more accurate approachwould be to multiply by the mole fraction of the toxiccomponent in the buried mixture. Thus, Farmer’sequation, as modified by Shen (1981) and Farino etal. (1983) is:

SC (2-3)

where

E i = emission rate of component i, (g/sec).D i = diffusion coefficient of component i in air,

(cm2/sec).Csi = saturation vapor concentration of component

i, (g/cm3).A = exposed area, (cm2).Pt

= total soil porosity, (dimensionless).M i = mole fraction of toxic component i in the

waste, (gmole/gmole).dsc = effective depth of soil cover, (cm).

Note that total soil porosity, rather than air-filled soilporosity, is used in this equation. The presence ofwater in a soil cover will tend to decrease the flux rateof a volatile compound by effectively decreasing theporosity, and also by increasing the geometriccomplexity of the soil pore system (because wateradheres to soil particles), thus effectively increasingthe vapor path (USEPA 1980b). Farmer et al.suggest, however, that when using their equation todesign a landfill cover, the total porosity value beused (USEPA 1980b), thereby designing for the worstcase (i.e., dry conditions). In most instances, it will beappropriate to apply this same worst-case logic tothe analysis of volatilization release from landfilledwastes, assume that landfill cover soils are dry, anduse a value for total porosity in Equation 2-3. It isrecognized, however, that there may be situationswhere it can be shown that cover soils exist in a wetcondition more often than in a dry one. In thesecases, the air-filled soil porosity (Pa) may be moreappropriate, and this value can be substituted for Ptin Equation 2-3 when analyzing volatilization release.

If not provided in existing literature, Di, a compound’sdiffusion coefficient (required for the above equation),can be calculated by Fuller’s Method (Perry andChilton 1973):

16

To estimate short-term (maximum) release rates,use a value for the temperature that reflects theexpected summer maximum temperatures. Annualaverage temperatures should be used to initiallyestimate long-term (average) release rates. Thisinitial estimated long-term release value will berevised as described in Section 2.3.3 to develop finallong-term release estimates.

Relevant atomic diffusion volumes for use inestimating Di are (Perry and Chilton 1973):

C = 16.5 Cl = 19.5 Aromatic ring = -20.2H = 1.98 Br = 35.0 Heterocyclic ring = - 20.2O = 5.48 F = 25.0*N = 5.69 S = 17.0

Table 2-3 presents diffusion coefficients that havebeen calculated for a variety of compounds, some ofwhich may be present at abandoned sites.

An alternative method (Shen 1981) for approximatingDi involves the identification of a compound listed inTable 2-3 that has a molecular weight and moleculardiffusion volume (calculated) similar to those of thetoxic substance under evaluation. The unknowndiffusion coefficient can then be calculated using:

where

(2-5)

D i = diffusion coefficient of the compound tobe estimated from the known D’.

D’ = diffusion coefficient of a compound thatcan be found in the table, the molecular

* This value is from Shen (1981).

weight and atomic diffusion, volume ofwhich are close to that of the unknown.

MW’ = molecular weight of the selectedcompound D’.

MWi= molecular weight of the compound to

be estimated.

Total soil porosity, Pt, can be calculated as follows(USEPA 1980b):

(2-6)

where

P t = total soil porosity, (dimensionless).B = soil bulk density,* (g/cm3): generally

between 1.0 and 2.0 g/cm3.P = particle density, (g/cm3): usually 2.65

g/cm3 used for most mineral material.

F o r e s t i m a t i o n , P t c a n b e a s s u m e d t o b eapproximately 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-f i l led soi l porosity (Pa) when analyzing thevolatilization release from soils with a high moisturecontent (Shen 1981). Alternatively, the local SoilConservation Service office can be contacted toobtain site-specific estimated air-filled soil porosityvalues for specific locations.

Saturat ion vapor concentrat ion, C s i , can bedetermined by (USEPA 1980b):

(2-7)

where

C s i= saturation vapor concentration of

component i, (g/cm3).P = vapor pressure of the chemical,” (mm

Hg).MWi = mole weight of component i, (g/mole).

R = molar gas constant, (62,361 mm Hg-cm 3/mole- 0K).

T = absolute temperature, (K).

Again, use maximum summer temperatures toestimate short-term release and annual averagetemperatures to initially estimate long-term release.

* Values for soil bulk density for specified locations can beobtained from the U.S. Soil Conservation Service, Soils 5 Filedata base.** If the vapor pressure of a chemical under consideration is notavailable in standard reference texts, estimate it as described inLyman et al. (1982).

1 7

Table 2-3. Diffusion Coefficients of Selected Organic CompoundsAtomic Diffusion coefficients (cm2/sec)

Molecular diffusionCompound Formula weight volume at 10°C at 30°C at 500C

Acetaldehyde C2H40Acetic acid C2H402Acetone C3H60Aniline C6H7NBenzene C6H6Bromoethane CH3BrBromoform CHBr3

Carbon tetrachloride CCI4Chlorobenzene C6H5CIChloroethane C2H5CIChloroform CHC13

Chloromethane CH3CICyclohexane C6H12Dichloroethane C2H4CI2Dichloroethylene C2H2Cl2DicchloropropaleneDimethylamrne

C3H6CI2C2H7N

EthanolEthyl acetateEthylamine

C2H60C4H802

C2H7NEthylbenzeneFluorotoluene

C8H10

C7H7FHeptaneHexanelsopropanolMethanolMethyl acetateMethyl chlorideMethylethyl ketonePCB (1 Cl)PentanePhenolStyreneTetrachloroethaneTetrachloroethyleneTolueneTricyhloroethaneTrichloroethyleneTrichlorofluoromethaneVinyl chlorideXylene

C7H16

C6H14

C3H80CH40C3H602CH2CI2C4H80C12HGCIC5H12

C6H60C8H8

C2H2CI4C2CI4C7H8

C2H3CI3C2HCl3CCI3F

C2H3CI

C 8 H 1 0

4 4 46.40 .11758 .13249 .14816

60 51.88 .10655 .12007 .13427

58 66.86 .09699 .10930 .12223

93 118.55 .07157 .08065 .09019

78 90.68 .08195 .09234 .10327

95 57.44 .09611 .10830 .12111

118 53.48 .09655 .10880 .12167

154 94.50 .07500 .08451 .09451

113 128.40 .06769 .07627 .08530

65 62.40 .09789 .11031 .12336

120 76.89 .08345 .09404 .10517

51 57.94 .10496 .11827 .13226

84 122.76 .07139 .08045 .08996

99 75.96 .08557 .09643 .10784

97 106.96 .07442 .08386 .09377

113 100.38 .07519 .08473 .09475

45 52.55 .11161 .12577 .14065

46 50.36 .11297 .12730 .14236

88 92.80 .07991 .09005 .10070

45 52.55 .11161 .12577 .14065

116 151.80 .06274 .07070 .07906

110 154.36 .06262 .07056 .07891

100 146.86 .06467 .07287 .08149

86 126.72 .07021 .07912 .08848

60 37.82 .12004 .13526 .15126

32 29.90 .14808 .16686 .18660

74 72.34 .09054 .10203 .11410

85 59.46 .09610 .10830 .12111

72 87.32 .08417 .09485 .10607

189 235.32 .04944 .05571 .06230

72 106.26 .07753 .08737 .09770

84 96.16 .07919 .08924 .09980

104 137.84 .06620 .07460 .08343

168 1143.96 .06858 .07729 .08643

166 111.00 .06968 .07852 .08781

92 111.14 .07367 .08301 .09283

133 97.44 .07496 .08447 .09446

131 93.48 .07638 .08606 .09625

138 100.00 .07391 .08329 .09314

63 58.44 .10094 .11375 .12720

106 131.60 .06742 .07597 .08495

Source: Shen 1981

18

See Section 2.3.3 for directions for calculating a final 14 provide means of estimating certain input variableslong-term release rate. required to solve Equations 2-9 and 2-11.

(2) Landfills with Internal Gas GenerationThibodeaux (1981) developed a method for estimatingtoxic vapor releases from co-disposal landfills.These facilities contain toxic wastes in combinationwith municipal or sanitary wastes that, because oftheir considerable organic content, generate landfillgases (e.g., H2, CH4, C02). In these cases, theupward movement (convective sweep) of the landfillgas becomes the significant controlling factor, greatlyaccelerating the upward migration and subsequentrelease to the atmosphere of the co-disposed toxicsubstances. In fact, review of Thibodeaux’s workindicates that the toxic gas migration accelerating theeffect of the landfill gas is so great that both soil andgas phase diffusion essentially become insignificant.The following simplified equation is recommended forestimating the volatilization of toxic substances fromco-disposal landfills:

As discussed in Farino et al. (1983) one can applyEquation 2-9 (adapted from Thibodeaux and Hwang1982) to estimate volatile releases resulting fromspills or leaks where a contaminant pool is visible onthe soil surface, or where soil is contaminated(saturated) from the surface down. The equation doesnot consider soil phase mass transfer resistance, andtherefore is not appropriate for use when spilledcontaminants have seeped into surface soils (in thiscase, use the landfarming equation that follows).Similarly, because it does not consider liquid phaseresistance, it is only useful for estimating releases ofpure compounds. The original equation presented inThibodeaux and Hwang (1982) has been modified toinclude a contaminated surface area term, therebyresulting in the calculation of a release rate ratherthan a flux rate value:

Ei = Ci*VyA

where

(2 -8 )

Ei = kiGCi*A

where

(2 -9 )

Ei = emission rate, (g/sec).= concentration of compound i in the soil

pore spaces, (g/cm3)*

Vy = mean landfill gas velocity in the soil

pore spaces, (cm/sec). Thibodeaux(1981) provides an average value of1.63 x 10-3 cm/sec for this factor.

A = area, (cm2).

E i = emission rate of chemical i, (g/s).k i G

= gas phase mass transfer coefficient ofchemical i, (cm/s).

C i* = vapor concentration of chemical i,

(g/cm3).*

A = area, (cm2).

Hwang (1982) has developed the following simplifiedmeans of estimating a compound’s gas phase masstransfer coefficient.

where

(3) Spills and LeaksEquations 2-9 and 2-11 will estimate the volatilereleases from fresh and old (respectively) chemicalspills on soil. Equations 2-10 and 2-12 through 2-

Ki, = g a s p h a s e m a s s t r a n s f e rcoefficient of chemical i, (cm/s).

MWH2O; MWi = molecu la r we igh t o f wa te r ;compound i, (g/mole).

T = temperature, (0K).k i G , H 2 0 = g a s p h a s e m a s s t r a n s f e r

coefficient for water vapor at250C, (cm/sec).

* For conservative analyses, one can assume that the actualcontaminant vapor concentration in the soil pore spaces is thesame as the equilibrium vapor concentration. In such cases,C can be used in place of Ci

*. Direct measurements of Ci*,

h&ever, may be developed during the site investigation. Whensuch data are available, their use is preferred.

* For conservative analyses, one can assume that the actualcontaminant vapor concentration in the soil pore spaces IS thesame as the equilibrium vapor concentration. In such cases, Csican be used in place of C i

*. Direct measurements of C i*,

however, may be developed during the site investigation. Whensuch data are available, their use is preferred.

19

When estimating short-term (maximum) releaserates, the highest (summer) seasonal temperatureexpected at the site can be used in calculating thegas phase mass transfer coefficient. For initialestimation of long-term release rates, the seasonalaverage temperature should be used. Final long-term release notes are developed as discussed inSection 2.3.3.

(4) LandfarmingIn cases where past spills, leaks, or intentionald isposa l d i rec t l y on to o r in to sur face so i l s(landfarming) have resulted in contaminated surfacesoils with liquids in the pore spaces, Equation 2-11can be used to estimate volatilization releases. Thisequation assumes that soil pore spaces connect withthe soil surface, that soil conditions are isothermal,and that there is no capillary rise of contaminant. Ita lso assumes tha t there i s su f f i c ien t l i qu idcontaminant in the pore spaces so that volatilizationwill not deplete the reservoir of contaminant to thepoint where it affects the rate of volatilization.Model ing the release from soi ls with sorbedcontaminants and no free liquids requires anothermodel.

Two models for predict ing the t ime-varyingvolatilization of sorbed contaminants on soil arepresented in USEPA (1986e). The equation presentedhere is adapted from Thibodeaux and Hwang (1982),which presents a volatilization release estimationequation designed for application to active or plannedlandfarms for petroleum wastes. Farino et al. (1983)determined it to be preferable to other approaches forestimating volatilization release of chemicals spilled orincorporated into soils, because it directly takes intoaccount the contaminant loss over time. It describesvapor diffusion as being soil-phase controlled, andessentially assumes that contaminant concentrationsin the soil remain constant (until all contaminant islost to the air), and that contaminant release occursby the “peeling away” of successive unimolecularlayers of contaminant from the surface of the “wet”(contaminated) zone. Thus, over time this processresults in a “dry zone” of increasing depth at the soilsurface, and a wet zone of decreasing depth belowthe dry zone. The original equation has been adjustedsomewhat for use at uncontrolled waste sites, andhas also been simplified as discussed in Farino et al.(1983), by assuming that the oil layer diffusion lengthvalue is low (i.e., that the spilled contaminant hasbecome incorporated into surface soils and is notpresent as a discrete film).

E i =2DCsA

(2-11)

where

D (cm2/sec) is related to the amount of contaminant ithat goes from liquid to gas phase, and then from gasphase to diffusion in air. It can be estimated asfollows:

D=D i

where

D = phase transfer coefficient, (cm2/sec).D i = diffusion coefficient of component i in

air, (cm2/sec).P t = total soil porosity, (dimensionless).

Again, use of total soil porosity in thisequation results in a worst-case (drysoil) estimate for D. As previouslydiscussed, however, in some cases(i.e., where soils are wet more oftenthan dry) it may be more appropriate touse air-filled soil porosity (Pa) in placeof Pt. See text addressing Equation 2-3 for a discussion of the application ofand values for these two terms.

Hi’ = Henry’s Law constant in concentrationform, (dimensionless).

Hi’, the Henry’s Law constant in concentration form(ratio of the boundary layer concentration ofcontaminant in air to the boundary layer concentrationof contaminant in “wet” soil) can be determined asfollows (Lyman et al. 1983):

(2-13)

Hi = Henry’s Law constant of contaminant i,(atm-m3/mol).

R = gas cons tan t , , (8 .2 x 10 - 5 a tm-m 3/mol- 0K).

T = absolute temperature, (0K).

Again, use summer maximum temperatures toestimate short-term release and annual average

20

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

Note tha t Equat ion 2 -11 assumes tha t thecontaminant concentration in the liquid and gasphases in the soil remains constant until all of thecontaminant has been released to air. Also, theequation holds from time zero (the time at which thesoil was sampled) to td (the time at which the soilbecomes dry, i.e., all contaminant has volatilized andthe release process stops). The formula forcalculating td (in seconds) is:

(2-14)

where

t d = the time at which all contaminant hasvolatilized from the soil, (sec).

h = depth from soil surface to the bottomof the contaminated region, (cm).

d = depth of dry zone at sampling time,(cm).

CBD = phase transfer coefficient, (cm2/sec).= bulk contaminant concentration in soil,

(g/cm3)Cs = contaminant liquid phase concentration

(g/cm3)(5) LagoonsMackay and Leinonen (1975) have developed anequation for estimating volatilization releases of lowsolubility compounds from waterbodies such ashazardous waste lagoons. This is presented asEquation 2-15. Equations 2-16 and 2-17 providemeans of calculating certain input parametersrequired by Equation 2-15. This approach assumesthat conditions are steady state (i.e., no constantaddition of contaminant), that diffusion is liquid statecontrolled, and that it occurs from a well-mixedwater phase to a well-mixed air phase across astagnant water/air interface. As pointed out in Farinoet al. (1983), if it can be assumed that atmosphericbackground levels of the contaminant of concern arenegl ig ib le, (as would usual ly be the case atabandoned hazardous waste facilities), then Mackayand Leinonen’s basic equation can be simplified tothe following form (which includes an area term toconvert flux rate to emission rate):

Ei = KiCsA (2-15)

where

EiK i

= emission rate, (g/sec).= overal l mass t ransfer coeff ic ient ,

(cm/sec).

Cs = contaminant liquid phase concentration,(g/cm3)

A = area, (cm2).

The overall mass transfer coefficient (K i) can becalculated via the following relationship:

1 1 RT

H i k i G

(2-16)

Ki = overal l mass t ransfer coeff ic ient ,(cm/sec).

k i L = liquid phase mass transfer coefficient,(cm/sec). See Equation 2-17.

R = ideal gas law constant, (8.2 x 10 -5

atm-m 3 /mol -0K).T = temperature, (0K).

H i = Henry’s Law constant of compound i,(atm-m3/mol).

kiG = gas phase mass transfer coefficient,(cm/sec). See Equation 2-10.

Hwang (1982) provides a simplified method fordetermining a compound’s liquid phase mass transfercoefficient for use in the above equation. To estimatekiL, use the following equation:

(2-17)

where

k i L = l i q u i d p h a s e m a s s t r a n s f e rcoefficient, (cm/sec).

MW02;MW i = mo lecu la r we igh t o f oxygen ;compound i.

T = temperature, (0K).kL,O2 = l i q u i d p h a s e m a s s t r a n s f e r

coefficient for oxygen at 250C,(cm/sec).

The value for kL,O2 can be obtained from chemicalreference texts or can be calculated (the preferredmethod) as described in Farino et al. (1983).

2.3.2.2 In-Depth AnalysisIn-depth analysis of volatile release can be executedin the same manner as that described for particulates.Subtract the monitored upwind (control) ambient toxicvapor concentration from the monitored downwindconcentration. Use the difference between these twovalues in an air dispersion model to estimate therelease rate at a “virtual point source” that wouldcorrespond with the source of the measureddownwind concentration.

The user of this manual should again refer to USEPA(1983c) and Seely et al. (1983) for detaileddiscussions of the planning and execution of airmonitoring studies. Refer to Chapter 3 of this manual

21

for a description of air contaminant dispersionmodeling tools.

2.3.3 Long-Term and Short-Term Release Cal-culationLong-term release values (70 years) for lagoons withdilute solutes can be estimated as follows:

Note that Vc and Ci must be based on the samevalue. (V c C i ) i s equa l to the to ta l mass o fcontaminant; it can be the total mass of contaminantin a lagoon.

For landfills and wind erosion of contaminatedparticulates, the release rate is assumed constant.The 70-year average annual release rate can becalculated by first ascertaining if contaminant willremain after 70 years. If so, then the release rateitself is the 70-year average annual release rate. Ifnot, then the 70-year average annual release rate isthe total initial mass divided by 70 years.

To estimate long-term release from contaminatedsurface soils, Equation 2-14 (converted to years bydividing by 3.16 x 107) is first used to determine thedry-out time. If no contaminant is expected toremain af ter 70 years ( i .e. , 70 > td) , s implydetermine the total amount of contaminant present atthe time of site investigation and divide by 70 years(in seconds) to get a conservative long-term releasevalue ( i .e. , AC s (h - d)/2.21 x 10 9). I f somecontaminant is expected to remain after 70 years (i.e.,70 < td), use the following equation to estimatelong-term release:

where

EAi = a v e r a g e l o n g - t e r m r e l e a s e o fcontaminant i, (g/yr).

Cs

= contaminated area, (cm2).= l i q u i d p h a s e c o n c e n t r a t i o n o f

contaminant i, (g/cm3).d = depth of dry zone at sampling time,

(cm).D = the amount of contaminant that goes

from liquid to gas phase, and then fromgas phase to diffusion in air (seeEquation 2-12).

C B = bulk contaminant concentration in soil(g/cm3)

N o t e t h a t t h i s a p p r o a c h d o e s n o t i n c l u d econsideration of contaminant loss caused by chemicaldegradation, and thus is conservative in nature.

Finally, for each chemical, sum the long-termvolatil ization release values from each on-sitesource to arrive at an overall long-term volatilizationrelease for each contaminant of concern.

Short-term maximum contaminant releases from anuncontrolled hazardous waste facility can be due to avariety of factors. For example, high summertemperatures, ambient pressures, rainfall, and breezyconditions can significantly effect the rate of volatilerelease, while high winds alone can greatly increasethe amount of contaminated particulate matter beingblown from the site. Therefore, short-term maximumcontaminant release is defined as that level of releasecalculated for release events during the first one-year period following site investigation.

The above approach to estimating long-term airreleases is inherently conservative in that it does notconsider site contaminant loss from other (non-airtransport) processes. Also, note that the short-termand long-term release values developed in thissection will be used, along with worst-case (short-term) and average (long-term) meteorological data,to develop short-term and long-term ambientconcentration values for later use in determiningexposure levels (see Chapter 3).

2.4 Quantitative Analysis of SurfaceWater ContaminationContaminated runoff , over land f low of toxiccontaminants from storage leaks and spills, or lagoonfailures will often constitute the sources of concernfor surface water contamination at uncontrolledhazardous waste sites. Projecting release rates forsuch contaminant sources can be very difficult,however. Releases from containers or impoundmentscan best be determined by on-site monitoring ofeach source. If this is not possible, engineeringjudgment, combined with a detailed evaluation of siteconditions, may provide a basis for developingrelease estimates. Releases by overland flow of

22

toxics can be quantified directly by measuring(sampling) the source material and determining thevolume and rate of release. Alternatively, runoffrelease estimation procedures, less costly thanmonitoring or modeling approaches, can also beapplied to uncontrolled sites.

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

This section reviews methods for estimating toxicreleases of uncontrolled hazardous waste sites tosurface waterbodies. Note, however, that only thesurface runoff component of release to surface wateris addressed here. Other sources must be estimatedfor each site based on judgment and experience.

2.4.1 Beginning Quantitative Analysis

2.4.1.1 Dissolved and Sorbed ContaminantMigrationMany of the organic substances of concern found atSuperfund sites are relatively nonpolar, hydrophobicsubstances (Delos et al., 1984). Such substances canbe expected to sorb to site soils and migrate from thesite more slowly than will polar compounds. Asdiscussed in Haith (1980) and Mills et al. (1982),estimates of the amount of hydrophobic compoundsreleased in site runoff can be calculated using theModified Universal Soil Loss Equation (MUSLE) andsorption partition coefficients derived from thecompound’s octanol-water partition coefficient. TheMUSLE allows estimation of the amount of surfacesoil eroded in a storm event of given intensity, whilesorption coefficients allow the projection of theamounts of contaminant carried along with the soil,and the amount carried in dissolved form.

(1) Soil Los CalculationEquation 2-20 is the basic equation for estimatingsoil loss. Equations 2-21 through 2-24 are used tocalculate certain input parameters required to applyEquation 2-20. The modified universal soil lossequation (Williams 1975), as presented in Mills et al.(1982), is:

Y(S)E = a(Vrqp)0.56 KLSCP

where

(2-20)

Y(S)E = sediment yield (tons per event, metrictons per event).

a = conversion constant, (95 English, 11.8metric).*

V r = volume of runoff, (acre-feet, m3).

q p= peak flow rate, (cubic feet per second,

m3/sec).K = the soil erodibility factor, (commonly

expressed in tons per acre perdimensionless rainfall erodibility unit). Kcan be obtained from the local SoilConservation 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 followingdiscussion for vegetated si te “C”values).

P = the erosion control practice factor,( d i m e n s i o n l e s s r a t i o : 1 . 0 f o runcontrolled hazardous waste sites).

Soil erodibility factors are indicators of the erosionpotential of given soil types. As such, they are highlysite-specific. K values for sites under study can beobtained from the local Soil Conservation Serviceoffice. The slope length factor, L, and the slopesteepness factor, S, are generally entered into theMUSLE as a combined factor, LS, which is obtainedf r o m F i g u r e s 2 - 4 t h r o u g h 2 - 6 . T h e c o v e rmanagement factor, C, is determined by the amountand type of vegetative cover present at the site. Itsvalue is “1” (one) for bare soils. Consult Tables 2-4and 2-5 to obtain C values for sites with vegetativecovers. The factor, P, refers to any erosion controlpractices used on-site. Because these generallydescribe the type of agricultural plowing or plantingpractices, and because it is unlikely that any erosioncontrol would be pract iced at an abandonedh a z a r d o u s w a s t e s i t e , u s e a w o r s t - c a s e(conservative) P value of 1 (one) for uncontrolledsites.

Storm runoff volume, Vr, is calculated as follows(Mills et al. 1982):

Vr = aAQr (2-21)

where

* Metric conversions presented in the following runoffcontamination equations are from Mills et al. (1982).

23

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

NOTE: Dashed lines are extension of LS formulae beyond valuestested in studies.

where

= the total storm rainfall, (in, cm).= water retention factor, (in, cm).

The value of SW , the water retention factor, isobtained as follows (Mockus 1972):

(2-23)

where

Sw = water retention factor, (in, cm).CN = the SCS Runo f f Curve Number ,

(dimensionless, see Table 2-6).a = conversion constant (1.0 English, 2.54

metric).

The CN factor is determined by the type of soil at thesite, its condition, and other parameters that establisha value indicative of the tendency of the soil toabsorb and hold precipitation or to allow precipitationto run off the surface. The analyst can obtain CN

Figure 2-6. Soil moisture-soil temperature regimes of thewestern United States (USDA 1974 aspresented in Mills et al. 1982).

Figure 2-6. Slope effect chart for areas where Figure 2-5 is not applicable (USDA 1974 as presentedin Mills et al. 1982).

Slope Length, Meters

10 20 40 60 100 200 400 600 1000 2000Slope Length, Feet

NOTE: Thedashed lines represent estimatesfor slopedimensionsbeyond the range of lengths and steepnesses for whichdata are available.

24

Table 2-4. “C” Values for Permanent Pasture, Rangeland, and Idle Land

Vegetal canopyType and heightof raised canopyb

No appreciable canopy

Canopy of tall weeds or shortbrush (0.5 m fall height)

Appreciable brush or brushes(2m fall height)

Trees but no appreciable lowbrush (4 m fall height)

Cover that contacts the surfaceCanopycoverc Percent ground cover

(%) Typed 0 2 0 4 0 6 0 8 0 9 5 - 1 0 0

G 0 . 4 5 0 . 2 0 0 . 1 0 0 . 0 4 2 0 . 0 1 3 0 . 0 0 3

w 0 . 4 5 0 . 2 4 0 . 1 5 0 . 0 9 0 0 . 0 4 3 0 . 0 1 1

2 5 G 0 . 0 3 6 0 . 1 7 0 . 0 9 0 . 0 3 8 0 . 0 1 2 0 . 0 0 3W 0 . 0 3 6 0 . 2 0 0 . 1 3 0 . 0 8 2 0 . 0 4 1 0 . 0 1 1

5 0 G 0 . 0 2 6 0 . 1 3 0 . 0 7 0 . 0 3 5 0 . 0 1 2 0 . 0 0 3W 0 . 0 2 6 0 . 1 6 0 . 1 1 0 . 0 7 5 0 . 0 3 9 0 . 0 1 1

7 5 G 0 . 1 7 0 . 1 0 0 . 0 6 0 . 0 3 1 0 . 0 1 1 0 . 0 0 3W 0 . 1 7 0 . 1 2 0 . 0 9 0 . 0 6 7 0 . 0 3 8 0 . 0 1 1

2 5 G 0 . 4 0 0 . 1 8 0 . 0 9 0 . 0 4 0 0 . 0 1 3 0 . 0 0 3W 0 . 4 0 0 . 2 2 0 . 1 4 0 . 0 8 5 0 . 0 4 2 0 . 0 1 1

5 0 G 0 . 3 4 0 . 1 6 0 . 0 8 5 0 . 0 3 8 0 . 0 1 2 0 . 0 0 3W 0 . 3 4 0 . 1 9 0 . 1 3 0 . 0 8 1 0 . 0 4 1 0 . 0 1 1

7 5 G 0 . 2 8 0 . 1 4 0 . 0 8 0 . 0 3 6 0 . 0 1 2 0 . 0 0 3W 0 . 2 8 0 . 1 7 0 . 1 2 0 . 0 7 7 0 . 0 4 0 0 . 0 1 1

2 5 G 0 . 4 2 0 . 1 9 0 . 1 0 0 . 0 4 1 0 . 0 1 3 0 . 0 0 3W 0 . 4 2 0 . 2 3 0 . 1 4 0 . 0 8 7 0 . 0 4 2 0 . 0 1 1

5 0 G 0 . 3 9 0 . 1 8 0 . 0 9 0 . 0 4 0 0 . 0 1 3 0 . 0 0 3W 0 . 3 9 0 . 2 1 0 . 1 4 0 . 0 8 5 0 . 0 4 2 0 . 0 1 1

7 5 G 0 . 3 6 0 . 1 7 0 . 0 9 0 . 0 3 9 0 . 0 1 2 0 . 0 0 3W 0 . 3 6 0 . 2 0 0 . 1 3 0 . 0 8 3 0 . 0 4 1 0 . 0 1 1

Source: Wischmeier 1972aAll 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.

Table 2-5. “C” Values for Woodland

Tree canopy per-Stand condition cent of areaa

Well stocked 100-75

Forest litter per-cent of areab

100-90

Undergrowthc

Managedd

Unmanagedd

“C” factor

0.0010.003-0.011

Medium stocked 7 0 - 4 0 8 5 - 7 5

Poorly stocked 3 5 - 2 0 7 0 - 4 9

Managed 0 . 0 0 2 - 0 . 0 0 4

Unmanaged 0.01-0.04

Managed 0.003-0.009Unmanaged 0.02-0.09 e

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.cundergrowth is defined as shrubs, weeds, grasses, vines, etc., on the surface area not protected by forest litter. Usuallyfound under canopy openings.dManaged - grazing and fires are controlled.

Unmanaged - stands that are overgrazed or subjected to repeated burningeFor unmanaged woodland with litter cover of less than 75 percent, C values should be derived by taking 0.7 of theappropriate 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 NumbersSoil

group DescriptionLowest runoff potential: Includes

Site typeOverall sitea Road/right of way Meadow Woods

59 74 30 45

C

D

74 84 58 66

82 90 71 77

deep sands wrth very little silt andclay, also deep, rapidly perme-able loess (infiltration rate =8-12 mm/h).Moderately low runoff potential:Mostly sandy soils less deep thanA, and loess less deep or lessaggregated than A, but the groupas a whole has above-averageinfiltration after thorough wetting(infiltration rate = 4-8 mm/h).Moderately high runoff potential:Comprises shallow soils and soilscontaining considerable clay andcolloids, though less than those ofgroup D. The group has below-average infiltration afterpresaturation (infiltration rate =1-4 mm/h).Highest runoff potential: Includesmostly clays of high swellingpercent, but the group also in-cludes some shallow soils withnearly impermeable subhorizonsnear the surface (infiltrationrate = O-1 mm/h).

86 92 78 83

Source: Adapted from Schwab et al. 1966.aValues taken from farmstead category, which is a composite including buildings, farmyard, road, etc.

2.4.3 Long-Term and Short- Term ReleaseCalculationFor surface runoff releases, the long-term releasevalue can be calculated as follows:

l Characterize an average storm event for the areai n t e r m s o f d u r a t i o n . T h i s c a n b e s t b eaccomplished by consulting local or regionalc l i m a t o l o g i c a l e x p e r t s , o r t h e N a t i o n a lClimatological Data Center in Asheville, NorthCarolina. Then, using USDC (1961) determinethe amount of rainfall corresponding to theselected duration rainfall event on a one year-return frequency basis. Divide this amount intothe mean annual rainfall for the area to obtain theaverage number of average rainfall events peryear.

l Use these data and the equations presented inthis section to calculate runoff contaminantrelease associated with each yearly averagestorm.

l Estimate the potential total long-term release forboth dissolved and sorbed runoff loss* as follows:

* This approach is overly conservative as it assumes that thecontaminant concentration in surface soil remains essentiallythe same during the entire 70-year period.

E A i = B N

where

(2-29)

E Ai = long-term release of contaminant i inrunoff, (mass/70 years).

B = dissolved or sorbed loss per stormevent, (i.e., PXi or PQi; see Equations2-27 and 2-28).

N = number of “average” storm events in70 years.

Determine the total amount of soil that will erodefrom the si te over 70 years. This can beaccomplished by applying the Universal Soil LossEquation (USLE-Wischmeier and Smith 1978).This equation, from which the MUSLE (seeEquation 2-20) was developed, estimates annualsoil losses in runoff. The USLE takes the sameform as the MUSLE, except that the stormevent-specific volume and flow rate variablesare replaced by a factor, R, the rainfall runofffactor. 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

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

Research based on the work of Haith et al. (1980) iscurrently underway at Cornell University* to developrunoff loading factors for organic chemicals in soils.After these factors are devised, the analyst will beable to obtain average loading values based solely ona chemical’s octanol/water partition coefficient andthe geographic location under study. This will greatlysimplify the generation of long-term average releaseestimates.

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

2.5 Quantitative Analysis of Ground-Water Contamination

Surface soils at uncontrolled hazardous waste sitesmay become contaminated with toxic materials as aresult of (1) the intentional placement of wastes onthe ground (dumping, landfarming), (2) spills, (3)lagoon failure (overland flow), or (4) contaminated siterunoff. Leaching of toxics from a contaminated soilsurface can carry contaminants into subsurfacelayers.

2.51 Beginning Quantitative Analysis

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

(1) Estimating Release Rate from Facilities Lined withClay or Natural SoilRelease rate estimation involves the determination ofboth the contaminant concentration in the leachateand the volumetric flux of leachate. The determinationof contaminant concentrat ion is made usingequilibrium conditions (steady state), whereas thevolumetric flux can be ascertained with instantaneoustime-varying models or with steady state equations.

Modeling the release rate of toxic constituents canthus be done in terms of either the instantaneoustime-varying releases or the annual average release(i.e., steady state release rate based on an annualaverage). This section discusses the determination ofthe 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 modelsnecessary for instantaneous time-varying releases.Analysts interested in performing instantaneoustime-varying release rate determinations are referredto Chapter 3, where the HELP and SESOIL modelsare discussed. HELP and SESOIL are appropriate formodeling dry solid waste in a landfill or landfarmsituation; they are not appropriate for modeling therelease rate of liquids from lagoons, landfills, orlandfarms. Rainstorms come in discrete intervalsseparated by dry periods. Using steady stateequations to model rainfall-induced leaching,however, assumes that 1/365th of the annualrecharge occurs each day. Although this is anassumption, it is felt to be a useful one for mostcases. Most abandoned hazardous waste sites havereceived liquids in the past; very few have receivedon ly d ry so l ids . Hence, the ques t ion o f theassumption of steady state conditions is relativelymoot. For the bulk of the modeling situations (liquidwastes), the steady state and the instantaneous ratesare the same, and since the steady state equationsare simpler, they are the method of choice.

For lagoons, the analyst should use the concentrationof contaminant in the lagoon as the concentration ofthe contaminant leaving the lagoon, since the“leachate” is the waste itself. The waste leaves thelagoon by percolating through the clay liner or thenative soil, or it permeates the flexible membraneliner (FML).

For landfills, the analyst should use the equilibriumsolubility of the solid waste, assuming that thecontaminant will have fully equilibrated with thepercolating rainwater. The use of the equilibriumsolubility concentration as the leachate concentrationis an assumption, it is based on a typical residencetime of 21 years for rain percolating through acovered (1 09-7 cm/sec ) secu re l and f i l l . Theassumption is that the time used for determining theequilibrium solubility of the chemical is much shorterthan the residence time in the fill. If the fill isuncovered (or covered with a permeable cover), thetravel time through the landfill may be too short forthe above assumptions to be valid. In these cases,the analyst should calculate the travel time andcompare it to the time used in the solubility test. If thetravel time is not longer than the test time, the analystshould estimate the leachate concentration as afraction of the equilibrium solubility concentration.Additionally, the above assumptions assume a landfillof only one waste stream, if the fill has only a smallquantity of the subject waste in it, the contact time isthe time for travel through the isolated material. Inthese conditions, the leachate concentration willtypically be a fraction of the equilibrium solubility. Theanalyst may wish, in some instances, to model thesolubility of the contaminant within a complexleachate. In this case, the solubility of a hydrophobic

29

contaminant can be increased by the organic fractionof the complex leachate.

For landfarms, the assumption that adequateresidence time is available for contaminants to reachequilibrium solubility may not be viable, and theanalyst should estimate the degree of solubilization.This can be done by dynamic modeling of the kineticsof dissolution, or it can be approximated based onexperience and engineering judgment. Because of thecomplexities of dynamic modeling, this approachusually is not worth the slightly increased accuracygained, especially since other parameters may affectthe accuracy of the final answer. Concentration istypically estimated as a fraction of the equilibriumsolubility.

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

(a) For landfilled solids, the only liquid present iswater percolating into the fill. For uncovered landfills,this can range from the infiltration fraction of therainfall, to the full precipitation (if no rain runs off ofthe fill before infiltrating), to larger flows of water if thesite is exposed to stormwater run-on from anadjacent area. For covered landfills, the infiltrationfraction may be limited by the permeability of thecover. Typically in wet climates the cover permeabilityis limiting, while in dry climates the permeability doesnot limit percolation, and normal soil percolation ratioscan be used.

The loading rate to ground water can be calculatedwith the following equation:

L c =q *A *C o

where

(2-32)

Lc = contaminant loading rate, (mass/time).q = percolation rate, see Equation 3-14

for calculation of q, (length/time).

c o

= area of landfill, (length squared).= s o l u b i l i t y o f s o l i d c h e m i c a l ,

(mass/volume).

(b) For lagooned or landfilled liquids, precipitationhas a minimal influence on leachate generation, asliquid waste will percolate to the watertable under theinfluence of gravity. The rate-determining step is thepermeability of the liner or underlying soil (if there isno liner). For liquids, the following form of Darcy’s lawshould be used to estimate the volumetric flux leavingthe site.

Q l = K s * i * A

where

(2-33)

Q1 = volume loading rate, (volume/time).KS = Darcy’s coefficient; for unlined lagoons

use native soil hydraulic conductivity;conductivity (length/time) (see Chapter3 for sources of hydraulic conductivity).

i = hydraulic gradient, (length/length).Equations 2-33 will handle situationswhere the liquids in the lagoon have afree depth. In many cases the depth ofthe free liquids is small, or it is smallwith respect to the distance betweenthe lagoon and the watertable (whenthe KS is for native soil). In thesecases the term “i” can be taken as 1.

A = area of lagoon, (length squared).

This Qi is then used to estimate mass loadingswith the following equation:

Lc=C s*Q 1 (2-34)

where

Q1

= 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 ratefrom a lagoon whether the flow through the vadosezone is saturated or unsaturated. For unlined activelagoons, the flow is typically saturated all the way tothe watertable. For clay-lined lagoons, the flow issaturated through the liner and unsaturated betweenthe liner and the water-table (assuming no breaches inthe liner). Equations 2-33 and 2-34 are appropriatewhen analyzing lagoon releases, but should not beused for spills or other conditions where thechemicals 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 aremostly water. For lagoons that contain organic fluids,however, the equations may need to be corrected.For liquids with a density or viscosity that differs fromwater, correct Ks for this different viscosity anddensity by calculating the term Kc, using thefollowing:

K, = Kw * Dc/Dw * Uw/uc (2-35)

where

Kc = c o r r e c t e d K s t e r m = h y d r a u l i cc o n d u c t i v i t y o f c o n t a m i n a n t ,(length/time).

K w = 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 withFlexible MembranesThe release rate from an intact lined landfill or lagooncan be calculated for a small group of contaminants.Failed liners can be modeled as a function of theextent of the failure using the modeling equations forclay or natural soil-lined facilities. Although a flexiblemembrane (FML) liner appears to allow no migrationthrough the barrier, it may indeed be penetrated byorganic compounds and contaminated water, althoughthe rate of permeation is understandably small. Therate at which a contaminant permeates through apolymeric material has been shown to be dependentupon various properties of the permeant, such assize, shape, polarity, and other factors (Steingiser etal. 1978).

Sa lame and o the rs p roposed the use o f apermeability equation to predict the rate of permeationof liquids and gases through various polymers(Salame 1961, 1973, 1985; Steingiser et al. 1978):

PS = ApØe-sH

where

(2-36)

p s= p e r m e a t i o n r a t e , ( g - m i l / 1 0 0

in2*day*cmHg).A p

= constant solely dependent on the typeo f p o l y m e r s u s e d , (g -m i l / 100

SHinn*day*cmHg).

= constant solely dependent on the typeof polymers used, (cc/cal).

Ø = the polymer “permachor” calculated foreach polymer-permeant pair, (cal/cc).

Salame lists values for these parameters obtainedfrom his extensive experimental work. These valuesare shown in Tables 2-7, 2-8, 2-9, and 2-10.

For permeation of water through FMLs, polymers arecategorized into five groups based on the values ofthe solubility parameter as shown in Table 2-8. Thisgrouping was achieved after examinat ion ofexperimental data for about 70 different polymers(Salame 1985). The solubility parameter provides anindication of polymer interaction with water, with moreinteraction occurring at higher values of the solubilityparameter. Examples of hydrogen bonding forpolymer group 5 include hydroxyl (OH) and amide(NHCO) radicals as in nylon and polyvinyl alcohol.The polymer with hydrogen bonding but with thevalue of “delta” less than 11 does not belong togroup 5. Permachor values for some selected organicliquids and for water are shown in Tables 2-9 and

2-10, respectively. The water “permachor” valuesfor various polymers given in Table 2-10 apply underdry conditions. For water permeation under wetconditions, permachor values may be reduced byabout 20 percent.

The term P can be used to calculate the release ratein grams/day. P is multiplied by the area of the liner,and then divided by its thickness. This assumes anormal water vapor pressure of 1 cm Hg at ambienttemperature. The equation is:

Lc = Ps*A*p/dl

where

(2-37)

LC = contaminant loading rate, (mass/time).PS = p e r m e a t i o n r a t e , ( g - m i l / l 0 0

in2*day*cmHg).A = area of liner, (in units of 100 in2).P = vapor pressure, (cmHg).

d e = thickness of the liner, (mils).

2.5.2 In-Depth AnalysisIn-depth analytical approaches for quantification ofbaseline contaminant release to ground water involvethe use of computerized models. Refer to Chapter 3of this manual for a detailed discussion of the natureand applications of such modeling tools.

2.5.3 Long-Term and Short-Term ReleaseCalculationFor toxic substance release to ground-waters y s t e m s , d i r e c t l y c a l c u l a t e t h e s h o r t - t e r m(maximum) release values from the measured surfaceand subsoil contaminant concentrations using thetools discussed in this section. Obtain long-term(average) values by applying the procedure previouslyoutlined for particulate releases to air (see Section2.3.3).

2.6 Soil Contamination

2.6.1 Beginning Quantitative AnalysisNo estimation methods are presented for analysis ofsurface soil contamination. Site soils will be sampledd i rec t l y and the degree and ex ten t o f the i rcontamination delineated during the RemedialInvestigation. Sampling and analysis may also havebeen conducted for subsurface soils. In certain cases,however, it may be desirable to project subsurfacecontamination without conducting unsaturated zonesampling. USEPA (1987a) covers soil samplingstrategies.

2.6.2 In-Depth AnalysisSurface soil monitoring, usually conducted during theRemedial Invest igat ion, const i tutes in-depthquantitative analysis. Subsurface (unsaturated zone)in-depth analysis will usually involve application ofsampling and modeling approaches. Sampling and

31

Table 2-8. Polymer Categorization for Permeation ofWater

analysis can provide a direct quantification of thedegree of contaminat ion in subsurface soi ls.Alternatively, computer models (e.g., SESOIL;Bonazountas and Wagner 1981) are used to projectthe level of unsaturated zone contamination over timefrom surface placement of toxics. Refer to Chapter 3of this manual for a detailed discussion of computermodels that can be applied to the unsaturated zonecontamination estimation.

Table 2-9. Permachor Values of Some Organic Liquids inPolyethylene and PVCa

In nonpolar polymer In polar polymerLiquid Ø Ø

Acetic acid 13.0 44.0Benzaldehyde 15.9 4.0Benzene 5.4 7.02-Butoxy ethanol 24.4 75.0Butyl acetate 13.0 5.0Butyl alcohol 18.0 50.0Butyl ether 10.4 46.0Butyraldehyde 13.5 0.0Capryllc acid 19.0 50.0Carbon tetrachloride 5.8 22.0p-chlorotoluene 7.6 7.5Cyclohexane 7.0 45.0Dibutylphthalate 31.4 17.0Diethylamine 10.0 5.7Ethanol 16.0 48.0Heptane 7.0 44.0Hexane 6.0 43.0Methyl ethyl ketone 12.5 1.0Methanol 15.0 47.0Nitroethane 15.4 7.0i-Pentyl propionate 15.0 7.0i-Propyl amine 11.0 6.7Trichloroethylene 5.4 3.0o-Xylene 9.4 11.0p-Xylene 7.4 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 DryPolymers

PermachorPolymer value (0)

Polyvinyl alcohol 160PolyacrylonitrileCellulose (dry)Polyvinylidene chloridePolycaprolactam (dry)Polyacrylonitrile styrene (70/30) (Lopac)Polyacrylonitrile styrene/butadiene(70/23/7) (Cycopac930)PolychlorotrifluoroethylenePolyethylene terephthalatePolyvinylidene fluoride (Kynar)Polyacrylonitrile styrene/ = tibutadiene(56/27/4/13) (Cycopa\c 920)Polyvinyl chloridePolyoxymethylene (Delrin)Polymethyl methacrylatePolyvinyl acetate (dry)Polystrene/acrylonitrile (74/26)Polyethylene (HD)PolysulfonePolypropylenePolycarbonate (Lexan3)PolystyrenePolyethylene (LD)PolyisobutylenePolyethylene/vinyl acetate (85/l5)PolybutadienePolymethyle pentene (TPX)Polydimethyl siloxane (dry)

1099787807675

71686765

62575545454034333328261715

88

-4

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

33

Chapter 3Contaminant Fate Analysis

3.1 Introduction

This chapter provides guidance for evaluating thetransport, transformation, and fate of contaminants inthe environment following their release from anuncontrolled hazardous waste site. The contaminantrelease rate estimates described in the previouschapter provide the basis for contaminant fateanalysis. The results form the basis for subsequentanalysis of exposed populations and estimation of thelevels of exposure incurred (see Appendix A). Thegoal of contaminant fate analysis is to identify off-site areas affected by contaminant migration and todetermine contaminant concentrations in these areas.

The fo l low ing sec t ions address ana lys is o fatmospheric fate, surface water fate, ground-waterfate, and biotic fate. Within each of those sections,contaminant transport is addressed (except for bioticfate analysis, which does not involve contaminanttransport). A screening analysis is conducted toprov ide an in i t ia l qua l i ta t i ve assessment o fcontaminant transport in the environment. It isdesigned to (1) identify each transport processgoverning the movement of various contaminantswithin and among environmental media, (2) determinethe direct ion and roughly gauge the rate ofcontaminant movement from the site, and (3) identifyareas to which contaminants have been or may betransported. Screening analysis is designed both toprovide ini t ial organizat ion and direct ion forsubsequent in-depth analysis of contaminantenvironmental transport, and to provide a consistentbasis for analysis from site to site.

When likely pathways of contaminant migration havebeen identified by screening analysis, those pathwaysrequiring further evaluation are quantitativelyaddressed. Like analysis of contaminant release, thisanalysis can involve either the use of “desktop”analytical solutions or numeric methodology.

Simplified environmental fate estimation proceduresare based on the predominant mechanisms oftransport within each medium, and they generallydisregard intermedia transfer or transformationprocesses. In general, they produce conservativeestimates (i.e., reasonable upper bounds) for final

ambient concentrations and the extent of hazardoussubstance migration. However, caution should betaken to avoid using inappropriate analytical methodsthat underestimate or overlook significant pathwaysthat impact human health.

When more in-depth analysis of environmental fateis required, the analyst must select the modelingp r o c e d u r e t h a t i s m o s t a p p r o p r i a t e t o t h ecircumstances. In general, the more sophisticatedmodels are more data-, t ime-, and resource-intensive.

The following criteria should be considered whenselecting an in-depth environmental fate model ormethod:

• Capability of the model to account for importanttransport , t r a n s f o r m a t i o n , a n d t r a n s f e rmechanisms;

? “Fit” o f t h e m o d e l t o s i t e - s p e c i f i c a n dsubstance-specific parameters;

?? Data requirements of the model, compared toavailability and reliability of site information; and

? Form and content of model output. This refers tothe model’s ability to address important questionsregarding human exposure or environmentaleffects and to provide all data required as input tofurther analysis.

Information regarding the major environmentalprocesses that may affect the fate of hazardoussubstances in each medium is provided. Theseprocesses include transformation and intermediatransfer mechanisms, as well as the more complextransport mechanisms that are not incorporated intoestimation procedures. By comparing the list ofimportant processes identified for the site with thesummary of model features presented at the end ofeach section, the analyst can select the model bestsuited 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 hasthe ability to produce sophisticated analyses ofenvironmental fate. GEMS consists of modelscapable of assessing contaminant fate in air, surfacewater, ground water, and soil. These fate modelscontain pertinent data files (including nationwide soil,land use, and meteorological data, and data on manymajor river systems, lakes, and reservoirs); user-input data manipulation and storage capabilities;statistical processing programs; and such graphicscapabilities as presentation of results in map form.

GEMS is designed to be user-friendly. Althoughenvironmental fate modeling experience is highlydesirable, personnel with no computer programmingbackground can also use the system because of itsprogressive menu and user prompting formats. Ateach decision point, the user is presented with a listof possible selections. When specific data arerequired to activate a program, the system requestseach type of data needed and the units required. Atany point in the procedure, the user can request helpfrom the system, and a clear explanation of thechoices or steps facing the user is provided.

The GEMS host computer is a Vax-11/780, which islocated at the EPA National Computer System atResearch Triangle Park, North Carolina. The systemcan be accessed and used with the following terminaltypes: DEC UT-100 series, Tektronix 4014 series,and ASCII.

Terminals must be capable of transmitting orreceiving ASCII data in full duplex mode, using evenpar i ty and seven-b i t da ta word length , w i thcommunication rates of 300 or 1200 bits per second.Most common acoustic modems are compatible(GSC 1982).*

Monitoring data can also be useful in analyzingcontaminant transport and fate. Monitoring results canprovide, however, only a measurement of the existingextent of contamination. In addition, monitoring dataalone may not allow the analyst to discriminate thecontributions of specific sources to measuredcontaminant loadings. In all assessments, somedegree of modeling contaminant movement within andamong environmental media will be necessary topredict the associated exposure over a 70-yearlifetime. Thus, a combination of monitoring andmodeling techniques will be necessary to conduct ananalysis of contaminant fate for exposure assessmentpurposes.

For in-depth guidance in selecting and running acomputer model to use in analyzing contaminant

* Contact personnel within the EED are Ms. Patricia Harrigan,Mr. Loren Hall, or Mr. Russell Kinnerson. They can be reachedat EPA, Washington, D.C., (202) 382-3931.

migration from a particular site, the analyst shouldreview the following guides:

- USEPA (1977a):

USEPA (l986b): Guidel ine on Air Qual i tyModels (Revised) 1986 andSupplement A (1987)

-USEPA (1987d): Surface Water Model Se-lection Criteria

-USEPA (1986a): Ground Water Model Se-lection Criteria

- USEPA (1985j): Modeling Remedial Actions

Guidelines for Air QualityMaintenance Planning andA n a l y s i s , V o l u m e 1 0(Revised): Procedures forEvaluating Air Quality Impactof New Stationary Sources

In addition, it is recommended that the analyst obtainthe user’s manual for any model selected beforeattempting its application.

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

To evaluate the retardation of contaminant plumescomposed of mixed wastes in ground-water systemsthe analyst is referred to the following references fordetailed guidance: Nkedi-Kizza et al. (1985), Rao etal. (1985), Woodburn et al. (1986).

3.2 Contaminant Fate Screening

Figures 3- l through 3-4 present the decisionnetworks for screening contaminant fate in air,surface water, ground water, and biota. Any migrationpathways (identified in the qualitative evaluation) thatwill require additional analysis are described inSections 3.3 through 3.6. These pathways will befurther evaluated to determine the likelihood ofpopulation exposure as described in Appendix A.

In Sections 3.2.1 through 3.2.4, brief guidance isprovided for the qualitative evaluation of contaminantmigration pathways. The paragraphs presented beloware keyed to the accompanying decision networksand are intended to provide further elaboration ofthose boxes in the decision networks.

3.2.1 Atmospheric FateThe following numbered paragraphs each refer toparticular numbered boxes in the Figure 3-1.

1. The atmospheric fate of contaminants must beassessed whenever it is determined that significant

36

gaseous or airborne particulate contaminants arereleased from the site. The atmospheric fate ofcontaminants released originally to other media, buteventually partitioning to the atmosphere beyond siteboundaries, must also be assessed whenever thisintermedia transfer is likely to be significant.

2. The predominant directions of contaminantmovement will be determined by relative directionalfrequencies of wind over the site (as reflected inarea-specific wind rose data). Atmospheric stabilityand wind speeds determine off-site areas affectedby ambient concentrations of gaseous contaminants.Usually, high stability and low wind speed conditionsresult in higher atmospheric concentrations ofgaseous contaminants close to the site. High stabilityand moderate wind speeds result in moderateconcentrations over a larger downwind area. Lowstability or high wind speed conditions cause greaterdispersion and dilution of contaminants, resulting inlower concentrations over larger areas.

For particulate contaminants (including thoseadsorbed to dust or soi l part ic les), ambientconcentrations in the atmosphere and areas affectedby airborne contaminants are determined bywindspeed and stability and also by particle sizedistribution. High winds result in greater dispersionand cause particulates to remain airborne longer(which may also increase release rates). Low windsand high stability will result in rapid settleout ofparticulates and in a more concentrated contaminantplume closer to the site. Larger particles will settlerapidly, decreasing the atmospheric concentrationswith distance from the site. Finer particles will remainairborne longer, and their behavior will more closelyapproximate that of gaseous contaminants, asdescribed above.

3. Settleout and rainout are important mechanismsof contaminant transfer from the atmospheric mediato both surface soils and surface waters. Rates ofcontaminant transfer caused by these mechanismsare difficult to assess qualitatively; however, theyincrease with increasing soil adsorption coefficients,solubility (for particulate contaminants or thoseadsorbed to part iculates), part ic le size, andprecipitation frequency.

A r e a s a f f e c t e d b y s i g n i f i c a n t a t m o s p h e r i cconcentrations of contaminants exhibiting the abovephysical /chemical propert ies should also beconsidered as potentially affected by contaminantrainout and settleout to surface media. Contaminantsdissolved in rainwater may percolate to ground water,run off or fall directly into surface waters, and adsorbto unsaturated soils. Contaminants settling to thesurface through dry deposition may dissolve in orbecome suspended in surface waters, or may beleached into unsaturated soils and ground water bysubsequent rainfall. Dry deposition may also result in

formation of a layer of relatively high contamination atthe soil surface. When such intermedia transfers arelikely, one should assess the fate of contaminants inthe receiving media.

4. If areas identified as likely to receive significantatmospheric contaminant concentrat ions includeareas supporting edible biota, the biouptake ofcontaminants must be considered as a possibleenvironmental fate pathway. Direct biouptake fromatmosphere is a potential fate mechanism forlipophilic contaminants. Biouptake from soil or waterfollowing transfer of contaminants to these mediamust also be considered as part of the screeningassessments o f these med ia ; fo r example ,hexachlorobenzene was found to accumulate inplants (Russell et al. 1971, Gillet 1980, Trabelka andGarten 1982).

3.2.2 Surface Water FateThe following numbered paragraphs each refer toparticular numbered boxes in the Figure 3-2.

1. The aquatic fate of contaminants released fromthe CERCLA site as well as those transferred tosurface water f rom other media beyond si teboundaries must be considered.

2. Direction of contaminant movement will usuallyonly be clear for contaminants introduced to riversand streams. Currents, thermal stratification oreddies, tidal pumping, and flushing in impoundmentsand es tuar ies render qua l i ta t i ve sc reen ingassessment of contaminant directional transporthighly conjectural for these types of waterbodies. Inmos t cases , e n t i r e w a t e r b o d i e s r e c e i v i n gcontaminants must be considered potent ia l lysignificant human exposure points. More in-depthanalyses or survey data may subsequently identifycontaminated and unaffected regions of thesewaterbodies.

3. Similarly, contaminant concentrations in rivers orstreams can be roughly assessed based on rate ofcontaminant introduction and dilution volumes.Estuary or impoundment concentration regimes arehighly dependent on the transport mechanismsenumerated above. Contaminants may be localizedand remain concentrated, or disperse rapidly andbecome d i lu ted to ins ign i f i can t leve ls . Theconservative approach is to conduct a more in-depthassessment and use model results or survey data asa basis for determining contaminant concentrationlevels.

4. Important intermedia transfer mechanisms thatmust be considered where significant surface watercontamination is expected include transfers to groundwater where hydrogeology of the area indicatessignificant surface-water/ground-water exchange;transfers to biota where waters contaminated with

lipophilic substances support edible biotic species;and transfer to the atmosphere where surface wateris contaminated by volat i le substances. Hightemperatures, high surface-area-to-volume ratios,high wind conditions, or turbulent stream flow alsoenhance volatilization rates.

Contaminant transfer to bed sediments representsanother significant transfer mechanism, especially incases where contaminants are in the form ofsuspended solids, or are dissolved, hydrophobicsubstances that can become adsorbed by organicmatter in bed sediments. For the purposes of thismanual, sediments and water are considered part of as i n g l e s y s t e m b e c a u s e o f t h e i r c o m p l e xinterassociation. Surface water/bed sediment transferis reversible; bed sediments often act as temporaryrepositories for contaminants and gradually re-release contaminants to surface waters. Sorbed orsettled contaminants are frequently transported withbed sediment migration or flow. Transfer of sorbedcontaminants to bottom-dwelling, edible biotarepresents a fate pathway potentially resulting inhuman exposure. Where this transfer mechanismappears likely, the biotic fate of contaminants shouldbe assessed.

3.2.3 Soil and Ground-water FateThe following numbered paragraphs each refer toparticular numbered boxes in Figure 3-3.

1. The fate of contaminants in the soil medium isassessed whenever the contaminant releaseatmospheric or fate screening assessments resultsshow that significant contamination of soils is likely.

2. The most significant contaminant movement insoils is a function of liquid movement. Dry, solublecontaminants dissolved in precipitation, run-on, orhuman-applied water will migrate through percolationinto the soil. Migration rates are a function of netwater recharge rates and contaminant solubility.

Liquid contaminants may percolate directly into soils.Organic liquids may alter soil permeabilities or may beof lower viscosity and/or higher density than water,resulting in percolation rates many times greater thanthat of water. Contaminants with high soil adsorptioncoefficients may bind to soils and become relativelyimmobile.

3. Important intermedia transfer mechanismsaffecting soil contaminants include volatilization orresuspension to the atmosphere and biouptake byplants and soil organisms. These, in turn, introducecontaminants to the food chain.

4. The fate of contaminants in ground water isassessed whenever si te contaminant releasescreening analysis indicates direct introduction ofcontaminants to ground water (e.g., through disposal

wells or fluid releases to an aquifer near the groundsurface), or whenever the screening assessments ofatmospheric, surface water, or soil contaminant fates(as outlined above) indicate potential contaminanttransfer to ground water.

5. The qualitative assessment of ground-water flowis often based on the assumption that subsurfacehydrologic gradients (which determine flow directionsand rates) approximate surface topography. Thisapproach is unreliable and should be used only in theabsence of hydrogeologic data. Ground-water flow isinfluenced by many factors including hydraulicconductivity of soils, hydraulic gradient, presence ofsubsurface impermeable barriers, presence ofdischarge areas (e.g., streams intercepting ground-water flow) and presence of fissures, cavities, ormacropores. Hydrogeologic survey data (whereavai lable) provide a more rel iable basis forcontaminant transport assessment than do surfacetopographs.

6. Site and surrounding community survey datadescribing the location of wells are compared with theexpected subsurface contaminant plume boundariesto identify locations of potential exposure points.

7. Important mechanisms of contaminant transferfrom ground water to other environmental mediainclude contaminated water exchange betweensurface waters and ground water and uptake ofcontaminants by edible biota. The former mechanismmust be considered whenever surface waters aredowngradient from the CERCLA site; it increases inlikelihood with closer proximity of these surfacewaters to the site. Available hydrogeologic informationfor the site and surroundings should be reviewed forany indication that the aquifer underlying the site isconnected to surface waters.

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

3.2.4 Biotic FateThe following numbered paragraphs each refer toparticular numbered boxes in Figure 3-4.

1. A screening environmental fate assessment forthe biotic medium is performed after the fate ofcontaminants in the atmosphere, surface waters, orground water has been assessed. Starting with theexpected distribution of contaminants in each of thesemedia, potent ial points of biot ic contact with

40

contaminated media and important affected bioticspecies are identified.

2. Important species are those used directly by man(game animals, sport or commercial fish, crustaceansand mollusks, agricultural crops and livestock;naturally-occurring fruits, herbs, other ediblevegetation), and those that introduce contaminants tospecies 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 groupslisted above).

3. Assessed mechanisms of transport in the bioticmedium include the food chain, natural animalmigration, or human commercial activity. Food chaintransport can result in high concentrations ofcontaminants in the tissue of edible species not indirect contact with contaminated air or water. Humancommercial transport and natural migratory behaviorof contaminated species can result in a widedistribution of edible species or tissue-containingcontaminants.

4. Edible tissue concentrations are a function of thelevel and type of biotic exposure to contaminants, thepartitioning of contaminants between organic tissueand substrate media, the biodegradabi l i ty ofcontaminants, organism-specific metabolic charac-teristics, and ecosystem characteristics.

3.3 Quantitative Analysis of AtmosphericFate3.3.1 Screening AnalysisThe atmospheric fate of substances released fromuncontrolled hazardous waste sites can be estimatedby using the following equation to estimate ground-level atmospheric concentrations of pollutants atselected points on a centerline of a plume directlydownwind from a ground-level source (Turner 1970):

where

(3-1)

C(X) = concentration of substance at distancex from site, (mass/volume).

Q = release rate of substance from site,(mass/time).

óy= dispersion coefficient in the lateral

(crosswind) direction, (distance).óz = dispersion coefficient in the vertical

direction, (distance).µ = mean wind speed, (distance/time).n = the value pi = 3.14.

The appropriate dispersion coefficients can beobtained from Figures 3-5 and 3-6. These figures

Figure 3-4. Environmental fate screening assessmentdecision network: food chain.

Contaminants

Food Chain

provide values for ó y and ó z, respectively, asfunctions of downwind distance, x, and stabilityclasses A though F. These stability classes are basedon the Pasquill stability classification system, whereClass A is very unstable and Class F is moderatelystable (Pasquill 1961). Table 3-1 presents a briefillustration 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).

10.1

Distance Downwind, km

*Lines designated A through F represent dispersion coefficient functions for atmospheric stability classes A through F. See textfor sources of atmospheric stability data.

43

Figure 3-6. Vertical dispersion coefficient as a function of downwind distance from the source (from Turner 1970).

Distance Downwind, km

“Curves designated A through F represent dispersion coefficient functions for atmospheric stability classes A through F. See textfor sources of atmospheric stability data.

44

Table 3-1. Key to Stability CategoriesNight

ThinlySurface wind overcast orspeed at a Day incoming Solar radiation > 4/8 Low < 3/8

Height of 10 (insolation) Cloud Cloudm (m/sec) Strong Moderate Slight Cover Cover

<2 A A-B B2-3 A-B B C E F3-5 B B-C C D E5-6 C C-D D D D>6 C D D D D

The neutral class (D) should be assumed for all overcastconditions during day or night.*Appropriate insolation categories may be determined through theuse of sky cover and solar elevation information as follows:

Solar SolarSolar elevation elevation

elevation angle < 600 angle < 350

Sky cover angle > 600 but > 350 but > 150

4/8 or Less or Strong Moderate SlightAny Amount ofHigh Thin Clouds5/8 to 7/8 Middle Moderate Slight SlightClouds (7000feet to 16,000foot base)5/8 to 7/8 LOW Slight Slight SlightClouds (less than7000 foot base)

Source: USEPA 1977b

To obtain the maximum hourly concentration, selectthe calculational methodology for coning and fanningplumes in USEPA (1977b). To obtain the estimatedmaximum concentration for a 3-, 8-, or 24-houraveraging time, multiply the l-hour maximum by thefactors given in USEPA (1977b).

T o e s t i m a t e l o n g - t e r m m e a n a t m o s p h e r i cconcentrations, obtain STAR (Stability Array) dataspecific to the site. These data provide seasonal orannual joint frequencies for each stability class, winddirection, and wind speed category. Assume anannual average wind speed of 3 meters/second, andca lcu la te the l ong - te rm mean a tmospher i cconcentration for each exposed population byapplying a weighted average, based on the relativefrequency of each stability class and of wind flowtoward selected exposure points. Equation 3-2provides a rough weighted average estimate (Turner1970):

CA(x)

f A

= concentration at point x during stabilityclass A (from Equation 3-1).

= relative annual frequency of stabilityclass A for the specified wind direction.

and subscripts B through F represent the variousstability classes.

Note that this estimate is a rough approximationbecause it is simplified by the assumption that themean wind speed is 3 meters/second for all stabilityclasses. A more sophisticated estimate can be madeby incorporating site-specific wind speed frequencydata, and performing similar weighted averagecalculation of ambient concentrations. This is a time-consuming procedure, however, and the use ofcomputer-based estimation procedures may bemore cost-effective if sophisticated estimates arerequired. STAR data are available from the NationalClimatic Center (NCC), Asheville, North Carolina(phone: (704) 259-0205) for all National WeatherService (NWS) locations in the U.S. The NWS Stationthat is most representative of the site should be used.

T h e a r e a w i t h i n w h i c h t h e g r o u n d - l e v e lconcentration of a hazardous substance is above apredetermined critical concentration (i.e., the plumeisopleth) can be described using the followingprocedures. Calculate the crosswind distance fromany po in t a long the p lume cen te r l i ne ( i .e . ,perpendicular to the plume centerline) to the isoplethboundary by Equation 3-3 (Turner 1970):

(3-3)

C ( C L )

y(x)

C(x)

ó y

= predetermined critical concentrationlevel, (mass/volume).

= perpendicular distance from point onplume centerline to the C(CL) isoplethboundary, (length units).

= concentration at plume centerline, xdistance from source, (mass/volume,as calculated by Equation 3-1).

= lateral dispersion coefficient, (lengthunits).

Vary the value for x (downwind distance from thesource) input into Equations 3-1 and 3-3, startingat a point near the site* and increasing this value untilthe value for C(x) (obtained from Equation 3-1)equals the predetermined critical concentrationC(CL). Values calculated for y describe the isoplethboundary on either side of the plume centerline.

* Equations are generally considered applicable to downwinddistances of at least 200 m.

45

Estimate the area within a plume isopleth using Figure3-7 wh ich p lo ts the va lue C(CL)µ ( re la t i veconcentration times wind speed versus isopleth area,for each stability class A through F).

All of the preceding simplified equations provideatmospheric fate estimates based on several simpleassumptions, one of which requires special mention.This is the assumption that the hazardous substancereleased from a site is in a form that can remainairborne indefinitely (i.e., either gaseous or consistingof particles less than 20 microns in diameter) (Turner1970).

In cases where fugitive dust blown from the siteincludes sol id hazardous substances (or soi lparticulates carrying adsorbed hazardous substance)of greater diameter than 20 microns, relatively rapidgravitational settling of the larger particles occurs.Consequently, much of the hazardous materialreaches the ground before advection and dispersioncan transport and dilute the plume as described bythe above equations. Thus, areas close to theuncontrol led hazardous si te may experiencesignificant soil contamination, and human exposurepoints farther from the site may experience loweratmospheric concentrations than estimated by theseequations. Hanna and Hosker (1980) present aprocedure for estimating the gravitational settling rate,distance of travel from the source, and deposition rateof airborne particulates.

All of the above simplified procedures incorporate thefollowing additional assumptions:

Steady-state condition, i.e., windspeed is steadyat rate u, and the hazardous substance release iscontinuous, at average rate Q. Wind direction isa lso assumed to be s teady ; shor t - te rmfluctuations are disregarded.

Longitudinal dispersion is negligible (substancetravels at wind speed in the downwind direction).

The substance is refractory (all removal anddecay processes are disregarded).

The substance is distr ibuted normal ly, oraccording to a Gaussian distribution, bothvertically and in the crosswind direction.

The air environment is homogeneous; windspeeds and stability are equal at all heights abovethe ground, and no obstructions to wind flow ordispersion exist other than at the ground.Complete reflection occurs at the ground/airinterface.

3.3.2 In-Depth AnalysisW h e r e e s t i m a t e s o f a m b i e n t a t m o s p h e r i cconcentrations of hazardous substances developed

by the preceding simplified procedures indicate thatthese concentrations pose potential health hazards,more accurate, in-depth analysis of atmospheric fatemay be required. Numerous computer models areavailable for this purpose and are listed in USEPA(1986b). These models vary in sophistication andcapability, and in their abi l i ty to incorporateexpressions describing the effect of var iousprocesses on the atmospheric fate of hazardoussubstances. The most important of these processesare briefly described below. Consider the importanceof each of these processes to the atmospheric fate ofthe substances under analysis before selecting acomputer model.

3.3.2.1 Intermedia TransferThe following are the most important processes thataffect the removal of hazardous substances from theair medium and their transfer to other sectors of theenvironment.

(1) DissolutionThis is the process whereby hazardous substances inthe gaseous state are dissolved into water dropletspresent in the atmosphere. This process, followed byprecipitation, distributes the substance over thesurface media, and percolation to ground water mayfollow. Direct dissolution may also occur betweengaseous substances in the atmosphere and surfacewaters at the air/water interface. Dissolution is aconstant, reversible process, the amount of haz-ardous substance in the aqueous phase is de-termined by the partition coefficient of the substancebetween the gas and aqueous phases. This partitioncoefficient is in turn a function of the vapor pressurea n d w a t e r s o l u b i l i t y o f t h e s u b s t a n c e , i t sconcentration in the air, and temperature. See Lymanet al. (1982) or Hanna and Hosker (1980) for methodsof estimating this partition coefficient and atmospherichalf-lives resulting from dissolution/ rainout.

(2) AdsorptionThrough the process of adsorption, hazardoussubstances in the vapor phase become attached toparticulate matter suspended in the air (aerosols), oronto soil particles at the air/soil media interface.Suspended aerosols settle to surface media, therebyremoving adsorbed substances from the airenvironment. The adsorption rate of a particularsubstance is principally a function of the number andsurface area of aerosols per volume of air, themolecular weight of the substance in question, itsconcentration in the air, and its saturation vaporpressure. Cupitt (1980) provides a method forestimating atmospheric contaminant removal ratesdue to adsorption to particulates and settleout.

(3) Gravitational SettlingThis mechanism is most important for particulatehazardous substances, or hazardous substances

46

*Curves designated A through F represent functions for atmospheric stability classes A through F. See text for sources of atmosphericstability data.

adsorbed onto suspended particulates, if theparticulate matter is more than 20 µm in diameter.These particles settle to the surface media at a ratethat is a function of their density, shape, anddiameter, and of wind speed (Hanna and Hosker1980).

(4) PrecipitationPrecipitation itself is a major mechanism for removalof particulate and aerosol matter. Raindrops requireparticulates or aerosols to serve as nuclei for theircondensat ion from the vapor state of water.Moreover, raindrops generally remove particulates

and aerosols > 1.0 µm in diameter as they fall belowthe cloud level.

3.3.2.2 lntramedia Transformation ProcessesMany hazardous substances are subject to decay ortransformat ion to other substances with newproperties while entrained in the air environment. Thetwo most important of these processes are describedbelow. While the product of such transformationprocesses will usually have different properties fromthose of the original hazardous substance, the newsubstance produced may also have hazardousproperties. Cupitt (1980) provides estimates of

47

cons tan ts tha t de te rmine the ra te o f eachtransformation process below, as well as of theimportance and likely products of these processes,for 46 hazardous materials. Hendry and Kenley (1979)provide rate constants and estimation procedures forthese processes.

(1) PhotolysisThis is the breakdown of substances because ofphotochemical reaction brought about by solarenergy. Photolysis can be direct, when the hazardoussubstance is itself affected by solar radiation, orindirect when the hazardous substance reacts withother substances that have been raised to a reactivestate by solar radiation. Photolysis rates depend onsolar radiation availability, the light absorptioncoefficient of the hazardous substances, and areaction yield constant (which describes the efficiencyof transformation of the hazardous substance with theavailable sun energy).

(2) OxidationThe reaction of substances with oxidants in theatmosphere can result in their transformation. Thetwo most important atmospheric oxidants are ozoneand the hydroxyl radical. Reaction rate constants foroxidation are chemical specific; the overall rate oftransformation of a hazardous substance by oxidationdepends on the concentration of the oxidant and thereaction rate constant.

3.3.2.3 The Effects of TerrainFeatures such as vegetation, large buildings, urbanareas, rough topography, hills, or mountains can allprofoundly affect the atmospheric fate of airbornesubstances, principally by altering the laminar flow oftransporting wind currents. The effects of terrain onwind currents may include increased turbulence,downwash in the lee of large obstacles, or localizedalterations in the direction of flow. Because therelease of substances from hazardous waste sitesusually occurs at ground level, the fate of thesesubstances is especially susceptible to the effects ofterrain. Select a model capable of accounting forthese effects in any case where these listed terrainfeatures exist between the site and points of humanexposure.

3.3.3 Computer ModelsT a b l e s 3 - 2 , 3 - 3 , a n d 3 - 4 p r o v i d e g e n e r a linformation about computer-based models that couldbe appropr ia te to in -dep th ana lys is o f theatmospheric fate of substances released fromCERCLA si tes. Table 3-2 contains resourcerequirements, references, and sources for eachmodel; Table 3-3 summarizes their features andcapabilities; and Table 3-4 discusses the datarequirements of each. By comparing the informationin these tables with identified site features, site dataavailability, final output requirements, and resource

availability, one can select the most applicable andcost-effective model.

The Industrial Source Complex (ISC) long-termmodel and the TOXBOX area source model arepresently integrated into the GEMS system. Thesemodels are accessed under a subsystem of GEMSreferred to as the GEMS Atmospheric ModelingSystem (GAMS). A brief description of ISC isprovided below.

The ISC (Bowers et al . 1979) is a Gaussiandispersion model, capable of est imat ing theconcentration and deposition rates of gaseous andparticulate pollutants around a point, area, or linesource. Because it is integrated into the GEMSsystem, it is especially useful for the analysis of theatmospheric fate of hazardous substances. Based ona user- input release locat ion ( in the form oflatitude/longitude coordinates or zip code), storedclimatological data from the nearest meteorologicalmonitoring stations are retrieved (GSC 1982).

The integration of ISC with a population distributionmodel called SECPOP gives it the capability ofexpressing atmospheric fate of pollutants in terms ofnumbers of people affected at various concentrationlevels (this capability is discussed in more detail inAppendix A, Exposed Populations).

The ISC model can estimate the concentration ofpollutants 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, andgravitational settling can be simulated based onuser-input half-life data (GSC 1982).

ISC can be used with IBM, CDC, or VAX computers.The model is implemented within GEMS on EPA’sVAX 11/780 and can be accessed with a variety ofuser terminal types. (See Section 3.1 for accessinstructions.)

3.3.4 Short- and Long-Term ConcentrationCalculationsLong-term average ambient air concentrations ofhazardous substances at human exposure points areestimated using the long-term average release rateover the time period of interest, and the weightedaveraging algorithm presented as Equations 3-1 and3-2. Annual average climatological data, or STARdata inc lud ing long- te rm f requenc ies o f a l lclimatological parameters, should be used as input tothese equations.

Where site-specific data are unavailable, short-term concentration levels are estimated using themaximum short-term release rate and climatologicalassumptions presented in Table 3-1. When using

48

Indu

stria

l Sou

rce

Com

plex

Tabl

e 3-

2.R

esou

rce

Req

uire

men

ts a

nd I

nfor

mat

ion

Sour

ces:

Atm

osph

eric

Fat

e M

odel

sR

efer

ence

s, s

ourc

es o

f doc

umen

tatio

n,M

odel

Des

crip

tion

Res

ourc

e R

equi

rem

ents

, co

mm

ents

softw

are

Box

Mod

el??

Are

a S

ourc

e.??

Ava

ilabl

e th

roug

h G

EM

S (

see

Sec

tion

??

Ver

tical

dis

pers

ion

or n

o ve

rtica

l3.

1).

disp

ersi

on o

ptio

n.??

Bas

ic b

ox m

odel

.C

limat

olog

ical

Dis

pers

ion

• Lo

ng-te

rm s

easo

nal

or a

nnua

l.??

Req

uire

s st

abilit

y ar

ray

data

.D

ocum

enta

tion:

Bus

se a

nd Z

imm

erm

anM

odel

(C

DM

)•

Poi

nt o

r ar

ea s

ourc

es.

??

FOR

TRA

N V

pro

gram

lang

uage

; ha

s19

76??

Gau

ssia

n pl

ume

mod

el.

been

Im

plem

ente

d on

the

UN

IVA

C 1

110.

Sof

twar

e: C

ompu

ter

Pro

duct

s, N

TIS

,•

Sim

ulat

es

nonc

onse

rvat

ive

pollu

tant

s.??

22 K

byt

es s

tora

ge r

equi

red.

Spr

ingf

ield

, V

A.

2216

1??

Can

sim

ulat

e tu

rbul

ence

ove

r ur

ban

?S

oftw

are

avai

labl

e as

par

t of

UN

AM

AP

area

s.pa

ckag

e fo

r $4

20.

• O

utpu

ts

long

-term

av

erag

eco

ncen

tratio

ns a

t us

er-s

peci

fied

Ram

CR

STE

R

rece

ptor

s.??

Ope

rate

s in

bot

h lo

ng-te

rm a

nd s

hort-

??

Inte

grat

ed in

to G

EM

S (

see

Sec

tion

3.1)

.D

ocum

enta

tion:

Bow

ers

et a

l. 19

79te

rm m

odes

.??

Sou

rce

data

: lo

catio

n, e

mis

sion

rat

e,S

oftw

are:

Com

pute

r P

rodu

cts,

NTI

S,

?A

ccou

nts

for

settl

ing

and

dry

depo

sitio

nph

ysic

al s

tack

hei

ght,

stac

k ga

s ex

itS

prin

gfie

ld,

VA

. 22

161

of p

artic

les;

dow

nwas

h, a

rea,

line

, and

velo

city

, st

ack

insi

de d

iam

eter

, an

d st

ack

volu

me

sour

ces;

plu

me

rise

as a

func

tion

gas

tem

pera

ture

. O

ptio

nal i

nput

s in

clud

eof

dow

nwin

d di

stan

ce:

sepa

ratio

n of

poi

ntso

urce

ele

vatio

n, b

uild

ing

dim

ensi

ons,

sour

ces;

and

lim

ited

terr

ain

adju

stm

ents

.??

App

ropr

iate

for

ind

ustri

al s

ourc

eco

mpl

exes

, ru

ral o

r ur

ban

area

s, f

lat

orro

lling

terr

ain,

tran

spor

t dis

tanc

es le

ssth

an 5

0 ki

lom

eter

s, a

nd o

ne h

our

toan

nual

ave

ragi

ng t

imes

.

??

Ste

ady-

stat

e G

auss

ian

plum

e m

odel

.??

Ava

ilabl

e co

de o

n U

NIM

AP

(V

ersi

on 6

).?

App

ropr

iate

for

poi

nt a

nd a

rea

sour

ces,

??

Sou

rce

data

: po

int

sour

ces

requ

ireur

ban

area

s, fl

at te

rrai

n tra

nspo

rtlo

catio

n, e

mis

sion

rat

e, p

hysi

cal s

tack

dist

ance

s le

ss th

an 5

0 ki

lom

eter

s, a

ndhe

ight

, st

ack

gas

exit

velo

city

, st

ack

one

hour

to

one

year

ave

ragi

ng t

imes

.in

side

dia

met

er a

nd s

tack

gas

?M

ay b

e us

ed t

o m

odel

prim

ary,

tem

pera

ture

. A

rea

sour

ces

requ

irepo

lluta

nts,

how

ever

set

tling

and

loca

tion,

siz

e, e

mis

sion

rat

e, a

nd h

eigh

tde

posi

tion

are

not

treat

ed.

of e

mis

sion

.

??

Ste

ady-

stat

e G

auss

ian

disp

ersi

onm

odel

.??

Des

igne

d to

cal

cula

te c

once

ntra

tions

from

poi

nt s

ourc

es a

t a s

ingl

e lo

catio

n.??

Hig

hest

and

hig

h-se

cond

hig

hco

ncen

tratio

ns a

re c

alcu

late

d at

eac

hre

cept

or.

??

App

ropr

iate

for

sin

gle

poin

t so

urce

s,ru

ral o

r ur

ban

area

s, tr

ansp

ort d

ista

nces

less

than

50

kilo

met

ers.

and

flat

or

rollin

gte

rrai

n.

parti

cle

size

, dis

tribu

tion-

with

corr

espo

ndin

g se

tting

vel

ociti

es,

and

surfa

ce re

flect

ion.

??

Met

eoro

logi

cal

data

: in

clud

es s

tabi

lity

win

d ro

se (

STA

R d

eck)

, av

erag

eaf

tern

oon

mix

ing

heig

ht,

aver

age

mor

ning

mix

ing

heig

ht,

and

aver

age

air

tem

pera

ture

.

??

Met

eoro

logi

cal

data

: ho

urly

sur

face

wea

ther

dat

a fro

m t

he p

repr

oces

sor

RA

MM

ET.

Act

ual a

nem

omet

er h

eigh

t is

also

requ

ired.

??

Ava

ilabl

e on

UN

IMA

P (

Ver

sion

6).

??

Sou

rce

data

: em

issi

on r

ate,

phy

sica

lst

ack

heig

ht,

stac

k ex

it ve

loci

ty,

stac

kin

side

dia

met

er a

nd s

tack

gas

tem

pera

ture

.??

Met

eoro

logi

cal

data

: ho

urly

sur

face

wea

ther

dat

a fro

m t

he p

repr

oces

sor

RA

MM

ET.

Act

ual

anem

omet

er h

eigh

t IS

also

requ

ired.

Ref

eren

ce:

Turn

er a

nd N

ovak

, 19

78.

Ref

eren

ce:

US

EP

A 1

977b

.

(Con

tinue

d)

Tabl

e 3-

2.(C

ontin

ued)

Ref

eren

ces,

sou

rces

of d

ocum

enta

tion,

Mod

elD

escr

iptio

nR

esou

rce

Req

uire

men

ts,

com

men

tsso

ftwar

eTe

xas

Clim

atol

ogic

al M

odel

Con

trol (

TCM

)*??

Long

-term

(se

ason

al o

r an

nual

).??

Req

uire

s st

abilit

y ar

ray

data

.D

ocum

enta

tion:

Tex

as A

ir C

ontro

l B

oard

• • •

Texa

s E

piso

dic

Mod

el (

TEM

)*

Mod

el M

PTE

R

VA

LLE

Y”

• • ? • •

Ga

us

sia

n d

isp

ers

ion

.•

Two

pollu

tant

s pe

r ru

n.In

clud

es o

ptio

n fo

r si

mul

atio

n of

urb

an•

area

turb

ulen

ce c

lass

es.

?

Han

dles

non

cons

erva

tive

pollu

tant

s.•

Poi

nt o

r ar

ea s

ourc

es.

Up

to 2

,500

rec

epto

r lo

catio

ns o

ndo

wnw

ind

user

-spe

cific

grid

.O

utpu

ts a

vera

ge c

once

ntra

tion

data

.S

tead

y-st

ate

mod

el.

?

Poi

nt o

r ar

ea s

ourc

es.

Sho

rt-te

rm -

10

min

utes

to 2

4 ho

urs.

Pro

duce

s m

axim

um a

nd a

vera

ge•

conc

entra

tions

ove

r tim

e pe

riods

sel

ecte

dby

use

r.•

Use

r ca

n se

lect

up

to 2

,500

dow

nwin

dre

cept

or p

oint

s. a

ccor

ding

to a

nau

tom

atic

or

spec

ified

grid

arr

ay.

Han

dles

non

cons

erva

tive

pollu

tant

s.U

p to

24

met

eoro

logi

c sc

enar

ios

can

bein

put f

or a

sin

gle

run.

Mul

tiple

poi

nt s

ourc

e al

gorit

hm u

sefu

l for

?

estim

atin

g ai

r qu

ality

con

cent

ratio

n of

rela

tivel

y no

n-re

activ

e po

lluta

nts.

App

ropr

iate

for

poi

nt s

ourc

es,

rura

l or

urba

n ar

eas,

flat

or

rollin

g te

rrai

n,tra

nspo

rt di

stan

ces

less

than

50

•ki

lom

eter

s, a

nd o

ne h

our

to o

ne y

ear

aver

agin

g tim

es.

Sho

rt- o

r lo

ng-te

rm.

•S

imul

ates

plu

me

impa

ct in

com

plex

terr

ain.

•P

rovi

des

scre

enin

g es

timat

es o

f wor

st-

case

sho

rt-te

rm c

once

ntra

tions

.?

Pro

vide

s an

nual

ave

rage

con

cent

ratio

ns.

12-r

ecep

tor

grid

.

FOR

TRA

N p

rogr

am la

ngua

ge;

has

been

1980

.Im

plem

ente

d on

Bur

roug

hs 6

810/

11.

Bat

ch m

ode.

17 K

byt

es m

emor

y re

quire

d.Te

chni

cal b

ackg

roun

d in

met

eoro

logy

, ai

rpo

llutio

n us

eful

.

FOR

TRA

N p

rogr

am a

pplic

able

to

a w

ide

Ref

eren

ce:

Chr

istia

nsen

197

6.ra

nge

of c

ompu

ter

type

s; h

as b

een

Impl

emen

ted

on B

urro

ughs

681

0/11

.R

equi

res

appr

oxim

atel

y 26

K b

ytes

mem

ory.

Eng

inee

ring,

met

eoro

logy

, at

mos

pher

ictra

nspo

rt ba

ckgr

ound

use

ful.

Sou

rce

data

: lo

catio

n, e

mis

sion

rat

e,D

ocum

enta

tion:

Pie

rce

and

Turn

er 1

980.

phys

ical

sta

ck h

eigh

t, st

ack

gas

exit

velo

city

, st

ack

insi

de d

iam

eter

, st

ack

gas

Chi

co a

nd C

atal

ano

1986

.te

mpe

ratu

re,

and

optio

nal g

roun

d le

vel

elev

atio

n.M

eteo

rolo

gica

l da

ta:

hour

ly s

urfa

cew

eath

er d

ata

from

the

pre

proc

esso

rR

AM

ME

T. A

ctua

l ane

mom

eter

hei

ght

isal

so re

quire

d.M

ay r

equi

re c

aref

ul a

naly

sis

of o

utpu

t by

Ref

eren

ce:

Bur

t 19

77.

expe

rienc

ed a

ir qu

ality

mod

eler

.S

oftw

are:

Com

pute

r P

rodu

cts,

NTI

S,

FOR

TRA

N V

pro

gram

, ap

plic

able

to

wid

e S

prin

gfie

ld,

VA

221

61.

rang

e of

com

pute

rs.

App

roxi

mat

ely

13 K

byt

es m

emor

yre

quire

d.

Sou

rces

: B

onaz

ount

as e

t al

. 19

82;

US

EP

A 1

979;

US

EP

A 1

982a

.*T

hese

mod

els

are

not E

PA

pre

ferr

ed m

odel

s. T

hey

can,

how

ever

, be

used

if it

can

be

dem

onst

rate

d th

at th

ey e

stim

ate

conc

entra

tions

equ

ival

ent t

o th

ose

prov

ided

by

the

pref

erre

dm

odel

s, e

.g.,

CD

M, R

AM

, IS

C, M

PTE

R. C

RS

TER

. fo

r a

give

n ap

plic

atio

n.**

Thus

mod

el is

reco

mm

ende

d fo

r scr

eeni

ng a

pplic

atio

ns o

nly.

Tabl

e 3-

3.Fe

atur

es o

f A

tmos

pher

ic F

ate

Mod

els

Sou

rce:

Bon

azou

ntes

et

al. 1

982;

US

EP

A 1

979;

US

EP

A 1

982a

.??Th

is m

odel

is

reco

mm

ende

d fo

r sc

reen

ing

appl

icat

ions

onl

y.??*

Thes

e m

odel

s ar

e no

t EP

A p

refe

rred

mod

els.

The

se m

odel

s ca

n be

use

d if

they

can

be

dem

onst

rate

d to

est

imat

e co

ncen

trat

ions

equ

ival

ent

to t

hose

pro

vide

d by

the

pre

ferr

edm

odel

s, e

.g.,

CPM

, R

AM

, IS

C,

MPT

ER,

CR

STER

. fo

r a

give

n ap

plic

atio

n.

site-specific data, the most stable atmospheric process as an extended screening tool to highlightconditions, lowest wind speed, and greatest percent contaminant releases to surface water that actuallyof wind flow toward the exposed population should be require detailed environmental fate (and subsequentused as input to Equation 3-1, along with maximum exposed populations) analysis. Contact the USEPArelease rate estimates for the duration of interest. Office of Toxic Substances, Exposure EvaluationUsually, the population nearest the point or area of a Division (Pat Kennedy, (202) 382-3916) for moreground-level release experiences the highest short- detailed information on accessing the Probabilisticterm exposure. Dilution Model.

As indicated in Table 3-2, several atmospheric fatemodels have the capability of producing short-termm a x i m u m a n d l o n g - t e r m a v e r a g e a m b i e n tconcentration estimates where in-depth analysis isdesirable.

3.4.1 Beginning Quantitative AnalysisThe following equation (adapted from Delos et al.1984) provides a rough estimate of the concentrationof a substance downstream from a point sourcerelease into a flowing waterbody, after dilution of thesubstance by the receiving waterbody:

3.4 Surface Water Fate Analysis

The environmental fate of hazardous materialsentering surface waterbodies is highly dependent onthe type of waterbody. The three major classificationsare rivers and streams, impoundments, and estuaries.Methods for estimating contaminant concentrations inthe first category are provided below.

(3-4)

As mentioned in the introduction to this chapter,contamination of flowing waterbodies will probably bea more common occu r rence w i th rega rd touncontrolled hazardous waste facilities than willcontamination of impoundments or estuaries. Thus, inthis section guidance for estimating contaminant fatein flowing waterbodies is presented. In those caseswhere contaminant fate in an impoundment or estuaryis necessary, the analyst is referred to Mills et al.(1982) for guidance.

C

Ce

QeQt

= 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 substancein the waterbody resulting from contaminant releasesfrom the subject site alone; it does not take intoaccount addit ional sources of contaminat ion(“background” concentrat ions) that may alsocontribute to the total level of contamination in thewaterbody.

The Probabilistic Dilution Model is an analytical toolthat can be used to extend the qualitative screeninganalysis presented in the previous section and that insome cases may make application of the quantitativea n a l y s e s d i s c u s s e d i n f o l l o w i n g s e c t i o n sunnecessary. This model has been adapted by theU.S. Environmental Protection Agency, Office ofTox ic Substances , to suppor t the exposureassessment process for contaminants in surfacewater. The model is based on the fact that, ingeneral, the most important process affecting acontaminant’s concentration in a surface waterbody isthe degree of its dilution. Thus, the model usesstreamflow data for a given subbasin and contaminantloading data (from the contaminant release analysisdiscussed in Chapter 2) to predict the number oftimes per year a given contaminant concentration willbe exceeded. For contaminants that have health-based concentration standards (or for which health-based concen t ra t ion cu t -o f f va lues can becalculated), the model can be used to predict theannua l number o f occur rences (days ) tha tunacceptable health risks may result for personsusing the affected waterbody. This model can beapplied to the Superfund exposure assessment

In cases where hazardous waste is introduced into astream through intermedia transfer from air, soil,ground water, or nonpoint source, or where therelease rate is known in terms of mass per unit timerather than per unit effluent volumes, in-streamconcentrations can be estimated by use of thefollowing equation:

(3-5)

Tr = i n t e r m e d i a t r a n s f e r r a t e ,(mass/time)

Qt = stream flow rate after intermediat r a n s f e r h a s o c c u r r e d ,(volume/time).

Assumptions implicit in these equations are thefollowing:

? Mixing of the hazardous substance in the water isinstantaneous and complete.

53

?? The hazardous material is refractive (i.e., alldecay or removal processes are disregarded).

?? Stream flow and rate of contaminant release tothe stream are constant (i.e., steady-stateconditions).

The assumption of complete mixing of a hazardoussubstance in a flowing water body is not valid within amixing zone downstream from the point or reachwhere the substance is introduced. Under certainconditions, this mixing zone can extend downstreamfor a considerable distance, and concentrations canbe considerably higher within the mixing zone thanthose estimated by the foregoing dilution equations.

found through this estimation procedure to be dilutedto concentrations below a predetermined level ofconcern, and no important exposure points existwithin the mixing zone, the fate of the substance inthis medium may need no further analysis. However,where the concen t ra t ion a f te r d i lu t ion o f anonconserva t ive subs tance is s t i l l above apredetermined critical level, it may be useful toestimate the distance downstream where theconcentration will remain above this level, as well asthe concentration of the substance at selectedexposure points downstream.

This type of estimation can be performed through useof an overall decay coefficient, which represents acombination of all decay and loss rates affecting theremoval of a substance from a waterbody. Theconcentration of a nonconservative substance at aselected point downstream from the release point andbelow the mixing zone (complete mixing is assumed)can be estimated by the following equation (fromDelos et al. 1984), which employs the concept of anoverall decay coefficient:

If the hazardous substance is introduced into aflowing waterbody over a length of that body, ratherthan from a point source, assume that the mixingzone begins at the downstream end of the reach overwhich introduction takes place. Neely (1982) presentsan estimation procedure for hazardous substanceconcentration at exposure points within a mixing zonethat incorporates an expression for dispersion.

The dilution equations (3-4, 3-5) and the procedurepresented by Neely (1982) assume that the introducedhazardous substance is conservative. Therefore, theypredict an estimated stream/river concentration thatremains constant from the downstream end of themixing zone throughout the remaining length of thestream, or decreases only with further dilutionresulting from additional stream flow from tributaries.This is useful as a basic model for the fate ofc o n s e r v a t i v e h a z a r d o u s s u b s t a n c e s ; fo rnonconservative substances, it provides a usefulworst-case estimate. If the released substance is

(3-7)

where

where

54

W(O) = concentration immediately below point ofintroduction, (from dilution Equations 3-4, 3-5).

This equation incorporates the following assumptions:

?? Mixing is complete.

? Conditions are steady state.

?? Longitudinal dispersion is negl igible; thesubstance transports downstream at streamvelocity.

?? All decay and transfer processes can bedescribed as first-order coefficients (i.e., decayrates are a direct funct ion of hazardoussubstance concentration).

Values for K can be derived empirically wheremonitoring data are available, or can be estimatedbased on decay rate constants available for manyhazardous substances in the technical literature.

Concentration data from immediately below the pointof substance release into a stream (after completemixing of waste stream into the waterbody), and fromat least one point downstream of the mixing zone arerequired for the empirical estimation of K. Note thatoverall decay coefficients are substance- and site-specific and can vary with climatic and hydrologicconditions. Care must be taken in calibrating thecoefficient empirically. Data covering seasonalfluctuations must be used, and seasonal values for Kcorresponding to the various observed conditions, ora worst-case K value (i.e., lowest reasonable value)for the purpose of conservative estimation, should bedeveloped.

For estimation of K through the summation ofpublished decay rate constants, the most importantremoval process affecting the compound of concernin the receiving waterbody must be known. For thisinformation, 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 constantvalues 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 fora given waterbody and hazardous substance underno-ac t ion cond i t ions ( i .e . , dur ing remed ia linvestigation), can be used to estimate the fate of thissame substance resulting from the release ratesprojected after implementation of various remedialaction alternatives.

3.4.2 In-Depth AnalysisWhen aquatic concentration estimates developed bythe above simplified methods (or methods covering

estuaries or impoundments provided by Mills et al.1982) indicate that these concentrations pose apotential human health hazard at one or moreexposure points, more accurate estimates of short-term and long-term concentrations of the hazardoussubstance may be required. A large number of in-depth methods and computer models exist to assessthe fate of substances in the aquatic environment.Each of these models differs in the number and typesof aquatic fate processes that it incorporates. Themost important of these aquatic processes aredescribed below, and information is provided to allowidentification of those processes most likely to besignificant at the site, and for the hazardoussubstances under analysis.

3.4.2.1 Intermedia TransfersThe major processes by which hazardous substancescan be transferred from surface water to otherenvironmental media are as follows:

(1) VolatilizationVolatilization of a substance from water depends onthe physicochemical properties of the substance andcharacteristics of the waterbody and body of airinvolved. Volatilization increases in importance forsubstances with higher vapor pressure, and forwaterbodies with higher surface area-to-volumeratios and higher turbulence (Delos et al. 1984).Callahan et al. (1979) stress the importance ofvolatilization as a route of intermedia transfer for 129priority pollutants. If volatilization is considered animportant process for the substance being studied, orif the importance of volatilization is unknown, the rateof volatilization can be estimated by the methodprovided by Mi l ls et al . (1982) for quiescentwaterbodies or by Delos et al. (1984) for turbulentbodies. Lyman et al. (1982) provide methods forestimating volatilization rates from water.

(2) SedimentationHazardous substances released to a surfacewaterbody in the solid, particulate form will settle outover time and become mixed into the bottomsediment. In addition, liquid hazardous substanceswith high affinities for adsorption to suspendedparticulates will settle out of surface waters with theseparticulates. The rate of sedimentation is governed bythe di f ference between sett l ing veloci ty andresuspension velocity. The former increases withmean particle size and density and with watertemperature, and can be estimated by the procedurepresented by Delos et al. (1984). Resuspensionvelocity is a function of bottom shear stress. Delos etal. (1984) provide a procedure to estimate this rate.Where sedimentation is considered to be an importantprocess, use a surface water fate model that has thecapability of accounting for bed-water exchange andsediment load transport.

55

(3) SorptionSubstances dissolved in surface waters can sorb ontosolids suspended in the water or onto bed sediments.This process, in effect, transfers the substances fromthe water to the sediment medium, and proceeds untilan equilibrium point is reached. This equilibrium point(and the resulting water and sediment concentrationsof the substance) is determined by the soil-waterpartition coefficient (a parameter that is a function ofsediment type, water pH, cation exchange capacity,and o rgan ic con ten t o f sed iment ) and thephysicochemical propert ies of the hazardoussubstance. In general, metals and hydrophobic,nonpolar organic compounds have a high tendency tosorb onto entrained or bottom sediment. See Lymanet al. (1982) for methods of estimating sedimentadsorption of waterborne contaminants.

3.4.2.2 lntramedia Transformation ProcessesThe following is a brief description of the importantintramedia transformation processes that may besignificant for the surface water fate of hazardoussubstances. Rate-controlling factors are stated foreach. Callahan et al. (1979), Mabey et al. (1982),Verschueren (1984), and Sax (1984) provide rateconstants for these processes for numerouscompounds.

(1) PhotolysisChemical transformation due to photolysis utilizesenergy from sunlight, and for some chemicals, canoccur by several processes. Direct photolysis ratesare a function of photon availability, light absorptioncoefficients for the chemical in question, and areaction yield constant (i.e., the efficiency ofsubstance transformation with the available solarenergy). Indirect photolysis occurs through the actionof intermediate substances naturally occurring in themedium. These intermediates absorb light energy byvarious processes and in this energized state, reactwith the hazardous substance. Indirect photolysis is afunction of photon availability, concentration and lightabsorption coefficient of the intermediate, and a rateconstant for the reaction between the energizedintermediate and the hazardous material.

(2) OxidationOxidation is the reaction of substances with oxidantspecies. Oxidation rates are a function of theconcentrat ions of the substance in quest ion,concentration of the oxidant, and a rate constant forreaction between them.

(3) HydrolysisHydrolysis is the nucleophilic displacement of anelectronegative substituent on a carbon atom by anhydroxyl group. The nucleophilic reactant can beeither a water molecule or an hydroxyl ion. Hydrolysisof most compounds is highly dependent on the pH ofthe waterbody medium and can be promoted by both

acid and base conditions. The rate of hydrolysis is afunction of the concentration of the hazardoussubstance and the rate constants for the acid- andbase-promoted processes at each pH value.

(4) BiodegradationBiodegradation is the breakdown of substancesthrough the enzymatic action of biota present in thewater. Most biodegradation is carried out by microbialbiota. It depends on the metabolic rates andcharacteristics and the population density of the bioticagents, which are in part functions of the availabilityof other nutrients, pH and temperature of themedium,, and sunlight availability, among otherfactors.

3.4.2.3 Computer ModelsTables 3-5, 3-6, and 3-7 summarize the features,data requirements, resource requirements, andreferences or contacts for selected computer-basedmodels appropriate to the in-depth analysis of theaquatic fate of hazardous releases from Superfundsites. Additional details for certain of the modelsaddressed in the tables are provided below:

Exposure Analysis Modeling System (EXAMS-II)(Burns et al., 1982) is a steady-state and dynamicmodel designed for rapid evaluation of the behavior orsynthetic organic chemicals in lakes, rivers, andestuaries. EXAMS-II is an interactive program thatallows the user to specify and store the properties ofchemicals and ecosystems, modify the characteristicsof either via simple English-like commands, andconduct rapid, efficient evaluations of the probablefate of chemicals. EXAMS-II simulates a toxicchemical and its transformation products usingsecond-order kinetics for all significant organicchemical reactions. EXAMS-II, however, does notsimulate the solids with which the chemical interacts.The concentration of solids must be specified foreach compartment; the model accounts for sorbedchemical transport based on solids concentrationsand specified transport fields. Benthic exchangeinc ludes pore-water advec t ion , pore-waterdiffusion, and solids mixing. The latter describes a netsteady-state exchange associated with solids that isproportional to pore water diffusion.

A da ta se t o f average or t yp ica l va lues fo rwaterbody-spec i f i c da ta i s p resent ly be ingdeveloped by Battelle Northwest Laboratories, undercontract to EPA. This data file will contain parametervalues for a number of major U.S. river systems,lakes, and reservoirs, and will be integrated with theEXAMS program. These values will be accessible forfate modeling of the waterbodies included (GSC1982).

MlNTEQA1 (Felmy et al., 1984; Brown and Allison,1987) is a geochemical model that is capable of

ca lcu la t ing equ i l i b r ium aqueous spec ia t ion ,adsorption, gas phase partitioning, solid phasesaturation states, and precipitation-dissolution of 11metals (arsenic, cadmium, chromium, copper, lead,mercury, nickel, selenium, silver, thallium, and zinc).MINTEQA1 contains an extensive thermodynamicdata base and contains six different algorithms forcalculat ing adsorpt ion. Proper appl icat ion ofMINTEQA1 requires applicable expertise, becausekinetic limitations at particular sites may preventcertain react ions even though they might bethermodynamically possible.

Hydrological Simulation Program - FORTRAN(HSPF) (Johanson et al., 1984; Donigian et al., 1984)is a comprehensive package for simulation ofwatershed hydrology and water quality for bothconventional and toxic organic pollutants. HSPFincorporates the watershed-scale ARM (AgriculturalRunoff Model) and NPS (Non-Point Source) modelsinto a basin-scale analysis framework that includespollutant transport and transformation in streamchannels.

The model uses information such as the time historyof rainfall, temperature, and solar radiation; landsurface characteristics such as land use patterns andsoil properties; and land management practices tosimulate the processes that occur in a watershed.The result of this simulation is a time history of thequantity and quality of runoff from an urban oragricultural watershed. Flow rate, sediment load, andnutrient and pesticide concentrations are predicted.The program takes these results, along withinformation about the stream network and pointsource discharges, and simulates instream processesto produce a time history of water quantity and qualityat any point in a watershed -- the inflow to a lake,for example. HSPF includes an internal data basemanagement system to process the large amounts ofsimulation input and output.

Water Analysis Simulation Program (WASP4)(Ambrose et al., 1986, 1987) is a generalizedmodeling framework for contaminant fate andtransport in surface water. Based on the flexiblecompartment modeling approach, WASP can beapplied in one, two, or three dimensions. WASP isdesigned to permit easy substitution of user-writtenroutines into the program structure. Problems thathave been studied using WASP include biochemicaloxygen demand, dissolved oxygen dynamics,nutrients and eutrophication, bacterial contamination,and toxic chemical movement.

A variety of water quality problems can be addressedwith the selection of appropriate kinetic subroutinesthat may be either selected from a library or writtenby the user. Toxics WASP (TOX14; Ambrose et al.,1987) combines a kinetic structure adapted fromEXAMS with the WASP transport structure and simple

sediment balance algorithms to predict dissolved andsorbed chemical concentrations in the bed andoverlying waters.

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

SARAH (Ambrose and Vandergrift, 1986) is as t e a d y - s t a t e m i x i n g z o n e m o d e l f o r b a c k -calculating acceptable concentrations of hazardouswastes discharged to land disposal or waste watertreatment facilities. For steady or batch wastestreams, S A R A H c o n s i d e r s t h e f o l l o w i n gconcentration reductions: dilution and loss duringtreatment, initial Gaussian mixing at the edge of astream, lateral and longitudinal diffusion in the mixingzone, sorption, volat i l izat ion, hydrolysis, andbioaccumulation in fish. The user must specify,appropriate in-stream criteria for protection of theaquatic community, and humans through consumptionof fish and water. The benthic community is notpresently considered. Treatment loss is handledempi r i ca l l y . The human exposure pa thwaysconsidered include ingestion of treated drinking waterand consumption of contaminated fish.

3.4.2.4 Short- and Long-Term ConcentrationCalculationsLong-term average ambient water concentrationsshould be calculated using (1) the average releaserate (from Chapter 3) projected for the time period ofinterest, and (2) the annual average stream flow rateas input to the above estimation procedures.

Short-term concentration levels are obtained throughuse of the short-term release rate developed duringcontaminant release analysis and the lowestreasonable 24-hour flow rate, or the 7-day, l0-year (7-Q-10) low flow rate for the period ofrecord, as presented in the above estimationprocedures.

Table 3-6 indicates several aquatic fate modelscapable of estimating both short- and long-termambient water concentrations that are appropriate toin-depth analysis of the aquatic fate of contaminantsreleased from Superfund sites.

3.5 Quantitative Analysis of Ground-Water Fate

To model the migration of contaminants in groundwater the following factors should be estimated:

57

Tabl

e 3-

5.R

esou

rce

Req

uire

men

ts a

nd In

form

atio

n So

urce

s: S

urfa

ce W

ater

Fat

e M

odel

sR

efer

ence

s, s

ourc

es o

f doc

umen

tatio

n,M

odel

Des

crip

tion

Res

ourc

e R

equi

rem

ents

, co

mm

ents

softw

are

• S

tead

y-st

ate,

1 -

dim

ensi

onal

mod

el?

Eas

y to

set

up

and

use

Ref

eren

ce:

Mills

et

al.

1982

Wat

er Q

ualit

y A

sses

smen

t M

etho

dolo

gy(W

QA

M)

Sim

plifi

ed L

ake/

Stre

am A

naly

sis

(SLS

A)

Mic

higa

n R

iver

Mod

el (

MIC

HR

IV)

Che

mic

al T

rans

port

and

Ana

lysi

s P

rogr

am(C

TAP

)

Exp

osur

e A

naly

sis

Mod

elin

g S

yste

m(E

XA

MS

-II)

?R

equi

res

only

des

k to

p ca

lcul

atio

ns??

Pro

vide

s ca

noni

cal i

nfor

mat

ion

??

Mod

els

lake

s, r

iver

s, a

nd e

stua

ries

??

Ste

ady-

stat

e, 1

-di

men

sion

al m

odel

?S

olut

ion

eith

er b

y de

sk t

op c

alcu

latio

ns o

rsi

mpl

e FO

RTR

AN

pro

gram

•S

uita

ble

for

sim

plifi

ed la

ke a

nd r

iver

syst

ems

??

Ste

ady-

stat

e, 1

-di

men

sion

al m

odel

??

Com

pute

r pr

ogra

m w

ritte

n in

FO

RTR

AN

?S

imila

r to

SLS

A,

but

can

mod

el m

ore

than

one

rea

ch??

Inte

nded

for

met

als

??

Mod

els

river

s an

d st

ream

s

??

Ste

ady-

stat

e, d

-dim

ensi

onal

com

partm

enta

l mod

el??

FOR

TRA

N I

V p

rogr

am s

uita

ble

for

num

erou

s co

mpu

ters

?S

imila

r to

SLS

A e

xcep

t m

ore

soph

istic

ated

; ea

ch C

TAP

com

partm

ent

is e

quiv

alen

t to

one

SLS

A “l

ake”

??

Mod

els

stre

ams,

stra

tifie

d riv

ers,

lake

s,es

tuar

ies,

and

coa

stal

em

baym

ents

??

Ste

ady-

stat

e, 3

-dim

ensi

onal

com

partm

enta

l m

odel

• C

ompl

ex c

ompu

ter

prog

ram

??

Con

tain

s co

mpr

ehen

sive

sec

ond-

orde

rde

cay

kine

tics

for

orga

nics

; mos

t mod

els

only

hav

e fir

st-o

rder

kin

etic

s??

Mod

els

orga

nic

chem

ical

s??

Sui

tabl

e fo

r fre

shw

ater

, no

n-tid

al a

quat

icsy

stem

s

??

No

com

pute

r pr

ogra

mm

ing

need

ed;

requ

ires

only

han

d ca

lcul

ator

??

Rec

omm

ende

d if

time,

cos

ts,

orin

form

atio

n ar

e re

stric

tive

??

Eas

y to

set

up

and

use

• C

ompu

ter

prog

ram

min

g no

t ne

cess

ary;

if

used

, on

ly 2

80 b

ytes

are

req

uire

d;su

itabl

e fo

r m

icro

com

pute

rs??

Wel

l doc

umen

ted

and

sugg

este

d fo

r us

ebe

fore

use

of a

mor

e so

phis

ticat

ed m

odel

??

May

be

used

with

han

d ca

lcul

ator

??

Eas

y to

set

up

and

use

??

Req

uire

s m

inim

al c

ompu

ter

prog

ram

min

g

??

Req

uire

s ex

tens

ive

data

inpu

t??

FOR

TRA

N p

rogr

am -

sui

tabl

e fo

r IB

M36

0/37

0, U

NIV

AC

108

, C

DC

660

0m

ainf

ram

e co

mpu

ters

??

Mic

roco

mpu

ter

vers

ion

avai

labl

e re

quiri

ng32

K b

ytes

sto

rage

??

One

of

the

bette

r do

cum

ente

d m

odel

s,w

hich

may

mak

e it

mor

e de

sira

ble

than

othe

r co

mpl

ex m

odel

s??

Req

uire

s ex

tens

ive

data

inpu

t??

Has

bee

n in

corp

orat

ed in

to E

PA

-OTS

GE

MS

sys

tem

(se

e S

ectio

n 4.

1)??

Wel

l doc

umen

ted

and

reco

mm

ende

d fo

rus

e ov

er m

ost o

ther

mod

els

• A

vaila

ble

on m

agne

tic t

ape

for

inst

alla

tion

on m

ainf

ram

e or

sm

all c

ompu

ters

(e.

g.,

PD

P-1

1 or

HP

300

0);

batc

h ve

rsio

nre

quire

s 64

K b

ytes

mem

ory

at a

min

imum

, mor

e fo

r co

mpl

ex m

odel

ing

??

Als

o av

aila

ble

in i

nter

activ

e ve

rsio

n,re

quiri

ng 1

64 K

byt

es m

emor

y pl

us 2

Kby

tes

for

each

che

mic

al a

nd 2

.5 K

byt

esfo

r ea

ch e

nviro

nmen

tA

n es

timat

ed 3

50 h

ours

req

uire

d fo

rin

stal

latio

n an

d se

tup,

ass

umin

g al

l dat

a

Doc

umen

tatio

n:O

RD

Pub

licat

ions

US

EP

A,

Cin

cinn

ati,

Ohi

o 45

268

(513

) 68

4-75

62R

efer

ence

: H

ydro

Qua

l 19

82D

ocum

enta

tion:

Will

iam

Gul

ledg

e25

81 M

Stre

et,

N.W

.W

ashi

ngto

n, D

C.

2003

7(2

02)

887-

1183

Ref

eren

ce:

Del

os e

t al

. 19

84Te

chni

cal

Ass

ista

nce

Ava

ilabl

e fro

m:

Bill

Ric

hard

son

US

EP

AE

nviro

nmen

tal

Res

earc

h La

bora

tory

-D

ulut

hLa

rge

Lake

s R

esea

rch

Sta

tion

Ref

eren

ce:

Hyd

roQ

ual

1982

Doc

umen

tatio

n:W

illia

m G

ulle

dge

Che

mic

al M

anuf

actu

rers

Ass

ocia

tion

2581

M S

treet

, N

.W.

Was

hing

ton,

D.C

. 20

037

(202

) 88

7-11

83

Ref

eren

ce:

Bur

ns e

t al

. 19

82D

ocum

enta

tion:

OR

D P

ublic

atio

ns,

Cen

ter

for

Env

ironm

enta

l Res

earc

h In

form

atio

nU

SE

PA

Cin

cinn

ati,

Ohi

o 45

268

(513

) 68

4-75

62C

ente

r fo

r W

ater

Qua

lity

Mod

elin

gE

nviro

nmen

tal

Res

earc

h La

bora

tory

US

EP

AA

then

s, G

a. 3

0613

(404

) 54

6-35

85

are

read

ily a

vaila

ble

(Con

tinue

d)

Tabl

e 3-

5.(C

ontin

ued)

Ref

eren

ces,

sou

rces

of d

ocum

enta

tion,

Mod

elD

escr

iptio

nR

esou

rce

Req

uire

men

ts,

com

men

tsso

ftwar

eM

etal

s E

xpos

ure

Ana

lysi

s M

odel

ing

Sys

tem

?(M

INTE

QA

1)• ? ?

Ste

ady-

stat

e, 3

-dim

ensi

onal

com

partm

enta

l mod

elC

ompl

ex c

ompu

ter

prog

ram

Des

igne

d fo

r m

odel

ing

of m

etal

load

ings

Sui

tabl

e fo

r fre

shw

ater

, no

n-tid

al a

quat

icsy

stem

s

Com

plex

met

al d

ynam

ics

requ

iring

exte

nsiv

e da

ta in

put

Can

be

used

with

mai

nfra

me

or s

mal

l(e

.g.,

PC

P 1

1/70

or

HP

300

0) c

ompu

ters

iner

activ

e fo

rmat

Furth

er i

nfor

mat

ion:

Yas

uo O

nish

iB

atte

lle P

acifi

c N

orth

wes

t La

bora

torie

sR

ichl

and,

WA

993

52(5

09)

376-

8302

Con

tain

s da

ta b

ase

with

the

rmod

ynam

icpr

oper

ties

of 7

met

als

Req

uire

s ex

tens

ive

data

inpu

tM

ost s

uita

ble

to m

inic

ompu

ters

(e.

g., H

P30

00,

PR

IME

. H

AR

RIS

) as

mod

elut

ilize

s di

rect

acc

ess

inpu

t-out

put,

whi

chca

n be

cos

tly o

n m

ainf

ram

e co

mpu

ters

Req

uire

s 25

0 K

byt

es o

f ov

erla

y-ty

pest

orag

eH

as b

een

used

on

IBM

370

ser

ies

com

pute

rs

Hyd

rolo

gica

l S

imul

atio

n P

rogr

amFO

RTR

AN

(H

SP

F)

Tran

sien

t O

ne-D

imen

sion

al D

egra

datio

n•

and

Mig

ratio

n M

odel

(TO

DA

M)

• • ?

Cha

nnel

Tra

nspo

rt M

odel

(C

HN

TRN

)? ? ? ?

Fini

te E

lem

ent T

rans

port

Mod

el (

FETR

A)

? ? ? ?

Tim

e-va

ryin

g, 1

-di

men

sion

al m

odel

Des

igne

d fo

r ye

ar-r

ound

sim

ulat

ion

Mod

els

orga

nic

pollu

tant

sS

econ

d-or

der

deca

y m

echa

nism

sM

odel

s no

n-tid

al r

iver

s, s

tream

s, a

ndm

ixed

lake

s

Tim

e-va

ryin

g, 1

-di

men

sion

al m

odel

Sec

ond-

orde

r de

cay

mec

hani

sms

Mod

els

river

and

est

uarin

e sy

stem

sR

equi

res

exte

rior

hydr

odyn

amic

mod

el(e

.g.,

EX

PLO

RE

) to

pro

vide

cha

nnel

and

flow

vel

ociti

es t

o TO

DA

M

Tim

e-va

ryin

g, 1

-di

men

sion

al m

odel

Mod

els

orga

nic

pollu

tant

sS

econ

d-or

der

deca

y m

echa

nism

sM

odel

s riv

ers,

lake

s, e

stua

ries,

and

coas

tal w

ater

sC

an b

e co

uple

d w

ith a

hyd

rody

nam

icm

odel

, C

HN

HY

D,

to e

stim

ate

flow

dyna

mic

s w

here

suc

h da

ta a

re n

otav

aila

ble

Tim

e-va

ryin

g, P

-dim

ensi

onal

mod

el(lo

ngitu

dina

l and

late

ral)

Sec

ond-

orde

r de

cay

mec

hani

sms

for

orga

nic

pollu

tant

sM

odel

s riv

ers,

est

uarie

s, c

oast

al s

yste

ms,

and

com

plet

ely

mix

ed la

kes

Can

be

coup

led

with

EX

PLO

RE

hydr

odyn

amic

mod

el t

o ge

nera

te f

low

velo

citie

s w

here

the

se a

re u

nkno

wn

? ? ? ?

Req

uire

s ex

tens

ive

data

inpu

tC

ompl

ex F

OR

TRA

N p

rogr

am,

writ

ten

inth

e pr

epro

cess

or la

ngua

ge F

LEC

S o

r in

FOR

TRA

N I

VA

pplic

able

to

VA

X o

r P

DP

11/

70co

mpu

ters

(ba

tch

mod

e)TO

DA

M h

as b

een

appl

ied;

how

ever

,do

cum

enta

tion

is c

urre

ntly

und

er r

evie

w;

rele

ase

date

unk

now

nR

equi

res

exte

nsiv

e da

ta in

put,

and

exte

nsiv

e se

tup

time

Has

not

bee

n fie

ld-te

sted

, and

docu

men

tatio

n is

cur

rent

ly u

nder

rev

iew

FOR

TRA

N I

V p

rogr

am la

ngua

geA

pplic

able

to

IBM

393

3 co

mpu

ter,

and

othe

rs

Inpu

t dat

a re

quire

men

ts a

re e

xten

sive

Com

pute

r pr

ogra

m w

ritte

n in

FO

RTR

AN

IV Can

be

used

on

IBM

. V

AX

, or

CD

C-

7600

com

pute

rsH

as b

een

field

-val

idat

edS

etup

and

exe

cutio

n tim

e re

quire

men

tsar

e ex

tens

ive

Ref

eren

ce: J

ohan

son

et a

l. 19

84So

ftwar

e:C

ente

r fo

r W

ater

Qua

lity

Mod

elin

gE

nviro

nmen

tal

Res

earc

h La

bora

tory

US

EP

AA

then

s, G

A 3

0613

(404

) 54

6-35

85

Ref

eren

ce: O

nish

i et a

l. 19

82Fu

rther

inf

orm

atio

n:Y

asuo

Oni

shi

Bat

telle

Pac

ific

Nor

thw

est

Labo

rato

ries

Ric

hlan

d, W

A 9

9352

(509

) 37

6-83

02

Ref

eren

ce: Y

eh 1

982

Doc

umen

tatio

n:D

r. G

. T. Y

ehE

nviro

nmen

tal S

cien

ces

Div

isio

nO

ak R

idge

Nat

iona

l La

bora

tory

P.O

. Box

XO

ak R

idge

, TN

378

30(6

15)

574-

7285

Ref

eren

ce:

Oni

shi

1981

Furth

er i

nfor

mat

ion:

Yas

uo O

nish

iB

atte

lle P

acifi

c N

orth

wes

t La

bora

torie

sR

ichl

and,

WA

993

52(5

09)

376-

8302

(Con

tinue

d)

Tabl

e 3-

5.(C

ontin

ued)

Mod

elS

edim

ent-C

onta

min

ant

Tran

spor

t(S

ER

ATR

A)

Des

crip

tion

??

Tim

e-va

ryin

g, P

-dim

ensi

onal

mod

elR

esou

rce

Req

uire

men

ts,

com

men

ts??

Req

uire

s ex

tens

ive

data

inpu

t(lo

ngitu

dina

l-and

ver

tical

)??

Com

plex

sed

imen

t tra

nspo

rt m

echa

nism

s??

Com

pute

r pr

ogra

m w

ritte

n in

FO

RTR

AN

prep

roce

ssor

lang

uage

FLE

CS

< in

??

Sec

ond-

orde

r de

cay

mec

hani

sms

for

ba

tc

h

mo

de

orga

nic

pollu

tant

s??

Has

bee

n fie

ld-te

sted

and

is a

vaila

ble

• M

odel

s riv

ers

and

lake

sfo

r us

e??

Req

uire

s an

est

imat

ed 7

50 m

an-h

ours

for

setu

p, a

ssum

ing

all r

equi

red

data

are

read

ily a

vaila

ble

Est

uary

and

Stre

am Q

ualit

y M

odel

(W

AS

P4)

??

Tim

e-va

ryin

g, 3

-dim

ensi

onal

mod

el??

Sop

hist

icat

ed s

econ

d-or

der

orga

nic

deca

y ki

netic

s??

Mod

els

river

s, la

kes,

and

est

uarie

s

??

Ver

y da

ta-in

tens

ive

mod

el??

Use

r m

ust

prov

ide

hydr

odyn

amic

flo

ws

betw

een

mod

el c

ompa

rtmen

ts?

App

licab

le t

o IB

M 3

70 o

r P

DP

11/

70sy

stem

s??

FOR

TRA

N I

V p

rogr

am r

equi

res

64 K

byte

s m

emor

y??

Req

uire

s 15

0-30

0 m

an-h

ours

for

setu

p

Ref

eren

ces,

sou

rces

of d

ocum

enta

tion,

softw

are

Ref

eren

ce:

Oni

shi

and

Wis

e 19

82a,

Gni

shi

and

Wis

e 19

82b

Doc

umen

tatio

n:O

RD

Pub

licat

ions

Cen

ter

for

Env

ironm

enta

l Res

earc

hIn

form

atio

nU

SE

PA

Cin

cinn

ati,

OH

452

68(5

13)

684-

7562

Tech

nica

l As

sist

ance

:R

ober

t A

mbr

ose

EP

A A

then

s E

nviro

nmen

tal R

esea

rch

Lab

Cen

ter

for

Wat

er Q

ualit

y M

odel

ing

Ath

ens,

GA

306

13(4

04)

546-

3546

Doc

umen

tatio

n an

d So

ftwar

e:D

r. Jo

hn C

onno

llyE

nviro

nmen

tal E

ngin

eerin

g an

d S

cien

ceM

anha

ttan

Col

lege

Bro

nx, N

.Y. 1

0471

(212

) 92

0-02

76 o

r:D

r. P

arm

ely

H. P

richa

rdE

nviro

nmen

tal

Res

earc

h La

bora

tory

Gul

f Bre

eze,

FL

3256

1(9

04)

932-

5311

Sur

face

Wat

er B

ack

Cal

cula

tion

Pro

cedu

re(S

AR

AH

)??

Ste

ady-

stat

e, 1

-di

men

sion

al a

naly

tical

solu

tion

??

FOR

TRA

N C

ode

??

Mod

els

cont

amin

ated

lea

chat

e pl

ume

feed

ing

the

dow

ngra

dien

t su

rface

wat

erbo

dy (

stre

am o

r riv

er)

??

Mon

te C

arlo

sim

ulat

ed g

ener

icen

viro

nmen

t??

Deg

rada

tion,

dilu

tion,

sor

ptio

n, a

ndvo

latiliz

ation

??

Gen

eric

env

ironm

ent,

min

imal

dat

a in

put

• FO

RTR

AN

m

odel

Rob

ert

Am

bros

eC

ente

r fo

r W

ater

Qua

lity

Mod

elin

gU

SE

PA

Ath

ens,

GA

306

13(4

04)

546-

3546

Doc

umen

tatio

n: J

an.

14,

1986

Fede

ral R

egis

ter,

Haz

ardo

us W

aste

Man

agem

ent

Sys

tem

, La

nd D

ispo

sal

Res

trict

ions

, Pro

pose

d R

ule

Softw

are.

Dav

id D

isne

y, E

nviro

nmen

tal R

esea

rch

Labo

rato

ry,

Env

ironm

enta

l P

rote

ctio

nA

genc

y, C

olle

ge S

tatio

n R

oad,

Ath

ens,

GA

306

13,

(404

) 54

6-54

32,

or (

404)

546-

3123

??

Bro

accu

mul

atio

n in

fish

Sou

rce:

Ver

sar

1983

a.

Direction - The direction of contaminant migration isimportant in predicting the potentially exposedpopulation.

Velocity - The migrating contaminant’s velocity isimportant in assessing when contamination will reacht h e e x p o s e d p o p u l a t i o n a n d h o w l o n g t h econtamination will be affecting that population.

Concentration - Concentration of the contaminant inground water at the exposure locations is used tocalculate dose to the population. This factor is usedto convert the amount of water consumed each dayto the mass of contaminant received each day. Themass information is then used to predict healtheffects associated with exposure to the contaminant(USEPA 1985d).

Volume - The contaminated region’s volume isi m p o r t a n t i n e v a l u a t i n g t h e e x t e n t o f t h econtamination, which is essential to estimating costsof remedial measures and viability of specificalternative remedial measures for the particular site. Itis also useful for determining how long a remedialmeasure will have to be taken.

The following ground-water discussions are dividedinto three sections:

1. The minimum technical foundation that is neededin order for the analyst to apply and interpret theequations and models for ground water. Thisdiscussion is meant to support the hydrologistfamiliar with water supply calculations, providingan introduction to contaminant hydrology.Readers needing a more complete introduction tohydrology may wish to read EPA’s Handbooktitled “Groundwater” (EPA/625/6-87/016).

2. Equations that can predict average contaminantvelocity and mass flux for dilute solute andconcentrated contaminant plumes. Knowing thetravel time and the degradation half-life, one canpredict contaminant attenuation. A nomograph isprovided for predicting dilution and contaminatedfront velocity of dilute solute plumes, as areequations that are useful in assessing the extentof contamination. The narrative contains guidancefor interpreting available monitoring data fromexisting wells and from monitoring wells. All of theequations apply to homogeneous and isotropicmedia; fractured rock flow and karstic terrain floware not addressed.

3 . C o m p u t e r m o d e l s t h a t p r e d i c t d i l u t i o n ,attenuation, and contaminated front velocity ofdilute solute plumes only. All of the computermodels assume homogeneous and isotropicmedia. Computer models that predict organic fluidmigration are not discussed, nor are models thatdescribe karstic terrain flow. The state of the art

63

for these models is not well-developed, and thusthey are considered beyond the scope of thisreport. The analyst wishing to model organic fluidmigration in porous media should use theequations in Section 3.5.2.

3.5.1 Discussion of Ground- Water Modeling

3.5.1.1 The Contamination CycleT h e t w o p r i m a r y t y p e s o f g r o u n d - w a t e rcontamination at uncontrolled hazardous waste sitesinvolve leaching of solid contaminants and percolationof liquid contaminations to the underlying aquifer.Solid material itself does not generally contaminateground water directly, because it does not movethrough the porous soil. Thus, it will not migrate untilprecipitation (or ground water) leaches (dissolves)some of it and carries it down to the water table.Ground-water contamination by this route dependson the precipitation rate and the solubility of the solidcontaminant. A variation of this route involvesdissolution of the solid contaminant by a complexleachate that contains organic constituents as well aswater. T h e e x i s t e n c e o f d i s s o l v e d o r g a n i cconstituents in the leaching fluid causes organiccontaminants to have a higher solubility. Theimportance of this phenomenon is greatest forcontaminants with a high octanol/water partitioncoefficient (Enfield 1984, Jaw-Kwei n.d.).

Liquids do not need infiltrating precipitation to carrythem down to the water table; they move on their ownwi th he lp f rom grav i ty . Thus , g round-watercontamination by liquids is not dependent on theprecipitation rate or the solubility of the contaminant.The viscosity and density of a liquid affect its rate ofmigration. After the liquid has percolated through thesoil, some will remain in the interstitial pore spaces;this material will dissolve into the percolatingprecipitation and migrate downward as a function ofits water solubility and the rainfall rate. Anothersource of contamination by liquid material arises fromintentional injection into the aquifer itself (deep-wellinjection) or “injection” into the vadose (unsaturated)zone (unlined lagoons).

Hazardous waste is often assumed to be primarilysolid waste; however, studies showing the relativeproportion of solid to pourable hazardous RCRAwastes indicate that pourable hazardous wasteconstitutes 60 to 95 percent of the total (Skinner1984). The equations for modeling liquid wastemigration pertain to a larger percentage of the wastemigration situations than the dilute solute transportmodels (computer models/nomograph).

Two other types of ground-water contamination mayalso occur. These are contamination by gaseouscontaminants and contamination by intermediatransfers. Gases constitute a relatively small sourceof ground-water contamination, since they are more

likely to contaminate air than ground water. The mainmechanism for gases contaminating ground water isequilibration of gases leaking from buried containersor injected into the ground, with percolating rainwatercausing subsequent downward migration and mixingof this contaminated water with ground water.Intermedia contamination of ground water can comefrom either air or surface water. Contamination fromair can result from two mechanisms: rain-out andwash-out. Rain-out occurs when airborne con-taminated particulates form condensation nuclei forthe formation of rain drops. Wash-out occurs whenfalling rain captures gaseous or particulate con-taminants as it falls to earth. The concentrations ofcontaminants entering ground water as a result ofgaseous contamination or intermedia transfers aregenerally very small, and these are not considered tobe significiant sources of ground-water contam-ination in most cases.

A third source of contaminat ion that may besignif icant at some si tes is through ground-water/surface-water system interconnections. Thatis, contaminated surface water may recharge aground-water system. This occurs only in reacheswhere the surface-waterbody is a “losing stream”(i.e., one that supplies water to the ground-watersystem). Frequently, ground water feeds surfacewater (gaining reaches). For gaining reaches, theground water, if contaminated, contaminates thesurface-waterbody into which it discharges.

One aspect of the contamination cycle that should beconsidered is the ratio of contaminant to con-taminated ground water. A very small quantity ofconcentrated contaminant can contaminate a largevolume of ground water to the ppm or ppb level.

3.5.1.2 Ground-Water Flow ConditionsAfter 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. Thehorizontal flow within the aquifer is saturated.

(1) Saturated ZoneA simplified flow equation is used to describe thevolumetric flow of water through a porous mediumunder saturated conditions. The volumetric flow (ordischarge) is proportional to the product of the drivingforce, the soil’s ability to transmit water, and thecross-sectional area perpendicular to the flowdirection. The driving force is the difference in theenergy (hydraulic head) between two points in theaquifer divided by the distance between the twopoints. This driving force is called the hydraulicgradient. A soi l’s abi l i ty to transmit water isrepresented by an empirically determined coefficientof hydraulic conductivity. This equation is calledDarcy’s law. The properties of the liquid (water orcontaminant) and the permeability of the porous

medium determine the hydraulic conductivity. The soilhas an intrinsic property of permeability, which isdetermined 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. Smalldiameter clay soil particles cause clay soil to have lowpermeability, while larger diameter sandy soil particlesresult in the high permeability of sandy soils. Thepermeability, and therefore the hydraulic conductivity,of a homogeneous soil is constant under conditions ofsaturated flow.

In cases where the vadose zone is saturated and theflow direction is vertical, the change in height of thewater per unit of vertical travel distance is alwaysone. Thus, the hydraulic gradient for vertical saturatedflow is unity, and the volumetric flows are proportionalto the permeability alone.

(2) Unsaturated ZoneDarcy’s law governs flow anywhere in the porousmedium, including the vadose, or unsaturated, zone.In the vadose zone, however, the pore spaces are notsaturated with water or any other liquid. The hydraulicconductivity of any liquid through a porous medium ispartly dependent on the amount of liquid in the porespaces, and hydraulic conductivity for unsaturated soilcan be expressed as a fraction of the hydraulicconductivity at saturation.

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

The unsaturated hydraulic conductivity at residualmoisture content is very small. When the soil is verydry, most of the moisture is tightly bound by capillaryforces in the void spaces, and the water will not floweasily. Unsaturated hydraulic conductivity increases,gradually at first and then more rapidly, as the degreeof saturation increases from the residual moisturecontent to the saturated moisture content. Since thehydraulic conductivity is dependent upon the moisturecontent, the specific discharge through the vadosezone varies with the degree of saturation at anydepth.

The rate of infiltration at the ground surface may belimited by the capacity of the soil to accept water orby the delivery rate of water at the ground surface(e.g., the precipitation rate). The infiltration rate intosoil cannot exceed the value for that soil’s saturatedhydraulic conductivity. When the hydraulic loading tothe surface of the ground is low, such as light rainfallalone, the flow of water through the vadose zone is

64

unsaturated; however, when the hydraulic loading islarge, such as beneath a lagoon, the flow of waterthrough the vadose zone can be saturated. When thehydraulic loading is small, it is limiting and the verticalflow through the vadose zone is unsaturated. Whenthe hydraulic loading is larger than the flow that canmove through the soil with saturated flow, thepermeability of the soil is limiting the flow, and thevertical flow through the vadose zone is saturated.

Generally, chlorinated hydrocarbons that contaminatewater are more dense than water.

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

351.3 Multiphase FlowThe water solubility of any particular chemical willdetermine whether it will be transported as a solute,as a colloid, or as a separate, concentrated phase.Many chemicals that have been identified ascontaminants in ground water are sparingly soluble inwater. When introduced to the ground-water systemas l iquids, such chemica ls can f l ow as anindependent species through the porous medium.When the immiscible contaminant comes into contactwith the water in the pore spaces of the vadose zoneor at the water table (phreatic surface), the liquids donot mix but essentially remain as two separatephases. Some of the chemical will go into solutionwith the water, but since the solubility of the chemicalis very low, the bulk of the contaminant will remain asa separate layer that could saturate the pore spaces itis flowing through. Thus, the migration of twoimmiscible liquids in porous media is called two-phase flow.

(3) Hydraulic Gradient for Immiscible FluidsThe hydraulic gradient, the difference in the hydraulicheads at two points divided by the distance (along theflow path) between the points, is the driving force forground-water movement in a porous medium. Withregard to an immiscible separate phase, however, thegradient that causes the immiscible liquid to flow isnot necessarily the same as that which influences theground water. If contaminants in an immiscible phasethat is more dense than water reach the bottom of theaquifer, that separate phase may alter its flowdirection to conform to the shape and slope of theaquitard surface. In some cases, the base of theaquifer may be sloped in a different direction from thedirection of flow determined by the hydraulic gradient.This possibility should be considered when theanalyst tr ies to identi fy the direct ion of thecontaminant plume’s migration.

Complete descriptions of two-phase flow require anadditional equation for each separate phase presentin the flow system. Several general rules that can beapplied in analyzing ground-water contaminationproblems involving immiscible chemicals, are asfollows:

The assumption that the hydraulic gradient of theseparate, immiscible phase approximates that ofground water is quite reasonable for the less denseimmiscible liquids. Since these contaminants float onthe water table, the hydraulic gradient of the phreaticsurface is probably also the gradient of the immisciblephase.

(1) FloatersThe specif ic gravi ty of an immiscible l iquidcontaminant will determine whether water will displaceit or it will be displaced by water. In downward flow,water can displace the lighter, immiscible liquid so thewater is found below the immiscible liquid. Inhorizontal flow, the less dense, immiscible liquid willtend to float upward until the separate immisciblephase floats on top of the water table. Thus, theimmiscible liquids that have a specific gravity of lessthan one are sometimes referred to as “floaters.” Asa general rule, immiscible hydrocarbons that arenonchlorinated are floaters (less dense than water).

(4) Hydraulic Conductivity of Immiscible FluidsIf the saturated hydraulic conductivity of waterthrough a porous medium is known, it is very easy tomodify that value to calculate the hydraul icconductivity of that same porous medium saturatedwith a different liquid, such as a separate layer of animmiscible phase.

3.5.1.4 Contaminant Flow and HydrodynamicDispersionIn contaminant transport, contaminants can bethought of as a mass flowing through a cross-sect ional area of the porous medium that isperpendicular to the flow direction. The discussionpresented here is for solute transport (mass that istransferred with the flowing ground water), but basicconcepts also apply to the flow of immiscible,separate phases.

(2) SinkersImmiscible contaminants more dense than water, The movement of contaminants in ground water canwhose specific gravity values are greater than one, be described by two principal mechanisms: grosscan displace water when flowing through the porous fluid movement (advective flow) and dispersion.medium. Gravity will cause dense immiscible liquids Gross fluid movement can be either ground-waterto sink as they flow horizontally through the porous movement or organic fluid movement (the waste itselfmedium. Thus, the immiscible liquids more dense moving as a concentrated liquid). Dispersion also canthan water are often referred to as “sinkers.” be described by two principal mechanisms: fluid

65

mixing (mechanical dispersion) and diffusion. Thenext section addresses the underlying mechanismsfor fluid mixing.

Fluid mixing is important for two reasons: (1) precisemodeling of contaminant movement and (2) modelingof dilution of the contaminant concentration betweensource and exposed population,

Dilution (mixing) in ground water is different fromdilution in air and in surface water. In both air andsurface water, dilution is a major phenomenon. Inground water, the magnitude of dilution is muchsmaller. Flow in both air and surface water can beturbulent. Turbulent flow means that all the flow pathsare not essentially parallel to the gross direction ofmotion; some flow paths are at right angles to thebulk fluid motion. The flow components that areperpendicular to the bulk fluid motion cause theplume to spread lateral ly. This reduces theconcentration in the plume, while making the plumecontaminate a larger volume of air or surface water.

In ground water, turbulent flow rarely exists. The slowspeed of ground water coupled with the straighteningeffect of many soil particles keeps the flow smoothand laminar. In an idealized conceptual model, theinterconnecting pore spaces can be thought of asforming flow channels or tubes; any tendency for theflow to eddy is resisted by the sides of the flowchannel. Since the interconnecting pore spaces donot make a continuous flow channel, some lateralmixing will occur in real soil.

Dispersion in air and surface water is caused by theeddy currents (and diffusion). If the flow is broken upinto two components, longitudinal flow and eddy flow,the gross motion is due to the longitudinal flow, andthe eddy f low is responsible for mixing. Themagnitude of the eddy currents is the same in alldirections (longitudinal, transverse, and vertical).Since the concentration gradients are weaker in thelongitudinal direction than they are in the transverseand vertical directions (for continuous steady statesources), the net effect of mixing in the longitudinaldirection is small compared to the effect of mixing inthe directions perpendicular to the flow direction.When air and surface water are modeled, thelongitudinal mixing is often neglected: lateral mixing ismodeled as the principal mixing phenomenon.

Dispersion in ground water is not caused by eddycurrents. Dispersion (neglecting diffusion for themoment) is caused by four principal phenomena:varying pore sizes, varying path length, variation invelocity gradient across pore space, and flow splittingaround soil particles with mixing within the porespace. The first three phenomena contribute tolongitudinal dispersion; the last phenomenon causeslateral dispersion. In ground water, the magnitude ofthe mixing is much greater for longitudinal mixing than

for lateral mixing. Researchers have reportedlongitudinal dispersivity values ranging from 2 to 25times higher than transverse dispersivity values(Gelhar et al. 1985).

In ground water, dilution occurs at a much slower ratethan it does in air or surface water. The overallmagnitude of mixing is smaller, and the component ofmixing that is most important to dilution (lateral) is thesmaller component of ground-water mixing. Forshort-term releases (spills), longitudinal mixing isuseful in diluting plume concentrations. This isbecause the plume can effectively mix with theuncontaminated water in front of and behind the slugof contamination, whereas continuous sources makethe length of the plume so long that its middle sectioncannot effectively mix with clean water in front orbehind it.

3.5.1.5 Transformation and RetardationMovement of contaminants can be modeled by fluidmovement, fluid mixing, and diffusion; however, formore accurate modeling, chemical transformation andretardation should also be considered. Somecontaminants are subject to transformation andretardat ion whi le others are not; the relat ivesignificance of transformation and retardation forspecific contaminants determines the need to modelthese mechanisms. Transformation is the term usedto describe loss of the contaminant from the plume.The mass of the contaminant is not lost; rather, themolecular structure is changed so that the toxicityassociated with the initial molecular structure is nolonger present. When the molecular structure ofdegradation products is more toxic than the originalcontaminant, degradat ion i s no t cons ideredattenuation. Attenuation is used to describe chemicalstructure changes that reduce or eliminate the toxicityof the contaminant, and to describe phenomena thatfunction as sinks for toxic contaminants. Phenomenathat a re revers ib le a re no t s inks fo r tox iccontaminants.

Chemical interactions between contaminants and thesoil matrix that are reversible delay the migration ofcontaminants but do not act as a sink. The effect ofthese chemical interactions is modeled as retardation.Retardation is modeled using a coefficient to scaledown the velocity of ground water to the slowereffect ive veloci ty of the contaminant mass.Attenuation reduces population risk; retardation delayspopulation risk.

Many reversible interactions can cause retardationphenomena; however , on ly two re ta rda t ionmechanisms apply to wide classes of contaminantsand are well enough understood to be modeled on aregular basis. Organic retardation and cationicretardat ion are the most f requent ly modeledphenomena. Organic retardation refers to hydrophobiccontaminants sorbing onto organic material in the soil

matrix. Cationic retardation refers to positive chargedions associating with the soil matrix. This associationcan be due to polar species in the ground water beingattracted to the ionic double layer surrounding clayparticles in the soil, or it can be due to ionic bondingwith the soil matrix.

(I) Retardation of OrganicsOrganic retardation, which refers to hydrophobiccontaminants sorbing onto organic material in the soilmatrix, is estimated in ground water by the use of aretardat ion coeff ic ient. The veloci ty of eachcompound in ground water is a function of thecharacteristics of the soil media and the compound’soctanol-water partition coefficient. The octanol-water partition coefficient measures the compound’sdegree of hydrophobicity. The parameter of the soilmedia that determines the presence of sorption sitesis the percent of organic carbon in the soil.

When the contaminant concentration in the water ishigh and the quantity of contaminant on the surfaceof the soil organic carbon is low, the net transfer isfrom the water to the soil. Since the transfer is anequi l ibr ium process, i t r e v e r s e s w h e n t h econcentration in the water is low and the quantity ofcontaminant on the surface of the soil organic carbonis high.

If the leachate contains sufficient quantities of organicmaterial to affect the solubility of the contaminant, themodeling of retardation is more difficult. The toxicconstituent flow will still be retarded, but not as much.Instead of partitioning between the water and soilorganic carbon, the contaminant will partition betweenthe polar-organic fluid and the soil organic carbon.The toxic contaminant will spend a smaller fraction oftime on the solid soil particles and a larger fraction oftime in the fluid; this will increase its migrationvelocity. Modeling this phenomenon, however, iscomplex and has already been well documentedelsewhere. The analyst interested in modelingretardation in complex leachates is referred toNkedi-Kizza et al. (1985), Rao et al. (1985), andWoodburn et al. (1986).

Once a contamination source stops contaminating theground water (either a one-time slug or the end of along-term loading), the saturated sorption sites startto lose contaminants to the clean ground water thatf lows after i t . This phenomenon causes thedevelopment of a plume shape that has a long tail ofdecreasing contaminat ion. Since the rate ofdesorption is high when the degree of saturation ishigh, and is lower as the quantity of contaminant onthe sorpt ion si tes diminishes, the desorpt ionphenomenon can provide a degrading influence onthe ground water for a long time.

(2) Retardation of CationsIn cationic retardation, positively charged ions’associate with the soil matrix (clay particles). There isa smaller effect for anion exchange. Anion exchangeis due to positive charges associated with hydrousoxides. Since soils typically have more negativelycharged clay particles than positively chargedhydrous oxides, cations flow with a more retardedvelocity than do anions. Contaminants that are notcharged are not subject to ionic retardation.Contaminants that are compounds or complexed ionsalso are not retarded by ionic retardation.

Cationic retardation is reversible, as is organicretardat ion, and i t forms a trai l of low-levelcontamination after the source of contaminationstops. Once a source stops contaminating the groundwater, the saturated ion exchange sites start to losecontaminants to the clean ground water. Thisphenomenon causes the development of a plumes h a p e t h a t h a s a l o n g t a i l o f d e c r e a s i n gcontamination. Since the rate of release is high whenthe degree of saturation is high, and lower as thequantity of contaminant on the ion exchange sitesdiminishes, the reversible ion exchange phenomenoncan provide a degrading influence on the groundwater for a long time.

(3) Transformation/AttenuationTransformation/attenuation is the term used to modelsinks for contaminants. The particular type ofchemical fate modeled depends on the contaminantand the soil characteristics. The following is a list ofdifferent fate mechanisms:

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

Attenuation is modeled with the use of a “half-life”parameter. Whether the degradation is due tohydrolysis or biodegradation, the time necessary forthe concentration to drop by half is the measure ofdegradability.

Appropriate individual decay rates or overall decaycoeff ic ients have been developed for somesubstances and are available in the technicalliterature. Sources for such data include: Callahan etal. (1979); Dawson et al. (1980); Mabey et al. (1982);Sax (1984); USCG (1974); and Verschueren (1984).Methods of estimating decay coefficients arepresented by Lyman et al. (1982).

67

3.5.1.6 Higher Velocity TransportSome situations can cause the migration velocity ofcontaminants to be faster than the ground-watervelocity. Macromolecules can themselves move fasterthan the ground water, and any hydrophobiccontaminants that are sorbed onto them will alsomove faster. Until recently, hydrophobic contaminantswere thought to flow with a retarded velocity onlybecause of preferential sorption onto stationaryorganic soil particles. In such cases, the time thecontaminant spends on the stationary soil particleslowers its average velocity. Conversely, the time ac o n t a m i n a n t s p e n d s o n a “h i g h s p e e d ”macromolecule raises the average velocity of thecontaminant. Since hydrophobic contaminants sorbo n t o b o t h s t a t i o n a r y a n d h i g h e r v e l o c i t ymacromolecules, both must be considered in orderfor the modeling of transport for hydrophobiccontaminants to be complete.

Macromolecules may be found in ground water inconcentrations ranging from 1 mg/l to 10 mg/l and arelarge enough that only the large pore spaces areavailable for migration. This means that their averagevelocity is the average velocity of the large porespaces, and not the average velocity of all the porespaces. The velocity of flow through each pore spaceis a function of the size of the pore space, and thelarger pore spaces allow faster flow than do the smallspaces. The velocity difference between the averagelarge pore space and the average pore space isapproximately one order of magnitude.

Macromolecules with large hydrophobic surface areaand small polar surface area will flow with a retardedvelocity because of reversible sorbtion onto soilcarbon. These macromolecules will not cause higherspeed transport. Macromolecules with large polarsurface areas and small hydrophobic surface areaswill travel faster than the ground water. Thesemolecules can speed up the migration velocity ofhydrophobic contaminants.

Macromolecular transport is not frequently modeled;however, when such modeling is necessary, theanalyst can refer to Enfield and Bengtsson (n.d.) fordetailed guidance.

3.52 Ground- Water Modeling Equations andNomographThis section provides a number of hydrologicmodeling equations and a nomograph. In no caseswill all equations be necessary; depending on theobserved chemical contaminant, a discrete subset ofthe equations will be useful in assessing the ground-wate r con tamina t ion p rob lem a t a spec i f i cuncontrolled 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 ofcontaminant can have dramatically different calculatedvelocities and concentrations; use of the appropriateanalytical techniques for each class is thus necessaryfor accuracy.

Est imating contaminant velocity is based onestimating water velocity. For those contaminants thatflow as water flows, contaminant velocity equals watervelocity (vertical or horizontal). For those that flow atrates different from water, the estimated watervelocity must be adjusted to approximate that of thecontaminant.

3.5.2.1 Calculating Ground-Water VelocityGround-water velocity can be determined for boththe saturated zone and the vadose (unsaturated)zone. Vadose zone velocity is discussed in the nextsection; saturated zone velocity is discussed in thissection.

Ground-water velocity in the saturated zone iscalculated using Darcy’s Law (Bouwer 1978):

v = Ksi

where

(3-9)

V = Darcy velocity of water, also termedsuper f i c ia l ve loc i ty , o r spec i f i cdischarge, (length/time).

KS = hydraulic conductivity of soil or aquifermaterial, (length/time).

i = hydraulic gradient, (length/length).

However, v, the Darcy velocity, is not the realmacroscopic velocity of the water, but the velocity asif the water were moving through the entire cross-sectional area normal to the flow, solids as well aspores (Bouwer 1978). The ground-water velocity iscalculated from the Darcy velocity by dividing it bysoil porosity, or, for more precise modeling, byeffective porosity (thus taking into account the factthat the entire cross-section of the pore is notflowing (i.e., due to boundary layer effects). For claysoils, the effective porosity also corrects for the effectof electro-osmotic counterflow and the developmentof electrokinetic streaming potentials (Bouwer 1978).The equation for calculating ground-water velocityfrom Darcy velocity using effective porosity is asfollows (Bouwer 1978):

vpw = v/Pe

where

(3-10)

V P W= ground water (pore water) velocity,

(length/time).V = Darcy velocity (superficial velocity,

specific discharge), (length/time).

68

Pe = effect ive porosi ty, (dimensionlessfraction).

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

The hydraulic gradient (the change in the elevation ofthe water table over distance from the site) shouldalso be taken from field data developed during siteinvestigation. Water levels in existing nearby wellscan also provide an indication of hydraulic gradient.Table 3-10 provides values for saturated moisturecontent, which is roughly equal to the effectiveporosity, or Pe, for several soil types.

It must be emphasized that site-specific data arehighly preferable to regional data, or data obtainedfrom any of the above-referenced tables. If site-specific information on effective porosity is available,it should be used; however, literature values for soilswith the same hydraulic conductivity provide sufficientaccuracy. Effective porosity (P,) can be approximatedby the difference between the moisture content atsaturation and at the wilting point (-15 bar)*. Theequation is as follows (Rawls 1986):

(3-11)

This estimation procedure addresses the fraction ofthe pore spaces that is contributing to flow, but doesn o t a d d r e s s t h e e f f e c t o f e l e c t r o - o s m o t i ccounterflow and the development of electrokineticstreaming potentials. For clays, this can be asignificant difference. Literature values listed in Table3-10 should be used for clay solids (these valuesincorporate the effects of the clays ionic double layer)(Rawls et al. 1982); either technique can be used forsand or loam soil.

The above method for predicting the average velocityof ground water is the most widely acceptedapproximation; however, it is only an approximation

*Wilting point is determined by drawing a suction of -15 bar todraw water out of the soil in a manner similar to the suction of aplant root. Bar is a measure of pressure (dynes/cm2).

and further refinement of this approach wouldimprove accuracy. Corrections for the path lengthdifference between the straight line distance versusthe tortuous path through which ground water flowscan improve the precision (Freeze and Cherry 1979),although the literature does not provide a consistentcorrection factor to apply. To provide a feel for themagnitude of this correction the analyst can reviewDas (1983) which suggests a correction of 1.41. Thisvalue can be used to correct the velocity or thedistance (not both) by dividing the number by 1.4.However, the analyst must interpret the resultsobtained through such correction with care, as thedegree to which the factor cited in Das applies to anygiven site is uncertain.

3.5.2.2 Calculating the Velocity of InfiltratingRainwaterThis section discusses the calculation of the velocityof percolating rainwater flowing through the vadosezone. Darcy’s law can be used to calculate theunsaturated flow velocity; however, the hydraulicconductivity must be corrected to reflect the effect ofpartially-filled pore spaces when the hydraulicloading is below that necessary to support saturatedflow.

Interstitial pore water velocity for unsaturatedtransport through the vadose zone can be calculatedas follows (Enfield et al. 1982):

(3-12)

69

Table 3-8. Representative Values of Saturated Table 3-9.Hydraulic Conductivity

Saturated Hydraulic Conductivity Rangesfor Selected Rock and Soil Types

Saturated Hydraulic Conductivity (cm/sec)Hydraulic

Soil textureconductivity

Number of soilsa(Ks; cm/sec)b

Sand 762 5.8 x 10-3

Loamy sand 338 1.7 x 10-3

Sandy loam 666 7.2 x 10-4

Loam 383 3.7 x 10-4

Silt loam 1,206 1.9 x 10-4

Sandy clay loam 498 1.2 x 10-4

Silt clay loam 366 4.2 x 10-5

Clay loam 689 6.4 x 1O-5

Sandy clay 45 3.3 x 1O-5

Salt clay 127 2.5 x 1O-5

Clay 291 1.7 x 10-5

Unweathered marineclay 5 x IO-ll -- IO-7

Glacial till 1O-10 -- 1O-4

Silt, loess IO-7 -- IO-3

Silty sand 1O-5 -- 1O-1

Clean sand 1O-4 __ 1

Gravel 10-1 -- 102

Unfracturedmetamorphic andIgneous rock 1O-2 -- 10-8

aNumber of tndividual soil samples included in datacompiled by Rawls et al. 1982.bpredicted values based on compiled soil properties.Source: Adapted from Rawls et al. 1982.

Shale 5 x IO-12 -- IO-7

Sandstone 10-8 -- 5 x IO-4

Limestone anddolomite 5 x 10-8 -- 5 x 1O-4

Fractured Igneous andmetamorphic rock 10-6 __ 1O-2

Permeable basalt 1O-5 __ 1

Karst limestone 1O-4 __ 1

Source: Adapted from Freeze and Cherry 1979.

Sandy loam 666 0.453 0.351 - 0.555 0.207 0.126 - 0.288

Loam

Silt loam

Sandy clayloam

Clay loam

Silty clay loam

Sandy clay

Silty clay

Clay

383 0.463

1,206 0.501

498 0.398

366 0.464

689 0.471

45 0.430127 0.479

291 0.475

0.375 - 0.551 0.270 0.195 - 0.345

0.420 - 0.582 0.330 0.258 - 0.402

0.332 - 0.464 0.255 0.186 - 0.324

0.409 _ 0.519 0.318 0.250 - 0.386

0.418 - 0.524 0.366 0.304 - 0.428

0.370 - 0.490 0.339 0.245 - 0.433

0.425 - 0.533 0.387 0.332 - 0.442

0.427 - 0.523 0.396 0.326 - 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

KS

b

= vo lumet r i c wate r con ten t in theunsaturated zone, (volume/volume orunitless).

= volumetric water content of soil undersaturated conditions, (volume/volumeor unitless).

= percolation rate (assumed to be equalt o t h e u n s a t u r a t e d h y d r a u l i cconductivity term in original Clapp andHornberger equation), (depth per unittime).

= saturated hydraulic conductivity, (depthper unit time).

= soil-specific exponential parameter,(unitless).

Representa t i ve va lues o f “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)

Sand 13 4.05 (1.78) 0.090 0 . 3 9 5 (0.056)

Loamy sand 30 4.38 (1.47) 0.085 0.410 (0.068)

Sandy loam 204 4.90 (1.75) 0.080 0.435 (0.086)

Silt loam 384 5.30 (1.87) 0.074 0.485 (0.059)

Loam 125 5.39 (1.87)) 0.073 0.451 (0.078)

Sandy clay 8 0 7.12 (2.43) 0.058 0.420 (0.059)loamSilt clay loam 147 7.75 (2.77) 0.054 0.477 (0.057)

Clay loam 262 8.52 (3.44) 0.050 0.476 (0.053)

Sandy clay 19 10.40 (1.64) 0.042 0.426 (0.057)

Silt clay 441 10.40 (4.45) 0.042 0.492 (0.064)Clav 140 11.40 (3.70) 0.039 0 . 4 8 2 (0.050)

aNumber of individual soil samples included in data compiled byClapp and Hornberger (1978).

bEmpirical parameter relating soil matric potential and moisturecontent; shown to be strongly dependent on soil texture.

cWolumetric soil moisture content (volume of water per volume ofsoil).

Source: Adapted from Clapp and Hornberger 1978.

3-11), in order to demonstrate the variability inestimates for these values.

The following equation provides an estimate of theterm q (Enfield et al. 1982):

q = H L + P r - E T - Q r

where

(3-14)

HL = hydraulic loading from manmadesources, (depth per unit time)

P r = precipitation, (depth per unit time)ET = evapotranspiration, (depth per unit

time)Qr = runoff, (depth per unit time).

Records of estimated percolation rates for the sitelocality during the time period in question (or annualaverage percolation rate estimates) are often availablefrom local climate or soil authorities, including regionalU.S. Geological Survey (USGS) and U.S. SoilConservation Service offices.

An estimation procedure can be used to evaluatepercolation rates (q) at sites where the sources listedabove cannot provide them directly. This estimationprocedure 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-cmTall, Well-Watered Grass Turf

Pan surrounded by a short green crop Pan surrounded by a dry surface ground

Upwind fetch of Average regional relative humidity, Average regional

crop (m from %* Upwind fetch ofdry fallow (m relative humidity, %*

Wind pan) 20-40 40-70 ›70 from pan) 2 0 - 4 0 4 0 - 7 0 › 7 0

0 0 . 5 5 0.65 0.75 0 0.7 0.8 0.85

Light < 170 km/day 10 0 . 6 5 0.75 0.85 10 0.6 0.7 0.8

100 0.7 0.8 0.85 100 0.55 0 . 6 5 0 . 7 5

r t i o o 0 . 7 0 . 8 5 0.85 1000 0.5 0.6 0.7

Moderate 170-425 km/day

Strong 425-700 km/day

Very strong > 700 km/day

0 0 . 5 0.6 0.65 0 0.65 0.75 0.8

10 0.6 0.7 0.75 10 0.55 0.65 0 . 7

100 0 . 6 5 0.75 0.8 100 0.5 0.6 0 . 6 5

1000 0 . 7 0 . 8 0 . 8 1000 0 . 4 5 0 . 5 5 0 . 6

0 0.45 0.5 0.6 0 0.6 0 . 6 5 0.7

10 0 . 5 5 0.6 0.65 10 0.5 0.55 0.65

100 0.6 0.65 0.7 100 0.45 0.5 0 . 6

1000 0.65 0.7 0.75 l 0 0 0 0.4 0 . 4 5 0.55

0 0.4 0 . 4 5 0.5 0 0.5 0.6 0 . 6 5

10 0.45 0.55 0.6 10 0.45 0.5 0.55

100 0.5 0.6 0.65 100 0.4 0.45 0.5

1000 0 . 5 5 0 . 6 0 . 6 5 1000 0.3 0.4 0 . 4 5

‘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 weatherstations (e.g., at local airports) are possible sourcesfor these three types of data.

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

ET is estimated by using measured Class A panevaporation rates (a measure of local evaporationrates under standardized conditions, available fromthe nearest NOAA gauging station) in the equation:

ET = EVAP x Cet x Cveg

where

(3-15)

EVAP = r e g i o n - s p e c i f i c o r s i t e - s p e c i f i cmeasured evaporation rates, (depth perunit time).

Cet = cor rec t ion fac to r fo r conver t ingmeasured pan evaporation rates toevapotranspiration rates from turfgrass, (unitless).

Cveg = c o r r e c t i o n f a c t o r f o r c o n v e r t i n gevapotranspiration from turf grass toe v a p o t r a n s p i r a t i o n f r o m o t h e rvegetative cover types, (unitless).

Values for Cet are taken from Table 3-12, whichrequires climatological and pan descriptive in-formation.

The term Cveg is available mainly for agriculturalcrops (Table 3-13), and varies with the thickness,depth, and characteristics of vegetative cover. Typicalvalues are 0.87 for shorter broadleaf plants (alfalfa) to0.6 for taller broadleaf plants (potatoes, sugar beets)and 0.6 for taller grains and grasses. Where crop-specific data are unavailable, a conservative defaultvalue for this term is the smallest reasonable value,or 0.6.

Qr, or the average runoff over the study period, isestimated by the method presented in Section 2.4 ofthis manual. A more reliable value for this term canbe obtained from local USGS gauging stations. Forrelatively level sites, a reasonable conservativedefault value for the purposes of this estimationprocedure is that Qr = 0, where site-specific dataare unavailable or cannot be estimated.

72

Table 3-13. Crop Coefficients for EstimatingEvapotranspiration

crop PeriodAlfalfa April 1 - October 10

Potatoes May 10 - September 15Small grains April 1 - July 20Sugar beets April 10 - October 15

Coefficient(Cveg)

0 . 8 7

0 . 6 5

0 . 6

0 . 6

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

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

3.5.2.3 Corrections for Viscosity and DensityWhen the movement of liquids other than water iscalculated, the saturated and the unsaturatedhydraulic conductivity must be corrected for thedensity and viscosity of the non-water liquid. Theequation for this correction is as follows:

Kc = Kc*(density of chemical/density of water) (3 - 1 6 )‘(viscosity of water/viscosity of chemical)

where

K w

K c

= hydraulic conductivity of water (Darcy’scoefficient), (saturated or unsaturated)

= hydraulic conductivity of chemical,(saturated or unsaturated).

When the migration velocity through the vadose zoneis calculated, density and viscosity should becorrected with the above equation. For horizontal flowbelow the water table, density and viscosity should befactored in when the hydraulic gradient is the slope ofthe chemical plume. In many cases, one can assumethat the thickness of the concentrated chemicalplume is relatively constant. For such situations, theslope of the concentrated chemical is zero and theanalyst should not correct for the density. The slope(hydraulic gradient) is that of water, and the Darcycoefficient reflects the density of water. However, theviscosity of the chemical is the viscosity of theflowing fluid of concern, and the analyst shouldcorrect for the viscosity.

3.5.2.4 Retardation EffectsHydrophobic or cationic contaminants that aremigrating as a dilute solute are subject to retardationeffects. Concentrated plumes are not subject to thisphenomenon. Contaminant migration as a dilute

solute in complex leachates containing organicconstituents will show some retardation, although notas much as in pure ground water.

When a hydrophobic contaminant flowing in a diluteplume flows past a soil particle that contains organiccarbon, the contaminant partitions between the polarsolvent (water) and the solid organic carbon. Whenthe concentration in the water is high and theconcentration on the soil particle low, the netmigration is from the water to the soil. When thereverse occurs and the concentration in the water islow 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 inequilibrium, there is no net migration. However, theflux from the soil to the water and the flux from thewater to the soil are not zero; rather, they are positivefluxes that are equal and are in opposite directions.When the partitioning is between concentratedchemical and soil particles, the contaminant does notprefer the solid “solvent” effects of the organiccarbon in the soil to the organic liquid solvent effectsof the concentrated chemical plume. Hence,hydrophobic contaminants partition out of polarsolvents (water) but not out of hydrophobic solvents,and thus, retardation effects are modeled for diluteplumes only.

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

The retardation protocol is based on the assumptionthat adsorption of hydrophobic contaminants is due tosorption to organic carbon in the soil. Basing theadsorption coefficients on soil organic carbon ratherthan total mass eliminates much, but not all, of thevariation in sorption coefficients between differentsoils. The remaining variation may be due to othercharacteristics such as surface area of soil particlesper mass of soil (function of particle size). Numerousstudies of the correlation of Kd with various soilvariables have found that the organic carbon contentusually gives the most significant correlation.Furthermore, this correlation often extends over awide range of organic carbon content -- from 0.1percent to nearly 20 percent of the soil in some cases(Lyman et al. 1982).

This protocol estimates hydrophobic retardationbased on soil organic carbon, but it should not betaken to imply that hydrophobic contaminants will notadsorb on minerals free of organic matter. Someadsorption will always take place, and it may besignificant under certain conditions, such as clay soils(high surface area per mass of soil) with very loworganic carbon content (no appreciable sorption tononexistent organic carbon). Unfortunately, methods

73

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

To simplify modeling, equilibrium conditions aremodeled as the contaminant velocity being a fractionof the ground-water velocity. If the analyst thinks ofthe time an individual portion of the contaminant massis in the water as the time it has ground-watervelocity, and the time the contaminant is on the soilparticles as the time the contaminant does not have avelocity, the contaminant velocity is related to theground-water velocity by the ratio of time on soilparticles to time in the water. The ratio of time in thewater to time on the soil particles is the same ratio asthe concentration ratio at equilibrium.

In complex leachates containing organics, the time ahydrophobic contaminant spends on the solid carbonis reduced because the ratio of the contaminant’ssolubility in the fluid to its solubility on soil carbon isincreased. The hydrophobic contaminant partitionsbetween the organics in the flowing fluid and theorganics that are solid.

The same logic applies to cation retardation, and thecontaminant velocity for cations is also modeled asfraction of ground-water velocity.

The equation used to calculate the retardation is asfollows (Kent et al. 1985):

Rd = 1 + (B* Kd)/pt (3-17)

where

Rd= retardation factor, (unitless).B = bulk density, (g/ml).P t = total porosity, (unitless).K d = distribution factor for sorption on

aquifer medium (from sorptionisotherm column studies, or fromregression equation based on theoctanol/water partition coefficient,(in ml/g).

The use of the retardation factor is described in thefollowing equation (Kent et al. 1985):

Rd = v p w / v d (3-18

where

Rd = retardation factor, (unitless).vpw = velocity of ground water, (same

units as Vc' length/time).vd = velocity of contaminant, (same

units as Vgw' length/time).

The term Kd is based on sorption isotherm columnstudies. While this is the more precise approach, theanalyst will typically have to work with estimatedparameters. For hydrophobic contaminants, the termKd can be estimated from the term Koc (Lyman et al.1982):

Koc = Kd/foc

where

Koc = partition coefficient for organiccarbon, (ml/g).

K d = distribution factor for soil, (ml/g).fo c = fraction of organic carbon in the

soil.

The term “fraction of organic carbon” (foc) is precisewhen taken from empirical measurements of the soilin the study area. For cases where this is notpossible, estimates can be made. For the vadosezone velocity, a value of foc from Rawls (1986)provides a good estimate. Rawls’ work focused onsoils near the surface, the area of interest toagriculture. For saturated zone velocity, the analysthas two choices. If the subsoil came from igneous ormetamorphic rock, the foc decreases with depth. Theactual value may be quite low; however, the model topredict retardation is only useful down to 0.1 percent.For this situation, the analyst should use 0.1 percentfor the foc. If the subsoil came from sedimentary rock,the foc distribution may be similar to the distribution foragricultural soils done by Rawls. The variation of focwith depth may be relatively constant. The carbonwas at the surface at one time, and has been buriedover geological time. Hence, the analyst should use avalue of foc from the Rawls (1986) distribution for thesaturated zone velocity determination (Trask andPatnode 1942). Soil/water partition coefficients havebeen developed for many contaminants of importance(Callahan et al. 1979 and Mabey et al. 1982).

If Koc is not known, it can be estimated fromregression equat ions that relate Koc to Ko w(octanol/water partition coefficient). There are sixregression equations that relate Koc to Kow. Theequation that was based on a chemical class closestto the subject contaminant should be used. If thecontaminant does not fit into a specific class, the firstregression equation should be used because it wasbased on the largest sample. The regressionequations are as follows (Lyman et al. 1982):

Log Koc = 0.544 log Kow + 1.377 (3-20)

based on a wide variety of contaminants, mostlypesticides

or

74

log Koc = 0.937 log Kow - 0.006 (3-21)

based on aromatics, polynuclear aromatics, triazinesand dinitroaniline herbicides

or

log Koc = 1.00 log Kow - 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

log Koc = l.029 log Kow -0.18 (3-24)

based on a variety of insecticides, herbicides, andfungicides

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 theoctanol/water partition coefficient (Kow), which relatesthe concentration in polar solvent (water) to theconcentration in hydrophobic solvent (octanols imu la t ing the so i l o rgan ic carbon) . I f thecontaminated plume has a large concentration oforganic chemicals dissolved in the ground water, theactual partitioning will be from a solvent/organicchemical system. This will raise the concentration inthe fluid and lower the concentration on the soilorganic carbon. This shift in partitioning will lower Rd,(i.e., the contaminant will migrate at a speed closer tothat of ground water). Much of the solubility ofextremely hydrophobic contaminants in the water ofan octanol/water partition coefficient test is due todissolution in the octanol that is dissolved in the waterrather than dissolution into water. This effect dependson the degree to which the water is not pure water;for most low-level contamination situations, thiseffect can be ignored. This manual does not presentequations for calculating a numerical correction forthis effect. The analyst should be aware of thegeneral influence of this effect, but not model theprecise numerical difference. For dilute plumes, theanalyst should model full retardation; for concentratedplumes, the analyst should model no retardation.

3.5.2.5 Contaminant VelocityThe velocity of concern is the actual contaminantvelocity. The determination of ground-water velocity

discussed earlier is done to provide a foundation forcalculating the contaminant velocity. The particularmethod used for determination of the contaminantvelocity is dependent on the type of ground-watertransport the chemical undergoes. Thus, the first stepin calculating the velocity is classifying the subjectcontaminants as to migration class.

Once the molecular identity of the contaminant isknown, three determining parameters can be takenfrom literature:

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

2 . Hydrophob ic i ty ( i .e . , i s i t hydroph i l i c o rhydrophobic?)

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

The five migration classes are as follows:

Migration Saturated zoneclass # Vadose zone transport tramsport

A) Solid/carried by Solute transport

B)

C)

D)

E)

precipitationHydrophilic Iiquid/waste percolationHydrophobic liquid/waste percolationHydrophobic liquid/waste percolationHydrophobic liquid/waste percolation

Solute transport

Low density/floater transportMedium density/buoyant transportHigh density/sinker transport

Although the specific chemical will migrate accordingto the above classes, it is important to note that theconcentrated plumes will also have a dilute plumenear them. For mass flux considerations, theconcentrated plume will dominate.

(1) Migration Class #A: Solid MaterialSol id mater ial wi l l d issolve into percolat ingprecipitation and migrate as a solute. Precipitationprovides the hydraulic loading that drives the rate ofrelease. The plume exists as a single plume (forsingle chemical contaminant) that has a singleaverage velocity. Unretarded contaminants move withthe ground water, and hence, the ground-watervelocity is the contaminant velocity. Retardedcontaminants move with a velocity that is slower thang r o u n d - w a t e r v e l o c i t y , a n d t h e r e f o r e t h econtaminant velocity is based on the ground-watervelocity adjusted for retardation. Typically, the velocityis a fraction of the ground-water velocity.

(2) Migration Class 49: Hydrophilic LiquidsLiquids will directly percolate into the soil (i.e., withoutwaiting for precipitation to cause leaching). The

75

hydraulic loading is due to the combination ofchemicals’ hydraul ic loading and that due toprecipitation. The velocity of transport through thevadose zone must be calculated with corrections forthe density and viscosity of the contaminant. Theplume exists as a single plume (for a single chemicalcontaminant) that can be considered to have a singleaverage velocity. Unretarded contaminants move withthe ground water, and hence, the ground-watervelocity is the contaminant velocity. This is only exactafter the plume has mixed with the ground-water tothe point that its density and viscosity are similar tothose of water. When the plume first reaches thewater table, it has not mixed with very much water,and its density and viscosity differences suggestcalculating a contaminant velocity that is differentfrom the ground-water velocity. Since the velocitydifference varies gradually from the source to thepoint downgradient where it is well mixed, thiscalculation is complex. Therefore, the analyst shouldcalculate as if the ground-water velocity representedthe contaminant velocity for the length of the plume.The analyst should be aware of the limitations of thismethod. Retarded (cationic) contaminants move witha velocity that is slower than ground-water velocity.In this case, contaminant velocity is based on theground-water velocity adjusted for retardation, and isa fraction of the ground-water velocity.

(3) Migration Class #C: Hydrophobic Liquids LowDensityOnce hydrophobic liquids reach the water table, theyform two distinct plumes (for a single chemicalcontaminant), with each having its own averagevelocity. The concentrated plume will float on thesurface of the water table and move in the samedirection as the ground-water flow. Its velocity is afunction of the contaminant’s viscosity. If mounding issignificant, the density must also be considered. Thedilute plume is formed by small amounts of thechemical dissolving in water as limited by thehydrophobic chemical’s solubility. This plume will befound below the concentrated plume, with the highestconcentration near the concentrated plume. From thep o i n t w h e r e t h e c o n t a m i n a t i o n l e a v e s t h econcentrated plume to form the dilute plume, thedilute plume will move with the ground-water flow (ata retarded velocity). The concentrated plume willhave a single average velocity, and it will start at thelocation of the source. The dilute plume will have asingle average velocity, but its starting point can befrom the location of the source, or it can form fromthe concentrated plume anywhere along the length ofthe concentrated plume.

Retarded contaminants in the dilute plume move witha velocity that is slower than ground-water velocity.Thus, contaminant velocity, based on the ground-water velocity adjusted for retardation, is typically afraction of the ground-water velocity. Contaminants

in the concentrated plume do not move with theground-water velocity; their velocity must bedetermined by consider ing the effect of theh y d r o p h o b i c c o n t a m i n a n t ’s v i scos i t y . Theconcentrated plume does not exhibit retardationeffects. If mounding is significant, the analyst alsomust factor in the density.

(4) Migration Class #D: Hydrophobic Liquids/MediumDensityThis class of compounds migrates similarly to Class#3, except that the concentrated plume will not floator sink, but will have more or less neutral buoyancy. Itwill move in the direction of ground-water flow, butits migration velocity will be a function of its viscosity.Again, the dilute plume will surround the concentratedplume, forming a transition zone between theuncontaminated water and the concentrated plumebody. From the point where the contaminant leavesthe concentrated plume to form the dilute plume, thedilute plume will move with the ground-water flow (ata retarded velocity). The concentrated plume willhave a single average velocity, and it will start at thelocation of the source, or it can form from theconcentrated plume anywhere along the length of theconcentrated plume.

Retarded contaminants in the dilute plume move witha velocity that is slower than ground-water velocity.Thus, the contaminant velocity, based on theground-water velocity adjusted for retardation, is afraction of the ground-water velocity. Contaminantsin the concentrated plume do not move with theground-water velocity; the i r ve loc i ty must bedetermined by consider ing the effect of theh y d r o p h o b i c c o n t a m i n a n t ’s v i scos i t y . Theconcentrated plume does not flow with retardationeffect.

(5) Migration Class #E: Hydrophobic Liquids/HighDensityAs with low and medium density hydrophobic% oncea high density plume reaches the water table, it formstwo d is t inc t p lumes ( fo r a s ing le chemica lcontaminant) with each having its own averagevelocity. The concentrated plume will sink to thebottom of the aquifer. Its velocity is a function of thecontaminant’s viscosity. If mounding on the aquitardis significant, the density must also be considered.The dilute plume will be above the concentratedplume, with the highest concentration near theconcentrated plume and the lowest concentration atthe farthest distances from the concentrated plume.The concentrated plume will have a single averagevelocity 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 thesource, or it can form from the concentrated plumeanywhere along the length of the concentrated plume.

76

Retarded contaminants in the dilute plume move witha velocity that is slower than ground-water velocity.Thus, the contaminant velocity, based on theground-water velocity adjusted for retardation, is afraction of the ground-water velocity. Contaminantsin the concentrated plume do not move with theground-water velocity; the i r ve loc i ty must bedetermined by consider ing the effect of thehydrophobic contaminant’s viscosity. If the sinkermounds above the aquitard significantly, the densityshould be taken into consideration. The concentratedplume does not flow with retardation effects.

3.5.2.6 Nomograph TechniqueThe following nomograph is based on a solution tosolute transport in an aquifer from a point source thatextends throughout the thickness of the aquifer.Contaminant transport from the source includesadvective flow with the ground water and longitudinaland transverse dispersion (see Wilson and Miller1978). The nomograph is taken from Kent et al.(1985); the analyst is referred to this sourcedocument for further discussion of the use of thenomograph 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 obtainan approximate solution. Scale factors are used totranslate Wilson and Miller (1978) to nomograph form.Dilution/dispersive mixing and retardation parametersare included in the solution.

Three scale factors that must be calculated beforeusing the nomograph are:

77

Two of the three ratios are computed directly, and thethird is found using the nomograph (Figure 3-8). Theprocedure for calculating the scaling factors and usingthe nomograph is presented as follows:

(1) Scale Factor DevelopmentThis nomograph models the same var iety ofconditions that the Wilson and Miller model (fromwhich it was derived) does, yet it does it with onlyone graph. This was achieved by scaling theparameters to make them dimensionless. Distance Xis made dimensionless by dividing by the distancescaling factor (XD, the characteristic dispersionl e n g t h ) . T h e m a s s f l u x ( Q * C , ) i s m a d e

dimensionless by dividing by the mass flux scalingfactor (QD). And time (T) is made dimensionless bydividing by the time scaling factor (TD). Obtain XDusing the following:

(3-29)

(3-30)

where variables are defined as in Figure 3-8,Definition of Terms.

where

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 FactorsUse the three scale factors and the nomograph(Figure 3-8) to calculate the concentration at time Tand 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 curveand X/Xd.

(d) Use this point and the point on the Q * Co/Qd lineto draw a straight line. Where this line intersectsthe concentration line, the concentration atdistance X and time T is indicated.

3.5.2.7 Extent of PlumeAs discussed earlier, a large volume of contaminatedground water can result from a small volume ofchemical release. For example, a IO-gallon spill ofsolvent can contaminate a billion gallons of groundwater to 10 ppb. Similarly, a 5000-gallon tankertruck can contaminate 500 billion gallons of groundwater to 10 ppb. The analyst must be aware of therelationship between volume of contaminant releasedand volume of contaminated ground water. The

Figure 3-8. (Continued)

Definition of Terms

Primary Variables : Units

C = concentration of leachate at a specific time and distance. (M/L3)X = distance from source where concentration of leachate is computed.

distance is measured in direction of ground-water flow (perpendicular to gradient). (L)Y = transverse distance measured from the centerline of ground-water flow

(assumed to be zero in the nomograph). (L)t = sample time from beginning of leachate source flow. (T)

Aquifer Parameters :

m = effective aquifer thickness or zone of mixing. (L)Pe = effective porosity of aquifer or zone of mixing. (Dimensionless)v = velocity of ground-water flow within voids, estimated directly from:

whereK = coefficient of permeability or hydraulic

conductivity of aquifer or zone of mixing.i = gradient of ground-water flow.

Transport Parameters :

Dx = longitudinal dispersion coefficient (mixing rate) with respect todistance in x direction and time, estimated directly or from:

Dx= axv+D*

where

ax = longitudinal dispersivityD* = molecular diffusion coefficient, which IS assumed

to be negligible for velocities typical of permeableaquifers. D* may be the dominant process inaquitards where ax V would be negible.

(Dimensionless)

(L2/T)

(L)(L2/T)

79

Figure 3-8. (Continued)

Definition of Terms

Dy = Transverse dispersion coefficient (mixing rate) with respect todistance in the y direction and time, estimated directly or from:

Dy=ayv+D*where

ay = transverse dispersivity,

(L2/T)

(L)

or estimated as:

Dy = Dx divided by a ratio, which commonly ranges between5 and 10 for medium to coarse sand aquifers.

Rd = Retardation factor estimated directly or from: (Dimensionless)

where

(1/T)

(T)

80

Figure 3-8. (Continued)

Definition of Terms

Units

Source Rate of Leachate :

QC0 = Mass flow fate:

where

Q = Volume flow rate estimated directly or from:

(M/T)

(L3/T)

A = area of source.q = recharge rate.

Co = Initial concentration

Intermediate Variables (used for nomograph only) :

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

TD= A characteristic dispersion time or scale factor given by:

QD = A characteristic dilution-dispersion flow given by:

(L2)(L/T)

(M/L3)

(L)

(T)

(L3/T)

81

equation is a simple mass balance equation and isexpressed as follows:

For liquid contaminants:

V 1 *C 1 = V g w *C g w (3-32)

where

V1 = volume of liquid chemical released.V g w = 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 thesame units.

For solid contaminants:

Mc *C c = C g w*V g w

where

(3-33)

To convert the quantity of contaminated ground waterto a volume of contaminated soil, the followingequation is used:

(Vgw*0.13368)/Pt = Vc

where

(3-34)

V g w = volume of contaminated ground water,(in gallons).

P t = total porosity, (dimensionless fraction).V c = volume of contaminated soil, (in cubic

feet).

Or alternatively:

Vgw/P t = Vc (3-35)

where both volumes are in the same units.

3.5.2.8 Use of Monitoring DataThe analyst should take care when using monitoringdata to assess the depth of contamination in order tocalculate volume or mass in the plume. Thedifference between monitoring and pumping wells willaffect the interpretation of the concentrations found inthe wells. Monitoring wells are the more desirable, butsince most existing wells will be pumping wells,monitoring wells will typically have to be installed. Thecost associated with drilling monitoring wells mostlikely will cause the analyst to rely on existingpumping wells.

Monitoring wells extract a small quantity of water (asample); this minimizes the well’s influence on theflow of the ground-water. They do not induce a largevertical component in the ground-water flow, andthus they sample a horizontal slice of the aquifer. Theconcentration in a sample removed from a monitoringwell represents a concentration at the depth of thewell screen. Thus, monitoring wells at various depthscan be used to assess the depth of contamination.

Pumping wells draw large quantities of water from anaquifer (a pumping well provides water). This causesa cone of depression to form on the water table andinfluences the flow direction above and beneath thewell screen. Pumping wells induce vertical flow in theaquifer near the well. This vertical movement causesthe concentration in the well to reflect the averageconcentration for a depth range that is substantiallygreater than the length of the well screen. Water willbe drawn from above and below the well screen. Thewell water does not reflect the concentration of aparticular depth, but rather reflects an averageconcentration from a range of depths. This makes anassessment of the depth of contamination difficult.However, it makes assessing the mass in the plumeeasier since the well draws a sample that representsthe concentrations at a wide range of depths near thewell screen depth.

3.5.2.9 VHS ModelIn addition to the nomographic technique, the Officeof Solid Waste (OSW) has developed a simplifiedmodel for its delisting program that relates leachateconcentration to receptor well concentration 500 feetdowngradient from the edge of a landfil l. Theapproach is called the VHS model (Vertical andHorizontal Spread model). The only reduction inconcentration provided by the VHS model is that dueto vertical and horizontal dispersion (OSW plans toadd hydrolysis and biodegration for organics). Theapproach involves back calculating from a health-based ground-water concentration at the exposedpopulation location to an acceptable leachateconcentration at the site. Wastes with leachate abovethis concentration must be managed as hazardouswastes. Those with leachate below this concentrationcan be managed in a mun ic ipa l l and f i l l o r

82

nonhazardous industrial landfill (i.e., outside thehazardous waste system).

The only data the VHS model requires are theleachate concentration and the annual volume ofwastes disposed of (constituent concentrations oftoxicants are also required in order to ensure thatthey are present in sufficient mass to sustainleaching). The model calculates a different dilutionfactor depending on the annual volume of wastedisposed of. For Superfund purposes, the totalvolume of waste at the site would be used as the“Annual Volume of Waste” term. A small volume ofwaste can rely on greater dilution, while a largevolume of waste is assigned a smaller dilutionpotential. All other input parameters are fixed atreasonable worst-case values. By f ix ing theenvironmental parameters, the model assumes ageneric environment that is consistent with OSW’srequirements. For CERCLA purposes, the model isconsidered to be useful as a simplified analyticalprocedure, and use of the VHS model for site-specific, in-depth analysis is not recommended.

When using th is model , one should keep i tslimitations in mind. The VHS model simulates solubletoxic consti tuents dissolving into percolat ingprecipitation and moving with the ground water. Itdoes not address solvent transport of organics (two-phase flow) or the percolation of organic fluids intothe ground.

Critics of the VHS model have pointed out twoweaknesses of the approach. The first point is thatthe model upon which the VHS model was based (theDomenico and Palciaukas model) does not relateleachate concentration to exposed population wellconcentration. This model relates the concentration inground water immediately below the hazardous wastesites to the exposed population well concentration.When leachate enters ground water, it will be mixedwith ground water. This contaminates ground waterand at the same time dilutes the concentration ofleachate. It is wrong to use the Co term in the modelas leachate concentration, because it represents theconcentration in ground water at the vertical pointwhere leachate enters. This concentration must bemeasured on a site-specific basis to make the useof the model consistent with the boundary conditionsused in the derivation of the model. The model isderived from the following assumptions:

1. Steady-state concentrations are achieved underthe conditions that the concentration C o inground water is maintained on a vertical plane offinite size.

2. No longitudinal dispersion occurs; dispersion onlyin the y and z directions is assumed.

3. Recharge or dilution mechanisms, other thanground-water flow and the above-mentioneddispersion, are ignored.

4. The contaminant velocity in ground water isknown.

The second weakness is the method used todetermine the cross-sectional area of the plume atthe edge of the landfill. The depth of the plume isdetermined by the horizontal velocity of ground waterand the vertical velocity of the contaminant. Themodel presumes that the vertical velocity of thecontaminant in the vadose zone is also the verticalvelocity of the contaminant in the saturated zone. Inthe vadose zone, the contaminants are under theinfluence of gravity; in the saturated zone, the verticalvelocity is much smaller because the effect of gravityis canceled by the buoyancy forces. The VHS modelassumes that the velocities are the same.

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

3.53 In-Depth Methods and ModelsSeveral references are available that provide detailedderivations and outline the application of moresoph is t i ca ted equat ions fo r the ana lys is o fcontaminant migrat ion in the saturated andunsaturated zones. The analyst is referred to thefollowing documents: USEPA 1985j; Van Genuchtenand Alves (1982) Walton (1984), and Javendel et al.(1984), USEPA (1986a), Geotrans (1986), and vander Heijde (1985 and 1987).

Tables 3-14, 3-15, and 3-16 provide informationregarding several modeling procedures for the in-depth assessment of the ground-water fate ofhazardous substances. Note that in order to providethe analyst with an indication of the large number ofcomputer models that could be applied to analysis ofcontaminant fate in ground water, Table 3-15(Features of Unsaturated Zone and Ground-WaterFate Models) provides data for 24 models in additionto the 11 for which more detailed information isprovided in Tables 3-14 and 3-16. Two of themodels addressed in these tables are part of GEMS:SESOIL and AT123D. The latter is described ingreater detail below, because it is more versatile andis applicable to a wide range of fate analysissituations. Additionally, following that discussionfurther detail for certain of the models addressed inTables 3-14, 3-15, and 3-16 is also provided.

AT123D (Ana ly t i ca l T rans ien t 1 - , 2 - , o r 3 -Dimensional Simulation Model) is capable ofsimulating the transport and fate of hazardousmaterial under 300 different user-selected situations(Yeh 1981). One of eight source configurations canbe selected: a point source; line sources aligned in

83

Tabl

e 3-

14.

Res

ourc

e R

equi

rem

ents

and

lnfo

rmat

ion

Sour

ces

:U

nsat

urat

ed Z

one

and

Gro

und-

Wat

er F

ate

Mod

els

Ref

eren

ces,

sou

rces

of d

ocum

enta

tion,

Mod

elD

escr

iptio

nR

esou

rce

requ

irem

ents

, com

men

tsso

ftwar

e

Uns

atur

ated

zon

e

Sea

sona

l S

oil

Com

partm

ent

Mod

el(S

ES

OIL

)

• Lo

ng-te

rm f

ate

sim

ulat

ions

??

inte

grat

ed in

to G

EM

S (

see

Sec

tion

3.1)

??

Acc

ount

s fo

r nu

mer

ous

hydr

olog

ic,

• V

ersa

tile,

eas

y to

use

met

eoro

logi

c ch

arac

teris

tics

of s

ite??

FOR

TRA

N p

rogr

am la

ngua

ge;

has

been

??

Acc

ount

s fo

r nu

mer

ous

trans

fer,

Impl

emen

ted

on I

BM

370

, V

AX

11/

780

trans

form

atio

n pr

oces

ses:

ads

orpt

ion

vola

tiliz

atio

n, d

egra

datio

n, b

rode

grad

atio

n•

Mod

els

orga

nics

, in

orga

nics

??

Pro

duce

s co

ntam

inan

t co

ncen

tratio

ndi

strib

utio

n in

uns

atur

ated

zon

e, q

ualit

y of

grou

nd-w

ater

run

off

?H

andl

es u

p to

thre

e la

yers

of s

oil t

ypes

,pe

rmea

bilit

ies

PR

ZM (

Pes

tcid

e R

oot

Zone

Mod

el)

PE

STA

N

• O

ne-d

imen

sion

al??

Org

anic

sub

stan

ces

??

Deg

rada

tion

IS s

imul

ated

??

Pro

vide

s po

lluta

nt v

eloc

ity,

dist

nbut

ion,

and

conc

entra

tion

data

??

Acc

omm

odat

es v

ario

us r

elea

se r

ates

,sc

hedu

les

??

One

-dim

ensi

onal

??

Org

anic

sub

stan

ces

??

Deg

rada

tion

IS s

imul

ated

??

Pro

vide

s po

lluta

nt v

eloc

ity,

dist

ribut

ion,

and

conc

entra

tion

data

??

Acc

omm

odat

es v

ario

us r

elea

se r

ates

,sc

hedu

les

??

PC

Bas

ed M

odel

??

Req

uire

s 25

6 K

RA

M m

inim

um,

640K

pref

erre

d,?

Inte

l 808

9 or

802

87 m

ath

copr

oces

sor

??

Has

bee

n fie

ld-v

erifi

ed w

ith p

estic

ides

??

FOR

TRA

N p

rogr

am l

angu

age

??

Con

side

red

a sc

reen

ing

mod

el??

Rap

id e

valu

atio

ns??

Inex

pens

ive,

eas

y to

use

; re

quire

s on

lyha

nd-h

eld

calc

ulat

or??

Has

bee

n fie

ld-v

erifi

ed w

ith p

estic

ides

Doc

umen

tatio

n: B

omaz

ount

as a

nd W

agne

r19

81

Con

tact

for

acc

ess

to G

EM

S s

yste

m:

Mr.

Lore

n H

all

US

. E

PA

, E

xpos

ure

Eva

luat

ion

Div

isio

nW

ashi

ngto

n, D

.C.

(202

) 38

2-39

31

Ref

eren

ce:

Car

sel

et a

l. 19

84

Info

rmat

ion:

Dav

id D

isne

yU

SE

PA

Env

ironm

enta

l R

esea

rch

Labo

rato

ryA

then

s, G

a. 3

0613

(909

) 54

6-31

32

Ref

eren

ce:

Enf

ield

et

al.

1982

Tabl

e 3-

14.

(Con

tinue

d)

Mod

elD

escr

iptio

nR

esou

rce

requ

irem

ents

, com

men

tsR

efer

ence

s, s

ourc

es o

f doc

umen

tatio

n,so

ftwar

e

Hyd

rolo

gic

eval

uatio

n of

land

fill

perfo

rman

ce (

HE

LP)

(as

mod

ified

by

And

erso

n-N

icho

ls)

?O

ne-d

imen

sion

alFo

ur o

ptio

ns a

llow

mod

elin

g w

ith a

vaila

ble

Info

rmat

ion:

?M

odel

s le

achi

ng fr

om la

ndfil

ls to

data

Bria

n B

ickn

ell

unsa

tura

ted

soil

bene

ath

land

fill

And

erso

n-N

icho

ls?

Has

four

opt

ions

to h

andl

e m

odel

ing

the

Pal

o A

lto,

Cal

if. 9

4303

solu

biliz

atio

n of

toxi

c co

nstit

uent

s?

Mod

els

orga

nics

/inor

gani

cs?

Use

s ra

infa

ll an

d w

aste

sol

ubilit

y to

mod

el l

each

ate

conc

entra

tions

lea

ving

land

fill

Sat

urat

ed z

one

??

One

- or

tw

o-di

men

sion

alR

ando

m W

alk

Sol

ute

Tran

spor

t M

odel

??

Tim

e-va

nant

rel

ease

rat

es(R

WS

TM)

??

Acc

omm

odat

es w

ell-i

njec

ted

rele

ase

(a.k

.a.

TRA

NS

)??

Inco

rpor

ates

dis

pers

ion,

ret

arda

tion

(req

uire

s P

LAS

M f

or f

low

mod

elin

g)??

Han

dles

non

cons

erva

tive

pollu

tant

s??

Acc

ount

s fo

r w

ell p

umpi

ng??

Pro

vide

s co

ntam

inan

t co

ncen

tratio

n at

user

-sel

ecte

d po

ints

Cou

pled

Flu

id,

Ene

rgy

and

Sol

ute

Tran

spor

t (C

FES

T) C

ombi

ned

with

UN

SA

T-ID

??

Thre

e-di

men

sion

al??

Acc

omm

odat

es h

eter

ogen

eous

,an

isot

ropi

c, m

ultil

ayer

ed s

oil

conf

igur

atio

ns?

Han

dles

sal

ine

aqui

fers

as

wel

l as

fresh

wat

er??

Tran

spor

t mec

hani

sms

of d

ispe

rsio

n,ad

vect

ion

sim

ulat

ed??

Sor

ptio

n, d

egra

datio

n m

echa

nism

s no

tin

corp

orat

ed??

Tim

e-va

riant

rel

ease

and

flo

w r

ates

??

Com

bina

tion

cove

rs u

nsat

urat

ed a

ndsa

tura

ted

zone

s

(415

) 49

3-18

64

??

Req

uire

s m

athe

mat

ical

pro

gram

min

g,hy

drog

eolo

gica

l kno

wle

dge

on p

art

ofus

er

Doc

umen

tatio

n: P

ricke

tt et

al.

1981

??

Has

bee

n fie

ld-v

alid

ated

??

Has

bee

n ap

plie

d fo

r ar

seni

c an

dor

gani

c w

aste

sD

ocum

enta

tion:

Gup

ta e

t al

. 19

87 (Con

tinue

d)

Tabl

e 3-

14. (

Con

tinue

d)

Mod

elD

escr

iptio

nR

esou

rce

requ

irem

ents

, co

mm

ents

Ref

eren

ces,

sou

rces

of d

ocum

enta

tion,

softw

are

San

dra

Was

te ls

olat

ion

Flow

and

•Tr

ansp

ort

Mod

el (

SW

IFT

and

SW

IFT

II) ?

Thre

e-di

men

sion

alTr

ansp

ort p

roce

sses

of a

dvec

tion,

disp

ersi

on s

imul

ated

Sor

ptio

n, d

egra

datio

n pr

oces

ses

acco

unte

d fo

rA

ppro

pria

te f

or w

aste

-infe

ctio

n,w

aste

-isol

atio

n m

odel

ing

Cod

e w

as b

ased

on

SW

IP M

odel

Leac

hate

P

lum

e M

igra

tion

Mod

el (

LPM

M)

Ana

lytic

al T

rans

ient

One

-, Tw

o-,

and

Thre

e-D

imen

sion

al S

imul

atio

n M

odel

(AT1

23D

)

??

Con

tinuo

us s

ourc

e m

odel

??

Dis

pers

ion

is s

imul

ated

??

Deg

rada

tion

proc

esse

s ac

coun

ted

for

??

Has

bee

n fie

ld v

erifi

ed?

A s

impl

istic

mod

el; r

esul

ts m

ay n

ot b

e as

soph

istic

ated

as

nece

ssar

y fo

r Le

vel I

IIw

ork

See

Sec

tion

4.4.

2 of

tex

t

Has

bee

n fie

ld-v

erifi

edD

ocum

enta

tion:

Ree

ves

and

Cra

nwel

lH

as a

ssoc

iate

d us

er’s

gui

de in

sel

f-19

81;

Finl

ey a

nd R

eeve

s 19

68te

achi

ng f

orm

atS

oftw

are:

FOR

TRA

N p

rogr

am; h

as b

een

impl

emen

ted

on v

ario

us C

DC

sys

tem

sin

clud

ing

CD

C 7

600

1986

ver

sion

has

bee

n re

leas

ed

Nat

iona

l E

nerg

y S

oftw

are

Cen

ter

Arg

onne

Nat

iona

l La

bora

torie

sA

rgon

ne,

Ill.

6043

9In

form

atio

n:ln

tera

Env

ironm

enta

l Con

sulta

nts,

Inc

.11

999

Kat

y Fr

eew

ay,

Sui

te 6

10H

oust

on,

Tex.

770

79

Can

be

used

in n

omog

raph

ic, h

and-

held

calc

ulat

or, o

r co

mpu

ter

form

Rel

ativ

ely

easy

to

use

FOR

TRA

N p

rogr

am a

pplic

able

to

wid

era

nge

of c

ompu

ters

May

req

uire

ext

ensi

ve s

etup

tim

eA

vaiIa

ble

thro

ugh

GE

MS

(se

e S

ectio

n4.

1)

Ref

eren

ces:

Ken

t et

al.

1982

Doc

umen

tatio

n: Y

eh 1

981

Tabl

e 3-

14. (

Con

tinue

d)

Mode

lDe

scrip

tion

Reso

urce

requ

irem

ents,

com

men

tsRe

fere

nces

, sou

rces

of d

ocum

enta

tion,

softw

are

Unsa

tura

ted

and

Satu

rate

d Zo

nes

Fini

te-E

lem

ent M

odel

of W

aste

(FEM

WAS

TE) a

ndFi

nite

Ele

men

t Mod

el o

f Wat

er N

ow(F

EMW

ATER

)

Solut

e Tr

ansp

ort a

nd D

isper

sion

Mod

el

Tw

o-di

men

siona

l??

Inte

rzon

e tra

nsfe

r IS m

odele

d??

Inco

rpor

ates

conv

ectro

n, di

sper

sion

??Si

mula

tes

degr

adat

ion o

f non

cons

erva

tive

subs

tance

s??

Abso

rptio

n IS

acc

ount

ed fo

r??

Capa

ble o

f mod

eling

laye

red,

hete

roge

neou

s soil

zone

s??

FEM

WAT

ER is

a m

odel

for g

roun

d-wa

ter f

low, w

hile

FEM

WAS

TE s

imula

tes

the

trans

port/

fate

of c

onta

mina

nts

??Ha

s be

en Im

plem

ente

d on

IBM

360

Docu

men

tatio

n: Y

eh a

nd W

ard

1981

??M

ay re

quire

bac

kgro

und

in hy

drog

eolog

y,dif

fere

ntial

equ

ation

s, pr

ogra

mm

ingIn

form

atio

n:??

Field-

verifi

edDr

. Geo

rge

T. Y

ehOa

k Ri

dge

Natio

nal L

abor

ator

yEn

viron

men

tal S

cienc

e Di

vision

P.O

. Box

xOa

k Ri

dge,

Ten

n. 3

7830

(615

) 57

4-72

85

??Tw

o-dim

ensio

nal

??Co

nser

vativ

e su

bsta

nces

(no

deca

ysim

ulat

ion)

??Fie

ld Ve

rified

??Re

lative

ly ine

xpen

sive,

eas

y to

use

??He

tero

gene

ous s

oil co

nditio

ns a

ccou

nted

for

??Pu

mpin

g or

rech

argin

g we

ll effe

ctsmo

deled

•Th

ickne

ss o

f sat

urat

ed z

one

may

var

y

Docu

men

tatio

n:Ko

wiko

w an

d Br

edeh

aeft

1974

Sou

rces

: U

SE

PA

198

2b;

Bro

wn

et a

l. 19

83;

Kuf

s et

al.

1983

; V

ersa

r 19

83.

88

Tabl

e 3-

15. (

Con

tinue

d)

‘IlOll-A

OUCO

US P

HASE

LIO

UIDS

1, F

OR U

WSAT

IHAT

ED Z

ONE

OWLI.

Table 3-15. (Continued)

90

one of three different ways with respect to ground-water flow; area sources, also aligned in one of threedifferent configurations; or a volume source (existingplume). Release types can be instantaneous,longer-term but finite, or constant. Aquitard locationscan be specified below or on both sides of the aquiferin any configuration; or the aquifer can be treated asinfinite in all directions. Advection and dispersiontransports are simulated. Losses resulting fromvolatilization, degradation, and adsorption aremodeled. The model predicts contaminant movementin one, two, or three dimensions (Yeh 1981).

Use of AT123D requires the following information:dispersion coefficients in horizontal, vertical, andlongitudinal direction: geometry of the aquifer,especially regarding configuration of aquitards; soilproperties, including bulk density, effective porosity,hydraulic conductivity (permeability); source type; andrelease duration and strength, soil-waste streampartition coefficient, hydraulic gradients, and anoverall decay constant (or soil half-life figures) forthe substance studied (Yeh 1981).

The model determines contaminant concentration atany point, at a downstream and lateral distance anddepth specified by the user, as a function of timefrom the beginning of source release.

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

In addition, the Office of Solid Waste (OSW) hasdeveloped a national model that uses the Monte Carlosimulation for relisting hazardous wastes on a genericbasis. This FORTRAN computer model is a three-dimensional advective-dispersive transport model.The model currently considers the mechanisms ofhydrolysis, dispersion, and rainfall recharge into theground-water plume. OSW is using the model toback-calculate from a health-based standard at theexposed population well to an acceptable on-siteleachate concentration. If a treated waste producesleachate with a contaminant concentration below theacceptable concentration, then it is consideredprotective of the public health.

The model currently uses the HELP model to provideleacha te re lease ra tes . Leacha te s t reng th(concent ra t ion) i s p rov ided by the Tox ic i tyCharacteristic Leaching Procedure (TCLP). OSWplans to add the geochemical model MINTEQ tohandle metal speciation. Biodegradation processesare being evaluated for incorporation into the model.

Since EPA’s model is a national model that uses ageneric environment, the data requirements areminimal. The model approximates an averageenvironment by making multiple runs (typically severalthousand runs for each chemical constituent) with

92

varying environmental data. By applying thisapproach, called a Monte Carlo simulation, one canmodel the dilution potential of all possible sites as acumulative frequency distribution versus expectedconcentration at an exposed population well. Theextent to which a particular CERCLA site matches theOSW model depends on the closeness of sitecharacteristics and the model assumptions. If aparticular CERCLA site has adequate hydrogeologicdata and satisfies the model assumptions, the modelcan be used for site-specific analyses. Before finalassessment of the desired level of cleanup, however,application of the model on a site-specific basis willtypically be required. Generic modeling is appropriatefor OSW’s purposes, but may suggest cleanup levelsbeyond those necessary at a part icular s i te.Preliminary work or screening-level efforts atCERCLA si tes where adequate, good qual i tyhydrogeologic data do not exist can benefit from themodel’s minimal data requirements for site-specificenvironmental parameters.

The model is being updated to incorporate flowthrough fractured media and the unsaturated zone.The data base for MINTEQ is being enlarged tohandle additional metals, and more data are beingcollected to validate the model results.

Since OSW’s model uses a Monte Carlo simulatedenvironment, it should be applied with this limitation inmind. Other limitations in the use of this model derivefrom two sources: (1) limitations in the scope of themodel, and (2) specific modeling choices made sothat the model would support OSW’s requirements.The model’s scope is limited by the leachate releasealgor i thm HELP, which models soluble toxicconstituents dissolving into percolating rainwater andmoving with that water. I t does not addresspercolation of organic fluids into the ground orassociated leaching by concentrated organics.

Additionally, the TCLP does not fully predict leachateconcentrations due to leaching with water containingdissolved solvents. It does assume the presence ofacetic acid in leach water, thereby providing somemeasure of hydrophobic solubility. Although HELPcan model a variety of landfill cover situations, OSW’srequirements were such that it modeled a landfill witha failed liner but an intact (aged) cover. Thepermeability of the hypothetical cover was chosen at1 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 permeabilitiesfor aged clay actually to be between 1.4 x 10-6 and43 x 10- 6 (USEPA 1986). For CERCLA sites,selection of a permeability within that range may bemore appropriate. Also, many CERCLA sites do nothave a cover, or the cover may be breached. In eithercase, the mass f lux leaving the si te wi l l beconsiderably larger. Even if the site has an intactcover, one may wish to predict long-term potential

re leases and a lso to cons ider the even tua l ?? The concentration of hazardous material insubsidence and breaching that may occur in the environmental media containing or supportingfuture. vector organisms.

Pesticide Root Zone Model (PRZM) (Carsel et al.,1984) simulates the vertical movement of pesticidesin unsaturated soil, both within and below the plantroot zone, and extending to the water table usinggenerally available input data that are reasonable inspatial and temporal requirements. The modelconsists of hydrology and chemical transportcomponents that simulate runoff, erosion, plantuptake, leaching, decay, foliar wash off, andvolatilization (implicitly) of a pesticide. Predictions canbe made daily, monthly, or annually.

?? The metabolic rate of the vector organisms.Metabol ic rates are funct ions of severalenvironmental parameters including temperatureand the availability of sunlight, oxygen, nutrients,and water or other factors.

3.5.4 Short- and Long-Term ConcentrationCalculationsLong-term average ground-water concentrations ofcontaminants at exposure points are a function of theconcentration profile over the time period of study,which are, in turn, a funct ion of hydrologicfluctuations, release rate fluctuations, and thee f fec t i veness o f remed ia l ac t ions . Averageconcentration values are obtained from steady-statemethods. Several of the in-depth analysis modelstabulated in Section 3.5.3 accept time-weightedinput data, and p rov ide long- te rm averageconcentrations, as well as the concentration profile asa function of time.

?? Substance bioavailability: the affinity of eachhazardous substance for partitioning into theorganic phase or its availability for other forms ofuptake. The bioavailability of each substancediffers, as does that of various chemical speciesof an individual substance: the octanol/waterpartition coefficient is an indication of thisparameter. Bioavailability of a given substancecan vary with environmental conditions. Factorsthat influence the physiochemical speciation ofsubstances, and thus their bioavailability, includesalinity, pH, Eh, organic carbon concentration,and temperature.

Short-term concentrations at exposure points areo b t a i n e d b y e x a m i n i n g t h e g r o u n d - w a t e rconcentration profile at the selected exposure pointover time, and identifiying of the period of maximumconcentration.

?? Characteristics of species metabolic processes.These characteristics differ among species andinclude feeding habits and ability to metabolicallydegrade, store, and eliminate the substance.Bioconcentration factors (or BCFs, the ratios oforganism t issue concentrat ion to ambientenvironmental concentration) for many speciesand hazardous substances have been empiricallydetermined and are discussed below.

3.6 Biotic Pathways

Consider the following transport mechanisms inassessing the distribution of hazardous substanceswithin the biologic medium and identifying thepotential points of human exposure:

3.6.1 Estimation ProceduresAfter the fate of a contaminant in air, water, andground water has been estimated, one can assess itsfate in biot ic populat ions. Using the ambientconcentration data developed for each of thesemedia, a determination is made whether any bioticpopulations that can potentially serve as pathways forhuman exposure to hazardous materials (i.e., vectororganisms) are within zones of elevated hazardousmaterial concentrations. Such vector populations mayinclude agricultural crops; agricultural livestock; fish,shel l f ish, or crustaceans that are importantcommercial or sport species; and game populations inhunting areas.

? Transport and distribution of vector organisms asa result of human commercial or sport activity.

?? Migration of organisms, or movement of theseorganisms with advective flow of environmentalsubstrate media.

?? Movement of contaminants through the foodchain. This mechanism often results in very highconcentrations of hazardous materials in thetissue of higher trophic level organisms within andwithout contaminated areas.

In assessing the biological fate of hazardousmaterials, the following processes, which determinethe rate of introduction of hazardous material to andthe final concentration of hazardous material withinvector organisms, should be considered:

General theoretical relationships between the abovefactors and concentrations of hazardous substancesat human exposure points are not available. This isbecause such relationships are highly specific toindividual ecologies, biotic species, hazardoussubstances, and human activities associated withinvolved biotic species.

93

For this reason, t h e a s s e s s m e n t o f b i o t i cconcentrations of hazardous substances at humanexposure points is l imited to the qual i tat iveidentification of major pathways, and the roughquantification of exposure levels wherever somemeans of relat ing ambient soi l , water, or airconcentrations to edible tissue concentrations areavailable.

The ava i lab le methods o f es t imat ing t i ssueconcentrations in aquatic animals, terrestrial animals,and terrestrial plants are discussed in the followingsections.

3.6.1.1 Aquatic AnimalsBecause aquatic animals are immersed in thecontaminated water medium to which they areexposed, it is commonly assumed that tissuecontaminant concentrations are a function ofcontaminant equilibrium partitioning between waterand organic tissue, and are therefore directly relatedto contaminant ambient water concentrations. Thebioconcentration factor (BCF) represents the ratio ofaquat ic animal t issue concentrat ion to waterconcentration. This ratio is highly contaminant-specific and is also dependent on the aquatic speciesand on site parameters.

The most reliable source of aquatic animal BCFvalues is monitoring data for the site. Wherever waterconcentrations and biotic tissue concentrations havebeen surveyed simultaneously, a site-specific BCFcan be calculated for the species and substanceinvolved (assuming water column concentrationvalues represent relatively steady concentrations overat least the previous several weeks, and not short-term high or low concentrations). This BCF can beused to project changes in tissue concentrationsresulting from projected changes in ambient waterconcentrations of the involved hazardous substance.

In cases where site monitoring data are insufficient fordevelopment of a BCF, one can use the BCF valuesreported in technical literature. A substantial amountof research is available regarding the bioconcentrationof hazardous substances, especially in aquaticorganisms (see USEPA Office of Water Regulationsand Standards: Ambient Water Quality Criteriadocuments, for a review of research current to 1980;or Verschueren 1984; Dawson, English, and Petty1980; Mabey at al. 1982; and Callahan et al. 1979 forBCF factors). Exercise care to match contaminants,species, and site conditions (e.g., temperature, pH,water salinity) for which reported BCF values weremeasured with conditions at the site. BCF values fordifferent species or contaminants or those measuredunder dissimilar conditions may not be applicable.

A third alternative for derivation of BCF values is tocalculate these values based on the structure orphysiochemical propert ies of the hazardous

substance. See Lyman et al. (1982), Kenaga andGoring (1978), and Veith et al. (1980) for instructionson BCF estimation procedures.

3.6.1.2 Terrestrial AnimalsLittle data are available allowing the quantification ofcontaminant concentrations in edible terrestrial animalt i s s u e b a s e d o n a m b i e n t e n v i r o n m e n t a lconcentrations. Kenaga (1980) compiled and studieddata comparing dietary concentrations of severalorganic compounds with the concentration of thesecompounds in the fat of beef cattle. He found that thefat /diet BCFs for these compounds correlatereasonably well with the water solubility (negativecorrelation) and octanol-water partition coefficient(positive correlation) of these compounds. BCFscould only be predicted within three to four orders ofmagnitude, however. This method of t issueconcentrat ion est imat ion must be consideredsemiquantitative at best.

Human exposure to contaminants through theterrestrial animal pathway can be reliably determinedonly through identification of potential vectororganisms and exposure points, and through asampling and analysis program for determining tissueconcentrations at these exposure points.

3.6.1.3 Terrestrial PlantsPlant adsorption of environmental contaminants hasbeen studied by various researchers, and some dataare available regarding the uptake of pesticides andother contaminants by edible crops. These data coverspecific crop uptake of specific contaminants (seeCDHS 1985 for a review of pesticide research),however, and no relationships allowing reliableextrapolation of soil/plant tissue concentration ratiosare presently identified. Where plant/soil BCF data areavailable in the technical literature for the specificplant species, contaminant, soil type, and tissue typeof concern in a Superfund exposure assessment,these BCF data can be used for a semiquantitativeestimation of edible tissue concentrations.

As is the case with terrestrial animals, the mostrel iable technique for assessing contaminantconcentrations at points of human exposure to planttissue is the identification of potential vectororganisms and exposure points, and the surveying oftissue contaminant concentration in these organisms.

94

Chapter 4Uncertainty in the Analysis

This chapter provides a brief introduction to theevaluation of uncertainties inherent in the exposureassessment process. When applying the exposureassessment tools outlined in the preceding sections,uncertainty may be a factor at each step. Suchuncertainty can involve variations in the values ofvariables used as input to a given model, theaccuracy with which the model itself represents actualenvironmental processes, and the manner in whichthe exposure scenario is developed. Each of thesecategories of potential uncertainty is discussed below.Once the exposure assessment is completed, itsresults must be reviewed and evaluated to identify thedegree of uncertainty involved. This factor shouldthen be considered when using the assessmentresults for remedial decisionmaking.

The following discussions focus on the uncertaintiesof assessing the average daily exposures to toxicchemicals; uncertainties related to the’ human healthresponse to these exposures are not discussed. Theinformation provided here does not constitute acomprehensive treatment of uncertainties in theexposure assessment process. It is intended to makethe analyst aware of the categories of uncertaintiesthat may be involved in exposure assessments. In-depth guidance for the execution of uncertaintyanalyses is provided in various references in theliterature. Specifically, the analyst may wish to reviewthe following sources of information concerningvarious aspects of uncertainty analysis pertinent tothe 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 UncertaintyMost of the analytical procedures presented in thismanual are quantitative in nature, and their resultsmay be highly dependent upon the accuracy of the

input var iables used. For example, hydraul icconductivity and other parameters that determine thevelocity of ground water and the contaminants that itmay carry can vary significantly over relatively shortdistances, thereby affecting one’s ability to estimateaverage contaminant velocities with confidence.Similarly, the presence of hydrogeologic hetero-gene i t ies can a f fec t the speed w i th wh ichcontaminants arrive at a given well from their point ofrelease and also their direction of travel. Often, thepresence of such heterogeneities may be unknown.Thus, the accuracy with which values for suchparameters can be quantified is critical to the degreeof confidence that the decisionmaker has in theassessment results.

Most scientific computation involves a limited numberof input variables, which are tied together by a modelto provide the desired output. The input parameterscan be broadly classified into the following categories:constants, state variables, and natural variables.

A constant has a single value irrespective of thenature of other variables. In some cases, thevariability of a parameter may be so small that it canbe considered constant. In other cases, even if thevalue varies, its effect on the final answer may beminimal. The results are not sensitive to variation inthat parameter’s value.

A state variable is one that has a fixed value, but thatvalue is not known accurately. The errors in suchvariables are due to limitations in experimentaltechniques. A relevant example is the octanol/waterpartition coefficient. While this has a single value for agiven system, some degree of uncertainty isintroduced through experimental errors. In someinstances the values of state variables are estimatedrather than measured; therefore, the uncertainties forsuch values are even higher.

A natural variable is one that can exhibit differentvalues. An example is soil porosity, which can exhibitdifferent values within a range because the soil matrixvaries with location, and because a given area mayinclude many soil types.

95

If the actual values for such variables are notaccurately known for the location in question, theestimated exposure may be significantly in error. Thisproblem is illustrated by a study where the values ofparameters needed to calculate the velocity of asolute in ground water were varied randomly, usingMonte Carlo simulation techniques (Mercer, Silka,and Faust 1985). This analysis determined that thevelocity estimates may vary over four orders ofmagnitude.

The selection of accurate input parameters isessential to estimate the contaminant velocity andother components of the exposure assessment.Often, however, the analyst will not be able todetermine the value of such parameters with absolutecertainty. It is important that one be aware of the typeand degree of uncertainties involved at each stage ofthe analysis, and interpret the results obtainedaccordingly.

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

Input parameter distributions can be used to generatethe output parameter distribution. The shape of theparameter distribution conveys the degree ofuncertainty of the parameter (input or output). This isthe most rigorous way to define the uncertainty of thepredicted output parameter; however, it is usedinfrequently in the environmental field due to the lackof input parameter distributions upon which to basethe predicted output parameter distribution. Thissubject will be discussed further in the section on theMonte Carlo technique.

In the environmental field, the methods used fordiscussing the degree of uncertainty are oftenqualitative rather than quantitative. Qualitativemethods involve discussing whether the data arethought to be representative or not. Some exposuremodeling is done based on literature values ratherthan measured values. In such cases the degree ofcertainty may be expressed as whether the estimatewas based on literature values or measured values,not on how well defined the distribution of theparameter is. Some exposure estimates are based onestimated parameters; the qualitative statement thatthe exposure was based on estimated parametersdefines the certainty in a qualitative manner.

4.2 Modeling Uncertainty

4.2.1 Model SimplificationThe degree to which a specific contaminant transportand fate model accurately represents the actualconditions that are present in the environmentconstitutes a large source of potential uncertainty.The analyst must choose the model that addressesthe appropriate aspects of interest.

Models are typically simplifications of the complexitiesof reality. There is some accuracy lost when makingthese simplifications. While such loss may be small insome cases, in others it may be unacceptably large.Two assumptions that illustrate this idea are theassumptions of homogeneous soils and isotropic soilsfor ground-water models. In most cases, theseassumptions do not materially change the answer. Ifthe soil under the site has layer cake stratigraphy, theassumption of homogeneity is invalid. Typically, mostcases will be in-between the two extremes ofh o m o g e n e o u s s o i l s a n d c o m p l e t e l y n o n -homogeneous soils. The analyst will have to decide ifthe assumptions are valid for each case.

In some cases the simplification of the real world intoan actual model is acceptable and, althoughproducing uncertainty, it is a necessary evil. There isa point at which the level of the discrepancy betweenthe model and the real world constitutes an error inthe use of the model and not an acceptables imp l i f i ca t ion tha t i s necessary to mode l acomplicated real world. At this point, the deviation isan error and not an uncertain prediction.

4.2.2 Averaging Hydraulic ConductivitiesAn example of this would be the modeling of groundwater flow by averaging the hydraulic conductivitiesacross all aquifer materials. For contaminant transportmodeling, this would constitute an error; however, formodeling well production, this is an acceptedpractice. Ground water modeling with numbers hasbeen occurring for the last 100 years. For the first 90years of this period, most of the modeling was forwater supply; contaminant migrat ion was notmodeled. The practice of averaging the hydraulicconductivities across the cross-sectional area of theaquifer produced answers that had high certaintieswhen predicting the volume of water that could beproduced by a well during a period of time. Somemodelers applied this technique to the problem ofmodeling contaminant migration and producederroneous results. Although they were accustomed tothis practice, it was not acceptable in this case.

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

96

layers separately. The result of the separate modelingwill show that the time of arrival in the sand is muchsooner than in the clay layer. Effectively the majorityof the contaminant mass would migrate through thesand layer and hardly any would use the clay layer formigra t ion . Assuming an average hydrau l i cconductivity would predict a time delay betweenrelease and arrival that is 100 to 1000 times too long.Such uncertainties, however, constitute an error ofapproach, and are not unresolvable uncertainties.

4.2.3 Dispersion SimulationDifferent ground-water models simulate dispersion indifferent ways. The degree to which a particularmodel accurately models the dispersion at a givensite affects the accuracy of using that model for thatsite. Ground water dispersion modeling is a youngfield and the state of the art is rapidly advancing. Theanalyst should become familiar with the dispersionsimulation technique for each model he/she uses.

Also, some ground water models presume an aquiferof infinite depth, while some model a finite aquiferdepth. Contaminants dispersing in an aquifer of finitedepth will effectively reflect off the lower aquitard andcause the resulting downstream concentrations to belarger. Use of a model appropriate to the constraintsof the site is necessary for accurate modeling of thedrop-off in contaminant concentration with traveldistance. Additionally, some models will simulatelateral constraints of the aquifer to model thisl i m i t a t i o n o n t h e r e d u c t i o n i n d o w n s t r e a mconcentrations.

Dispersion modeling in air and surface water hasbeen performed for a much longer time, and as such,the methods for modeling dispersion have coalescedinto a consistent approach. However, limitations onthe extent of dispersion for air modeling can vary. Forexample, a valley model will simulate the constraint oflateral dispersion by the valley walls. A model thathandles inversions will simulate the build-up ofcontaminant concentration due to limited verticalmixing. Surface water models may vary on theapproach they take to modeling initial mixing. Somesurface water models use compartments to managethe modeling task. If the modeler uses a smallnumber of large compartments, small scale effectsmay not be accurately modeled and the results will beless certain.

4.2.4 Numerical Models and Analytical ModelsDifferent types of models provide varying accuracy indifferent situations. Two types of models arenumerical (finite-element) and analytical models.Neither is best in all cases, but one is usually betterin a given situation. The numerical models aretypically more difficult to use, and thus ease of usemay enter into the decision of model selection.

Analytical models often involve mathematicalsimplifications. These simplifications are made inorder to find a closed-form solution. In most casesthe accuracy lost is negligible; however, in extremecases the inaccuracy will be large.

Typically, analytical models require less computertime than do numerical models. If the grid is large, anumerical model requires a substantial amount ofcomputer time for each run. Numerical modelstypically require more input data. Different programneeds cause different questions to be raised. Apreliminary scoping problem will rarely require anumerical model; conversely, a problem that requiresmaximum defensibility will suggest that the additionaldata and operational burdens of a numerical modelare justified in light of the greater certainty of theoutput.

In cases where the question involves simulating whatwill happen in typical generic situations across thecountry, an analytical model will give a better picturethan a numerical model. Numerical models addresssite-specific conditions better than do analyticalmodels: they do not necessarily model a typicalsituation with any increased accuracy.

4.2.5 Chemical Degradation SimulationSome models do not describe all of the processesthat may potentially occur. For example, degradationis not accounted for in some models. If the con-taminant is extremely refractory (i.e., does notdegrade), this limitation will not materially affect theanswer. If the contaminant degrades quickly,however, this limitation will cause the model results tobe in substantial error. Some models simulate theeffect on the reaction rate kinetics of two con-centrations while some use only one concentration.The simpler approach of 1st order reaction kinetics isacceptable if the other concentration does not varyappreciably, and is less accurate if both the con-centrations vary substantially. The analyst must relyon his/her judgment to ensure that the uncertainty isminimized.

4.2.6 Model Operational ParametersCertain modeling parameters specified by the analystcan have a profound effect on the accuracy andviability of the output. An example is the parameter“time step.” Time step is used on iterative models.Models may either calculate an answer explicitly orthey may determine their solution with a successiveiteration approach. For iterative models, the analystwill have to make many model runs, and not stop untilhe/she has a good run. The challenge of choosing anappropriate time step is that both too large and toosmall time steps cause inaccuracies. The analystmust find the optimum size for the time step. A timestep that is too small causes numerical error

97

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

Numerical error propagation in iterative models cancause inaccurate answers. If the analyst uses toomany iterations, the truncation error of digitalrepresentation of numbers can build upon thesuccessive iterations and produce output that istotally erroneous, The degree of error that can bepresent can make the output totally meaningless. Forexample, the estimated output concentrations caninclude concentrations that are greater than a millionPPM or concentrations that are negative. Clearly,concentrations in these ranges signify a bad run. Theanalyst must also watch out for iteration errors thatproduce errors that are less obvious and, hence,there is the possibility that the analyst will not beaware of their occurrence. Conversely, the analystmay choose too few iterations and the resulting timestep between each iteration then becomes too large.In this case, the model will inaccurately calculateeach step. The analyst must become familiar with themodels he/she is using so as to stay in the safe areabetween the two extremes. Knowing the precise limitsis difficult, but staying between them is important.

4.2.7 Source ShapeThe degree to which the shape of the source ismodeled can effect uncertainty. For example, if theanalyst uses a point source model to model an areasource, the nearby concentrations will be lessaccurate than they would be if the analyst used anarea source model. Line sources and volume sourcescan provide the same problem. At large distancesfrom the source, the effect of the shape of the sourceis less important, and may often be neglected. Somesources are best modeled as a vertical line sourceand some are best modeled as a horizontal linesource; hence, orientation is a factor as well asshape. It is a matter of fit between the model and theactual site rather than choosing the best sourceshape for all cases.

4.2.8 Steady State ModelingUse of a steady-state model to model a truesteady-state scenario provides accurate results. Useof a steady-state model to model a truly dynamicscenario can produce inaccurate answers. In mostcases, the analyst will have to make a judgment as towhether the actual scenario is close enough to steadystate to justify using a steady-state model. Theanalyst must match the model to the question beingasked, and to the details of the specific site, in orderto minimize the uncertainty of the output.

4.2.9 Number of Dimensions Addressed by theModelChoice of a one-, two- or three-dimensionalmodel can affect the uncertainty of the results.Neither is best in all cases and, typically, one ispreferred in a given site-specific scenario. The

three-dimensional model general ly has lessuncertainty than the one- -or two-dimensionalmodels, but, this is not always the case. For example,when modeling the migration of contaminants inground water through a IO-foot thick aquifer, atwo-dimensional model will produce more certainresults than the blind application of a three-dimensional model. It is not just a trade-off betweendifficulty of the model and quality of the output, but amatching situation as well.

4.3 Scenario Uncertainty

The analyst needs to be aware of uncertainties thatresult from using conservative assumptions whendata are lacking. While it is traditional in exposureassessment to make conservative assumptions in theabsence of data, such assumptions must bereasonable and the assessment results must bein te rp re ted w i th cau t ion . Use o f reasonab lyconservative assumptions at each step may producecumulative assessment results that are overlyconservative and thus unreasonable.

In addition, conceptual errors may result in the use ofassumptions that affect the selection of the modelingtechnique applied to the exposure assessment. Forexample, using a three-dimensional model insituations where the aquifer thickness is not “large”in relation to the areal extent of contamination wouldnot be appropriate. Thus, the concepts upon whichthe exposure scenario is based must be carefullyconsidered to make sure that they adequately reflectthe situation under evaluation.

Quantitative descriptions of scenario uncertainty areoften impractical, and qualitative descriptions of thelevel of uncertainty are more common for the young,and developing, field of exposure assessment. Anyexposure prediction has cases of overstatement andunderstatement of r isk. Where possible, theunderstatements and overstatements of risk areminimized. Where this is not possible, the analystattempts to balance them so as to produce aprediction that is most realistic.

4.4 Approaches for Dealing withUncertainty4.4.1 Sensitivity AppraisalsVariation in the values of input parameters causesvariation in the values of the output parameters. Theratio of the input parameter variation to the outputparameter variation will be different for parameters indifferent parts of the equation. Sensitivity appraisalsinvolve assessing which parameters have the highestratios and which have the lowest. The accuracy ofparameters that have the largest effect on theaccuracy of the output parameters should be high,while parameters that have only a small effect on theaccuracy of the output parameters can be estimated

98

or determined by less accurate and less costlymethods.

Sensit ivi ty appraisals can be quanti tat ive orqualitative. A quantitative sensitivity appraisal involvesplotting the output parameter as a function of variationof a single input parameter, while holding all of theother input parameters constant. As one can imagine,there may be a different functional relationshipbetween the output parameter and the varying inputparameter for each combination of fixed inputparameters. For complex models the approach canbecome overwhelming. Typically, the analyst will beable to interpret the equation and set up the fixedinput variables so as to minimize the number offunctional relationships produced. However, it maystill be burdensome, and it may produce results thatare more precise than necessary.

dependent. The analyst could assume independencybecause the two variables represent different factorsthat do not have a direct functional relationshipbetween them. But, if the analyst looked at enoughsets of data, the sites with high conductivity wouldhave more gradients that are flat; conversely, thesites with low conductivity would have more gradientsthat are steep. Thus, the two variables would exhibitcovariation and cannot be considered strictlyindependent. This weakens the validity of using theMonte-Carlo approach.

In the environmental modeling field, the qualitativeapproach has strong advantages over a quantitativeapproach. The qual i tat ive approach involvesinspecting the model’s equations, and ascertainingwhich input variables are the most sensitive. This isusua l ly done by v isua l inspect ion , w i th anunderstanding of the mathematical relationships in theequation. For example, if one input parametermultiplies all the other terms, the analyst can expectthe input parameter to have a sensitivity ratio of one.If the input parameter is the exponent of the otherterms, the analyst can expect this parameter to havea very high sensitivity ratio. If the input parameter ispart of a separate term that is added to the rest of theequation, and it is multiplied by a constant of lowvalue, the input parameter can be assumed to have alow sensitivity ratio. A qualitative appraisal is usuallythe most efficient technique for determining the inputparameter accuracy needs.

While i t is possible to use input parameterdistributions to generate model output distributionsusing Monte-Carlo simulations, it is usually notpossible to get the input parameter distributions. Theinput parameter distribution shows the variation ofparameter values. It must be based on a largenumber of observations (actual measurements). Theenvironmental field is young and growing. As such,most sampling (to date) falls short of providing themass of data necessary to generate an inputparameter distribution. Faced with this dilemma, someanalysts have fallen back on assuming suchdistribution. Since they do not have a way to gaugethe distribution, a uniform distribution from the lowestto highest possible value is assumed. This distributionstates that there is an even probability that the valuecould be any value between the lowest and thehighest value of the range.

4.4.2 Monte-Carlo SimulationsThe Monte-Carlo technique involves running amodel a large number of times with varying inputparameters. The values for the input parameters arechosen from the parameter distributions, with itsrelative frequency of a particular value being usedbeing equal to the relative frequency in the parameterdistribution. This is based on the assumption that theinput variables vary independently from each other.This technique generates an output parameterdistribution, which provides a mode and a statementof the uncertainty associated with the prediction.

Assuming an input parameter distribution does nothelp to reduce uncertainty, however, as the certaintyof the output is then a function of the assumedcertainty of the input parameter. For example, if youassume that the input parameters are very precise,then the certainty of the output is high. Conversely, ifyou assume the parameters may have an equalprobability to be any value across the range ofpossible values, the certainty of the output will be low.Using a Monte-Carlo approach with assumed inputparameter distributions that are uniform only indicateshow accurate the model is at predicting the outputparameter when you have no idea what the inputparameters are, since models predict output based onthe relationship to the input parameters. Thus, usingthe Monte-Carlo technique to assess the certainty ofa model’s predictions cannot be done with assumedinput parameter distributions.

4.4.3 Using Monitoring Data to Calibrate theModel

One difficulty with this technique is the assumption of One of the best ways to reduce the uncertainty of theindependent variation. The input variables are chosen predicted parameter is to use monitoring data toas if there were no relationship among them. If the ca l ib ra te the mode l . I f you have measuredvariables are truly independent, the results are contaminant concentrations that are comparable toaccurate. Typically, however, the variables are related modeled contaminant concentrations, the analyst canto each other and are, thus, dependent variables. For correct for over- or under- predictions. If forexample, if the two input variables are hydraulic example, the measured values are always 90% of theconductivity and hydraulic gradient, the analyst could predicted values, the analyst can multiply all of theassume that they are ei ther independent or output values by 90%.

99

The difficulty of this technique is that the values mustbe comparable. In many cases the model is beingused to predict future events. Current contaminantconcentrations can be determined more accurately bymonitoring, thus the need for modeling is reduced.

In air and surface water modeling the differencebetween current and future events is much smallerthan for ground water modeling. Air and surface watermove more quickly than does ground water. Hence,calibration is a more useful technique in air andsurface water modeling than in ground watermodeling.

If the ground water model predicts a certaincontaminant concentration 1 mile from the sourcea f t e r 2 0 y e a r s , a n d m o n i t o r i n g s h o w s n ocontamination at 1 mile from the source, this cannotbe used to calibrate the model. The plume may nothave reached the point 1 mile away, as of yet. In 20years, monitoring may very well show the samecontaminant concentration that was predicted by themodel. Care should be taken to ensure that themonitoring data used to calibrate the model arecomparable in time and space.

4.5 Level of Uncertainty Appropriate forExposure ModelingThere is no one level of certainty that is appropriatefor all situations. Each program has different needs,and various parts of a program have diverse needs. Ascreening level study has less need for accuracy thana court case that will require a substantial sum ofmoney from a PRP. The level of defensibility requiredwill vary from one situation to another

EPA program offices have developed a multi-tieredapproach. A desk top model may be sufficient for afirst-tier analysis, an analytical model may besufficient for a second-tier analysis, and a numericalmodel may be required for a third-tier analysis. Forexample, the method of screening sites for inclusionon the National Priorities List should be less rigorousthan the method of supporting a decision on varioussite clean-up options. Data requirements will alsovary.

Although it would be nice to have maximum accuracyin all cases, it would also imply maximum difficulty inall cases. Clearly, a balance must be found betweendifficulty 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 theanalyst’s supervisor. This information should includethe predictions of exposure over time, and it shouldinclude some communication regarding the level of

uncertainty associated with the prediction. The levelof uncertainty can be expressed in a quantitative orqual i tat ive form. F u r t h e r g u i d a n c e o n r i s kcommunication can be found in USEPA (1987e).

A quantitative appraisal of the uncertainty is the mostpreferable way to express the uncertainty. Aquantitative presentation may be an output parameterdistribution which tells the most probable value(mode) and the relative probability that the value islarger or smaller than the mode. Or, the presentationmay consist of the predicted value and a standarddeviation. The standard deviation provides the level ofprecision or uncertainty. Another approach involvesproviding the predicted value and the 95% confidencelimits. The 95% confidence limits express that 95%of the possible values of the parameter will bebetween the upper and lower confidence limits, Themain catch to precise numerical expression of theuncertainty is the lack of sufficient data upon which tobase the quantitative expression of the uncertainty. Inthe future, it may be possible to use this preciseapproach.

A qualitative appraisal of the uncertainty is the mostviable way to express the level of uncertainty. Aqualitative presentation will describe the significantfactors that determine the level of uncertainty. Thequality of the prediction is a function of the quality ofthe inputs to the prediction. Major inputs that affectquality are: data precision, model sophistication, anddefensibility of the scenario.

Expressing the quality of the data would entaildescribing the sources of the data. For example: Didthe data come from literature values or were the datataken from actual site measurements? Were the datameasured by the best available techniques or werethey sampled by another technique? Were replicatesamples taken? Was the sampling protocol sufficientto obtain representative samples? Are the costs ofthe sampling program appropriate for the use of theresults, or could more expensive data gatheringtechniques be used?

Expressing the quality of the model used would entaila description of the type of model. For example: Isthe model a desk-top calculation, an analyticalmodel, or a numerical model? Has the model been inuse for some time or is it new? Is the model astandard model used by the agency or is it new to theagency? Have other people used the model? Doesthe model address all of the important facets of thesituation, or does it neglect some potentially importantfactors? Has the model been used in court casesbefore? How good is the model relative to otherpossible models? Is it the best available model at thispoint in time? Is the model the most defensible modelavailable? Were monitoring data used to calibrate themodel predictions? How comparable were themonitoring data to the model predictions?

100

Expressing the quality of the scenario is more difficult.Reasonableness of the scenario is important. Use ofsimilar scenarios by the agency in the past is usefulinformation. Questions to ask would include: Was thescenario used in court cases, for rulemaking activitythat has been published in the Federal Register,and/or did it receive public comment? Was the publiccomment favorable or did it bring out potentialdifficulties? Does the scenario neglect certainexposure routes that have been neglected by theagency in the past?

The important aspect to consider is how good theprediction 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, andnothing more.

101

Chapter 5References

Ambrose, R.B. and Vandergrift, S.B. 1986. SARAH, asurface water assessment model for backcalculating reductions in abiotic hazardous waste.Athens, GA: U.S. Environmental ProtectionAgency. EPA-600/3-86/058.

Ambrose, R.B., Vandergrift, S.B. and Wool, T.A.1986. WASP3, A hydrodynamic and water qualitymodel-model theory, a user’s manual, a pro-grammer’s guide. Athens, GA: U.S. EnvironmentalProtection Agency. EPA-600/3-86-034.

Ambrose, R.B., Wool, T.A., Connolly, J.P. andSchantz, R. 1987. WASP4, a hydrodynamic andwater quality model-model theory, a user’smanual, a programmer’s guide. Athens, GA: U.S.Environmental Protection Agency.

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

Beljin, M.S. 1985. A program package of analyticalmodels for solute transport in groundwater‘SOLUTE’. Indianapolis, IN: International GroundWater Modeling Center, Holcomb ResearchInstitute, Butler University.

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

B o n a z o u n t a s , M . , F i k s e l , J . , e t a l . 1 9 8 2 .Environmental mathematical pollutant fate modelinghandbook/catalogue (Draft). Washington, DC: U.S.Environmental Protection Agency, Office of Policyand 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 andII. Washington, DC: U.S. Environmental ProtectionAgency. EPA 45014-79-030. As reviewed in:Bonazountas M, Fiksel J., et al. 1982. En-

vironmental mathematical pollutant fate modelinghandbook/catalogue (Draft). Washington, DC: U.S.Environmental Protection Agency, Office of Policyand Resource Management. Contract No. 68-01-5146.

Brown, D.S. and Allison, J.D. 1987. MINTEQA1,equilibrium metals speciation model: a user'smanua l . A thens , GA: U .S . Env i ronmenta lProtection Agency.

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

Burns, L.A., Cline, D.M., Lassiter, R.R. 1982.Exposure analysis modeling system (EXAMS) usermanual and system documentation. Athens, GA:Environmental Research Laboratory, Office ofResearch and Development, U.S. EnvironmentalProtect ion Agency. EPA-600/3-82-023. Asreviewed in: Versar Inc. 1983. Methodology forassessing exposures to chemical substances viathe ingestion of drinking water. Washington, DC:U.S. Environmental Protection Agency. ContractNo. 68-01-6438.

Burt , E. 1977. VALLEY model user’s guide.Washington, DC: U.S. Environmental ProtectionAgency. EPA-450/2-77-018. As reviewed in:U.S. EPA. 1982. Env i ronmenta l Mode l ingCatalogue. Washington, DC: U.S. EnvironmentalProtection Agency. Information Clearing House.PM-211A.

Busse, A.D., Zimmerman, J.R. 1976. User’s guide forthe climatological dispersion model. ResearchTriangle Park, NC: U.S. Environmental ProtectionAgency . EPA-R4-73-024. As rev iewed in :B o n a z o u n t a s M , F i k s e l J . , e t a l . 1 9 8 2 .Environmental mathematical pollutant fate modelinghandbook/catalogue (draft). Washington, DC: U.S.Environmental Protection Agency, Office of Policyand Resource Management. Contract No. 68-01-5146.

103

Callahan, M., Slimak, M., Gabel, N., et al. 1979.Water-related environmental fate of 129 prioritypollutants. Washington, DC: U.S. EnvironmentalProtection 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 thepesticide root zone model (PRZM): release 1.Athens, Georgia: Environmental ResearchLaboratory, U.S. Environmental Protection Agency,EPA-600/3-84- 109.

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

Christiansen, J.H. 1976. Design and application of theTexas episodic model. Proceedings of theConference on Environmental Modeling andSimulation. Washington, DC: U.S. EnvironmentalProtection Agency. EPA-600/9-76-016. Asreviewed in: U.S. EPA. 1982. EnvironmentalModeling Catalogue. Washington, DC: U.S.Environmental Protection Agency. InformationClearing House. PM-211 A.

Clapp, R.B. and Hornberger, G.M. 1978. Empiricalequations for some soil hydraulic properties. WaterResources Research. 14:601-604.

Codell, R.B., Key, K.T. and Whelan, G. 1982. Acollection of mathematical models for dispersion insurface water and ground water. Washington DC:U . S . N u c l e a r R e g u l a t o r y C o m m i s s i o n .NUREGICR-0868.

104

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

Cowherd, C., Muleski, G.E., Englehart, P.J. andGillette, D.A. 1985. Rapid assessment of exposureto particulate emissions from surface contaminationsites. Kansas City, MO. Midwest ResearchInstitute. PB85-192219.

Craig, D.G., Turelle, J.W. 1964. Guide for winderosion control on cropland in the Great Plainsstates. U.S. Department of Agriculture. SoilConservation Service. Washington, DC.

Cupitt, L.T. 1980. Fate of toxic and hazardousmaterials in the air environment. Research TrianglePark, NC: U.S. Environmental Protection Agency.EPA-600/3-80-084.

Das, B.M. 1983. Advanced Soil Mechanics. NewYork, McGraw-Hill Book Company.

Davis, L.A. and Segol, G. 1985. Documentation anduser’s guide: GS2 and GS3 - variably saturatedflow 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 wasteconstituents. Athens, GA: U.S. EnvironmentalProtection Agency, Environmental ResearchLaboratory.

Delos, C.G., Richardson, W.L., DePinto, J.V., et al.1984. Technical guidance manual for performingwasteload allocations, book II: streams and rivers.U.S. Environmental Protection Agency. Office ofWater Regulations and Standards. Water QualityAnalysis 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 hydrologicalsimulation program - Fortran (HSPF). U.S.Environmental Protection Agency. EPA-600/3-84-065.

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

Enfield, C.G. and Bengtsson, G. Not dated. Macromolecular transport of hydrophobic contaminants inaqueous environments. EPA/ORD, Ada, OK.

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

Enfield, C.G. 1984. Chemical transport facilitated bymul t iphase f low sys tems. Rober t S . Ker rEnvironmental Research Laboratory. U.S.Environmental Protection Agency, Ada, OK.Presented at seminar on Degradation, Retentionand 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 forestimating air emissions from hazardous wastetreatment, storage and disposal facilities. RevisedD r a f t F i n a l R e p o r t . G C A C o r p o r a t i o n .G C A / T e c h n o l o g y D i v i s i o n . B e d f o r d ,Massachusetts, prepared for U.S. EnvironmentalProtection Agency, Office of Solid Waste. LandDisposal Branch. Contract No. 68-02-3168.

Farmer, W.J., Yang, M.S., Letey, J., Spencer, W.F.and Roul ier, M.H. 1978. Land disposal ofhexachlorobenzene waste: controlling vapormovement in soils. San Antonio, TX. 4th AnnualSymposium on Land Disposal.

Felmy, A.R., Girvin, D.C., Jenne, E.A. 1984.MINTEQ -- a computer program for calculatingaqueous 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: NuclearRegulatory Commission. NUREG/CR-1968,SAND 81-0410. As reviewed in: Lo TYR, ScottBH, Benjamin RR. 1983. Remedial act ionassessment models for hazardous waste sites.Review draft. Athens, GA: U.S. EnvironmentalProtection 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 coastalwaters. 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 SoftwareCorporation. Washington, DC: Office of Pesticidesand Toxic Substances, U.S. EnvironmentalProtection Agency. Contract No. 68-01-6618.

Gelhar, L.W., Mantoglou, A., Welty, C. and Rehfeldt,K.R. 1985. A review of field-scale physical solutetransport processes in saturated and unsaturatedporous media. Palo Alto, CA: Electric PowerResearch Institute. EPRI EA-190.

Geotrans. 1986. Analytical models for evaluatingleachate migration in groundwater systems.Washington, DC: U.S. Environmental ProtectionAgency, Office of Solid Waste.

105

Gillett, J.W. 1980. Terrestrial microcosm technologyin assessing fate, transport and effects of toxicchemicals. In Dynamics, Exposure and HazardAssessment of Toxic Chemicals, R. Hauge, editor.Ann Arbor Science, Ann Arbor, Ml. As quoted inTrabalka and Garten (1982).

Gupta, S.K., Cole, C.R., Kincaid, C.T., and Monti,A.M. 1987. Coupled fluid, energy, and solutetransport (CFEST) model: formulation and user’smanual. Columbus, Ohio: Office of Nuclear WasteIsolation, Battelle Memorial Institute.

Gureghian A.B. 1983. TRIPM: a two-dimensionalf in i te-e lement model for the s imul taneoustransport of water and reacting solutes throughsaturated and unsaturated porous media.Columbia, OH: Battelle Memorial Institute, Officeof Nuclear Waste Isolation. BMI/ONWI-465.

Haith, D.A. 1980. A mathematical model forestimating pesticide losses in runoff. Journal ofEnvironmental Quality. 9(3):428-433.

Hanna, S.R., Hosker, R.P., Jr. 1980. Atmosphericremoval processes for toxic chemicals. SilverSpring, MD: National Oceanic and AtmosphericAdministration. Technical Memorandum ERLARL-102.

Haynes, W.A. 1966. Guide for wind erosion control inthe northeastern United States. U.S. Department ofAgriculture. Soil Conservation Service. Washington,DC.

Hendry, D.G., Kenley, R.A. 1979. Atmosphericreac t ion p roduc ts o f o rgan ic compounds .Washington, DC: U.S. Environmental ProtectionAgency. EPA-56011 2-79-001.

Henrion, M. and Morgan, M.G. 1984. A computer aidfor risk and other policy analysis. Pittsburgh, PA:Carnegie-Mel lon Universi ty, Department ofEngineering and Public Policy.

Hilsmeier, W.F., Gifford, F.A. 1962. Graphs forestimating atmospheric diffusion. Oak Ridge, TN:Atomic Energy Commission. ORO-545. Asreviewed in: Turner DB. 1970. Workbook ofatmospheric dispersion estimates. ResearchTriangle Park, NC: U.S. Environmental ProtectionAgency, Office of Air Programs. AP-26.

Hoffman, F.O., Miller, C.W., Little, C.A. 1984.Uncertainties associated with predictions derivedfrom models and parameters. In: Models andparameters for environmental radiologicalassessments. C.W. Miller, Ed. Oak Ridge NationalLaboratory, Health and Safety Research Division.

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

Huyakorn, P.S., Ungs, M. J., Mulkey, L.A. andSud icky , S .A. 1987. A three-dimensionalanalytical 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. Methodologyfor assessing exposures to chemical substancesvia 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.,

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

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

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

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

Johnanson, R.C., Imhoff, G.C., Davis, H.H. 1984.Hydrocomp Inc. Users manual for hydrologicalsimulation 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 chemicalsubstances via the ingestion of drinking water.Washington, DC: U.S. Environmental ProtectionAgency. Contract No. 68-01-6271.

Kenaga, E.E. and Goring, C.A.I. 1978. Relationshipbetween water solubility, soil sorption, octanol-water partitioning, and concentration of chemicalsin biota. In: Eaton JG, Parrish PR, Hendricks AC,eds. Aquatic toxicology. Proc. of third annualsymposium on aquatic toxicology, New Orleans,LA, 17-18 Oct. ASTM special technical publication707.

Kenaga, E.E. 1980. Correlation of bioconcentrationfactors of chemicals in aquatic and terrestrialorganisms with their physical and chemicalproperties. Environ. Sci. Technol. 14(5): 553-556.

Kent, D.C., Pettyjohn, W.A. and Prickett, T.A. 1985.Analytical methods for the prediction of leachateplume 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 andmixing in ground water. Solid and Hazardous WaterResearch and Development Annual Symposiumproceedings. Columbus, OH: National Water WellAssociation. As reviewed in: Versar Inc. 1983.Theoretical evaluation of sites located in the zoneof saturation. Draft final report. Chicago, IL: U.S.Environmental Protection Agency. Contract No.68-01-6438.

Kinzelbach, W. 1986. Groundwater modeling: anintroduction with sample programs in BASIC.Amsterdam, The Netherlands: Elsevier SciencePublishers.

Kipp, K.L. Jr., 1987. HST3D: A computer code forsimulation 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 insimulation. New York. NY: Marcel Dekker, Inc.

Konikow, L.F., and Bredehoeft, J.D. 1978. Computermode l o f two-d imens iona l t ranspor t anddispersion in ground water. Washington, DC: U.S.Geological Survey. Techniques of Water ResourceInvestigation, Book 7, Chapter 2. As reviewed in:Versar Inc. 1983. Theoretical evaluation of siteslocated in the zone of saturation. Draft final report.Chicago, IL: U.S. Environmental ProtectionAgency. Contract No. 68-01-6438.

Kufs, C., Repa, E., Rogoshewski, P., et al. 1983.Leachate plume migration control (unpublisheddraft) . Cincinnat i , OH: U.S. EnvironmentalProtection Agency. Contract No. 68-03-3113.

Liu, H. 1977. Predicting dispersion coefficient ofstreams. J. Environmental Engineering Division.Proceedings of the American Society of CivilEngineers. Vol. 103.

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

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

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

Martinez, M.J. 1985. A finite element computerprogram for simulating radionuclide transportthrough porous media. Albuquerque, NM: SandiaNational 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. Sizeconsideration for establishing a standard for

106

inhalable particles. In Journal of the Air PollutionControl Association.

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

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

Morgan, M.G., Henrion, M., Nair, I., Mullin, T. andWiecha, C. 1984. A generic “pre-HEED” oncharacterizing and dealing with uncertainty inhealth and environmental risk assessment.Pittsburgh, PA: Carnegie-Mellon University,Department of Enginering and Public Policy.

Neely, W.B. 1982. The definition and use of mixingzones. Environ. Sci. Technol. 16(9):520A-521A.

Nkedi-Kizza, P., Rao, P.S.C. and Hornsby, A.G.1985. Influence of Organic cosolvents on sorptionof hydrophobic organic chemicals by soils.Environmental Science and Technology Vol. 19,No. 10, p. 975-979

107

Onishi, Y., Whelan, G. and Skaggs, R.L. 1982.Development of a mult imedia radionucl ideexposure assessment methodology for low-levelwaste management. Athens, GA: Off ice ofResearch and Development, U.S. EnvironmentalProtection Agency. As reviewed in: Versar Inc.1983. Methodology for assessing exposures tochemical substances via the ingestion of drinkingwater. Washington, DC: U.S. EnvironmentalProtection Agency. Contract No. 68-01-6271

Onishi, Y. and Wise, S.E. 1982a. Mathematical model,SERATRA, for sediment-contaminant transport inrivers and its application to pesticide transport inFour 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 chemicalsubstances via the ingestion of drinking water.Washington, DC: U.S. Environmental ProtectionAgency. Contract No. 68-01-6271.

Onishi, Y. and Wise, S.E. 1982b. User’s manual forthe instream sediment-contaminant transportmodel SERATRA. Athens, GA: Office of Researchand Development, U.S. Environmental ProtectionAgency. EPA-60013-82-055. As reviewed in:

Versar Inc. 1983. Methodology for assessingexposures to chemical substances via the ingestiono f d r ink ing water . Wash ing ton , DC: U.S .Environmental Protection Agency. Contract No.68 -01-6438 .

Onishi, Y. 1981. Sediment-contaminant transportmodel. Journal of the Hydraulics Division, ASCE.107(HY9):1089 - 1107. Proc. Paper 16505. Asreviewed in: Versar Inc. 1983. Methodology forassessing exposures to chemical substances viathe ingestion of drinking water. Washington, DC:U.S. Environmental Protection Agency. ContractNo. 68-01-6271.

Pasquill, F. 1961. The estimation of the dispersion ofwindborne materials. Meteorol. Mag. 90, 1063,33-49. As reviewed in: Turner DB. 1970.Workbook of atmospheric dispersion estimates.Research Triangle Park, NC: U.S. EnvironmentalProtection Agency, Office of Air Programs. AP-26.

Perry, R. and Chilton, C. 1973. Chemical engineershandbook. 5th edition. New York: McGraw-Hill.

Pierce, D. and Turner, D.B. 1980. User’s guide forMPTER. 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 forselected groundwater qual i ty evaluat ions.Champaign, IL: Illinois Department of Energy andNatural Resources. ISWS/BUL-65181. Asreviewed in: Versar Inc. 1983. Theoret icalevaluat ion of s i tes located in the zone ofsaturation. 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. Sorpt ion and transport ofhydrophobic organic chemicals in aqueous andmixed solvent systems: model development andpreliminary evaluation. Journal of EnvironmentalQuality, Vol. 14, No. 3, July-September. RawlsWJ, Brakensiek DL, Saxton KE. 1982. Estimationof Soil Water Properties. Transactions of theAmerican Society of Agricultural Engineers.25(5):1316-1320 and 1328.

Rawls, W.J. 1986. Computer printouts from the soilsdata base, dated August 28, 1986, from W.J.Rawls, Beltsville Agricultural Research Center,Beltsville, Maryland.

Reeves, M. and Cranwell, R.M. 1981. User’s manualfor the Sandia Waste-Isolation Flow Transportmodel (SWIFT). Washington, DC: Nuclear

Regulatory Commission. NUREG/CR-2324,SAND81-2516. As reviewed in: Lo TYR, ScottBH, Benjamin RR. 1983. Remedial actionassessment models for hazardous waste sites.Review draft. Athens, GA: U.S. EnvironmentalProtection Agency. Contract No. 68-03-3116.

Reeves, M., Ward, D.S., Johns, N.J. and Cranwell,R.M. 1986. The Sandia waste-isolation flow andtransport 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. andHaasbeek, J.F. 1987. BIOPLUME II; computermodel of two-dimensional contaminant transportu n d e r t h e i n f l u e n c e o f o x y g e n l i m i t e dbiodegradation in ground water. Houston, TX:National Center for Ground water Research, RiceUniversity.

Rubinstein, R.Y. 1981. Simulation and the MonteCarlo method. New York, NY: Wiley SciencePublishers.

Runchal, A., Sagar, B., Baca. R.G. and Kline, N.W.1985. PORFLO - a continuum model for fluidflow, heat transfer, and mass transport in porousmedia. Richland, WA: Rockwel l HanfordOperations. RHO-BW-CR-150P.

Russell, R.S., Bartlett, B.0. and Bruce, R.S. 1971.The significance of long-lived nuclides after anuclear war. In D.W. Bensen and A.H. Sparrow,eds. Survival of food crops and livestock in theevent of nuclear war. Oak Ridge, TN: US AtomicEnergy Commission. As quoted in Trabalka andGarten (1982).

Salame, M. (no date) Permeabi l i ty-structurerelationships of high polymers. Obtained by privatecommunication. Monsanto Co., Bloomfield, CT.

Salame, M. 1961. The prediction of liquid permeationin polyethylene and related polymers. SPE Trans.1(4):153.

Salame, M. 1973. Transport properties of nitrilepolymers. J. Polymer Sci. 41:1-15.

Salame, M. 1985. Private communication. MonsantoCo., Bloomfield, CT.

Sanford, W.E. and Konikow, L.F. 1985. A two-constituent solute transport model for groundwaterhaving variable density. Reston, VA: U.S.Geological Survey. WRI 85-4279.

Sax, N.I. 1984. Dangerous properties of industrialmaterials, 6th edition. New York, NY: VanNostrand Reinhold.

Schnoor, et al. 1987. Processes, coefficients, andmodels for simulating toxic organics and heavymetals in surface waters. Athens, GA: U.S.Environmental Protection Agency. EPA-600/3-87/015.

Schwab, G.O., Frevert, R.K., Edminster, T.W. andBarnes, K.K. 1966. Soil and water conservationengineering. 2nd edition. New York: John Wileyand Sons.

Seely, D., Turina, P., Pangaro, N., et al. 1983.Development of protocols for ambient air samplingand monitoring at hazardous waste facilities:methods summary report. Draft Report. GCACorporation. GCA/Technology Division. NewBedford, MA. Prepared for U.S. EnvironmentalProtection Agency, Office of Solid Waste, LandDisposal Branch. Contract No. 68-02-3168.

Sehmel, G.A. 1980. Particle resuspension: a review.In Environment International. Vol. 4. PergamonPress, Ltd.

Shen, T. 1981. Estimating hazardous air emissionsfrom disposal sites. Pollution Engineering 13(8):31-34.

Skidmore, E.L., Woodruff, N.P. 1968. Wind erosionforces in the United States and their use inpredicting 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 andEmergency Response at the First Public Briefingon the 1984 Amendments to the ResourceConservation and Recovery Act. December 11,1984, Washington, DC.

Smith, W.J., Whicker, F.W., Meyer, H.R. 1982.Review and categorization of saltation, suspensionand resuspension models. Nuclear Safety. 23(6).

Steingiser, S., Nemphos, S.P., Salame, M. 1978.Barrier polymers. In: Kirk-Othmer encyclopediaof chemical technology, 3rd edition. John Wileyand Sons, New York, NY.

Texas Air Control Board. 1980. User’s guide to theTexas climatological Model. Austin, TX: Texas AirControl 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, S.T. 1982. Landfarming ofpetroleum wastes - modeling the air emissionproblem. In Environmental Progress. Vol. 1, No. 1.

Thibodeaux, L.J. 1981. Estimating the air emissionsof chemicals from hazardous waste landfills. InJournal of Hazardous Materials. Vol. 4.

Trabelka, J.R., Garten, C.T., Jr. 1982. Developmentof predictive models for xenobiotic bioaccumulationin terrestrial ecosystems. Oak Ridge, TN: OakRidge National Laboratory, Environmental SciencesDivision, (NTIS DE83 003171). ORNL-5869.

Turner, D.B. 1970. Workbook of atmosphericdispersion estimates. Research Triangle Park, NC:U.S. Environmental Protection Agency, Office ofAir Programs. AP-26.

Tracy, J.V. 1982. User’s guide and documentation foradsorption and decay modifications to the USGSsolute transport model. Washington, D.C.: U.S.Nuclear Regulatory Commission. NUREGiCR-2502.

Trask, P.D., Patnode, H.W. 1942. Source beds ofpetroleum. Tulsa, OK: The American Association ofPetroelum Geologists.

Travis, B. 1984. TRACR3D: A model of flow andtransport 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 forRAM. Publication No. EPA-60018-78-016 Volsa and b. Research Triangle Park, NC: U.S.Environmental Protection Agency.

USCG. 1974. Hazardous chemical data. WashingtonDC: United States Coast Guard CG-446-2.

USDA. 1974. Department of Agriculture. Universalsoil loss equation. Agronomy technical note no.32. Portland, Oregon. U.S. Soil ConservationService. West Technical Service Center.

USDC. 1961. Rainfall frequency atlas of the UnitedStates. Washington, DC: U.S. Department ofCommerce. Weather Bureau. Technical PaperNumber 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 maintenanceplanning and analyses, Vol . 10 (Revised).Procedures for evaluating air quality impact of newstationary sources. Research Triangle Park, NC:

109

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 PublicationNo. EPA-45012-77-013.

USEPA. 1979. Environmental modeling catalogue.Washington, DC: U.S. Environmental ProtectionAgency. Contract No. 68-01-4723.

USEPA. 1980a. Hazardous waste backgrounddocument for the control of hazardous wasteleachate. Cincinnati, OH: Municipal EnvironmentalResearch Laboratory, Solid and Hazardous WasteResearch Division, Briefing document for ReviewCommittee Meeting held at the U.S. EnvironmentalProtection Agency, Washington, DC. January 23,1980.

USEPA. 1980b. Land disposal of hexachlorobenzenewastes: controlling vapor movement in soil. U.S.Environmental Protection Agency. MunicipalEnvironmental Research Laboratory. Office ofResearch and Development. Cincinnati, OH.EPA-60012-80-119.

USEPA. 1982a. Environmental modeling catalogue.Washington, DC: U.S. Environmental ProtectionAgency. Information Clearing House. PM-211 A.

USEPA. 1982b. Establishment of guidelines formodel ing groundwater contaminat ion f romhazardous waste facilities. Preliminary groundwatermodeling profile (discussion draft). Washington,D.C.: U.S. Environmental Protection Agency, Officeof Solid Waste.

USEPA. 1983a. Compilation of air pollutant emissionfactors: AP-42. Research Triangle Park, NC:U.S. Environmental Protection Agency. Office ofAir, Noise and Radiation. Office of Air QualityPlanning and Standards.

USEPA. 1983b. Methods for assessing exposure towindblown particulates. Washington, DC: U.S.Environmental Protection Agency. Office of Healthand Environmental Assessment. Off ice ofResearch and Development. EPA-60014-83-007.

USEPA. 1983c. Technical assistance document forsampling and analysis of toxic organic compoundsin ambient air. Research Triangle Park, NC: U.S.Environmental Protection Agency, EnvironmentalMonitoring Systems Laboratory. EPA-60014-83-027.

USEPA. 1985a. Guidance on remedial investigationsunder CERCLA. Washington, D.C.: U.S.Environmental Protection Agency, Office ofEmergency and Remedial Response.

Washington, D.C.: O f f i c e o f H e a l t h a n dEnvironmental Assessment, U.S. EnvironmentalProtection Agency.

USEPA. 1985b. Guidance on feasibility studies underCERCLA. Washington, D.C.: U.S. EnvironmentalProtection Agency, Office of Emergency andRemedial Response.

USEPA. 1986b. Guideline on air quality models.Research Triangle Park, N.C.: Office of Air QualityPlanning and Standards. U.S. EnvironmentalProtection Agency. EPA-450/2-78-027R.

USEPA. 1985c. Superfund public health evaluationmanual. Draft. Prepared by ICF, Inc. for the PolicyAnalysis Staff, Office of Emergency and RemedialResponse. Washington, D.C.: U.S. EnvironmentalProtection Agency. October 1, 1985.

USEPA. 1986c. Guidelines for exposure assessment.Washington, D.C.: Office of Research andDevelopment, U.S. Environmental ProtectionAgency. 51 FR 34042.

USEPA. 1985d. Methods for assessing exposure tochemical substances. Versar Inc. Washington,D.C.: U.S. Environmental Protection Agency, Officeof Toxic Substances. EPA 56015-85-001.

USEPA. 1986d. Office of Solid Waste. Hazardouswaste management system; land disposedrestrictions; proposed rule. Federal Register Vol.51, No. 9. January 14, 1986.

USEPA. 1985e. Methods for assessing exposure tochemical substances in the ambient environment.Versar Inc. Washington, D.C.: U.S. EnvironmentalProtection Agency, Office of Toxic Substances.EPA 56015-85-002.

USEPA. 1986e. Development of advisory levels forpolychlor inated biphenyls (PCB) c leanup.Washington, DC: U.S. Environmental ProtectionAgency, Exposure Assessment Group, Office ofResearch and Development. (NTIS #PB86-232774) May 1986. EPA/600/6-86-002.

USEPA. 1985f. Methods for assessing exposure fromdisposal of chemical substances. Versar Inc.Washington, D.C.: U.S. Environmental ProtectionAgency, Office of Toxic Substances. EPA 560/5-85-003.

USEPA. 1987a. Data quality objectives for remedialresponse activities. Washington, D.C.: Office ofEmergency and Remedial Response and Office ofWaste Programs Enforcement, Office of SolidW a s t e a n d R e m e d i a l R e s p o n s e , U . S .Environmental Protection Agency.

USEPA. 1985g. Methods for enumerating and USEPA. 1987b. Compendium of superfund fieldcharacterizing populations exposed to chemical operating methods. Washington, D.C.: Office ofsubstances. Versar Inc. Washington, D.C.: U.S. Solid Waste and Emergency Response, U.S.Environmental Protection Agency, Off ice of Toxic Environmental Protection Agency. EPA-540/P-Substances. EPA 560/5-85-004. 87/001. OSWER Directive 9355.0-14.

USEPA. 1985h. Methodology for assessing exposuresto chemical substances via the ingestion ofdrinking water. Versar Inc. Washington, D.C.: U.S.Environmental Protection Agency, Off ice of ToxicSubstances. EPA 560/5-85-005.

USEPA. 1985i. Methods for assessing environmentalpathways of food contamination. Versar Inc.Washington, D.C.: U.S. Environmental ProtectionAgency, Office of Toxic Substances. EPA 560/5-85-008.

USEPA. 1987c. Proposed criteria for selection ofgroundwater exposure assessment models.Washington, D.C.: O f f i c e o f H e a l t h a n dEnvironmental Assessment. U.S. EnvironmentalProtection Agency. OHEA-E-219.

USEPA. 1987d. Mathematical model selection criteriafor performing exposure assessments: surfacewater models. Washington, D.C.: Office of Healtha n d E n v i r o n m e n t a l A s s e s s m e n t . U.S.Environmental Protection Agency. OHEA-E-245.

USEPA. 1985j. Modeling remedial actions atuncontrolled hazardous waste sites. Cincinnati OH:Hazardous Waste Eng ineer ing ResearchLaboratory, Gffice of Research and Development,U . S . E n v i r o n m e n t a l P r o t e c t i o n A g e n c y .EPA/540/2-85/001.

USEPA. 1987e. R isk Communica t ion , R iskAssessment, Management Communication, Guideto Selected Sources. USEPA Office of InformationR e s o u r c e M a n a g e m e n t , O f f i c e o f T o x i cSubstances, March 1987. EPA-IMSD/87/002.

USEPA. 1986a. Mathematical model selection criteriafor performing exposure assessments: groundwatercontaminants from hazardous waste facilities.

van der Heijde, Paul K.M., Bachmat, Y., Bredehoeft,J., Andrews, B., Holtz, D., Sebastian, S. 1985.Groundwater management: The use of numericalmodels. Washington, DC: American Geophysical

110

Union, Water Resources Monograph #5, 2ndEdition.

van der Heijde, P.K.M. and Srinivasan, P. 1986. A“random-walk” model for solute transport ingroundwater ‘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. Modelassessment for delineating well head protectionareas. Washington, DC: Office of GroundwaterProtection, U.S. Environmental Protection Agency.International Groundwater Modeling Center,Holcomb Research Institute. Butler University,Indianapolis, Indiana.

van Genuchten, M.T. 1978. Mass transport insaturated-unsaturated media: one-dimen-sional solutions. Princeton, NJ: Department of CivilEngineering. Princeton University. 78/WR-11.

van Genuchten, M.T. and Alves, W.J. 1982.Analytical solutions of the one-dimentionalconvective-dispersive solute transport equation.Riverside, CA: U.S. Salinity Laboratory, AgriculturalResearch Service, U.S. Department of Agriculture.Technical Bulletin 1661.

Veith, G.D., DeFoe, D.L., Bergstedt, B.V. 1980.Measuring and estimating the bioconcentrationfactor of chemicals in fish. J. Fish Res Board Can.36: 1040-1048.

Versar. 1983. Theoretical evaluation of sites locatedin the zone of saturation. Draft final report. Versar,Inc. Chicago IL: U.S. Environmental ProtectionAgency. Contract No. 68-01-6438.

Verschueren, K. 1984. Handbook of environmentaldata on organic chemicals. New York: VanNostrand/Reinhold Press.

Voss, C.I. 1984. SUTRA: A finite element simulationmodel for saturated-unsaturated fluid density-dependent ground water flow with energy transportor chemical react ive single species solutetransport. Reston, VA: U.S. Geological Survey,Water Resources Investigation. 84-4369.

Walton, W.C. 1984. Handbook of analytical groundwater models. International Ground WaterModeling Center, Holcomb Research Institute,Butler University, Indianapolis, Indiana.

Walton, W.C. 1985. Thirty-five BASIC groundwaterp r o g r a m s f o r d e s k t o p m i c r o c o m p u t e r s‘WALTON84-35BASIC’. Ind ianapo l is , IN:International Ground Water Modeling Center,Holcomb Research Institute, Butler University.

Williams, J.R. 1975. Sediment-yield prediction withthe universal equation using runoff energy factor.In Present and prospect ive technology forpredicting sediment yields and sources. U.S.Department of Agriculture. ARS-S-40.

Wilson, J.L., Miller, P. J. 1978. Two-dimensionalplume in uniform ground-water flow. Journal ofthe Hydraulics Division, ASCE 104(4): 503-514.

Wischmeier, W.H., Smith, D.D. 1978. Predictingrainfall erosion losses - a guide to conservationplanning. Washington, DC: U.S. Department ofAgriculture. Agriculture Handbook No. 537.

Wischmeier, W.H. 1972. Estimating the cover andmanagement factor on undisturbed areas. U.S.D e p a r t m e n t o f A g r i c u l t u r e . O x f o r d , M S :Proceedings of the USDA Sediment YieldWorkshop.

Woodburn, K.B., Rao, P.S.C., Fukui, M., Nkedi-Kizza, P. 1986. Solvophobic approach forpredict ing sorpt ion of hydrophobic organicchemicals on synthetic sorbents and soils. J. ofContam. Hydrology. 1: 227-241.

Yeh, G.T. and Huff, D.D. 1985. FEMA: A finiteelement model of material transport throughaquifers. Oak Ridge, TN: Oak Ridge NationalLaboratory, ORNL-6063.

Yeh, G.T. 1981. AT123D. Analytical transient one-,two-, and three-dimensional simulation of wastetransport in the aquifer system. Oak Ridge, TN:Oak Ridge National Laboratory, EnvironmentalSciences Division Publication No. 1439. ORNL-

Yeh, G.T. 1982. CHNTRN: a chemical transportmodel for simulating sediment and chemicaldistribution in a stream/river network. Washington,DC: Office of Pesticides and Toxic Substances,U.S. Environmental Protection Agency. ContractNo. W-7405-eng-26. As reviewed in: Versar1983. Methodology for assessing exposures tochemical substances via the ingestion of drinkingwater. Washington, DC: U.S. EnvironmentalProtection Agency. Contract No. 68-01-6271.

Yeh, G.T. 1987. FEMWATER: A finite element modelof water flow through saturated-unsaturatedporous media - first revision. Oak Ridge, TN: OakRidge Natational Laboratory, ORNL-5567/RI.

Yeh, G.T., Ward, D.S. 1981. FEMWASTE: A finite-element model of waste t ransport throughsaturated-unsaturated porous media. Oak RidgeNational Laboratory, Environmental Services

111

Division: Publication No. 1462, ORNL-5602. 137p. As reviewed in: Versar Inc. 1983. Theoreticalevaluat ion of s i tes located in the zone ofsaturation. 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 lntegration

Table of Contents

Chapter Page

1 QUANTITATIVE ANALYSIS OF EXPOSED POPULATIONS . . . . . . . . . . . . . . . . . . . . . . .

1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1.2 Exposed Populations Screening . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1.3 Quantitative Exposed Populations Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1.4 Identification and Enumeration of Exposed Human Populations . . . . . . . . . . . . . . . .

1.4.1 Populations Exposed Through Air . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1.4.2 Populations Exposed Through Surface Water or Ground Water . . . . . . . . . .1.4.3 Populations Exposed Through Food . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1.4.4 Populations Exposed Through Soil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1.5 Population Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1.6 Activity Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 EXPOSURE CALCULATION AND INTEGRATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.1 Inhalation Exposure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2.2 Dermal Exposure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2.3 Ingestion Exposure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.3.1 Food/Soil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2.3.2 Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.4 Exposure Integration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3 APPENDIX A REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

114

114114116118118119119120120120

121

122123128128128129

131

113

Chapter 1Quantitative Analysis of Exposed Populations

1.1 Introduction

The results of contaminant release and fateanalyses provide the basis for assessing exposedpopulations. This assessment compares environ-mental contamination data with populations data todetermine the likelihood of human contact withcontaminants of concern. This chapter detailsmethods use fu l i n eva lua t ing the fo l low ingcomponents of exposed populations analysis:

1. Identification and enumeration of exposedpopulations;

2. Characterization of exposed populations; and

3. Analysis of activities that bring populationsinto contact with contaminants.

Each of these components is detailed in the followingsubsections.

As with other evaluations, exposed populationsanalysis begins with a screening assessment, whichidentifies exposure pathways that are incomplete, i.e.,those situations where contaminants are released andmigrate from a site, but do not contact humanpopulations and are not likely to do so in the future.Such situations require no further analysis. At thesame time, exposed populations screening also pointsout those exposure pathways that are complete andthat will require quantitative analysis to estimate theextent of human exposure.

Data needed to quanti fy potent ial ly exposedpopulations are readily available. In essence, allquantitative exposed populations evaluations can beconsidered in-depth analyses. For each populationsegment identified in this portion of the exposureassessment process, exposures are quantified andintegrated as described in Chapter 2 of this Appendix.

1.2 Exposed Populations ScreeningExposed populations screening is primarily qualitative.This evaluation draws on the results of contaminantfate analysis (presented in Chapter 3) to determine

the likelihood and extent of human population contactwith contaminants.

Exposed populations screening is guided by thedecision network provided in Figure A-l. Thefollowing numbered paragraphs each refer toparticular numbered boxes in the figure.

1. Human exposure through inhalation should beevaluated for contaminants that have migrated or maymigrate from the site into air. The assessment shouldconsider both contaminated dust and volatilecompounds. For screening purposes, comparingcontaminant concentration isopleths with maps of thelocal area will identify the potential for such humanpopulation inhalation exposure. The user shouldrealize, however, that exposure can occur inrecreat ional areas as wel l as in resident ia l ,commercial, or industrial areas, and should interpretlocal area maps accordingly.

2. In cases where surface waterbodies have beencontaminated by toxics migrating from a site, thewater’s potential commercial use as a fish or shellfishsource should be evaluated. If the waters arecommercially fished, fishermen may be exposedthrough dermal contact with contaminated water,a l t h o u g h s u c h e x p o s u r e w i l l g e n e r a l l y b eovershadowed by other exposure mechanisms.

3. In cases where recreationally or commerciallycaught fish/shellfish are taken from contaminatedwaters, persons consuming the catch may beexposed. For chemicals that tend to bioaccumulate,consumers may be exposed to contaminantconcentrations in fish/shellfish tissue that are manytimes greater than those present in the water columnor sediments. When performing exposed populationsscreening, the analyst need only determine whetherwaters identified in the environmental fate analysis ashaving received contaminants from the hazardouswaste site are used commercially or recreationally.

4. Individuals who swim in contaminated waters canexperience dermal exposure to toxics over their entirebody. In add i t ion , s ign i f i cant quant i t ies o fcontaminated water may be ingested inadvertentlywhile swimming, and swimmers will be exposed to

114

volatile contaminants in the water through inhalation.Other screening should evaluate the existing orpotential degree to which the local population usescontaminated water-bodies (fresh or marine) forswimming.

5. If contaminated ground water or surface water isa source of potable water, the population served mayexperience considerable ingestion exposure. Similarly,the population may also be exposed to toxics throughboth dermal absorption and inhalation (of volatiles)while showering or bathing. When undertaking ascreening analysis, it is only necessary to determinewhich residences or commercial/institutional estab-lishments are likely to obtain their potable water fromcontaminated water sources.

6. If contaminants migrate to off-site soils, personscontacting such soil may be exposed. Individuals whogrow their own fruit or vegetables at home mayexperience additional exposure from ingesting foodgrown in contaminated soils, as do those consumingcontaminated commercially-grown foods. Similarly,l ivestock that have grazed on contaminatedvegetation may constitute a source of ingestionexposure for consumers. Screening analysis shouldstrive to correlate areas of human habitation withareas of contaminated soil, as defined in theenvironmental fate analysis.

7. Similarly, if direct access to the site is possible,children may be attracted to the location and maydirectly contact hazardous materials or contaminatedsoil. Such activity may result in inhalation or dermalexposure, as well as intentional or inadvertentingestion of contaminated soil. For screeningpurposes, the proximity of residential areas to the siteshould indicate the potential for direct access bychildren.

1.3 Quantitative Exposed PopulationsAnalysis

Quantitative analyses of potentially exposed humanpopulations comprises three distinct steps, which areillustrated in Figure A-2. First, the results ofenvironmental fate analysis are compared with dataiden t i f y ing and enumera t ing nearby humanpopulations to provide boundaries and quantify thepopulation(s) potentially or actually coming intocontact with contaminated air, water, and soil.Populations consuming contaminated food (homegrown vegetables, fish) can similarly be identifiedonce the areal extent of contamination is known.

Population characterization, the second step, involvesidentifying those groups within the exposed populationthat, because of the specific health effects of somepollutants or factors related to the population itself,would experience a higher risk than would theaverage population as a result of a given level of

exposure . Indeed, the hea l th e f fec ts o f thecontaminants under evaluation will often dictate theneed for population characterization. For example, ifmutagenic or teratogenic substances are involved,women of childbearing age should be considered ahigh-risk group. In addition, factors relating to theexposed population may cause certain groups toconstitute high-risk subpopulations. These include:

Persons with a genetic predisposition to certainhealth effects;

Persons whose health or resistance to disease isimpaired by behavioral factors such as smoking,use of alcohol or drugs, etc;

Infants, children, and the elderly, who are moresusceptible to health impacts from a givenexposure than are persons of other ages;

Persons who are already suffering from diseaseand may be more suscep t ib le to fu r the rimpairment as a result of a given level ofexposure than are healthy persons;

Persons who are exposed to naturally highbackground levels of contaminants (e.g., seleniumor arsenic) and may be at greater risk to smallincremental increases of hazardous substancesthan are persons who are not exposed to suchbackground levels; and

Nutritionally deficient populations who may beless resistant to exposure than those withadequate diet.

While most studies will consider only the exposedpopulation as a whole and not as separate discretesubpopulations, in certain cases, such detailedpopulation analysis may be warranted for in-depthstudies.

Age and sex influence the average inhalation rate, therate of food and water intake, the body area subjectto dermal exposure, and the types of food consumed,all of which can affect the level of exposure actuallyexperienced. Some quantitative assessments mayrequire further characterization of populations todetermine age- and sex-specific exposure factors.

The third step is activity analysis. Once populationidentification and characterization have answered thequestion “Who may be exposed?“, the question“How and to what level are component portions ofthis population exposed?” may next be asked in orderto refine the evaluation. This refinement involvesdetermining the exposed population’s activities.Comprehensive analysis can encompass the range ofindoor, outdoor, and in-car activities. For SuperfundFeasibility Studies, however, average values foractivity-related considerations usually suffice.

116

The activity analysis can also help to identify high-risk groups. For example, those groups that mayexperience a significantly higher frequency or durationof exposure as compared with the general populationcan also be considered high-risk groups.

1.4 Identification and Enumeration ofExposed Human PopulationsThe major population data base that can be accessedto determine the size, distribution, and demographiccharacteristics of a geographically defined populationis the Census of Population.

The data collected in the Census are organizedaccording to geographic areas. Within these areas,data are further broken down into Census-definedstatistical areas and government units. Populationdata are available within Standard MetropolitanStatistical Areas (SMSAs) down to the level of the“block” and in non-SMSAs to the level of theEnumeration District (ED).

These data are especially useful in quantifying andcharacterizing populations exposed as a result of theirpresence in a specific locale (e.g., those exposed totoxics in ambient air or soil). An isopleth map ofvarying concentrations around a source can beoverlaid with Census maps. Such maps are availablefor areas within SMSAs and can be purchased fromthe Bureau of the Census. Also, Census Tracts(Series PHC80-2) contains detailed characteristicsof the population (e.g., age, sex, race, education)within each tract, a division of an SMSA containing4,000 residents each. Census Tracts is currentlyavailable on microfiche by SMSA and on computertape.

Many Super-fund sites are not within SMSAs. Censusdata for non-SMSA areas are not available on maps,but can be transcribed from Census publications.

The most useful Census publications for this type ofdata are Number of inhabitants (Series PC80-1-A)and General Population Characteristics (SeriesPC80-1-B). Each series is currently available andconsists of a separate volume for each state, togetherwith a nat ional summary volume. Number ofinhabitants provides only population counts, with nodemographic data. It provides data down to the levelof county subdivision and incorporated town. GeneralPopulation Characteristics provides population countsby age, sex, and other demographic data, andcontains data down to the level of small towns (1,000or more inhabitants).

All printed Census information is available forpurchase through the Government Printing Office(GPO); all series issued on microfiche, maps,computer tapes, and technical documentation areavailable directly from the Customer Services Branch

at the Bureau of the Census, Department ofCommerce, Washington, D.C., and can be ordered bycalling (202) 763-4100. Alternatively, it may be moreconvenient to contact one of the Census Bureauregional offices. Cities where such offices are locatedand phone numbers for the public information servicewithin each regional office are listed in Table A-l.

Table A-l. Regional Census BureauOffices

Atlanta. Ga. (404) 881-2274

Boston,Mass. (617) 223-0226

Charlotte, N.C. (704) 371-6144

Chicago, IL. (312) 353-0980

Dallas, Tex. (214) 767-0625

Denver,Colo. (303) 234-5825

Detroit, Mich. (313) 226-4675

Kansas City, Kans. (913) 236-3731

Los Angeles, Calif. (213) 209-6612New York, N.Y. (212) 264-4730

Philadelphla, Pa. (215) 597-8313

Seattle, Wash. (206) 442-7080

7.4.7 Populations Exposed through AirA convenient means of accessing quantitativepopulation data for a specific area impacted by aircontaminants is to directly link environmental fate andexposed populations analysis through use of anintegrated computer-based fate model , andpopulation data retrieval program called ATM-SECPOP. Developed by the EPA Office of ToxicSubstances, Exposure Evaluation Division (OTS-EED), this model primarily analyzes point sourceemissions, but can also be adapted to area or linesource analyses. ATM-SECPOP integrates theoutput of a concentration prediction model (ATM)(Patterson et al. 1982); a population distribution database (the proprietary 1980 Census Master AreaReference File (MARF)), which is accessed via apopulation distribution model called SECPOP; andgraphic and mapping information displays. Thisintegration affords a rapid and efficient means ofgenerating and presenting exposure data relating tothe airborne release of chemical substances. Thegraphic display functions can be used to illustrate therelationship of variables such as the distribution ofexposure or concentration versus distance for any orall directions around a facility. Graphic displays maybe in the form of bar charts, scatter plots, rosediagrams, or maps. Because of the proprietary natureof the data contained in MARF, ATM-SECPOP’s useis restricted to personnel and contractors of EPA,Office of Toxic Substances (EPA-OTS). Specialarrangements can be made for others to use thedata. Inquiries should be directed to the Modeling

118

Section of the Exposure Assessment Branch ofEPA-OTS in Washington, D.C. A detailed discussionof ATM is presented in Chapter 3 of this manual.

Where sites are accessible, the possibility thatchildren may enter and explore or play on the siteshou ld be eva lua ted . On-s i te , ch i ld ren mayexperience inhalation exposure to contaminated dust,volatiles, or both. In some cases, the site boundarymay adjoin residential properties, and the area ofcontamination may actually include such residences.Accurate estimation of the potentially exposedpopulation in such a case is difficult; it can beassumed that each household with children in theimmediate vicinity of the site has one child who mayfind the site inviting. This should provide an upperbound estimate on the actual number of children whomay 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 beapplied to the total number of local households toenumerate those in the area with children. Theanalyst must decide which households are closeenough to the site to be considered.

Similarly, workers conducting activities at the site mayalso experience inhalation exposure. Local authorities(e.g., Zoning Board) may be able to supplyinformation on the likelihood of on-site work-relatedactivities that can be used to estimate the number ofworkers who may become exposed. Remediationworkers are not included in this estimated exposedpopulation.

7.4.2 Populations Exposed through Surface Wateror Ground WaterEnvironmental fate analysis results can be used toiden t i f y geograph ica l l y -de f ined sources o frecreational (aquatic) dermal exposure, such as riverreaches downstream of an uncontrolled hazardoussite. The exposed population comprises swimmers inthose specific contaminated waters. The localgovernment agency concerned with recreation shouldbe able to provide estimates of the populationsswimming in local waters; this will usually be thestate, city, or county Department of Parks orRecreation. Alternatively, one can use the followingnational average value from the Bureau of OutdoorRecreation (USDOI 1973): 34 percent of the totalpopulation swims outdoors in natural surfacewaterbodies (including oceans, lakes, creeks, andrivers).

All persons served by a water supply system thatdraws water from a contaminated water source mustbe considered as potentially exposed throughingestion and dermal exposure while bathing.Information concerning local surface drinking watersources and populations served can be obtained fromthe local Department of Public Works, Planning

Department, or Health Department. These sourcesshould be able to provide information on publicdepartments or private drinking water treatmentcompanies that use ground water as their raw watersupply, and also on the number of householdsdrawing water from private wells.

1.4.3 Populations Exposed through FoodExposure to contaminated food will usually beassociated with fruit and vegetables grown in homegardens or with game res id ing in or us ingcontaminated areas. In order to identify the number ofpersons consuming contaminated home grown fruitand vegetables, first consult General PopulationCharacteristics, Series PC80-1-B to learn the totalnumber of households in a given geographic area.Then the data presented in Table A-2, which provideestimates of the percent of households in urban andrural areas that have fruit and vegetable gardens andthe average number of persons per household, canbe applied to the local population data to estimate thenumber of persons likely to consume contaminatedhome grown produce.

The USDA Food Consumption of Households reportseries can be consulted to estimate the localpopulation using a given food item for urban, ruralnon-farm, and rural farm locales. These reportspresent seasonal food use survey data on thefollowing bases: Northeast (USDA 1983a), NorthCentral (USDA 1983b), South (USDA 1983c), andWest (USDA 19834). More aggregated data are alsoprovided for the entire United States in a companionreport (USDA 1983e). The percent of householdsusing a given food item can be obtained from thesereports. The product of this value and the totalresident population of an area is an estimate of thelocal exposed population. Similar national level dataare also provided on the basis of age and sex in Foodand Nutrient Intakes of Individuals in 1 Day in theUnited States (USDA 1980). In addition, the U.S.Food and Drug Administrat ion (FDA) can becontacted for data concerning daily intakes of variousfood items. Such data have been compiled for theFDA Total Diet Study (Pennington 1983).

Table A-2. U.S. Home Fruit and Vegetable GardenUse, 1977

Percent ofhouseholds Household Percent of

with size (no. of total U.S.Urbanization gardens persons) populationUrban 43 3.17 32

Rural non-farm 41 3.44 9

Rural farm 84 3.86 3

Source: USEPA 1980.

119

Monitoring data may indicate whether fish and gameare contaminated in the subject area. One canestimate the fishing population by contacting the localagency responsible for issuing fishing licenses; thismay be the state fish and game commission or thestate department of natural resources. Since there are2.69 persons in the average household (Bureau of theCensus 1986) one can estimate the actual exposedpopulation by multiplying 2.69 by the number oflicensed hunters or fishermen in the area.

1.4.4 Populations Exposed through SoilExposure to contaminated soil constitutes a potentialexposure route for workers or children playingoutdoors. Neighborhood children playing at the sitecan be exposed to high levels of contaminants. Soil-related exposure in such cases would be throughdirect dermal contact with the contaminated soil.Another potentially significant, but infrequentlyencountered, exposure mechanism involves childrenwho eat dirt; this eating behavior, known as pica, maylead to their actually ingesting contaminated soil.Hand-to-mouth contact during normal play is amore common means of ingesting soil, however. Forany site located near residential areas, the degree ofaccessibility to children should be considered. Bureauof the Census data can be used as described inSection A-1.4.1 to estimate the number of localchildren who may have access to the site.

In addition, workers conducting activities at the site(other than remediation) may have direct dermalcontact with contaminated soils. Section A-1.4.1provides general guidance to identify and enumerateexposed worker populations.

1.5 Population Characterization

After exposed populations have been identified andenumerated, they can be characterized by age andsex factors. The physiological parameters thatdetermine the dose received per a given level ofexposure (e.g., breathing rate, skin surface area, andingestion rate) are often age- or sex-specific. Also,from a toxicity standpoint, subpopulations defined byage or sex, such as the elderly or women ofchildbearing age, may be especially susceptible to achemical substance. Superfund studies will generallyuse average values, but by characterizing exposedpopulations, one can determine exposure distributionswithin the population at large and delineate specifichigh-risk subpopulations.

The Census Publication series General PopulationCharacteristics (PC80-1-B) cites figures for theage and sex structure of the population residing in aspecific area. Separate volumes for each statecontain age and sex breakdowns at the level ofcounty subdivisions and small towns. If more detail isrequired, the Census Bureau microfiches containing

this information at the Census tract level (onlyavailable by SMSAs).

In the case of exposure resulting from ingestion offood, the food consumption surveys of the USDA(1983a-e) record age and sex data for the sampledpopulation. These data are contained in five separateregional reports; the appropriate one should beconsulted.

In lieu of obtaining site-specific data, one can usethe population characteristics of the U.S. as a whole,provided in the yearly Statistical Abstract of theUnited States (for example, see Bureau of theCensus 1986), to approximate the populationdistribution in the area of concern.

1.6 Activity Analysis

Activities engaged in by members of a givenpopulation or subpopulation can dramatically affectthe level of human exposure to environmentalcontaminants. For example, persons whose lifestyleor employment involves frequent strenuous activitywill inhale larger volumes of air per unit time than willthose living a less strenuous life, and will experiencea higher level of exposure to airborne contaminants.

Activity analysis allows refinement of certainparameters used in the calculation of exposure,including:

?? Inhalation rate;

? Frequency of exposure; and

?? Duration of exposure.

The procedure for integrating activity-relatedinhalation, frequency, and duration data into theexposure assessment process is detailed in thefollowing chapter.

120

Chapter 2Exposure Calculation and Integration

This chapter provides guidance for calculating andintegrating exposures to all populations affected bythe various exposure routes associated with a givenuncontrolled hazardous waste site. Specifically,guidance is provided to estimate exposure from:

The goal of this analysis is to quantify the amount ofcontaminant contacted within a given time interval.

1. Inhalationa. Ambient airb. Indoor air (contaminants released during

showering)

2. Dermal contacta. Water (swimming)b. Soil

3. Ingestiona. Foodb. WaterC. Soil

Short-term and long-term exposures are calculatedin the same manner. First, for each exposure scenariounder consideration, an exposure per event isestimated. This exposure value quantifies the amountof contaminant contacted during each exposureevent, with “event” being defined differentlydepending on the nature of the scenario underconsideration (e.g., each day spent swimming in acontaminated river is a single swimming exposureevent, each day’s inhalation of contaminated air is aninhalation exposure event). Event-based exposureestimates take into account the concentration ofcontaminant in the medium through which exposureoccurs, the rate of contact with such media(inhalation rate, ingestion rate, etc.), and the durationof each event.

This analysis is based on the results of all previousanalyses, and is the final stage of the exposureassessment. This guidance is complete; no additionaldocumentation is required to finish the analysis.

Integrated exposure analysis is conducted for onlythose contaminants having complete exposurepathways (i.e., those contaminants that are releasedand migrate from the site and that do contact humanpopulations). Therefore, no screening evaluation isincluded in the exposure integration process. Whilecalculating the exposure incurred is traditionally thefinal step in the quantitative exposure assessmentprocess, it can also be viewed as a component of thehuman health risk assessment. Therefore, thematerial detailed in this chapter is also discussed inthe Superfund Public Health Evaluation Manual(USEPA 1985).

The analyst can convert event-based exposurevalues to final exposure values by multiplying theexposure per event by the frequency of exposureevents over the timeframe being considered. Short-term exposure is based on the number of exposureevents that occur during the short-term timeframe(10 to 90 days), while long-term exposures arebased on the number of events that occur within anassumed 70-year lifetime. The 70-year assumedaverage lifetime is traditionally used in exposureassessments, and it provides a conservative upperbound of lifetime exposure. Certain exposurescenarios, however, may only apply to short-termexposure. Whenever practical, the analyst shouldstrive to determine the timeframe over which a givenexposure pathway would be expected to affect theexposed population. Once determined, the timeframewill indicate whether that pathway should beevaluated on a short- or a long-term basis.

Exposure is defined as the amount of pollutant Exposure estimates are expressed in terms of masscontact ing body boundaries (skin, lungs, or of contaminant/unit of body mass/day by dividing dailygastrointestinal tract). Exposure calculation considers exposure by the value for total body mass of anhow often populations come into contact with average individual in the exposed population. Forcontaminants in specific environmental media, the Superfund studies, an average adult body mass of 70mode o f such con tac t , and the amoun t o f kg will usually be adequate for this conversion. Incontaminated medium that contacts the internal or cases where exposure to specified subpopulationsexternal body surface during each exposure event. must be evaluated, values for other than average

121

adults may be required. Consult Anderson et al.(1984) to obtain alternative body mass values.Similar ly, average values for act iv i ty-relatedparameters (e.g., inhalation rate) generally will beadequate for Superfund site evaluations. For specialsituations and detailed exposure analysis, analystscan refer to the discussion of activity data in Freed etal. (1985). An exposure factors handbook is currentlyunder development (USEPA 1987), and the analystperforming exposure assessments after publication ofthis manual should consult that document for themost up-to-date exposure factors.

The following sections address the exposurecalculation process specific to each exposuremechanism. Data management forms designed toorganize and tabulate the data in the exposurecalculation process are presented in Appendix C.

2.1 Inhalation Exposure

Inhalation exposure per event is estimated based onthe hours per event, the inhalation rate of theexposed individual during the event, and theconcentration of contaminant in the air breathed. Theformula for calculating event-based exposure is thefollowing:

(A-1 )

Short-term exposure is calculated using the short-term contaminant air concentration, and long-termexposure is based on the long-term concentration.

Inhalation exposures are keyed to geographiclocations delineated during the Environmental FateAnalysis. Ambient concentration is generally assumedto be homogeneous throughout a limited area orsector (within an isopleth); however, this assumptionis not always well-founded. Numerous studies haveshown that there can be marked differences in indoorand outdoor concentrations of pollutants (Budiansky1980, Moschandreas et al . 1978) or amongmicroenvironments in the same area (Ott 1981). To

account for these differences when calculatingexposure, several investigations have coined the term“microenvironment,” which refers to a type ofphysical setting where concentrations of pollutantscan be expected to be similar. For Superfund studies,it is usually unnecessary to disaggregate analysis ona microenvironment basis. Instead, it can generally beassumed that contaminants have been present longenough for indoor-to-outdoor concentrations tohave reached equilibrium.

To calculate exposure duration, the analyst considersthe amount of time exposed persons actually spend inthe contaminated area. For example, if a site is in aresidential area, one can conservatively estimateexposure by assuming that all residents spend theentire 24-hour day within the contaminated zone. If asite is located in an industrialized area, it may bemore appropriate to base duration on an 8-hourworkday, if it can be reasonably assumed thatworkers do no t a lso l i ve in the immed ia teindustrialized area. Such factors must be evaluatedon a case-by-case basis. For inhalation exposure,frequency is assumed to be daily.

For a general application, use an average adult valuefor inhalation rate. An example of an adult averagederived from experimental results (USEPA 1981) is aninhalation rate of 1 m3/hour. This value can be usedto conservatively estimate exposure regardless ofmicroenvironments or activity.

Generating time-weighted average inhalation ratesprovides a more precise estimate of inhalation rate.This calculation is based on microenvironment-related data and activity stress levels/ventilation ratesassociated with the individual microenvironment. Ifthis level of detail is warranted, the inhalation ratespresented in Table A-3 can be used. Freed et al.(1985) cite directions for developing time-weightedaverage inhalation rates.

To calculate ambient inhalation exposure, one shouldobtain contaminant air concentration values from theresults of the environmental fate analysis. In onecase, however, concentration values will have to becalculated in the exposure integration stage of theexposure assessment. Persons showering or bathingin potable water contaminated with toxics may beexposed through inhalation if the contaminants arevolatile. This is especially true of showering, since thehigh turbulence, combined with the elevatedtemperature of the shower water, can produce asignificant release of volatile components.

Various approaches are available to estimatecon taminan t concen t ra t ions indoors . Theseapproaches depend on a number of factors, includingthe room air volume, air exchange and mixing factors,contaminant concentration in the water, the amount ofwater used, and the manner in which a contaminant

122

Table A-3. Summary of Human inhalation Rates forMen, Women and Children by Activity Level(m3/hour)a

Restingb Lightc Moderated Heavye

Adult male 0.6 1.3 2.6 7.1

Adult female 0.6 1.3 2.4 4.9

Average adultf 0.6 1.3 2.6 6.0

Child, age 6 0.4 1.4 2.1 2.4

Child, age 10 0.4 1.7 3.3 4.2

aValues of inhalation rates for males, females, and childrenpresented in this table represent the midpoint of ranges ofvalues reported for each activity level in Anderson et al. (1984)

blncludes watching television, reading, and sleeping.clncludes most domestic work, attending to personal needs and

care, hobbies, and conducting minor indoor repairs and homeimprovements.

dlncludes heavy indoor cleanup and performance of majorindoor repairs and alterations and climbing stairs.

elncludes vigorous physical exercise and climbing stairs carryinga load.

fDerived by taking the mean of the adult male and adult femalevalues for each activity level. A representative 24-hourbreathing rate for an average adult is 1.1 m3/hour. This valueIS based on the assumption that the average adult spends 93.2percent of the time at the light/resting level of activity, 5.8percent at a moderate level of activity, and 0.9 percent at aheavy level of activity. Values for the percent of time spent ateach activity level are from Freed et al. (1985).

is released into room air (instantaneously, con-tinuously, time-dependent). If showering/bathingexposure estimation is required for a Superfundexposure assessment, the analyst is referred toVersar (1984) for a detailed discussion of techniquesto estimate indoor air contaminant concentration. Forboth showers and baths, the analyst should assume acontinuous contaminant release during the bathing/showering period. Values for the other variable factorsmentioned above can be obtained from Versar(1985).

To evaluate inhalation exposure to contaminantsvolatilizing from potable water while showering, theanalyst should again assume frequency to be daily.Each shower is assumed to last 15 minutes.

Inhalation exposure to swimmers can be based onmonitored or estimated ambient air concentrationsabove a contaminated water body. To estimateconcentrations, calculate the rate of volatilization ofthe contaminant from the water body and use thisvalue as the input to a “box model” air migrationmodel. The dynamic release rate can be calculatedusing Equations 2-10, 2-15, 2-16, and 2-17. Therecommended air model is BOXMOD (in EPA’sGEMS system, see Chapter 3).

2.2 Dermal Exposure

Dermal exposure is determined by the concentrationof hazardous substance in a contaminated medium

that is contacted, the extent of contact (i.e., the bodysurface area contacted), and the duration of suchcontact. For exposure to contaminated water, dermalexposure per event is calculated as follows:

DEX=t exAVxCxPCxFxl liter/l000 cm 3

When possible, it is important to consider the degreeto which a given contaminant is actually able to enterthe body. Some compounds will not readily penetratethe skin, while others may do so at a rapid rate. Theabove equation can only be used in cases wheredermal permeability constants for the contaminant(s)of concern are known. Table A-4 lists dermalpermeability constants for selected compounds. Formany compounds, however, dermal permeabilityconstants will not be available. In such cases, theanalyst must assume that contaminants are carriedthrough the skin as a solute in water which isabsorbed (rather than being preferentially absorbedindependently of the water), and that the contaminantconcentration in the water being absorbed is equal tothe ambient concentration. Thus, the permeation rateof water across the skin boundary is assumed to bethe factor controlling the contaminant absorption rate.Short-term dermal exposure per event is calculatedusing the short-term contaminant concentrations inwater or soil, and long-term exposure is based onthe long-term contaminant concentrations.

The local recreation department may have detaileddata quantifying the duration and frequency of wateruse for swimming. When such locale-specific dataare not available, the following national averagefigures, based on data from the Bureau of OutdoorRecreation (USDOI 1973), can be applied:

? Frequency of exposure = 7 days/year.

123

Table A-4. Permeability Constants for Various CompoundsPermeability

Compound constanta (cm//hr) Reference

SURFACTANTS

Decanoic acid

Dodecanoic acid

Tetradecanoic acid

Hexadecanoic acid

Octadecanoic acid

Sodium dodecyl sulfate

Sodium dodecyl isothionate

Sodium p-1-dodecylbenzenesulphonate

Sodium laurate

1.00E-03

2.00E-03

6.00E-04

1.20E-05

6.00E-06

2.00E-03

5.40E-05

6.00E-06

1.00E-03

IONS

Aluminum

Potassium

Bromide

Palmitate

Laurate

7.20E-06

6.70E-05

1.80E-05

4.20E-05

3.00E-03

DRUGS

Methotrexate 6.00E-10

Benzoyl peroxide 5.10E-07

Estradiol 3.90E-03

Amphetamine 1.40E-05

Ouabain 3.90E-06

Burimamide 1.70E-07

Metramide 1.10E-07

Cimetidine 3.30E-07

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

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

3.31E-01

3.49E-01

3.60E-01

3.60E-01

3.60E-01

5.28E-01

5.50E-01

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

Nacht et 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. 1977

Roberts et at. 1977

Roberts et al. 1977

Roberts et al. 1977

Roberts et al. 1977

(Continued)

124

Table A-4. (Continued)Permeability

Compound constanta (cm//hr) Reference

PHENOLS (Continued)

Chloroxylenol

1,4,6-Trichlorophenol

2,4-Dichlorophenol

STEROIDS

5.90E-01

5.94E-01

6.01E-01

Progesterone

Pregnenolone

Hydroxypregnenolone

Hydroxyprogesterone

Cortexone

Testosterone

Cortexolone

Corticosterone

Cortisone

Hydrocortisone

Aldosterone

Estrone

Estradiol

Estriol

Dihydroepiandrosteroneb

Dihydrotestosteroneb

ALCOHOLS

1.50E-03

1.50E-03

6.00E-04

6.00E-04

4.50E-04

4.00E-04

7.50E-05

6.00E-05

1.00E-05

3.00E-06

3.00E-06

3.60E-03

3.00E-04

4.00E-05

1.70E-04

3.90E-04

Methanol 5.00E-04

Ethanol 8.00E-04

Propanol 1.20E-03

Butanol 2.50E-03

Pentanol 6.00E-03

Hexanol 1.30E-02

Heptanol 3.20E-02

Octanol 520E-02

Nonanol 6.00E-02

Decanol 8.00E-02

GLYCOL ETHERS

2-Methoxyethanol

2-Ethoxyethanol

2-Ethoxyethanol acetate

2-n-Butoxyethanol

1-Methoxypropan-2-0l

2-(2-Methoxyethoxy)ethanol

2-(2-Ethoxyethoxy)ethanol

2-(2-n-Butoxyethoxy)ethanol

2.89E-03

8.42E-04

8.07E-04

2.14E-04

1.25E-03

2.06E-04

1.32E-04

3.60E-05

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. 1982

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 A-4. (Continued)

CompoundPermeability

constanta (cm//hr) Reference

PESTICIDESc

Azodrin

Ethion

Guthion

Malathion

Parathion

Baygon

Carbaryl

Aldrin

Dreldrin

Lindane

24-D

Diquat

OTHER

9.80E-04

2.20E-04

1.06E-03

5.50E-04

650E-04

1.31E-03

4.90E-03

5.20E-04

5.10E-04

6.20E-04

3.90E-04

2.00E-05

Water 8.00E-04

Ethylbenzene 1.00E-03

Styrene 6.00E-04

Toluene 9.00E-04

Anilrneb 2.00E-02

N-nitrosodiethanolamine

Ethyl ether

2-Butanone

1-Butanol

2-Ethoxyethanol

2,3-Butanediol

Benzeneb

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

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-Dutkiewicz1982

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-Dutkiewicz1982

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-Dutkrewicz1982

a Permeability constants are for contaminants as a dilute solution in water, exceptas 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. Thesevalues should be used for dermal exposure to pure wastes.

d Permeability constants are for contaminants in gel. These values should not beused for dermal exposure due to contact with contaminated water. These valuesmay be used for dermal exposure to pure wastes.

126

? Duration of exposure = 2.6 hours/day.

Dermal absorption of waterborne contaminants maybe a significant exposure route. The factors thatinfluence dermal absorption of chemicals are: (1) thenature of the compound (molecular weight ,lipophilicity), (2) the presence of other compoundsthat might facilitate passage of a chemical though theskin (e.g., chelating or complexing agents), and (3)the permeability of the skin. Generally only lipid-soluble, non-ionized compounds are absorbedsignificantly through the skin. Also, the skin isnormally permeable only to compounds whosemolecular weights are less than 500 Daltons. Thepermeability of the skin to larger molecular weightcompounds and to less lipophilic compounds can beincreased when corrosive agents such as acids arepresent or when there are skin abrasions. Forwaterborne chemicals, exposure through the skin isalmost directly proportional to concentration.

Brown, Bishop, and Rowan (1984) recently reportedthat when compared with ingestion, dermal absorptionof volatile organic contaminants in drinking wateraccounted for approximately 29 to 91 percent of thetotal dose incurred, with the average being about 64percent. The dermal exposure route becomesespecially pertinent when organic contaminants arepresent in very dilute aqueous solution, as may oftenbe the case at Super-fund sites. In certain cases, then,dermal exposure to contaminants contained in groundor surface water may actually overshadow ingestionexposure.

When persons become exposed to contaminants indrinking water, the dermal exposure associated withbathing or showering should also be considered. Onec a n u s e t h e s a m e a p p r o a c h t o a s s e s sbathing/showering as was used for swimming.Generally, an average frequency of one bath orshower per day can be assumed, and each event canbe estimated to last 15 minutes.

For swimming or bathing exposure, the surface areaavailable for dermal exposure is assumed to equal thetotal amount of human skin surface area. Averageavailability values are given below for adults andchildren. If the exposed population is not separated byage groups, both availability values should be used torepresent a general range of exposure for the totalswimming or bathing population. Both availabilityfigures 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 insoil is calculated as follows:

D E X = C i x A V x D A x F

(A-3)

DEXC i

AV

DAF

BW

= dermal exposure, (mg/kg/day).= weight fraction of chemical substance

in soil, (unitless).= skin surface area available for contact,

(cm*).= dust adherence, (mg/cm2).= frequency of exposure events per

lifetime.= average adult body weight, (70 kg).

Values for contaminant weight fraction in thecontaminated soil will be available from the sitesurvey. Skin surface availability depends on thenature of activity being conducted, and can vary for agiven activity depending on the season of the year.Anderson et al. (1984) provide data on skin surfaceareas of different parts of the body for adults andchildren. Based on a projection of the type of activityat 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 foruse 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 ToxicSubstance Control Commission of the State ofMichigan (Harger 1979):

? Commercial potting soil adheres to hands at 1.45mg/cm2.

? Dust of the clay mineral kaolin adheres to handsat 2.77 mg/cm2.

The degree to which these values represent dustadherence at any given site is uncertain, as suchadherence will depend on a variety of site-specificfactors. Therefore, instead of selecting one of theabove values to estimate direct dermal exposure, it issuggested that the analyst use both values andgenerate an exposure range. The lifetime frequencyof direct dermal exposure will also vary considerablyand will depend on the nature of the site, its ease ofaccess, and a variety of other factors. Thus, contactfrequency should be estimated on a case-by-casebasis, based on knowledge of the site and itsenvirons.

127

Note that this approach is conservative in that itassumes that all of the contaminant adsorbed to thesoil (dust) particles is available for absorption throughthe skin. In fact, only a percentage of the totaladsorbed contaminant mass may actually be availablefor such absorption, as some percentage may remainbound to the soil particle.

The s i te survey w i l l p rov ide va lues fo r thecontaminant weight fraction in the contaminated soil.Skin surface availability depends on the nature of theactivity being conducted, and can vary for a givenactivity depending on the season of the year.Anderson et al. (1984) provide data on skin surfaceareas of different parts of the body for adults andchildren. Based on a projection of the type of activityat the site and the age of the exposed population(i.e., workers or children), one can use the data inAnderson et al. to develop skin surface estimates foruse in estimating direct dermal exposure.

The lifetime frequency of direct dermal exposure willvary considerably and will depend on the nature ofthe 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 andits environs.

2.3 Ingestion Exposure

2.3.1 Food/SoilFood ingestion exposure is estimated as the productof contaminant concentration in the food consumedand the amount of food consumed per day.Frequency is daily for foods that are a regular part ofthe diet. For recreationally caught fish, frequency canbe estimated based on the seasonal nature of fishinginvolved.

USDA source materials listed in Section A-1.4.3 arealso useful in quantifying the amount of contaminatedfood ingested. The Food Consumption of Householdsreport series provides data quantifying the amount ofvarious food categories consumed by households ona seasonal basis. Similar data are presented in foodand Nutrient Intakes of Individuals in 1 Day in theUnited States. The first source can be used to deriveestimates of the amount of various foods consumedby the overall exposed population by applyingseasonal percentage use values to local populationcensus data. The second source is used insubpopulation analyses by applying sex- and age-specific consumption values to census data for theexposed population.

Consumption of fish caught in contaminated watersmay be an important ingestion route, since certaincontaminants of concern tend to biomagnify in thefood chain. This phenomenon results in tissueconcentrations of contaminants in predator fish

exhibiting levels that greatly exceed the ambientconcentration in the waterbody. An average daily fishingestion rate for the U.S. population has beenestimated as 6.5 grams per day (USEPA 1980b).Persons for whom fish constitutes a major portion ofthe overall diet may consume up to 124 grams perd a y ( U S D A 1 9 8 0 ) . A W e s t C o a s t s t u d y o fconsumption of fish caught in contaminated waters bysport fishermen (Puffer et al. 1981) reports a medianfish ingestion rate of 37 grams/day. This report alsolists a maximum rate of 225 grams/day.

Ingestion exposure estimates are calculated in thesame manner, regardless of the type of foodinges ted . Mu l t ip l i ca t ion o f the con taminantconcentration in the ingested food by the amount ofcontaminated food ingested per day yields exposureper day.

Children may ingest soil during play both inadvertentlyand in ten t iona l l y (p ica behav io r ) . In thoseassessments where the exposed populations analysishas found that children may have access to areas ofcontaminated soil, this exposure route should beevaluated. Data quantifying the amount of soilingested by children are conflicting and varyconsiderably. 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 thisrange, reasonable typical values can be identified andassociated with various age groups, if desired. Forstudies warranting such detail, the daily soil ingestionrate values presented in Table A-5 can be used. Forstudies that do not require such detail, one can usean overall average soil ingestion value of 100 mg/day.

Table A-5. Typical Daily Soil IngestionRates for Children by AgeGroup Soil ingestion range

Age (mg/day)0-9 months 0

9-18 months 50

1.5 - 3.5 years 200

3.5 - 5 years 50

5 - 18 years 10

Source: Calabrese et al. 1987.

2.3.2 WaterEvent-based water Ingestion exposure equals thedaily total amount of contaminant ingested from eithersurface or ground waters affected by the Superfundsite. This exposure is determined by the contaminantconcentration in the water and the amount of wateringested per day. On average, an adult ingestioncoefficient of 2.0 liters per day (USEPA 1980b) canbe used for Superfund site analyses. Frequency ofdrinking water exposure is daily.

128

When contaminated surface waters are usedrecreationally, it may be appropriate to estimateexposure that results from inadvertently ingestingcontaminated water while swimming. For this analysis,the same values for event frequency and durationpreviously presented in Section A-2.2 should beused. In addit ion, to est imate the amount ofcontaminated water ingested per event, an assumedvalue of 50 ml per hour can be used.

2.4 Exposure Integration

The final step in the exposure assessment processfor uncontrolled hazardous waste sites is theintegration of all exposures experienced by individualexposed populations. This simply involves organizingthe results of the previous analyses to total allexposures to a g iven hazardous subs tanceexperienced by each population segment. Becausedifferent chemicals exhibit different toxicologicalproperties, exposures to each contaminant of concernare considered separately. Note that in some cases,individual populations may be exposed to a givenchemical in a particular medium through more thanone exposure scenario. For example, persons whoswim in contaminated waters may obtain theirdr inking water f rom the same contaminatedwaterbody. In such cases, the dermal exposureexperienced while swimming can be added to thatexperienced during bathing or showering to generatean overall dermal exposure value for that populationsegment. The data management forms supplied inAppendix C are designed to help organize the resultsof exposure calculation and integration.

129

Chapter 3Appendix A References

Anderson E, Browne N, Duletsky S, et al. 1984.Development of statistical distribution or ranges ofstandard factors used in exposure assessments.Revised draft final report. Washington, DC: Officeof Health and Environmental Assessment, U.S.Environmental Protection Agency. Contract No.68-02-3510.

Dutkiewicz T, Tyras H. 1967. A study of skinabsorption of ethylbenzene in man. British JournalInd. Med. 24:330-32.

Dutkiewicz T, Tyras H. 1968. Skin absorption oftoluene, styrene, and xylene by man. Departmento f Tox ico log ica l Chemis t ry and Indus t r ia lToxicology, Medical Academy, Lodz, Poland.

Baranowska-Dutkiewicz B. 1982. Skin absorption ofaniline from aqueous solutions in man. ToxicologyLetters 10:367-72.

Blank IH, Moloney J, Ernslie AG, et al. 1984. Thediffusion of water across the stratum corneum as afunction of its water content. The Journal ofInvestigative Dermatology. 82:188-194.

Feldman RJ, Maibach HI. 1974. Percutaneouspenetration of some pesticides and herbicides inman. Toxicology and Applied Pharmacology28: 126-32.

Bronaugh R, Condon E, Scheuplein RJ. 1981. Theeffect of cosmetic vehicles on the penetration ofn-nitrosodiethanolamine through excised humanskin. Journal of Investigative Dermatology 76:94-6.

Freed JR, Chambers T, Christie WN, Carpenter CE.1985. Methods for assessing exposure to chemicalsubstances: volume 2 - methods for assessingexposure to chemical substances in the ambientenvironment. Washington, DC: U.S. EnvironmentalProtection Agency, Office of Toxic Substances,Exposure Evaluation Division. EPA 560/5-83-015.

Brown HS, Bishop DR, Rowan CA. 1984. The role of Galey WR, Londdale HK, Nacht S. 1976. The in vitroskin absorption as a route of exposure for volatile permeability of skin and buccal mucosa to selectedorganic compounds (VOCs) in drinking water. drugs and tritiated water. J. Invest. Dermatol.Amer. J. of Public Health. 74 (5). 67:713-717.

Budiansky S. 1980. Indoor air pollution. Environ. Sci.Technol. 14(g): 1023-1027.

Bureau of the Census. 1986. Statistical abstract of theUnited States: 1986 (106th edition). Washington,DC: U.S. Department of Commerce, U.S.Government Printing Office.

Harger JRE. 1979. A model for the determination ofan action level for removal of curene contaminatedsoil. Memorandum to P.S. Cole, Executive Director.Lansing, Ml: Toxic Substance Control Commission.

Howes D. 1975. The percutaneous absorption ofsome anionic surfactants. J. Soc. Cosmet. Chem.26:47-63.

Calabrese EJ, Kostecki PT, Gilbert CE. 1987. Howmuch so i l do ch i l d ren ea t? An emerg ingconsideration for environmental health riskassessment. Paper submitted to Comments inToxicology.

McCullough JL, Snyder DS, Weinstein GD, FriedlandA, Stein B. 1976. Factors affecting humanpercutaneous penetration of methotrexate and itsanalogues in vitro. J. Invest. Dermatol. 66:103-107.

Dugard PH, Walker M, Mawdsley SJ, Scott RC. 1984.Absorption of some glycol ethers through humanskin in vitro. Environment Health Perspectives57: 193-97.

Moschandreas DJ, Stark JWC, McFadden JF, MorseSS. 1978. Indoor pollution in the residentialenvironment - 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: percutaneouspenetration and metabolic disposition. JournalAmerican Academy Dermatology 4:31-7.

Ott WR. 1981. Exposure estimates based oncomputer-generated activity patterns. Paperpresented at the 74th annual meeting of the AirPollution 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 transportmodel. Draft report. Oak Ridge, TN: Oak RidgeNational Laboratory. ORNL-TM-8182. IEG-A D - 8 9 - F - 1 - 3 9 9 9 - 0 .

Pennington, JAT. 1983. Revision of the total dietstudy. In J. Amer. Dietetic Assoc. 82 (2).

Puffer H., Azen SP, Young DR, et a l . 1981.Consumption rates of potentially hazardous marinefish caught in the metropolitan Los Angeles area.California Department of Fish and Game. EPAGrant No. R 807 120010.

Roberts MS, Anderson RA, Swarbrick J. 1977.Permeability of human epidermis to phenoliccompounds. Journal Pharm. and Pharmacol.29:677-83.

132

Schaefer H, Zesch A, Stuttgen G. 1982. Skinpermeability. New York: Springer-Verlag.

Scheuplein RJ, Blank IH, Brauner GJ, MacFarlaneDJ. 1969. Percutaneous absorption of steroids.Journal of Investigative Dermatology 54(1):63-70.

Scheuplein RJ, Blank IH. 1971. Permeability of theskin. Physiological Reviews 51(4):702-47.

Sutton TJ. 1973. Dermal toxicity and penetrationstudies following topical application of threehistamine H2-receptor antagonists with acomparison with an H1-receptor antagonist.Toxicology and Applied Pharmacology 50(3):459-65.

Tregear RT. 1966. Physical functions of skin. NewYork: Academic Press.

USDA. 1980. Food and nutrient intakes of individualsin 1 day in the United States, spring 1977,nationwide food consumption survey 1977-78,preliminary report no. 2. Washington, DC: Scienceand Education Administration.

USDA. 1983a. Food consumption of households inthe Northeast, seasons and year 1965-66, reportno. 13. Washington, DC: Agricultural ResearchService. August 1972.

USDA. 1983b. Food consumption of households inthe North Central region, seasons and year 1965-66, report no. 14. Washington, DC: AgriculturalResearch Service. September 1972.

USDA. 1983c. Food consumption of households inthe South, seasons and year 1965-66, report no.15. Washington, DC: Agricultural Research Service.January 1973.

USDA. 1983d. Food consumption of households inthe West, seasons and year 1965-66, report no.16. Washington, DC: Agricultural Research Service.January 1973.

USDA. 1983e. Food consumption of households inthe United States, seasons and year 1965-66.Washington, DC: Agricultural Research Service.March 1972.

USDOI. 1973. Outdoor recreation: a legacy forAmerica. Washington, DC: U.S. Department ofInterior.

USEPA. 1980a. Dietary consumption distributions ofselected food groups for the U.S. population.Washington, DC: U.S. Environmental ProtectionA g e n c y . O f f i c e o f P e s t i c i d e s a n d T o x i cSubstances, Office of Testing and Evaluation. EPA560/11-80-012.

USEPA. 1980b. Water quality criteria documents.Federal Register, Vol.\45 No. 231, November 28,1980.

USEPA. 1981. The exposure assessment group’shandbook for performing exposure assessments(draft report). Washington, DC: U.S. EnvironmentalProtection Agency.

USEPA. 1985. Superfund public health evaluationmanual. Draft. Washington, DC: ICF, Inc. Preparedfor the Policy Analysis Staff, Office of Emergencyand Remedial Response, U.S. EnvironmentalProtection Agency. October 1, 1985.

USEPA. 1987. Exposure fac to rs handbook .Washington, DC: Exposure Evaluation Division,U . S . O f f i c e o f T o x i c S u b s t a n c e s , U . S .Environmental Protection Agency. Contract No.68-02-4254, Task No. 83.

Versar. 1984. Methods for estimating concentrationsof chemicals in indoor air. Draft final report. VersarInc. Washington, DC: Prepared for the ExposureAssessment Branch, Exposure Evaluation Division,

Office of Toxic Substances, U.S. EnvironmentalProtection Agency.

Versar. 1 9 8 5 . E x p o s u r e a s s e s s m e n t f o rperchloroethylene. Revised draft report. Versar Inc.Exposure Assessment Branch, ExposureEvaluation Division, Office of Toxic Substances,U.S. Environmental Protection Agency.

133

II

Appendix B

Possible Exposure Assessment Data Requirements for Uncoltrolled Hazardous WasteSites 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

Contaminant release

Type of Site Area of Concern Area Subclass Parameter

Contaminated surface soil Particulate release Wind erosion • Soil erodibility indexa

(includes spills and leaks)• Soil ridge roughness

factora

• Field length alongprevailing wind direction

??Vegetative cover factor??Concentrations of

contaminantsb

??Volume of contaminatedregionb

Unpaved roads ??Silt content??Mean speed of vehicles

traversing contaminatedaread

??Mean weight of vehiclestraversing contaminatedaread

??Mean number of wheelsof vehicles traversingcontaminated aread

Volatilization Short-term release

Runoff to surface water

Excavation and transfer of ??Silt contentc

soil• Mean wind speede

• Drop height• Material moisture content• Dumping device capacity

• Vapor concentration ofcontaminants in soil porespacesf

Long-term release • Depth from soil surface tobottom of contaminatedregionb

??Area of contaminationb

• Depth of “dry”(uncontaminated) zone atsampling timeb

• Concentrations ofcontaminants in soil andin liquid phaseb

??Soil porosityb,c

??Absolute temperatureb,e

• Time measured fromsampling time

• Soil erodibility factorg

• Slope - length factor• Vegetative cover factord

Erosion control practicefactord

??Area of contamination??Soil bulk densityc

??Total areal concentrationsof contaminants

Release to ground water - See Chapter 3.6 ofManual

(Continued)

136

Table B-1. (Continued)

Type of Analysis Type of Site Area of Concern Area Subclass Parameter

Landfill Volatilization NO internal gas generation ??Area of contamination• Soil porosityc

??Effective depth of soilcover

• Mole fractions ofcontaminants in waste

??Absolute ambienttemperaturee

??Absolute ambientpressuree,h

??Soil bulk densityC,I

??Concentration ofcontaminants in soilb

??Volume of contaminatedregionb

Lagoon

Release to groundwater

Volatilization

Migration intoground water

Contaminant fate Contaminatedsurface soil,landfill, lagoon

Atmospheric fate

With internal gasgeneration

??Vapor concentration ofcontaminants in soil porespacesf

• Area of contamination

- See Chapter 3.6 ofManual

??Liquid-phaseconcentrations ofcontaminants

??Area of contamination??Absolute ambient

temperaturee

??Volume of contaminatedregronb

- See Chapter 3.6 ofManual

• Distance from site toselected exposure point

??Mean wind speede

• Relative annual frequencyof wind flow towards pointxe

• Relative annual frequencyof stability class for windflow towards point xe

??Stability classes(A = unstable, F = stable);according to Pasquillclassification systeme

• Vegetative cover factord

137

Table B-1. (Continued)

Type of Analysis Type of Site Area of Concern Area Subclass Parameter

Surface water fate ??Combined effluent andstream flow data

Exposed populations

Ground water fate Saturated zone

Unsaturated zone

All General

Contaminated surfacewater

??intermedia substancet r a n s f e r r a t e f

??Width of water bodyt

• Stream velocityi

??Stream depth’??Slope of stream channeli

??Soil hydraulicconductivityk

??Hydraulic gradientl

??Effective soil porositym

??Average percolation orrecharge ratem

• Volumetric water contentof soil in unsaturatedzonef

??Hydraulic loading frommanmade sourcesf,n

• Precipitation raten,o

??Evapotranspiration ratef,n

• Runoff ratef,n

??Average depth ofcontaminated arean

??Evaporation rate0

• Location of population??Number of persons??Age/sex distribution

• Recreation patterns(fishing, hunting,swimming)

??Commercial fisheriespresent

??Drinking water intakelocations and populationsserved

Contaminated ground water ??Drinking water intakelocations and populationsserved

abcdef

g

i

j

kI

Some values can be obtained from existing literature.For calculation of long-term release ( > 70 years).Can be obtained from Soil Conservation Service (SCS) “Soils 5 File” data base.Estimated indirectly from site survey information.Can be estimated based on existing meteorological station data.Can be calculated.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.Can be obtained from USGS data.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.n Needed to calculate average percolation/recharge rate when not measured at site.o Available from local or National Weather Service.

138

Table B-2. Index to Variable Terms

TermUsed Definition

E Potential annual wind erosion soil lossI’

Units(mass/area/time)

Equation(s) inwhich term is used Source

2-1 calculated

Soil erodibility index (dimensionless) 2-1 site data andliterature

site data andliterature

literature

site data andliterature

site data

calculated

K’

C’L’

V

EVT

k Particle size multiplier

s Sift content

S p

W

W

D P

Ei

Di Diffusion coefficient of component i (cm2/sec)

A

C s i Saturation vapor concentration of component i (g/cm3)P t Soil porosity (dimensionless)

dsc Effective depth of soil cover (cm)Mi Mole fraction of toxic component i in the waste (g/g)T Temperature (K,C)

MWi

MW a

Pa

D’

MW’

B

Soil ridge roughness factor

Climatic factor

Field length along the prevailing wind direction

Vegetative cover factor

Emission factor for vehicular traffic

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

Contaminated area

Molecular weight of contaminant i

Molecular weight of air

Molecular diffusion volumes of toxic contaminantV1) and air (V2)Absolute pressure

Known diffusion coefficient of a compound withmolecular weight and molecular diffusion volumeclose to that of the unknown (Di)

Molecular weight of the selected compoundcorresponding to D’

Soil bulk density

(dimensionless)

(dimensionless)

(feet)

(dimensionless)

(kg/vehicle kilometertraveled; lb/vehicle miletraveled)

(dimensionless)

(%)

(kph; mph)

(Mg; tons)

(dimensionless)

(dimensionless)

(g/sec)

(cm2; areas; ha)(100 in2)

(g/mole)

(g/mole)

(g/mole)

(g/cm3)

2-1

2-1

2-1

2-12-2

2-22-2

2-22-22-22-2

2-3; 2-8; 2-9;2-11; 2-15

2-3; 2-4; 2-5;2-122-3; 2-8; 2-9;2-11; 2-15; 2-19; 2-21; 2-24;2-25; 2-26;2-30; 2-32;2-33; 2-372-3; 2-72-3; 2-6; 2-12;3-17; 3-34;3-352-32-32-4; 2-7;2-10;2-13; 2-16;2-172-4; 2-5; 2-7;2-10; 2-172-42-4

see text

site data, SCSSoils 5 Fife

site data

site data

site data

Figure 2-3

calculated

calculated

site data

calculated

site data; SCSSoils 5 File

site data

site data

site data

literature

2-42-5

see text

literature andcalculated

site data

see text

2-5 literature

2-6; 2-25;2-26; 2-27;3-17

site data; SCSSoils 5 File

139

Table B-2. (Continued)

Term Equation(s) in whichUsed Definition Units term is used Source

P Particle density (g/cm3) 2-6

P Vapor pressure of the chemical (mm Hg) 2-7; 2-37

R Gas constant (62.3 mm Hg-liter/k-mol; 8.2 x 10-5

atm-m3/-mol-k)

Ci* Vapor concentration of compound i (g/cm3)

VyMean landfill gas velocity in the soil pore spaces (cm/sec)

k iGGas-phase mass transfer coefficient of (cm/s)chemical i

2-7; 2-13; 2-16

see text

literature orestimated (see text)

see text

MWH20 Molecular weight of water (g/mole)kiG,H2O Gas phase mass transfer coefficient for water

vapor at 25°C

C sThe liquid-phase concentration of component i (g/cm3)

C BBulk contaminant concentration in soil (g/cm3)

t Time measured from sampling time (seconds)

d Depth of dry zone at sampling time (cm)D Related to the amount of contaminant i that goes (cm2/sec)

from liquid to gas phase, and then from gas phaseto diffusion in air

Hi’ Henry’s Law constant in concentration form

HiHenry’s Law constant

h Depth from soil surface to the bottom of thecontaminated region

tdThe time at which all contaminant has volatizedfrom the soil

KiOverall mass transfer coefficient

k iLLiquid phase mass transfer coefficient

MW o 2Molecular weight of oxygen

kL, o2 Liquid phase mass transfer coefficient for oxygenat 25°C

EAiVcC i

E

Y(S)Ea

Vr

q p

K

L

S

C

Average release of contaminant i

Volume of contaminated region

Concentration of contaminant i in soil

Total release rate of contaminant i obtained bysumming all above-listed releases of thecontaminant at the site

Sediment yield in tons per event

Conversion constant

Volume of runoff

Peak flow rate

The soil-erodibility factor. Obtained from thelocal Soil Conservation Service OfficeThe slope-length factor

The slope-steepness factor

The cover factor

(dimensionless) 2-12; 2-13

(atm-m3/mol) 2-13; 2-16

(cm) 2-14

(sec) 2-14 calculated

(cm/sec)

(cm/sec)

(g/mole)

2-15; 2-16 calculated

2-16; 2-17 calculated

2-17 see text

2-17 literature

(mass/time)

(cm3)(g/cm3, kg/ha, lb/acre)

(g/sec)

2-18; 2-19; 2-29

2-18; 3-34; 3-35

2-18; 2-25;2-28; A-3

2-18

(metric tons)

(acre-feet, m3)

(ft3/sec, m3/sec)

2-20; 2-27

2-20; 2-21;2-23; 2-24

2-20; 2-21

2-20; 2-24

2-8; 2-9 site data

2-8 see text

2-9; 2-10; 2-16 calculated

2-10 see text

2-10 calculated

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

site data

site data

site data

site data

calculated

calculated

literature

site data

calculated

site data

site data

calculated

calculated

see text

calculated

calculated

site data, literature(commonly expressed in 2-20; 2-30tons per acre per R unit)(dimensionless) 2-20; 2-30

(dimensionless) 2-20; 2-30

(dimensionless) 2-20; 2-30

see Figures 2-4through 2-6see Figures 2-4through 2-6see text and Table2-4

(Continued)

140

Table B-2. (Continued)

TermUsed Definition

P The erosion control practice factorUnits

(dimensionless)

Equation(s) inwhich term is used Source2-20; 2-30 see text

Qr

Rt

Depth of runoff

The total storm rainfall

(in, cm)

(in, cm)

2-21; 2-22;2-24; 2-282-22; 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

NationalClimatological DataCenter, Asheville,NC; USDC (1961)

Sw Water retention factor (in, cm)

CNTr

The SCS Runoff Curve Number

Storm duration

(dimensionless)

(hour)

Table 2-6

NationalClimatological DataCenter, AshevilIe,NC; USDC (1961)

calculated

calculated

calculated(see text)

Sorbed substance quantity

Dissolved substance quantity

Available water capacity of the top cmof soil

Sorption partition coefficient

(kg, lb)(kg, lb)(dimensionless)

(cm3/g)

PXi

PQi

B

Sorbed substance loss per event

Dissolved substance loss per event

Dissolved or sorbed loss per stormevent

Number of “average” storm events in70 years

(kg, lb)(kg, lb)(kg, lb)

calculated

calculated

calculated(see text)

NationalClimatological DataCenter, Asheville,NC; USDC (1961)(see text)

N (dimensionless)

Y(S)A

Rr

Sd

Dd

Annual soil loss in runoff

Rainfall and runoff factor

Sediment delivery ratio

Overland distance between site andreceiving waterbody

Contaminant Loading rate

(tons)

(dimensionless)

(dimensionless)

(Ft)

calculated

site data

(mass/time) calculatedLc

C0Qib

Ks

Solubility of solid chemical

Volume loading rate

Soil specific exponential function

Soil hydraulic conductivity

(mass/volume)

(volume/time)

(dimensionless)

(length/time)

literature

calculated

Table 3-11

site data,Table 3-8; 3-9

site data

calculatedHydraulic gradient

Hydraulic conductivity of liquidcontaminant in site soil

Hydraulic conductivity of water in sitesoil (same as K,)

Density of liquid contaminant

Density of water

(dimensionless)

(length/time)Kc

Kw(length/time) site data

(mass/volume)

(ma&volume)

literature

literatureDcDwUcuw

Dynamic viscosity of liquid contaminant [mass/(length x time)] 2-35

Dynamic viscosity of water [mass/(length x time)] 2-35

literature

literature

141

Table B-2. (Continued)

TermUsed Definition

Sl Slope of stream bedUnits

(dimensionless)

Equation(s) inwhich term is used Source3-6 site data

gW(X)

W(CL)

W(O)

Gravitational acceleration constant

Water concentration of substance atdownstream distance X

Predetermined critical waterconcentration level

Water concentration of substanceimmediately below point of introductionto stream

K

e

V p w

q

Overall aquatic decay coefficient

Exponential function

Interstitial pore-water velocity orground-water velocity

Average percolation or recharge rate

V Darcy velocity

p eSoil Effective Porosity

Saturated Water Content soil (equal toPt)Wilting Point Moisture Content

Volumetric water content of soil

HL Hydraulic loading from manmadesources

Pr

ET

Q r

EVAP

Cet

Cveg

Precipitation rate

Evapotranspiration rate

Runoff rate

Evaporation rate

Rd

Correction factor for converting panevaporation rate to evapotranspirationrate for turf grass

Correction factor for converting turfgrass evapotranspiration to that forother vegetative cover

Retardation factor

VdRetarded velocity of hydrophobic

KocPartition coefficient for organic carbon

focKow

Fraction of organic carbon in soil

Octanol/water partition coefficient

X d

T d

Qd

Nomograph factor

Nomograph factor

Nomograph factor

(32 ft/sec2) 3-6

(mass/volume) 3-7

(mass/volume)

(mass/volume)

(time-l)

3-8

3-7; 3-8

3-7; 3-8

2-36; 3-73-10; 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; 3-313-11; 3-13

3-11

3-12; 3-13

3-14

(length&me)

(depth/time)

(Iength/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)

3-14

3-14; 3-15

3-14

3-153-15

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

3-193-20; 3-21;3-22; 3-23;3-24; 3-25

3-26; 3-293-27

3-28; 3-31

______calculated

water qualitycriteria

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-l2

Table 3-13, seetext

calculated

calculated

calculated

site data, literature

literature

calculated

calculated

calculated

143

Table B-2. (Continued)

TermUsed DefinitionDx Longitudinal Dispersion Coefficient

Units(length2/time)

Equation(s) inwhich term is used Source3-26: 3-27; calculated

Dymax

ayYA

T ½

Vl

C l

Transverse Dispersion CoefficientAquifer thicknessLongitudinal dispersivityTransverse dispersivityCoefficient for decayDecay constantHalf-lifeVolume of liquid chemical releasedAverage concentration of chemicalcontaminant in released liquidVolume of contaminated ground water

(length2/time)(length)(length)(length)(dimensionless)(1/time)(time)(lengths)(mass/lengths)

3-28: 3-29;3-313-28; 3-31(Fig. 3-8)(Fig. 3-8)(Fig. 3-8)(Fig. 3-8)(Fig. 3-8)(Fig. 3-8)3-323-32

calculatedsite dataliteratureliteratureliteraturecalculatedcalculatedsite datasite data

Vgw

Mc

C c

IEXteI

BwF

DEXAVPC

DAC

Average concentration of contaminantin ground waterMass of solid wasteConcentration Expressed as MassFractionInhalation exposureDuration of exposure eventAverage Inhalation rateAverage adult body weightFrequency of exposure eventDermal exposureSkin surface area availableDermal permeability constant forsubject contaminantDust adherenceContaminant concentration

(lengths)

(mass/length3)

(mg)(dimensionless)

(mg/kg/day)(hours/event)(m3/hr)(70 kg)(number/lifetime)(mg/kg/day)(cm2)(cm/hr)

(mg/cm2)

3-32; 3-33;3-34; 3-353-32; 3-33

site data

site data

3-33 site data3-33 site data

A-1 calculatedA-1; A-2 estimatedA-1 Table A-3A-1; A-2; A-3 Eq A-1A-1; A-2; A-3 estimatedA-2; A-3 calculatedA-2; A-3 estimatedA-2 Table A-4

A-3 See text

144

Appendix CData Management Forms

This appendix presents master copies of datamanagement forms designed for use when applyingthe various analyses described in this manual. Theforms are intended to provide easy, consistentorganizat ion of the resul ts of each analysiscomponent in the human exposure assessmentprocess (qualitative analysis, quantitative contaminantrelease analysis, etc.) for ready use in subsequentanalytical components. In addition, these forms willalso organize exposure assessment output in a formmost useful for conducting a risk assessment(executed following and based on the results of theexposure assessment) as well as the development ofa site Endangerment Assessment for enforcementpurposes.

These forms are included as master copies, thatshould be photocopied for use in a given siteinvestigation. In many cases, a number of copies ofcertain forms will be required to tabulate all results ofthe exposure assessments. For example, Form No. 7:Exposure Integration requires that the exposedpopulation segment be logged into the upper leftcorner of the form, and exposure information for thatpopulation segment be entered into the remainingcolumns for each chemical to which the population isexposed. If four distinct exposed population segmentsare affected at the site, four copies of the form will berequired.

145

Form

1:

Qua

litat

ive

Expo

sure

Ana

lysi

sS

ite N

ame:

Dat

e:A

naly

st:

1.

Che

mic

alO

n-si

te R

elea

seR

elea

seS

ourc

eLi

kelih

ood

Mag

nitu

de*

Rel

ease

Mec

hani

smR

ecei

ving

Med

ium

Pot

entia

lly E

xpos

edP

opul

atio

n S

egm

ent

Exp

osur

e M

echa

nism

2. 4. 5.

Form

1. Q

ualit

ativ

e E

xpos

ure

Ana

lysi

s (C

ontin

ued)

Site

Nam

e:D

ate:

Ana

lyst

:

Che

mic

alO

n-si

te R

elea

seR

elea

seS

ourc

eLi

kelih

ood/

Mag

nitu

de*

Rel

ease

Mec

hani

smR

ecei

ving

Med

ium

Pot

entia

lly E

xpos

edP

opul

atio

n S

egm

ent

Exp

osur

e M

echa

nism

6. 7. 8. 10.

*Cod

e ea

ch s

ourc

e as

to: (

1) L

ikel

ihoo

d of

rel

ease

and

(2)

Pot

entia

l mag

nitu

de o

f rel

ease

. Use

H, M

, L (

high

, med

ium

, low

) de

sign

atio

n an

d pr

ovid

e a

lette

r co

de fo

r lik

elih

ood

and

mag

nitu

de,

each

sep

arat

ed b

y a

“/”.