risk assessment guidance for superfund volume i human

United States Office of Emergency andEnvironmental Protection

EPA/540/1 -89/002Remedial Response December 1989

Agency Washington, DC 20460 P B 9 0 - 1 5 5 5 8 1

Superfund

Risk AssessmentGuidance for SuperfundVolume IHuman HealthEvaluation Manual(Part A)

Interim Final

EPA/540/l -89/002December 1989

Risk AssessmentGuidance for Superfund

Volume 1Human Health Evaluation Manual

(Part A)

Interim Final

Office of Emergency and Remedial ResponseU.S. Environmental Protection Agency

Washington, DC 20450

Printed on Recycled Paper

Page ii

NOTICE

The policies and procedures set forth here are intended solely as guidance to EPA and othergovernment employees and contractors. This guidance does not constitute rulemaking by the Agency, andcannot be relied on to create a substantive or procedural right enforceable by any party in litigation withthe United States. EPA may take action that is at variance with the policies and procedures in this manualand may change them at any time without public notice.

This interim final guidance is based on policies in the proposed revisions to the National Oil andHazardous Substances Pollution Contingency Plan (NCP), which were published on December 21, 1988 (53Federal Register 51394). The final NCP may adopt policies different than those in this manual and should,when promulgated, be considered the authoritative source. A final version of this manual will be publishedafter the revised NCP is promulgated.

Following the date of its publication, this manual is intended to be used as guidance for all humanhealth risk assessments conducted as part of Superfund remedial investigations and feasibility studies.Issuance of this manual does not invalidate human health risk assessments completed before (or in progressat) the publication date and based on previously released Agency guidance.

Page iii

ABOUT THE REVISION . . .

WHAT ITIS

WHO IT'SFOR

WHAT’SNEW

DISTRIBU-TI0N PLAN

WHERETO SENDCOMMENTS

EPA’s Human Health Evaluation Manual is a revision of the Superfund PublicHealth Evaluation Manual (SPHEM; October 1986); it is Volume I of the two-volume setcalled Risk Assessment Guidance for Superfund. This manual has three main parts: thebaseline risk assessment (Part A); refinement of preliminary remediation goals (Part B); andevaluation of remedial alternatives (Part C). (Only Part A is included in the firstdistribution see below.)

Risk assessors, risk assessment reviewers, remedial project managers (RPMs), and riskmanagers involved in Superfund site cleanup activities will benefit from this revision.

This revision builds upon the process established in SPHEM and provides more detailedguidance on many of the procedures used to assess health risk. New information andtechniques are presented that reflect the extensive Superfund program experience conductinghealth risk assessments at Superfund sites. Policies established and refined over the years- - especially those resulting from the proposed National Oil and Hazardous SubstancesPollution Contingency Plan (NCP) -- have been updated and clarified. Additionally, thelinks between the human health evaluation, the environmental evaluation, and the remedialinveatigation/feasibility study (RI/FS) have been strengthened.

In Part A you will find:

For the risk assessor -- Updated procedures and policies, specific equations andvariable values for estimating exposure, and a hierarchy of toxicity data sources.

For the risk assessment reviewer -- A baseline risk assessment outline for consistentpresentation of risk information and format, and a reviewer’s checklist to ensureappropriate quality and content of the risk assessment.

For the RPM -- A comprehensive overview of the risk assessment process in theRI/FS, a checklist for RPM involvement throughout the process, and a completeindex for quick reference.

For the risk manager -- An expanded chapter on risk characterization (Chapter 8)to help summarize and present risk information for the decision-maker, and moredetailed descriptions of uncertainties in the assessment.

This manual is being distributed as an interim final document while the proposed NCP isbeing finalized. After the final NCP is published, the manual will be updated and finalized.Parts B and C -- which were not distributed as interim final because they are highlydependent on possible revisions to the NCP -- will be added. Periodically, updates ofportions of the manual will be distributed.

Toxics Integration BranchOffice of Emergency and Remedial Response

401 M Street, SW (0S-230)Washington, DC 20460

Phone 202-475-9486

Page iv

WORKGROUP

EPA HEADQUARTERS

Office of Emergency and Remedial Response Marlene BergDavid CooperLinda CullenCarla DempseySteve GolianBruce MeansPat MundySandra Panetta

Office of Solid Waste

Office of Waste Programs Enforcement:

Office of Solid Waste and Emergency Response:

Office of Policy, Planning, and Evaluation

Office of General Counsel

Office of Research and Development

Office of Water:

Region I:

Region V:

Region VI:

Region X:

EPA REGIONAL OFFICES

Stephanie Irene

Georgia Valaoras

Larry Zaragoza

Charlotte WhiteCraig Zamuda

J o e F r e e d m a n

Rebecca MadisonSue Norton

Frank GostomskiRobert Zeller

Sarah Levinson

Dan BicknellPamela Blakley

Fred Reitman

Dana DavoliDavid Tetta

OTHER EPA OFFICES

Great Lakes National Program Office, IL: Cynthia Fuller

Office of Health and EnvironmentalAssessment, OH: Chris DeRosa

Office of Air Quality Planning andStandards, NC Fred Hauchman

Page v

TABLE OF CONTENTS

INTRODUCTION

CHAPTER 1 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1.1 OVERVIEW OF THE HUMAN HEALTH EVALUATION PROCESSIN THE RI/FS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1.1.1 Project Scoping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1.1.2 Site Characterization (RI) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1.1.3 Feasibility Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1.2 OVERALL ORGANIZATION OF THE MANUAL . . . . . . . . . . . . . . . . . . . . .

CHAPTER 2 STATUTES, REGULATIONS, GUIDANCE, AND STUDIES RELEVANT TO T HE HUMAN HEALTH EVALUATION . . . . . . . . . . . . . . . . . . . . . . . . . .

2.1 STATUTES, REGULATIONS, AND GUIDANCE GOVERNING HUMANHEALTH EVALUATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.1.1 CERCLAand SARA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2.1.2 National Contingency Plan (NCP) . . . . . . . . . . . . . . . . . . . . . . . . . . . .2.1.3 Remedial Investigation/Feasibility Study Guidiance . . . . . . . . . . . . . . . . . .2.1.4 ARARs Guidance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.5 Superfund Exposure Assessment Manual . . . . . . . . . . . . . . . . . . . . . . . . .

2.2 RELATED SUPERFUND STUDIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.2.1 Endangerment Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2.2.2 ATSDRHealth Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2.2.3 ATSDR Health Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

CHAPTER 3 GETTING STARTED PLANNING FOR THE HUMAN HEALTHEVALUATION I N T H E RI/FS . . . . . . . . . . . . . . . . . . . . . . . . .

3.1 Goa l of the RI/FS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3.2 Goal of the RI/FS Human Health Evaluation . . . . . . . . . . . . . . . . . . . . . . . . .3.3 Operable Units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3.4 RI/FS Scoping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5 Level of Effort/Level of Detail of the Human Health Evaluation . . . . . . . . .

PART A -- BASELINE RISK ASSESSMENT

CHAPTER 4 DATA COLLECTION . . . . . . . . . . . . . . . . . . . . . . . . . . . .

4.1 BACKGROUND INFORMATION USEFUL FOR DATA COLLECTION . . . . . .

4.1.1 Types of Data . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1.2 Data Needs a n d the RI/FS . . .. . . . . . . . . . . . . . . . . . . . . . . . .

Page

1-1

1-3

1-41-41-8

1-10

2-1

2-1

2-12-42-52-72-8

2-8

2-82-9

2-10

3-1

3-13-13-23-23-3

4-1

4-1

4-14-2

Page vi

4.2

4.3

4.4

4.5

4.6

4.7

4.8

4.9

4.1.3 Early Identification of Data Needs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-34.1.4 Use of the Data Quality Objectives (DQO) Guidance . . . . . . . . . . . . . . . . . 4-44.1.5 Other Data Concerns . . . . . . . . . . . . . . . . .

REVIEW OF AVAILABLE SITE INFORMATION .

ADDRESSING MODELING PARAMETER NEEDS . . .

DEFINING BACKGROUND SAMPLING NEEDS . . .

4.4.14.4.24.4.34.4.4

Types of Background . . . . . . . . . . . . . . . . .Background Sampling Locations . . . . . . . . . .Background Sample Size . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . 4-4

. . . . . . . . . . . . . . . . . . . . . . 4-4

. . . . . . . . . . . . . . . . . . . . . 4-5

. . . . . . . . . . . . . . . . . . . . . . 4-5

. . . . . . . . . . . . . . . . . . . . . . 4-5

. . . . . . . . . . . . . . . . . . . . . . 4-8

. . . . . . . . . . . . . . . . . . . . . . 4-8Comparing Background Samples to Site-Related Contamination . . . . . . . . . . 4-9

PRELIMINARY IDENTIFICATION OF POTENTIAL HUMAN EXPOSURE . . . . . . . . . . . . . 4-10

4.5.1 General Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-104.5.2 Soil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-114.5.3 Ground Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-124.5.4 Surface Water and Sediment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-134.5.5 Air . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-144.5.6 Biota . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-15

DEVELOPING AN OVERALL STRATEGY FOR SAMPLE COLLECI’ION . . . . . 4-16

4.6.14.6.24.6.34.6.44.6.54.6.6

Determine Sample Size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-17Establish Sampling Locations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-18Determine Types of Samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-19Consider Temporal and Meteorological Factors . . . . . . . . . . . . . . . . . . . . . . 4-19Use Field Screening Analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-20Consider Time and Cost of Sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-21

QA/QC MEASURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-21

4.7.1 Sampling Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-214.7.2 Sampling Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-214.7.3 QC Samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-224.7.4 Collection Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-224.7.5 Sample Preservation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-22

SPECIAL ANALYTICAL SERVICES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-22

TAKING AN ACTIVE ROLE DURING WORKPLAN DEVELOPMENT ANDDATA COLLECTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-22

4.9.1 Present Risk Assessment Sampling Needs at Scoping Meeting . . . . . . . . . . . . 4-224.9.2 Contribute to Workplan and Review Sampling and Analysis Plan . . . . . . . . . 4-234.9.3 Conduct Interim Reviews of Field Investigation Outputs . . . . . . . . . . . . . . . 4-24

Page vii

CHAPTER 5 DATA EVALUATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

5.1

5.2

5.3

5.4

5.5

5.6

5.7

5.8

5.9

COMBINING DATA AVAILABLE FROM SITE INVESTIGATIONS

EVALUATION OF ANALYTICAL METHODS . . . . . . . . . . . . . . . . . . . . . . . .

EVALUATION OF QUANTITATION LIMITS . . . . . . . . . . . . . . . . . . . . . .

5.3.1

5.3.25.3.35.3.45.3.5

Sample Quantitation Limits (SQLs) That Are Greater ThanReference Concentrations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Unusually High SQLs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

5-1

5-2

5-5

5-7

5-75-10

When Only Some Samples in a Medium Test Positive for a Chemical . . . . . . . . . . 5-10When SQLs Are Not Available . . . . . . . . . . . . . . . . .. . . . . . . . .. . . . . . . . . . . . . . . . . 5 -11When Chemicals Are Not Detected in Any Samples in a Medium . . . . . . . . . 5-11

EVALUATION OF QUALIFIED AND CODED DATA . . . . . . . . . . . . . . . . . . . . . . .

5.4.1 Types of Qualifiers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5.4.2 Using the Appropriate Qualifiers . . . . . . . . . . . . . . . . . . . . . . . . . . . .

COMPARISON OF CONCENTRATIONS DETECTED IN BLANKS WITHCONCENTRATIONS DETECTED IN SAMPLES . . . . . . . . . . . . . . . . . . . . .

EVALUATION OF TENTATIVELY IDENTIFIED COMPOUNDS . . . . . . . . .

5.6.1 When Few TICs Are Present . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5.6.2 When Many TICs Are Present . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

COMPARISON OF SAMPLES WITH BACKGROUND . . . . . . . . . . . . . . . .

5.7.1 Use Appropriate Background Data . . . . . . . . . . . . . . . . . . . . . . . . . .5.7.2 Identify Statistical Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5.7.3 Compare Chemical Concentrations with Naturally Occurring Levels . . . . .5.7.4 Compare Chemical Concentrations with Anthropogenic Levels . . . . . . . .

DEVELOPMENT OF A SET OF CHEMICAL DATA AND INFORMATIONFOR USE IN THE RISK ASSESSMENT . . . . . . . . . . . . . . . . . . . . . . . . . . .

FURTHER REDUCTION IN THE NUMBER OF CHEMICALS (OPTIONAL)

5.9.15.9.25.9.35.9.45.9.5

Conduct Initial Activities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Group Chemicals by Class . . . . . . . . . . . . . . . . . . . . . . . . .Evaluate Frequency of Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Evaluate Essential Nutrients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Use a Concentration-Toxitity Screen . . . . . . . . . . . . . . . . . . . . . . . . . .

5.10 SUMMARY AND PRESENTATION OF DATA . . . . . . . . . . . . . . . . . . . . .

5-11

5-115-16

5-16

5-17

5-185-18

5-18

5-195-195-195-19

5-20

5-20

5-205-225-225-235-23

5-24

5-275.10.1 Summarize Data tillection and Evaluation Results in Text . . . . . . . . . .5.10.2 Summarize Data Collection and Evaluation Results in Tables and Graphics . . 5-27

Page viii

CHAPTER 6 EXPOSURE ASSESSMENT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-1

6.4

6.1

6.2

6.3

6.5

6.6

6.7

6.8

6.9

BACKGROUND . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-1

6.1.16.1.2

STEP

6.2.16.2.2

STEP

6.3.16.3.26.3.36.3.4

6.3.5

Components of an Exposure Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . 6-1Reasonable Maximum Exposure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-4

1: CHARACTERIZATION OF EXPOSURE SETTING . . . . . . . . . . . . . . . 6-5

Characterize Physical Setting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-5Characterize Potentially Exposed Populations . . . . . . . . . . . . . . . . . . . . . . . 6-6

2: IDENTIFICATION OF EXPOSURE PATHWAYS . . . . . . . . . . . . . . . . 6-8

Identify Sources and Receiving Media . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-8Evaluate Fate and Transport in Release Media . . . . . . . . . . . . . . . . . . . . . . 6-11Identify Exposure Points and Exposure Routes . . . . . . . . . . . . . . . . . . . . . . 6-11Integrate Information on Sources, Releases, Fate and Transport, ExposurePoints, and Exposure Routes Into Exposure Pathways . . . . . . . . . . . . . . . . . 6-17Summarize Information on All Complete Exposure Pathways . . . . . . . . . . . . 6-17

STEP 3: QUANTIFICATION OF EXOSURE: GENERALCONSIDERATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-19

6.4.1 Quantifying the Reasonable Maximum Exposure . . . . . . . . . . . . . . . . . . . . . 6-196.4.2 Timing Considerations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-23

QUANTIFICATION OF EXPOSURE: DETERMINATION OF EXPOSURECONCENTRATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-24

6.5.16.5.26.5.36.5.46.5.56.5.66.5.76.5.8

General Considerations for Estimating Exposure Concentrations . . . . . . . . . . 6-24Estimate Exposure Concentrations in Ground Water . . . . . . . . . . . . . . . . . . 6-26Estimate Exposure Concentrations in Soil . . . . . . . . . . . . . . . . . . . . . . . . . 6-27Estimate Exposure Concentrations in Air . . . . . . . . . . . . . . . . . . . . . . . . . . 6-28Estimate Exposure Concentrations in Surface Water . . . . . . . . . . . . . . . . . . 6-29Estimate Exposure Concentrations in Sediments . . . . . . . . . . . . . . . . . . . . . 6-30Estimate Chemical Concentrations in Food . . . . . . . . . . . . . . . . . . . . . . . . 6-31Summarize Exposure Concentrations for Each Pathway . . . . . . . . . . . . . . . . 6-32

QUANTIFICATION OF EXPOSURE: ESTIMATION OF CHEMICALINTAKE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-32

6.6.1 Calculate Ground-water and Surface Water Intakes . . . . . . . . . . . . . . . . . . . 6-346.6.2 Calculate Soil, Sediment, or Dust Intakes . . . . . . . . . . . . . . . . . . . . . . . . . . 6-396.6.3 Calculate Air Intakes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-436.6.4 Calculate Food Intakes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-43

COMBINING CHEMICAL INTAKES ACROSS PATHWAYS . . . . . . . . . . . . . . . 6-47

EVALUATING UNCERTAINTY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-47

SUMMARIZING AND PRESENTING THE EXPOSURE ASSESSMENTRESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-50

Page ix

CHAPTER 7 TOXICITY ASSESSMENT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-1

7.1

7.2

7.3

7.4

7.5

7.6

7.7

TYPES OF TOXICOLOGICAL INFORMATION CONSIDERED INTOXICITY ASSESSMENT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-3

7.1.1 Human Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-37.1.2 Animal Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-57.1.3 Supporting Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-5

TOXICITY ASSESSMENT FOR NONCARCINOGENIC EFFECTS . . . . . . . . . . . 7-5

7.2.17.2.27.2.37.2.47.2.57.2.67.2.7

Concept of Threshold . . . . . . . . . . . .Derivation of an Oral RfD(RfDo) . . .Derivation of an Inhalation RfD (RfD i)Derivation of a Subchronic RfD (RfDs)Derivation of a Developmental ToxicantOne-day and Ten-day Health AdvisoriesVerification of Reds . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . 7-6

. . . . . . . . . . . . . . . . . . . . . . . . . . . 7-6

. . . . . . . . . . . . . . . . . . . . . . . . . . . 7-8. . . . . . . . . . . . . . . . . . . . . . . . . 7-8RfD (RfD dt) . . . . . . . . . . . . . . . . . 7-9. . . . . . . . . . . . . . . . . . . . . . . . . . . 7-9. . . . . . . . . . . . . . . . . . . . . . . . 7-10

TOXICITY ASSESSMENT FOR CARCINOGENIC EFFECTS . . . . . . . . . . . . . 7-10

7.3.1 Concept of Nonthreshold Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-107.3.2 Assigning a Weight of Evidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-117.3.3 Generating a Slope Factor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-117.3.4 Verification of Slope Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-13

IDENTIFYING APPROPRIATE TOXICITY VALUES FORSITE RISK ASSESSMENT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-13

7.4.1 Gather Toxicity Information for Chemicals Being Evaluated . . . . . . . . . . . . . 7-137.4.2 Determine Toxicity Values for NonCarcinogenic Effects (RfDs) . . . . . . . . . . . 7-157.4.3 Determine Toxicity Values for Carcinogenic Effects (Slope Factors) . . . . . . . . 7-16

EVALUATING CHEMICALS FOR WHICH NO TOXICITY VALUES AREAVAILABLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-16

7.5.1 Route-to-Route Extrapolation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-167.5.2 Dermal Exposure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-167.5.3 Generation of Toxicity Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-16

UNCERTAINTIES RELATED TO TOXICITY INFORMATION . . . . . . . . . . . . . 7-19

SUMMARIZATION AND PRESENTATION OF THE TOXICITY INFORMATION 7-20

7.7.1 Toxicity Information for the Main Body of the Text . . . . . . . . . . . . . . . . . . 7-207.7.2 Toxicity Information for Inclusion in an Appendix . . . . . . . . . . . . . . . . . . . . 7-20

Pagc x

CHAPTER 8 RISK CHARACTERIZATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-1

8.1 REVIEW OF OUTPUTS FROM THE TOXICITY AND EXPOSUREASSESSMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-1

8.1.1 Gather and Organize Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-48.1.2 Make Final Consistency and Validity Check . . . . . . . . . . . . . . . . . . . . . . . . 8-4

8.2 QUANTIFYING RISKS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-6

8.2.1 Calculate Risks for Individual Substances . . . . . . . . . . . . . . . . . . . . . . . . . . 8-68.2.2 Aggregate Risks for Multiple Substances . . . . . . . . . . . . . . . . . . . . . . . . . . 8-11

8.3 COMBINING RISKS ACROSS EXPOSURE PATHWAYS. . . . . . . . . . . . . . . . . . 8-15

8.3.1 Identify Reasonable Exposure Pathway Combinations . . . . . . . . . . . . . . . . . 8-158.3.2 Sum Cancer Risks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-168.3.3 Sum Noncancer Hazard Indices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-16

8.4 ASSESSMENT AND PRESENTATION OF UNCERTAINTY . . . . . . . . . . . . . . . 8-17

8.4.1 Identify and Evaluate Important Site-Specific Uncertainty Factors . . . . . . . . . 8-178.4.2 Identify and Evaluate Toxicity Assessment Uncertainty Factors . . . . . . . . . . . 8-22

8.5 CONSIDERATION OF SITE-SPECIFIC HUMAN STUDIES . . . . . . . . . . . . . . . 8-22

8.5.1 Compare with ATSDR Health Assessment . . . . . . . . . . . . . . . . . . . . . . . . 8-248.5.2 Compare with Other Available Site-Specific Epidemiological or Health Studies 8-24

8.6 SUMMARIZATION AND PRESENTATION OF THE BASELINE RISKCHARACTERIZATION RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-25

8.6.1 Summarize Risk Information in Text. . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-258.6.2 Summarize Risk Information in Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-26

CHAPTER 9 DOCUMENTATION, REVIEW, AND MANAGEMENT TOOLS FOR THE RISK

9.1

9.2

9.3

ASSESSOR, REVIEWER AND MANAGER . . . . . . . . . . . . . . . . . . . . . . 9-1

DOCUMENTATION TOOLS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-1

9.1.1 Basic Principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-19.1.2 Baseline Risk Assessment Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-29.1.3 0ther Key Reports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-3

REVIEW TOOLS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-3

MANAGEMENT TOOLS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-14

Pagc xi

CHAPTER 10 RADIATION RISK ASSESSMENT GUIDANCE . . . . . . . . . . . . . . . . . . . . . . .10-1

10.1

10.2

10.3

10.4

10.5

10.6

RADIATION PROTECTION PRINCIPLES AND CONCEPTS . . . . . . . . . . . . . . 10-3

REGULATION OF RADIOACTIVELY CONTAMINATED SITES . . . . . . . . . . 10-8

DATA COLLECI’ION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-10

10.3.110.3.210.3.310.3.410.3.510.3.610.3.7

Radiation Detection Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-10Reviewing Available Site Information. . . . . . . . . . . . . . . . . . . . . . . . . . . 10-14Addressing Modeling Parameter Needs . . . . . . . . . . . . . . . . . . . . . . . . . . 10-14Defining Background Radiation Sampling Needs . . . . . . . . . . . . . . . . . . . 10-14Preliminary Identification of Potential Exposure . . . . . . . . . . . . . . . . . . . . 10-15Developing a Strategy for Sample Collection . . . . . . . . . . . . . . . . . . . . . . 10-15Quality Assurance and Quality Control (QA/QC) Measures . . . . . . . . . . . . 10-16

DATA EVALUATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-16

10.4.110.4.210.4.310.4.410.4.5

10.4.610.4.710.4.8

10.4.9

Combining Data from Available Site Investigations . . . . . . . . . . . . . . . . . .Evaluating Analytical Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Evaluating Quantitation Limits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Evaluating Qualified and Coded Data. . . . . . . . . . . . . . . . . . . . . . . . . . .Comparing Concentrations Detected in Blanks with ConcentrationsDetected in samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Evaluating Tentatively Identified Radionuclides . . . . . . . . . . . . . . . . . . . .Comparing Samples with Background. . . . . . . . . . . . . . . . . . . . . . . . . . .Developing a Set of Radionuclide Data and Information forUse in a Risk Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Grouping Radionuclides bv Class. . . . . . . . . . . . . . . . . . . . . . . . . . . .

10-1710-1710-1710-20

10-2010-2110-21

10-2110-21

10.4.10 Further Reduction in the Number of Radionuclides . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-2110.4.11 Summarizing and Presenting Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-22

EXPOSURE AND DOSE ASSESSMENT . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-22

10.5.110.5.210.5.310.5.410.5.510.5.610.5.710.5.8

Characterizing the Exposure Setting.. . . . . . . . . . . . . . . . . . . . . . . . . . .Identifying Exposure Pathways . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Quantifying Exposure General Considerations . . . . . . . . . . . . . . . . . . . .Quantifying Exposure Determining Exposure Point Concentrations . . . . . .Quantizing Exposure Estimating Intake and Dose Equivalent . . . . . . . . .Combining Intakes and Doses Across Pathways . . . . . . . . . . . . . . . . . . . .Evaluating Uncertainty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Summarizing and Presenting Exposure Assessment Results . . . . . . . . . . . .

10-2310-2310-2410-2510-2610-2710-2710-27

TOXICITY ASSESSMENT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . l0-27

10.6.1 Hazard Identification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-2810.6.2 Dose-Response Relationships . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-30

Page xii

10.7

10.8

PART B

PART C

RISK CHARACI’ERIZATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

10.7.1 Reviewing Outputs from the Toxicity and Exposure Assessments . . . . . .10.7.2 Quantifying Risks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10.7.3 Combining Radionuclide and Chemical Cancer Risks . . . . . . . . . . . . . . . .10.7.4 Assessing and Presenting Uncertainties . . . . . . . . . . . . . . . . . . . . . . . . . .10.7.5 Summarizing and Presenting the Baseline Risk Characterization Results . . .

DOCUMENTATION, REVIEW, AND MANAGEMENT TOOLS FOR THE RISKASSESSOR, REVIWER, AND MANAGER . . . . . . . . . . . . . . . . . . . . . . . . . . .

- - REFINEMENT OF PRELIMINARY REMEDIATION GOALS[Reserved]

RISK EVALUATION OF REMEDIAL ALTERNATIVES[Reserved]

- -

APPENDICES

APPENDIX A ADJUSTMENTS FOR ABSORPTION EFFICIENCY . . . . . . . . . . . . . . . . . .

A.1 ADJUSTMENTS OF TOXICITY VALUE FROM ADMINISTERED TOABSORBED DOSE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

A.2 ADJUSTMENT OF EXPOSURE ESTIMATE TO AN ABSORBED DOSE . . . . .

A.3 ADJUSTMENT FOR MEDIUM OF EXPOSURE . . . . . . . . . . . . . . . . . . . . . . .

APPENDIX B INDEX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

10-32

10-3210-3210-3310-3310-34

10-34

A-1

A-1

A-3

A-3

B-1

Page xiii

Exhibit

1-1 1-2

2-1 2-2

4-1 4-2

5-15-2 5-3

5-4

5-5

5-6

5-7

6-16-26-3

6-46-5

6-6 6-76-86-96-106-11

6-126-136-146-15

6-166-17

6-18

LIST OF EXHIBITS

Risk Information Activities in the RI/FS Process . . . . . . . . . . . . . . . . . . . . .Part A: Baseline Risk Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Relationship of Documents Govening Human Health Evaluation . . . . . . . . . . . . . .Role of the Human Health Evaluation in the Superfund Remedial Process . . . . . . .

Elements of a Conceptual Evaluation Model . . . . . . . . . . . . . . . . . . . . . . . . . . . .Examples of Modeling Parameters for Which Information May Need To BeObtained During a Site Sampling Investigation . . . . . . . . . . . . . . . . . . . . . . . . . .

Data Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Example of Output Format for Validated Data . . . . . . . . . . . . . . . . . . . . . . . . . .Examples of the Types of Data Potentially Unsuitable for a QuantitativeRisk Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .CLP Laboratory Data Qualfiers and Their Potential Use in QuantitativeRisk Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Validation Data Qualifers and Their Potential Use in QuantitativeRisk Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Example of Table Format for Presenting Chemicals Sampled inSpecific Media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Example of Table Format for Summarizing Chemicals of PotentialConcern in All Media Sampled . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

The Exposure Assessment Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Illustration of Exposure Pathways . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Common Chemical Release Sources at Sites in the Absence ofRemedial Action . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Important Physical/Chemical and Environmenal Fate Parameters . . . . . . . . . . . . . .Important Considerations for Determining the Environmental Fate andTransport of the Chemicals of Potential Concern at a Superfund Site . . . . . . . . . . .Flow Chart for Fate and Transport Assessments . . . . . . . . . . . . . . . . . . . . . . . . . .Matrix of Potential Exposure Routes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Example of Table Format for Summarizing Complete Exposure Pathways at a Site . .Generic Equation for Calculating Chemical Intakes . . . . . . . . . . . . . . . . . . . . . . .Example of Table Format for Summarizing Exposure Concentrations . . . . . . . . . . .Residential Exposure Ingestion of Chemicals in Drinking Water(and Beverages Made Using Drinking Water) . . . . . . . . . . . . . . . . . . . . . . . . . . .Residential Exposure:Residential Exposure:Residential Exposure:Residential Exposure:Residential Exposure:Residential Exposure:and Shellfish . . . . . .Residential Exposure:Fruits and Vegetables

Ingestion of Chemicals in Surface Water While Swimming . . .Dermal Contact with Chemicals in Water . . . . . . . . . . . . . .ingestion of Chemicals in Soil . . . . . . . . . . . . . . . . . . . . . .Dermal Contact with Chemicals in Soil . . . . . . . . . . . . . . . .Inhalation of Airborne (Vapor Phase) Chemicals . . . . . . . . .Food Pathway -- Ingestion of Contaminated Fish. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Food Pathway -- Ingestion of Contaminated. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1-51-7

2-22-6

4-6

4-7

5-35-4

5-6

5-12

5-14

5-26

5-26

6-36-9

6-106-12

6-136-146-186-206-216-33

6-356-366-37

6-41

6-45

6-40

6-44

6-46

Page xiv

6-19

6-20

6-216-22

7-17-27-3

8-18-28-38-48-5

8-6

8-7

8-8

9-19-29-3

10-110-210-310-4

10-5

Residential Exposure: Food Pathway -- Ingestion of ContaminatedMeats, Eggs, and Dairy Products. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Example of Table Format for Summarizing Values Used to EstimateExposure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Example of Uncertainty Table for Exposure Assessment . . . . . . . . . . . . . . . . . . .Example of Table Format for Summarizing the Results of theExposure Assessment -- Current Land Use . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Steps in Toxicity Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Example of Table Format for Toxicity Values: Potential Noncarcinogenic Effects . .Example of Table Format for Toxicity Values: Potential Carcinogenic Effects . . . .

6-48

Steps in Risk Characterization . . . . . . . . . . . . . . . . . . . . . . . .Example of Table Format for Cancer Risk Estimates . . . . . . . .Example of Table Format for Chronic Hazard Index Estimates . .Example of Table Format for Subchronic Hazard Index EstimatesExample of Presentation of Impact of Exposure Assumptions onCancer Risk Estimate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Example of Presentation of Impact of Exposure Assumptions onHazard Index Estimate . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Example of Presentation of Relative Contribution of IndividualChemicals to Exposure Pathway and Total Cancer Risk EstimatesExample of Presentation of Relative Contribution of Individual

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .. . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . .

Chemicals to Exposure Pathway and Total Hazard Index Estimates . . . . . . . . . . . .

Suggested Outline for a Baseline Risk Assessment Report . . . . . . . . . . . . . . . . . .Reviewer Checklist . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Checklist for Manager Involvement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Radiological Characteristics of Selected Radionuclides Found at Superfund Sites . .Typesof Field Radiation Detection Instruments . . . . . . . . . . . . . . . . . . . . . . . . .Types of Laboratory Radiation Detection Instruments . . . . . . . . . . . . . . . . . . . . .Examples of Lower Limits of Detection (LLD) For Selected RadionuclidesUsing Standard Analytical Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Summary of EPA’s Radiation Risk Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . .

6-496-51

6-52

7-47-177-18

8-38-78-88-9

8-21

8-23

8-27

8-28

9-49-9

9-15

10-510-1110-13

10-1810-31

Page xv

PREFACE

The Comprehensive Environmental Response,Compensation, and Liability Act (CERCLA)requires that actions selected to remedy hazardouswaste sites be protective of human health and theenvironment. CERCLA also mandates that whena remedial action results in residual contaminationat a site, future reviews must be planned andconducted to assure that human health and theenvironment continue to be protected. As part ofits effort to meet these and other CERCI.Arequirements, EPA has developed a set ofmanuals, together entitled Risk AssessmentGuidance for Superfund. The Human HealthEvaluation Manual (Volume I) provides guidancefor developing health risk information atSuperfund sites, while the EnvironmentalEvaluation Manual (Volume II) provides guidancefor environmental assessment at Superfund sites.Guidance in both human health evaluation andenvironmental assessment is needed so that EPAcan fulfill CERCLA’S requirement to protecthuman health and the environment.

The Rirk Assessment Guidance for Superfundmanuals were developed to be used in theremedial investigation/feasibility study (RI/FS)process at Superfund sites, although the analyticalframework and specific methods described in themanuals may also be applicable to otherassessments of hazardous wastes and hazardousmaterials. These manuals are companiondocuments to EPA’s Guidance for ConductingRemedial Investigations and Feasibility StudiesUnder CERCLA (October 1988), and users shouldbe familiar with that guidance. The twoSuperfund risk assessment manuals were developedwith extensive input from EPA workgroupscomprised of both regional and headquarters staff.These manuals are interim final guidance, finalguidance will be issued when the revisionsproposed in December 1988 to the National Oiland Harzardous Substances Pollution ContingencyPlan (NCP) become final.

Although human health risk assessment andenvironmental assessment are different processes,they share certain common information needs andgenerally can use some of the same chemical

sampling and environmental setting data for a site.Planning for both assessments should begin duringthe scoping stage of the RI/FS, and site samplingand other data collection activities to support thetwo assessments should be coordinated. Anexample of this type of coordination is thesampling and analysis of fish or other aquaticorganism, if done properly, data from suchsampling can be used in the assessment of humanhealth risks from ingestion and in the assessmentof damages to and potential effects on the aquaticecosystem.

The two manuals in this set target somewhatdifferent audiences. The Environmental EvaluationManual is addressed primarily to remedial projectmanagers (RPMs) and on-scene coordinators(OSCs), who are responsible for ensuring athorough evaluation of potential environmentaleffects at sites. The Environmental EvaluationManual is not a detailed “how-to” type ofguidance, and it does not provide “cookbook”approaches for evaluation. Instead, it identifiesthe kinds of help that RPMs/OSCs are likely toneed and where they may find that help. Themanual also provides an overall framework to beused in considering environmental effects. Anenvironmental evaluation methods compendiumpublished by EPA’s Office of Research andDevelopment, Ecological Assessments of HazardousWaste Sites: A Field and Laboratory ReferenceDocument (EPA/600/3-89/013), is an importantreference to be used with the manual.

The Human Health Evaluation Manual isaddressed primarily to the individuals actuallyconducting healh risk assessments for sites, whofrequently are contractors to EPA other federalagencies, states, or potentially responsible parties.It also is targeted to EPA staff, including thoseresponsible for review and oversight of riskassessments (e.g., technical staff in the regions)and those responsible for ensuring adequateevaluation of human health risks (i.e., RPMs).The Human Health Evaluation Manual replaces aprevious EPA guidance document, The SuperfundPublic Health Evaluation Manual (October 1986),which should no longer be used. The new manual

Page xvi

incorporates lessons learned from application ofthe earlier manual and addresses a number ofissues raised since the earlier manual’s publication.Issuance of the new manual does not invalidatehuman health risk assessments completed before(or in progress at) the publication date.

The Human Health Evaluation Manualprovides a basic framework for health riskassessment at Superfund sites, as theEnvironental Evaluation Manual does for

Evaluation Manual differs, however, by providingmore detailed guidance on many of the proceduresused to assess health risk. This additional levelof detail is possible because of the relatively largebody of information, techniques, and guidanceavailable on human health risk assessment and theextensive Superfund program experienceconducting such assessments for sites, Eventhough the Human Health Evaluation Manual isconsiderably more specific than the EnvironmentalEvaluation Manual, it also is not a "cookbook,"

environmental assessment. The Human Health and proper application of the guidance requiressubstantial expertise and professional judgment.

ACKNOWLEDGEMENTS

This manual was developed by the Toxics Integration Branch (TIB) of EPA’s Office of Emergencyand Remedial Response, Hazardous Site Evaluation Division. Linda Cullen provided overall project

management, contract supervision, and technical coordination for the project under the direction of BruceMeans, Chief of TIB’s Health Effects Program.

The EPA Workgroup (comprised of members listed on page iv) provided valuable input regarding the organization, content, and policy implications of the manual throughout its development. The projectmanager especially wishes to acknowledge the assistance of the Workgroup Subcommittee Chairpersons:Rebecca Madison, Bruce Means, Sue Norton, Georgia Valaoras, Craig Zamuda, and Larry Zaragoza.

Other significant contributors to the manual included Joan Fisk, Michael Hurd, and Angelo Caraseaof the Analytical Operations Branch (Office of Emergency and Remedial Response); Paul White, AnneSergeant, and Jacqueline Moya of the Exposure Assessment Group (Office of Research and Development);and Barnes Johnson of the Statistical Policy Branch (Office of Policy, Planning, and Evaluation). Inaddition, many thanks are offered to the more than 60 technical and policy reviewers who providedconstructive comments on the document in its final stages of development.

ICF Incorporated provided technical assistance to EPA in support of the development of this manual,under Contract No. 68-01-7389.

Robert Dyer, Chief of the Environmental Studies and Statistics Branch, Office of Radiation Programs,served as project manager for Chapter 10 (Radiation Risk Assessment Guidance), with assistance from staffin the Bioeffects Analysis Branch and the regional Radiation Program Managers. Chapter 10 was preparedby S. Cohen and Associates, Incorporated (SC&A), under Contract No. 68-02-4375.

CHAPTER 1

INTRODUCTION

The Comprehensive Environmental Response,Compensation, and Liability Act of 1980, asamended (CERCLA, or “Superfundn), establishesa national program for responding to releases ofhazardous substances into the environment.1 T h eNational Oil and Hazardous Substances PollutionContingency Plan (NCP) is the regulation thatimplements CERCLA.2 Among other things, theNCP establishes the overall approach fordetermining appropriate remedial actions atSuperfund sites. The overarching mandate of theSuperfund program is to protect human healthand the environment from current and potentialthreats posed by uncontrolled hazardous substancereleases, and the NCP echoes this mandate.

To help meet this Superfund mandate, EPA’sOffice of Emergency and Remedial Response hasdeveloped a human health evaluation process aspart of its remedial response program. Theprocess of gathering and assessing human healthrisk information described in this manual isadapted from well-established chemical riskassessment principles and procedures (NAS 1983;CRS 1983; OSTP 1985). It is designed to beconsistent with EPA’s published risk assessmentguidelines (EPA 1984 EPA 1986a-e; EPA 1988a;EPA 1989a) and other Agency-wide riskassessment policy. The Human Health EvaluationManual revises and replaces the Superfund PublicHealth Evaluation Manual (EPA 1986f).3 Itincorporates new information and builds onseveral years of Superfund program experienceconducting risk assessments at hazardous wastesites. In addition, the Human Health EvaluationManual together with the companionEnvtionmental Evaluation Manual (EPA 1989b)replaces EPA’s 1985 Endangerment AssessmentHandbook, which should no longer be used (seeSection 2.2.1).

The goal of the Superfund human healthevaluation process is to provide a framework fordeveloping the risk information necessary to assistdecision-making at remedial sites. Specificobjectives of the process are to:

provide an analysis of baseline risks4

and help determine the need for actionat sites;

provide a basis for determining levelsof chemicals that can remain onsite andstill be adequately protective of publichealth;

provide a basis for comparing potentialhealth impacts of various remedialalternatives; and

provide a consistent process forevacuating and documenting public healththreats at sites.

The human health evaluation processdescribed in this manual is an integral part of theremedial response process defined by CERCLAand the NCP. The risk information generated bythe human health evaluation process is designedto be used in the remedial investigation/feasibilitystudy (RI/FS) at Superfund sites. Although riskinformation is fundamental to the RI/FS and tothe remedial response program in general,Superfund site experience has led EPA to balancethe need for information with the need to takeaction at sites quickly and to streamline theremedial process. Revisions proposed to the NCPin 1988 reflect EPA program managementprinciples intended to promote the efficiency andeffectiveness of the remedial response process.Chief among these principles is a bias for action.EPA’s Guidance for Conducting Remedial

2

Investigations and Feasibility Studies UnderCERCLA (EPA 1988b) also was revised in 1988to incorporate management initiatives designed tostreamline the RI/FS process and to makeinformation collection activities during the RImore efficient. The Risk Assessment Guidance forSuperfund, of which this Human Health EvaluationManual is Volume I,5 has been developed toreflect the emphasis on streamlining the remedialprocess. The Human Health Evaluation Manualis a companion document to the RI/FS guidance.It provides a basic framework for developinghealth risk information at Superfund sites and alsogives specific guidance on appropriate methodsand data to use. Users of the Human HealthEvaluation Manual should be familiar with theRI/FS guidance, as well as with other guidancesreferenced throughout later chapters of thismanual.

The Human Health Evaluation Manual isaddressed primarily to the individuals actuallyconducting human health evaluations for sites(frequently contractors to EPA, other federalagencies, states, or potentially responsible parties).It also is targeted to EPA staff responsible forreview and oversight of risk assessments (e.g.,technical staff in the regions) and thoseresponsible for ensuring an adequate evaluation ofhuman health risks (i.e., remedial projectmanagers, or RPMs). Although the terms riskassessor and risk assessment reviewer are used inthis manual, it is emphasized that they generallyrefer to teams of individuals in appropriatedisciplines (e.g., toxicologists, chemists,hydrologists, engineers). It is recommended thatan appropriate team of scientists and engineers beassembled for the human health evaluation ateach specific site. It is the responsibility ofRPMs, along with the leaders of human healthevaluation teams, to match the scientific supportthey deem appropriate with the resources at theirdisposal.

Individuals having different levels of scientifictraining and experience are likely to use themanual in designing, conducting, and reviewing

human health evaluations. Because assumptionsand judgments are required in many parts of theanalysis, the individuals conducting the evaluationare key elements in the process. The manual isnot intended to instruct non-technical personnelhow to perform technical evaluations, nor to allow

professionals trained in one discipline to performthe work of another.

The Human Health Evaluation Manualadmittedly cannot address all site circumstances.Users of the manual must exercise technical andmanagement judgment, and should consult withEPA regional risk assessment contacts andappropriate headquarter staff when encounteringunusual or particularly complex technical issues.

The first three chapters of this manualprovide background information to help place thehuman health evaluation process in the context ofthe Superfund remedial process. This chapter(Chapter 1) summarizes the human healthevaluation process during the RI/FS. The threemain parts of this process -- baseline riskassessment, refinement of preliminary remediationgoals, and remedial alternatives risk evaluation-- are described in detail in subsequent chapters.Chapter 2 discusses in a more general way therole of risk information in the overall Superfund

3

remedial program by focusing on the statutes,regulations, and guidance relevant to the humanhealth evaluation. Chapter 2 also identifies andcontrasts Superfund studies related to the humanhealth evaluation. Chapter 3 discusses issuesrelated to planning for the human healthevaluation.

1.1 OVERVIEW OF THE HUMANHEALTH EVALUATIONPROCESS IN THE RI/FS

Section 300.430 of the proposed revised NCPreiterates that the purpose of the remedial processis to implement remedies that reduce, control, oreliminate risks to human health and theenvironment. The remedial investigation andfeasibility study (RI/FS) is the methodology thatthe Superfund program has established forcharacterizing the nature and extent of risks posedby uncontrolled hazardous waste sites and fordeveloping and evaluating remedial options. The1986 amendments to CERCLA reemphasized theoriginal statutory mandate that remedies meet athreshold requirement to protect human healthand the environment and that they be cost-effective, while adding new emphasis to thepermanence of remedies. Because the RI/FS is ananalytical process designed to support riskmanagement decision-making for Superfund sites,the assessment of health and environmental riskplays an essential role in the RI/FS.

This manual provides guidance on the humanhealth evaluation activities that are conductedduring the RI/FS. The three basic parts of theRI/FS human health evaluation are:

(1) baseline risk assessment (described inPart A of this manual);

(2) refinement of preliminary remediationgoals (Part B); and

(3) remedial alternatives risk evaluation(Part C).

Because these risk information activities areintertwined with the RI/FS, this section describesthose activities in the context of the RI/FSprocess. It relates the three parts of the human

health evaluation to the stages of the RI/FS,which are

project scoping (before the RI);

site characterization (RI);

establishment of remedial actionobjectives (FS);

development and screening ofalternatives (FS); and

detailed analysis of alternatives (FS).

Although the RI/FS process and related riskinformation activities are presented in a fashionthat makes the steps appear sequential anddistinct, in practice the process is highlyinteractive. In fact, the RI and FS are conductedconcurrently. Data collected in the RI influencesthe development of remedial alternatives in theFS, which in turn affects the data needs and scopeof treatability studies and additional fieldinvestigations. The RI/FS should be viewed as aflexible process that can and should be tailored tospecific circumstances and information needs ofindividual sites, not as a rigid approach that mustbe conducted identically at every site. Likewise,the human health evaluation process describedhere should be viewed the same way.

Two concepts are essential to the phasedRI/FS approach. First, initial data collectionefforts develop a general understanding of the site.Subsequent data collection effort focuses on fillingpreviously unidentified gaps in the understandingof site characteristics and gathering informationnecessary to evaluate remedial alternatives.Second, key data needs should be identified asearly in the process as possible to ensure. thatdata collection is always directed toward providinginformation relevant to selection of a remedialaction. In this way, the overall sitecharacterization effort can be continually scopedto minimize the collection of unnecessary data andmaximize data quality.

The RI/FS provides decision-makers with atechnical evaluation of the threats posed at a site,a characterization of the potential routes ofexposure, an assessment of remedial alternatives(including their relative advantages and

4

disadvantages), and an analysis of the trade-offs inselecting one alternative over another. EPA’sinterim final Guidance for Conducting RemedialInvestigations and Feasibility Studies underCERCLA (EPA 1988b) provides a detailedstructure for the RI/FS. The RI/FS guidanceprovides further background that is helpful inunderstanding the place of the human healthevaluation in the RI/FS process. The role thatrisk information plays in these stages of the RI/FSis described below, additional background can befound in the RI/FS guidance and in a summary ofthe guidance found in Chapter 2. Exhibit 1-1illustrates the RI/FS process, showing where in theprocess risk information is gathered and analyzed.

1.1.1 PROJECT SCOPING

The purpose of project scoping is to definemore specifically the appropriate type and extentof investigation and analysis that should beundertaken for a given site. During scoping, toassist in evaluating the possible impacts of releasesfrom the site on human health and theenvironment, a conceptual model of the siteshould be established, considering in a qualitativemanner the sources of contamination, potentialpathways of exposure, and potential receptors.(Scoping is also the starting point for the riskassessment, during which exposure pathways areidentified in the conceptual model for furtherinvestigation and quantification.)

PROJECT SCOPING

Program experience has shown that scoping is a veryimportant step for the human health evaluation process,and both the health and environmental evaluation teamsneed to get involved in the RI/FS during the scopingstage. Planning for site data collection activities isnecessary to focus the human health evaluation (andenvironmental evaluation) on the minimum amount ofsampling information in order to meet time and budgetconstraints, while at the same time ensuring that enoughinformation is gathered to assess risks adequately. (SeeChapter 3 for information on planning the human healthevaluation.)

The preliminary characterization duringproject seeping is initially developed with readilyavailable information and is refined as additionaldata are collected. The main objectives of scopingare to identify the types of decisions that need tobe made, to determine the types (includingquantity and quality) of data needed, and todesign efficient studies to collect these data.Potential site-specific modeling activities shouldbe discussed at initial scoping meetings to ensurethat modeling results will supplement the samplingdata and effectively support risk assessmentactivities.

1.1.2 SITE CHARACTERIZATION (RI)

During site characterization, the sampling andanalysis plan developed during project scoping isimplemented and field data are collected andanalyzed to determine the nature and extent ofthreats to human health and the environmentposed by a site. The major components of sitecharacterization are

collection and analysis of field data tocharacterize the site

development of a baseline riskassessment for both potential humanhealth effects and potentialenvironmental effects; and

treatability studies, as appropriate.

Part of the human health evaluation, thebaseline risk assessment (Part A of this manual)is an analysis of the potential adverse healtheffects (current or future) caused by hazardoussubstance releases from a site in the absence ofany actions to control or mitigate these releases(i.e., under an assumption of no action). Thebaseline risk assessment contributes to the sitecharacterization and subsequent development,evaluation, and selection of appropriate responsealternatives. The results of the baseline riskassessment are used to:

help determine whether additionalresponse action is necessary at the site;

modify preliminary remediation goals

5

EXHIBIT 1-1

RISK INFORMATION ACTIVITIES IN THE RI/FS PROCESS

6

help support selection of the "no-action"remedial alternative, where appropriate;and

document the magnitude of risk at asite, and the primary causes of that risk.

Baseline risk assessments are site-specific andtherefore may vary in both detail and the extentto which qualitative and quantitative analyses areused, depending on the complexity and particularcircumstances of the site, as well as the availabilityof applicable or relevant and appropriaterequirements (ARARs) and other criteria,advisories, and guidance. After an initial planningstage (described more fully in Chapter 3), thereare four steps in the baseline risk assessmentprocess: data collection and analysis; exposureassessment toxicity assessment; and riskcharacterization. Each step is described brieflybelow and presented in Exhibit 1-2.

Data collection and evaluation involvesgathering and analyzing the site data relevant tothe human health evacuation and identifying thesubstances present at the site that are the focusof the risk assessment process. (Chapters 4 and5 address data collection and evaluation.)

An exposure assessment is conducted toestimate the magnitude of actual and/or potentialhuman exposures, the frequency and duration ofthese exposures, and the pathways by whichhumans are potentially exposed. In the exposureassessment, reasonable maximum estimates ofexposure are developed for both current andfuture land-use assumptions. Current exposureestimates are used to determine whether a threatexists based on existing exposure conditions at thesite. Future exposure estimates are used toprovide decision-makers with an understanding ofpotential future exposures and threats and includea qualitative estimate of the likelihood of suchexposures occurring. Conducting an exposureassessment involves analyzing contaminantreleases; identifying exposed populations;identifying all potential pathways of exposure;estimating exposure point concentrations forspecific pathways, based both on environmentalmonitoring data and predictive chemical modelingresults; and estimating contaminant intakes forspecific pathways. The results of this assessmentare pathway-specific intakes for current and future

exposures to individual substances. (Chapter 6addresses exposure assessment.)

The toxicitv assessment component of theSuperfund baseline risk assessment considers: (1)the types of adverse health effects associated withchemical exposures; (2) the relationship betweenmagnitude of exposure and adverse effects; and (3)related uncertainties such as the weight ofevidence of a particular chemical’s carcinogenicityin humans. Typically, the Superfund site riskassessments rely heavily on existing toxicityinformation developed on specific chemicals.Toxicity assessment for contaminants found atSuperfund sites is generally accomplished in twosteps: hazard identification and dose-responseassessment. The first step, hazard identification,is the process of determining whether exposure toan agent can cause an increase in the incidence ofan adverse health effect (e.g., cancer, birth defect).Hazard identification also involves characterizingthe nature and strength of the evidence ofcausation. The second step, dose-responseevaluation, is the process of quantitativelyevaluating the toxicity information andcharacterizing the relationship between the doseof the contaminant administered or received andthe incidence of adverse health effects in theexposed population. From this quantitative dose-response relationship, toxicity values are derivedthat can be used to estimate the incidence ofadverse effects occurring in humans at differentexposure levels. (Chapter 7 addresses toxicityassessment.)

The risk characterization summarizes andcombines outputs of the exposure and toxicityassessments to characterize baseline risk, both inquantitative expressions and qualitative statements.During risk characterization, chemical-specifictoxicity information is compared against bothmeasured contaminant exposure levels and thoselevels predicted through fate and transportmodeling to determine whether current or futurelevels at or near the site are of potential concern.(Chapter 8 addresses risk characterization.)

The level of effort required to conduct abaseline risk assessment depends largely on thecomplexity of the site. In situations where theresults of the baseline risk assessment indicatethat the site poses little or no threat to humanhealth or the environment and that no further (or

7

EXHIBIT 1-2

PART A BASELINE RISK ASSESSMENT

8

limited) action will be necessary, the FS should bescaled-down as appropriate.

The documents developed during sitecharacterization include a brief preliminary sitecharacterization summary and the draft RI report,which includes either the complete baseline riskassessment report or a summary of it. Thepreliminary site characterization summary may beused to assist in identification of ARARs and mayprovide the Agency for Toxic Substances andDisease Registry (ATSDR) with the data necessaryto prepare its health assessment (different frombaseline risk assessment or other EPA humanhealth evaluation activities; see Chapter 2). Thedraft RI report is prepared after the completionof the baseline risk assessment, often along withthe draft FS report.

1.1.3 FEASIBILITY STUDY

The purpose of the feasibility study is toprovide the decision-maker with an assessment ofremedial alternatives, including their relativestrengths and weaknesses, and the trade-offs inselecting one alternative over another. The FSprocess involves developing a reasonable range ofalternatives and analyzing these alternatives indetail using nine evaluation criteria. Because theRI and FS are cortducted concurrently, thisdevelopment and analysis of alternatives is aninteractive process in which potential alternativesand remediation goals are continually refined asadditional information from the RI becomesavailable.

Establishing protective remedial actionobjectives. The first step in the FS processinvolves developing remedial action objectives thataddress contaminants and media of concern,potential exposure pathways, and preliminaryremediation goals. Under the proposed revisedNCP and the interim RI/FS guidance, preliminaryremediation goals typically are formulated firstduring project scoping or concurrent with initialRI activities (i.e., prior to completion of thebaseline risk assessment). The preliminary

remediation goals are therefore based initially onreadily available chemical-specific ARARs (e.g.,maximum contaminant levels (MCLS) for drinkingwater). Preliminary remediation goals forindividual substances are refined or confirmed atthe conclusion of the baseline risk assessment

(Part B of this manual addresses the refinementof preliminary remediation goals). These refinedpreliminary remediation goals are based both onrisk assessment and on chemical-specific ARARs.Thus, they are intended to be protective and tocomply with ARARs. The analytical approachused to develop these refined goals involves:

identifying chemical-specific ARARs;

identifying levels based on riskassessment where chemical-specificARARs are not available or situationswhere multiple contaminants or multipleexposure pathways make ARARs notprotective

identifying non-substance-specific goalsfor exposure pathways (if necessary); and

determining a refined preliminaryremediation goal that is protective ofhuman health for all substance/exposurepathway combinations being addressed.

Development and screening of alternatives.Once remedial action objectives have beendeveloped, general response actions, such astreatment, containment, excavation, pumping, orother actions that may be taken to satisfy thoseobjectives should be developed. In the process ofdeveloping alternatives for remedial action at asite, two important activities take place. First,volumes or areas of waste or environmental mediathat need to be addressed by the remedial actionare determined by information on the nature andextent of contamination, ARARs, chemical-specificenvironmental fate and toxicity information, andengineering analyses. Second, the remedial actionalternatives and associated technologies arescreened to identify those that would be effectivefor the contaminants and media of interest at thesite. The information developed in these twoactivities is used in assembling technologies intoalternatives for the site as a whole or for aspecific operable unit.

The Superfund program has long permittedremedial actions to be staged through multipleoperable units. Operable units are discreteactions that comprise incremental steps toward thefinal remedy. Operable units may be actions thatcompletely address a geographical portion of a site

9

or a specific site problem (e.g., drums and tanks,contaminated ground water) or the entire site.Operable units include interim actions (e.g.,pumping and treating of ground water to retardplume migration) that must be followed bysubsequent actions to fully address the scope ofthe problem (e.g., final ground-water operableunit that defines the remediation goals andrestoration timeframe). Such operable units maybe taken in response to a pressing problem thatwill worsen if unaddressed, or because there is anopportunity to undertake a limited action that willachieve significant risk reduction quickly. Theappropriateness of dividing remedial actions intooperable units is determined by considering theinterrelationship of site problems and the need ordesire to initiate actions quickly. To the degreethat site problems are interrelated, it may be mostappropriate to address the problems together.However, where problems are reasonablyseparable, phased responses implemented througha sequence of operable units may promote morerapid risk reduction.

In situations where numerous potentialremedial alternatives are initially developed, it maybe necessary to screen the alternatives to narrowthe list to be evaluated in detail. Such screeningaids in streamlining the feasibility study whileensuring that the most promising alternatives arebeing considered.

Detailed analysis of alternatives. During thedetailed analysis, each alternative is assessedagainst specific evaluation criteria and the resultsof this assessment arrayed such that comparisonsbetween alternatives can be made and key trade-offs identified. Nine evaluation criteria, some ofwhich are related to human health evaluation andrisk, have been developed to address statutoryrequirements as well as additional technical andpolicy considerations that have proven to beimportant for selecting among remedialalternatives. These evaluation criteria, which areidentified and discussed in the interim final RI/FSguidance, serve as the basis for conducting thedetailed analyses during the FS and forsubsequently selecting an appropriate remedialaction. The nine evaluation criteria are asfollows:

(1) overall protection of human health andthe environment;

(2) compliance with ARARs (unless waiverapplicable);

(3) long-term effectiveness and permanence;

(4) reduction of toxicity, mobility, or volumethrough the use of treatment;

(5) short-term effectiveness;

(6) implementability

(7) cost;

(8) state acceptance; and

(9) community acceptance.

Risk information is required at the detailedanalysis stage of the RI/FS so that each alternativecan be evaluated in relation to the relevant NCPremedy selection criteria.

The detailed analysis must, according to theproposed NCP, include an evaluation of eachalternative against the nine criteria. The first twocriteria (i.e., overall protectiveness and compliancewith ARARs) are threshold determinations andmust be met before a remedy can be selected.Evaluation of the overall protectiveness of analternative during the RI/FS should focus on howa specific alternative achieves protection over timeand how site risks are reduced.

The next five criteria (numbers 3 through 7)are primary balancing criteria. The last two(numbers 8 and 9) are considered modifyingcriteria, and risk information does not play adirect role in the analysis of them. Of the fiveprimary balancing criteria, risk information is ofparticular importance in the analysis ofeffectiveness and permanence. Analysis of long-term effectiveness and permanence involves anevaluation of the results of a remedial action interms of residual risk at the site after responseobjectives have been met. A primary focus of thisevaluation is the effectiveness of the controls thatwill be applied to manage risk posed by treatmentresiduals and/or any untreated wastes that may beleft on the site, as well as the volume and natureof that material. It should also consider thepotential impacts on human health and theenvironment should the remedy fail. An

10

evaluation of short-term effectiveness addressesthe impacts of the alternative during theconstruction and implementation phase untilremedial response objectives will be met. Underthis criterion, alternatives should be evaluated withrespect to the potential effects on human healthand the environment during implementation of theremedial action and the length of time untilprotection is achieved.

1.2 OVERALL ORGANIZATION OFTHE MANUAL

The next two chapters present additionalbackground material for the human healthevaluation process. Chapter 2 discusses statutes,regulations, guidance, and studies relevant to theSuperfund human health evaluation. Chapter 3discusses issues related to planning for the humanhealth evaluation. The remainder of the manualis organized by the three parts of the humanhealth evaluation process:

the baseline risk assessment is coveredin Part A of the manual (Chapters 4through 10);

refinement of preliminary remediationgoals is covered in Part B of the manual

(not included as part of this interim finalversion); and

the risk evaluation of remedialalternatives is covered in Part C of themanual (not included as part of thisinterim final version).

Chapters 4 through 8 provide detailedtechnical guidance for conducting the steps of abaseline risk assessment, and Chapter 9 providesdocumentation and review guidelines. Chapter 10contains additional guidance specific to baselinerisk assessment for sites contaminated withradionuclides. Sample calculations, sample tableformats, and references to other guidance areprovided throughout the manual. All material ispresented both in technical terms and in simplertext. It should be stressed that the manual isintended to be comprehensive and to provideguidance for more situations than usually arerelevant to any single site. Risk assessors neednot use those parts of the manual that do notapply to their site.

Each chapter in Part A includes a glossary ofacronyms and definitions of commonly used terms.The manual also includes two appendices:Appendix A provides technical guidance formaking absorption adjustments and Appendix Bis an index.

11

ENDNOTES FOR CHAPTER 1

1. References made to CERCLA throughout this document should be interpreted as meaning “CERCLA, as amended by the SuperfundAmendments and Reauthorization Act of 1986 (SARA).”

2. 40 CFR Part 300. Proposed revisions to the NCP were published on December 21, 1988 (53 Federal Register 51394).

3. The term “public health evaluation” was introduced in the previous risk assessment guidance (EPA 1986f) to describe the assessmentof chemical releases from a site and the analysis of public health threats resulting from those releases, and Superfund site risk assessmentstudies often are referred to as public health evaluations, or PHEs. The term "PHE" should be replaced by whichever of the three partsof the revised human health evaluation process is appropriate "baseline risk assessment," "documentation of preliminary remediationgoals," or "risk evaluation of remedial alternatives."

4. Baseline risks are risks that might exist if no remediation or institutional controls were applied at a site.

5. Volume II of the Risk Assessment Guidance for Superfund is the Environmental Evaluation Manual (EPA 1989b), which providesguidance for the analysis of potential environmental (i.e., not human health) effects at sites.

12

REFERENCES FOR CHAPTER 1

Congressional Research Service (CRS), Llbrary of Congress. 1983. A Review of Risk Assessment Methodologies. Washington, D.C.

Environmental Protection Agency (EPA). 1984. Risk Assessment and Management: Framework for Decisionmaking. EPA/600/9-85/002.

Environmental Protection Agency (EPA). 1986a. Guidelines for Carcinogen Risk Assessment. 51 Federal Register 33992 (September24, 1986).

Environmental Protection Agency (EPA). 1986b. Guidelines for Exposure Assessment. 51 Federal Register 34042 (September 24,1986).

Environmental Protection Agency (EPA). 1986c. Guidelines for Mutagenicitv Risk Assessment. 51 Federal Register 34006 (September24, 1986).

Environmental Protection Agency (EPA). 1986d. Guidelines for the Health Assessment of Suspect Developmental Toxicants. 51Federal Register 34028 (September 24, 1986).

Environmental Protection Agency (EPA). 1986e. Guidelines for the Health Risk Assessment of Chemical Mixtures. 51 FederalRegister 34014 (September 24, 1986).

Environmental Protection Agency (EPA). 1986f. Superfund Public Health Evaluation Manual. Office of Emergency and RemedialResponse. EPA/540/l-86/060. (OSWER Directive 9285.4-l).

Environmental Protection Agency (EPA). 1988a. Proposed Guidelines for Exposure-related Measurements. 53 Federal Register 48830(December 2, 1988).

Environmental Protection Agency (EPA). 1988b. Guidance for Conducting Remedial Investigations and Feasibility Studies UnderCERCLA. Interim Final. Office of Emergency and Remedial Response. (OSWER Directive 9355.3-01).

Environmental Protection Agency (EPA). 1989a. Proposed Amendments to the Guidelines for the Health Assessment of SuspectDevelopmental Toxicants. 54 Federal Register 9386 (March 6, 1989).

Environmental Protection Agency (EPA). 1989b. Risk Assessment Guidance for Superfund: Environmental Evaluation Manual.Interim Final. Office of Emergency and Remedial Response. EPA/540/l-89/001A. (OSWER Directive 9285.7-01).

National Academy of Sciences (NAS). 1983. Risk Assessment in the Federal Government: Managing the Process. National AcademyPress. Washington, D.C.

Office of Science and Technology Policy (OSTP). 1985. Chemical Carcinogens A Review of the Science and Its Associated Principles.50 Federal Register 10372 (March 14, 1985).

CHAPTER 2

STATUTES, REGULATIONS,GUIDANCE, AND

STUDIES RELEVANT TOTHE HUMAN HEALTH

EVALUATION

This chapter briefly describes the statutes,regulations, guidance, and studies related to thehuman health evaluation process. Thedescriptions focus on aspects of these documentsmost relevant to human health evaluations andshow how recent revisions to the documents bearupon the human health evaluation process.Section 2.1 describes the following documents thatgovern the human health evaluation

the Comprehensive EnvironmentalResponse, Compensation, and LiabilityAct of 1980 (CERCLA or Superfund)and the Superfund Amendments andReauthorization Act of 1986 (SARA);

the National Oil and HazardousSubstances Pollution Contingency Plan(National Contingency Plan, or NCP);

Guidance for Conducting RemedialInvestigations and Feasibility Studies UnderCERCLA (RI/FS guidance);

CERCLA Compliance with Other LawsManual (ARARs guidance); and

Superfund Exposure Assessment Manual(SEAM).

Exhibit 2-1 shows the relationship of thesestatutes, regulations, and guidances governing

human health evaluation. In addition, Section2.2 identifies and briefly describes other Superfundstudies related to, and sometimes confused with,the RI/FS human health evaluation. The types ofstudies discussed are:

2.1

laws

endangerment assessment,ATSDR health assessments; andATSDR health studies.

STATUTES, REGULATIONS,AND GUIDANCE GOVERNINGHUMAN HEALTHEVALUATION

This section describes the major Superfundand program documents relevant to the

human health evaluation process.

2.1.1 CERCLA AND SARA

In 1980, Congress enacted the ComprehensiveEnvironmental Response, Compensation, andLiability Act (CERCLA) (42 U.S.C 9601 et seq.),commonly called Superfund, in response to thedangers posed by sudden or otherwiseuncontrolled releases of hazardous substances,pollutants, or contaminants into the environment.CERCLA authorized $1.6 billion over five yearsfor a comprehensive program to cleanup the

2

EXHIBIT 2-1

RELATIONSHIP OF DOCUMENTS GOVERNINGHUMAN HEALTH EVALUATION

3

worst abandoned. or inactive waste sites in thenation. CERCLA funds used to establish andadminister the cleanup program are derivedprimarily from taxes on crude oil and 42 differentcommercial chemicals.

The reauthorization of CERCLA is knownas the Superfund Amendments andReauthorization Act (SARA), and was signed bythe President on October 17, 1986. (All furtherreferences to CERCLA in this appendix should beinterpreted as “CERCLA as amended by SARA”)These amendments provided $8.5 billion for thecleanup program and an additional $500 millionfor cleanup of leaks from underground storagetanks. Under SARA, Congress strengthenedEPA’s mandate to focus on permanent cleanupsat Superfund sites, involve the public in decisionprocesses at sites, and encourage states andfederally recognized Indian tribes to activelyparticipate as partners with EPA to address thesesites. SARA expanded EPA's research,development (especially in the area of alternativetechnologies), and training responsibilities. SARAalso strengthened EPA’s enforcement authority.The changes to CERCLA sections 104 (ResponseAuthorities) and 121 (Cleanup Standards) havethe greatest impact on the RI/FS process.

Cleanup standards. Section 121 (CleanupStandards) states a strong preference for remediesthat are highly reliable and provide long-termprotection. In addition to the requirement forremedies to be both protective of human healthand the environment and cost-effective, otherremedy selection considerations in section 121(b)include:

a preference for remedial actions thatemploy (as a principal element of theaction) treatment that permanently andsignificantly reduces the volume, toxicity,or mobility of hazardous substances,pollutants, and contaminants;

offsite transport and disposal withouttreatment as the least favored alternativewhere practicable treatment technologiesare available; and

the need to assess the use of alternativetreatment technologies or resource

recovery technologies and use them tothe maximum extent practicable.

Section 121(c) of CERCLA requires aperiodic review of remedial actions, at least everyfive years after initiation, for as long as hazardoussubstances, pollutants, or contaminants that maypose a threat to human health or the environmentremain at the site. If during a five-year review itis determined that the action no longer protectshuman health and the environment, furtherremedial actions will need to be considered.

Section 121(d)(2)(A) of CERCLAincorporates into law the CERCLA CompliancePolicy, which specifies that Superfund remedialactions meet any federal standards, requirements,criteria, or limitations that are determined to belegally applicable or relevant and appropriaterequirements (i.e., ARARs). Also included is thenew provision that state ARARs must be met ifthey are more stringent than federal requirements.(Section 2.1.4 provides more detail on ARARs.)

Health-related authorities. Under CERCLAsection 104(i)(6), the Agency for Toxic Substancesand Disease Registry (ATSDR) is required toconduct a health assessment for every site includedor proposed for inclusion on the NationalPriorities List. The ATSDR health assessment,which is fairly qualitative in nature, should bedistinguished from the EPA human healthevaluation, which is more quantitative. CERCLAseetion 104(i)(5)(F) states that:

the term “health assessments” shall includepreliminary assessments of the potential riskto human health posed by individual sites andfacilities, based on such factors as the natureand extent of contamination, the existence ofpotential pathways of human exposure(including ground or surface watercontamination, air emissions, and food chaincontamination), the size and potentialsusceptibility of the community within thelikely pathways of exposure, the comparisonof expected human exposure levels to theshort-term and long-term health effectsassociated with identified hazardoussubstances and any available recommendedexposure or tolerance limits for suchhazardous substances, and the comparison ofexisting morbidity and mortality data on

diseases that may be associated with theobserved levels of exposure. TheAdministrator of ATSDR shall useappropriate data, risk assessments, riskevaluations and studies available from theAdministrator of EPA.

There are purposeful differences between anATSDR health assessment and traditional riskassessment. The health assessment is usuallyqualitative, site-specific, and focuses on medicaland public health perspectives. Exposures to sitecontaminants are discussed in terms of especiallysensitive populations, mechanisms of toxicchemical action, and possible disease outcomes.Risk assessment, the framework of the EPAhuman health evaluation, is a characterization ofthe probability of adverse effects from humanexposures to environmental hazards. In thiscontext, risk assessments differ from healthassessments in that they are quantitative, chemical-oriented characterizations that use statistical andbiological models to calculate numerical estimatesof risk to health. However, both healthassessments and risk assessments use data fromhuman epidemiological investigations, whenavailable, and when human toxicological data areunavailable, rely on the results of animaltoxicology studies.

2.1.2 NATIONAL CONTINGENCY PLAN(NCP)

The National Contingency Plan provides theorganizational structure and procedures forpreparing for and responding to discharges of oiland releases of hazardous substances, pollutants,and contaminants. The NCP is required bysection 105 of CERCLA and by section 311 of theClean Water Act. The current NCP (EPA 1985)was published on November 20, 1985, and asignificantly revised version (EPA 1988a) wasproposed December 21, 1988 in response toSARA. The proposedfollowing subparts:

Subpart A --

Subpart BOrganization

NCP is organized into the

Introduction

-- Responsibility andfor Response

Subpart C -- Planning and Preparedness

Subpart D -- Operational ResponsePhases for Oil Removal

Subpart E -- Hazardous SubstanceResponse

Subpart F -- State Involvement inHazardous Substance Response

Subpart G -- Trustees for NaturalResources

Subpart H -- Participation by OtherPersons

Subpart I -- Administrative Record forSelection of Response Action

Subpart J -- Use of Dispersants andOther Chemicals

Subpart E, Hazardous Substance Response,contains a detailed plan covering the entire rangeof authorized activities involved in abating andremedying releases or threats of releases ofhazardous substances, pollutants, andcontaminants. It contains provisions for bothremoval and remedial response. The remedialresponse process set forth by the proposed NCPis a seven-step process, as described below. Riskinformation plays a role in each step.

Site discovery or notification. Releases ofhazardous substances, pollutants, or contaminantsidentified by federal, state, or local governmentagencies or private parties are reported to theNational Response Center or EPA. Upondiscovery, such potential sites are screened toidentify release situations warranting furtherremedial response consideration. These sites areentered into the CERCLA Information System(CERCLIS). This computerized system serves asa data base of site information and tracks thechange in status of a site through the responseprocess. Risk information is used to determinewhich substances are hazardous and, in somecases, the quatities that constitute a release thatmust be reported (i.e., a reportable quantity, orRQ, under CERCLA section 103(a)).

Preliminary assessment and site inspection(PA/SI). The preliminary assessment involvescollection and review of all available information

5

and may include offsite reconnaissance to evaluatethe source and nature of hazardous substancespresent and to identify the responsible party(ies).At the conclusion of the preliminary assessment,a site may be referred for further action, or adetermination may be made that no further actionis needed. Site inspections, which follow thepreliminary assessment for sites needing furtheraction, routinely include the collection of samplesand are conducted to help determine the extentof the problem and to obtain information neededto determine whether a removal action iswarranted. If, based on the site inspection, itappears likely that the site should be consideredfor inclusion on the National Priorities List(NPL), a listing site inspection (LSI) is conducted.The LSI is a more extensive investigation than theSI, and a main objective of the LSI is to collectsufficient data about a site to support HazardRanking System (HRS) scoring. One of the mainobjectives of the PA/SI is to collect risk-relatedinformation for sites so that the site can be scoredusing the HRS and priorities may be set for moredetailed studies, such as the RI/FS.

Establishing priorities for remedial action.Sites are scored using the HRS, based on datafrom the PA/SI/LSI. The HRS scoring process isthe primary mechanism for determining the sitesto be included on the NPL and, therefore, thesites eligible for Superfund-financed remedialaction. The HRS is a numerical scoring modelthat is based on many of the factors affecting riskat a site. A revised version of the HRS (EPA1988b) was proposed December 23, 1988.

Remedial investigation/feasibility study(RI/FS). As described in Section 1.1, the RI/FSis the framework for determining appropriateremedial actions at Superfund sites. AlthoughRI/FS activities technically are removal actionsand therefore not restricted to sites on the NPL(see sections 101(23) and 104(b) of CERCLA),they most frequently are undertaken at NPL sites.Remedial investigations are conducted tocharacterize the contamination at the site and toobtain information needed to identify, evaluate,and select cleanup alternatives. The feasibilitystudy includes an analysis of alternatives basedon the nine NCP evacuation criteria. The humanhealth evaluation described in this manual, andthe environmental evaluation described elsewhere,

are the guidance for developing risk informationin the RI/FS.

Selection of remedy. The primaryconsideration in selecting a remedy is that it beprotective of human health and the environment,by eliminating, reducing, or controlling risks posedthrough each pathway. Thus, the risk informationdeveloped in the RI/FS is a key input to remedyselection. The results of the RI/FS are reviewedto identify a preferred alternative, which isannounced to the public in a Proposed Plan.Next, the lead agency reviews any resulting publiccomments on the Proposed Plan, consults with thesupport agencies to evaluate whether the preferredalternative is still the most appropriate, and thenmakes a final decision. A record of decision(ROD) is written to document the rationale forthe selected remedy.

Remedial design/remedial action. Thedetailed design of the selected remedial action isdeveloped and then implemented. The riskinformation developed previously in the RI/FShelps refine the remediation goals that the remedywill attain.

Five-year review. Section 121(c) of CERCLArequires a periodic review of remedial actions, atleast every five years after initiation of suchaction, for as long as hazardous substances,pollutants, or contaminants that may pose a threatto human health or the environment remain atthe site. If it is determined during a five-yearreview that the action no longer protects humanhealth and the environment, further remedialactions will need to be considered.

Exhibit 2-2 diagrams the general steps of theSuperfund remedial process, indicating where inthe process the various parts of the human healthevaluation are conducted.

2.1.3 REMEDIAL INVESTIGATION/FEASIBILITY STUDY GUIDANCE

EPA’s interim final Guidance for ConductingRemedial Investigations and Feasibility StudiesUnder CERCLA (EPA 1988c) provides a detailedstructure for conducting field studies to supportremedial decisions and for identifying, evaluating,and selecting remedial action alternatives underCERCLA. This 1988 guidance document is a

6

EXHIBIT 2-2

ROLE OF THE HUMAN HEALTH EVALUATION INTHE SUPERFUND REMEDIAL PROCESS

a The RI/FS can be undertaken prior to NPL listing.

7

revision of two separate guidances for remedialinvestigations and for feasibility studies publishedin 1985. These guidances have been consolidatedinto a single document and revised to:

reflect new emphasis and provisions ofSARA;

incorporate aspects of new or revisedguidance related to RI/FSs;

incorporate management initiativesdesigned to streamline the RI/FSprocess; and

reflect experience gained from previousRI/FS projects.

The RI/FS consists of the following generalsteps:

project scoping (during the RI);

site characterization (RI);

establishment of remedial actionobjectives (FS);

development and screening ofalternatives (FS); and

detailed analysis of alternatives (FS).

Because Section 1.1 describes each of these steps,focusing on the role that risk information plays inthe RI/FS, a discussion of the steps is notrepeated here. The RI/FS guidance provides thecontext into which the human health evaluationfits and should be used in conjunction with thismanual.

2.1.4 ARARS GUIDANCE

The interim final CERCLA Compliance withOther Laws Manual (EPA 1988d; EPA 1989a), orARARs guidance, was developed to assist in theselection of onsite remedial actions that meet theapplicable or relevant and appropriaterequirements (ARARs) of the ResourceConservation and Recovery Act (RCRA), CleanWater Act (CWA), Safe Drinking Water Act(SDWA), Clean Air Act (CM), and other federaland state environmental laws, as required by

CERCLA section 121. Part I of the manualdiscusses the overall procedures for identifyingARARs and provides guidance on theinterpretation and analysis of RCRA requirements.Specifically

Chapter 1 defines “appliable” and"relevant and appropriate," providesmatrices listing potential chemical-specific, location-specific, and action-specific requirements from RCRA, CWA,and SDWA, and provides generalprocedures for identifying and analyzingrequirement,

Chapter 2 discusses special issues ofinterpretation and analysis involvingRCRA requirements, and providesguidance on when RCRA requirementswill be ARARs for CERCLA remedialactions;

Chapter 3 provides guidance forcompliance with CWA substantive (foronsite and offsiite actions) andadministrative (for offsite actions)requirements for direct discharges,indirect discharges, and dredge and fillactivities;

Chapter 4 provides guidance forcompliance with requirements of theSDWA that may be applicable orrelevant and appropriate to CERCLAsites; and

Chapter 5 provides guidance onconsistency with policies- for ground-water protection.

The manual also contains a hypothetical scenarioillustrating how ARARs are identified and used,and an appendix summarizing the provisions ofRCRA, CWA and SDWA.

Part II of the ARARs guidance covers theClean Air Act, other federal statutes, and staterequirements. Specifically:

Chapter 1 provides an introduction toPart II of the guidance, and also includesextensive summary tables;

8

2.1.5

Chapter 2 describes Clean Air Actrequirements and related RCRA andstate requirements;

Chapters 3 and 4 provide guidance forcompliance with several other federalstatutes;

Chapter 5 discusses potential ARARs forsites contaminated with radioactivesubstances;

Chapter 6 addresses requirements specificto mining, mil1ing, or smelting sites; and

Chapter 7 provides guidance onidentifying and complying with stateARARs.

SUPERFUND EXPOSUREASSESSMENT MANUAL

The Superfund Exposure Assessment Manual(EPA 1988e), which was developed by theSuperfund program specifically as a companiondocument to the original Superfund Public HealthEvaluation Manual (EPA 1986), provides RPMsand regional risk assessors with the guidancenecessary to conduct exposure assessments thatmeet the needs of the Superfund human healthrisk evaluation process. Specifically, the manual:

provides an overall description of theintegrated exposure assessment as it isapplied to uncontrolled hazardous wastesites; and

serves as a source of referenceconcerning the use of estimationprocedures and computer modelingtechniques for the analysis ofuncontrolled sites.

The analytical process outlined in theSuperfund Exposure Assessment Manual providesa framework for the assessment of exposure tocontaminants at or migrating from uncontrolledhazardous waste sites. The application of bothmonitoring and modeling procedures to theexposure assessment process is outlined in themanual. This process considers all contaminantreleases and exposure routes and assures that an

adequate level of analytical detail is applied tosupport the human health risk assessment process.

The exposure assessment process described inthe Superfund Exposure Assessment Manual isstructured in five segments:

(1)

(2)

(3)

(4)

(5)

analysis of contaminant releases from asubject site into environmental media;

evaluation of the transport andenvironmental fate of the contaminantsreleased;

identification, enumeration, andcharacterization of potentially exposedpopulations;

integrated exposure analysis; and

uncertainty analysis.

Assessment Methods Handbook (EPA 1989c),

Two recent publications from EPA’s Officeof Research and Development, the ExposureFactors Handbook (EPA 1989b) and the Exposure

provide useful information to supplement theSuperfund Exposure Assessment Manual. All threeof these key exposure assessment references shouldbe used in conjunction with Chapter 6 of thismanual.

2.2 RELATED SUPERFUNDSTUDIES

This section identifies and briefly describesother Superfund studies related to, and sometimesconfused with, the RI/FS human health evaluation.It contrasts the objectives and methods andclarifies the relationships of these other studieswith RI/FS health risk assessments. The types ofstudies discussed are endangerment assessments,ATSDR health assessments, and ATSDR healthstudies.

2.2.1 ENDANGERMENT ASSESSMENTS

Before taking enforcement action againstparties responsible for a hazardous waste site,EPA must determine that an imminent andsubstantial endangerment to public health or the

9

environment exists as a result of the site. Such alegal determination is called an endangermentassessment. For remedial sites, the process foranalyzing whether there may be an endangermentis described in this Human Health EvaluationManual and its companion EnvironmentalEvaluation Manual. In the past, an endangermentassessment often was prepared as a study separatefrom the baseline risk assessment. With thepassage of SARA and changes in Agency practice,the need to perform a detailed endangermentassessment as a separate effort from the baselinerisk assessment has been eliminated.

For administrative orders requiring a remedialdesign or remedial action, endangermentassessment determinations are now based oninformation developed in the site baseline riskassessment. Elements included in the baselinerisk assessment conducted at a Superfund siteduring the RI/FS process fully satisfy theinformational requirements of the endangermentassessment. These elements include the following:

identification of the hazardous wastesor hazardous substances present inenvironmental media;

assessment of exposure, including acharacterization of the environmentalfate and transport mechanisms for thehazardous wastes and substances present,and of exposure pathways;

assessment of the toxicity of thehazardous wastes or substances present;

characterization of human health risks;and

characterization of the impacts and/orrisks to the environment.

The human health and environmentalevaluations that are part of the RI/FS areconducted for purposes of determining thebaseline risks posed by the site, and for ensuringthat the selected remedy will be protective ofhuman health and the environment. Theendangerment assessment is used to supportlitigation by determining that an imminent andsubstantial endangerment exists. Informationpresented in the human health and environmental

evaluations is basic to the legal determination ofendangerment.

In 1985, EPA produced a draft manualspecifically written for endangerment assessment,the Endangerment Assessment Handbook. EPAhas determined that a guidance separate from theRisk Assessment Guidance for Superfund (HumanHealth Evaluation Manual and EnvironmentalEvaluation Manual) is not required forendangerment assessment therefore, theEndangerment Assessment Handbook will not bemade final and should no longer be used.

2.2.2 ATSDR HEALTH ASSESSMENTS

CERCLA section 104(i), as amended, requiresthe Agency for Toxic Substances and DiseaseRegistry (ATSDR) to conduct health assessmentsfor all sites Iisted or proposed to be fisted on theNPL. A health assessment includes a preliminaryassessment of the potential threats that individualsites and facilities pose to human health. Thehealth assessment is required to be completed “tothe maximum extent practicable" beforecompletion of the RI/FS. ATSDR personnel,state personnel (through cooperative agreements),or contractors follow six basic steps, which arebased on the same general risk assessmentframework as the EPA human health evaluation:

(1)

(2)

(3)

(4)

(5)

(6)

evaluate information on the site’sphysical, geographical, historical, andoperational setting, assess thedemographics of nearby populations, andidentify health concerns of the affectedcommunity(ies);

determine contaminants of concernassociated with the site;

identify and evaluate environmentalpathways;

identify and evaluate human exposurepathways;

identify and evaluate public healthimplications based on available medicaland toxicological information; and

develop conclusions concerning thehealth threat posed by the site and make

10

recommendations regarding furtherpublic health activities.

The purpose of the ATSDR healthassessment is to assist in the evaluation of dataand information on the release of toxic substancesinto the environment in order to assess anycurrent or future impact on public health, develophealth advisories or other health-relatedrecommendations, and identify studies or actionsneeded to evaluate and prevent human healtheffects. Health assessments are intended to helppublic health and regulatory officials determine ifactions should be taken to reduce human exposureto hazardous substances and to recommendwhether additional information on humanexposure and associated risks is needed. Healthassessments also are written for the benefit of theinformed community associated with a site, whichcould include citizen groups, local leaders, andhealth professionals.

Several important differences exist betweenEPA human health evaluations and ATSDRhealth assessments. EPA human healthevaluations include quantitative, substance-specificestimates of the risk that a site poses to humanhealth. These estimates depend on statistical andbiological models that use data from humanepidemiologic investigations and animal toxicitystudies. The information generated from a humanhealth evaluation is used in risk managementdecisions to establish cleanup levels and select aremedial alternative.

ATSDR health assessments, although theymay employ quantitative data, are more qualitativein nature. They focus not only on the possiblehealth threats posed by chemical contaminantsattributable to a site, but consider all healththreats, both chemical and physical, to whichresidents near a site may be subjected. Healthassessments focus on the medical and publichealth concerns associated with exposures at a siteand discuss especially sensitive populations, toxicmechanisms, and possible disease outcomes. EPAconsiders the information in a health assessment along with the results of the baseline riskassessment to give a complete picture of healththreats. Local health professionals and residentsuse the information to understand the potentialhealth threats posed by specific waste sites.Health assessments may lead to pilot health effects

studies, epidemiologic studies, or establishment ofexposure or disease registries.

EPA’s Guidance for Coordinating ATSDRHealth Assessment Activities with the SuperfundRemedial Process (EPA 1987) provides informationto EPA and ATSDR managers for use incoordinating human health evaluation activities.(Section 2.1, in its discussion of CERCLA,provides further information on the statutory basisof ATSDR health assessments.)

2.2.3 ATSDR HEALTH STUDIES

After conducting a health assessment,ATSDR may determine that additional healtheffects information is needed at a site and, as aresult, may undertake a pilot study, a full-scaleepidemiological study, or a disease registry. Threetypes of pilot studies are predominant:

(1)

(2)

(3)

a symptom/disease prevalence studyconsisting of a measurement of self-reported disease occurrence, which maybe validated through medical records ifthey are available;

a human exposure study consisting ofbiological sampling of persons who havea potentially high likelihood of exposureto determine if actual exposure can beverified, and

a cluster investigation study consistingof an investigation of putative diseaseclusters to determine if the cases of adisease are excessively high in theconcerned community.

A fill-scale epidemiological study is ananalytic investigation that evaluates the possiblecausal relationships between exposure tohazardous substances and disease outcome bytesting a scientific hypothesis. Such anepidemiological study is usually not undertakenunless a pilot study reveals widespread exposureor increased prevalence of disease.

ATSDR, in cooperation with the states, alsomay choose to follow up the results of a healthassessment by establishing and maintainingnational registries of persons exposed to hazardoussubstances and persons with serious diseases or

11

illness. A registry is a system for collecting andmaintaining, in a structured record, information onspecific persons from a defined population. Thepurpose of a registry of persons exposed tohazardous substances is to facilitate developmentof new scientific knowledge through identificationand subsequent follow-up of persons exposed toa defined substance at selected sites.

Besides identifying and tracking of exposedpersons, a registry also is used to coordinate theclinical and research activities that involve theregistrants. Registries serve an important role inassuring the uniformity and quality of thecollected data and ensuring that data collection isnot duplicative, thereby reducing the overallburden to exposed or potentially exposed persons.

12


Environmental Protection Agency (EPA). 1985. National Oil and Harardous Substances Pollution Contingency Plan. Final Rule. 50Federal Register 47912 (November 20, 1985).

Environmental Protection Agency (EPA). 1986. Superfund Public Health Evaluation Manual. Office of Emergency and RemedialResponse. EPA/540/l-86/060. (OSWER Directive 9285.4-l).

Environmental Protection Agency (EPA). 1987. Guidance for Coordinating ATSDR Health Assessment Activities with the SupcrfundRemedial Process. Office of Emergency and Remedial Response. (OSWER Directive 9285.442).

Environmental Protection Agency (EPA). 1988a. National Oil and Hazardous Substances Pollution Contingency Plan. Proposed Rule.53 Federal Register 51394 (December 21, 1988).

Environmental Protection Agency (EPA). 1988b. Hazard Ranking System (HRS) for Uncontrolled Hazardous Substance Releases.Proposed Rule. 53 Federal Register 51962 (December 23, 1988).

Environmental Protection Agency (EPA). 1988c. Guidance for Conducting Remedial Investigations and Feasibility Studies UnderCERCLA Interim Final. Office of Emergency and Remedial Response. (OSWER Directive 9355.3-01).

Environmental Protection Agency (EPA). 1988d. CERCLA compliance with Other Laws Manual. Part I. Interim Final. Office ofEmergency and Remedial Response. (OSWER Directive 9234.1-01).

Environmental Protection Agency (EPA). 1988c. Superfund Exposure Assessment Manual. Office of Emergency and RemedialResponse. EPA/540/l-88/001. (OSWER Directive 9285.5-l).

Environmental Protection Agency (EPA). 1989a. CERCLA Compliance with Other Laws Manual. Part II. Interim Final. Office ofEmergency and Remedial Response. (OSWER Directive 9234.1-02).

Environmental Protection Agency (EPA). 1989b. Exposure Factors Handbook. Office of Health and Environmental Assessment.EPA/600/8-89/043.

Environmental Protection Agency (EPA). 1989c. Exposure Assessment Methods Handbook. Draft. Office of Health andEnvironmental Assessment.

CHAPTER 3

GETTING STARTED: PLANNINGFOR THE HUMAN HEALTHEVALUATION IN THE RI/FS

This chapter discusses issues related toplanning the human health evaluation conductedduring the RI/FS. It presents the goals of theRI/FS process as a whole and the human healthevaluation in particular (Sections 3.1 and 3.2). Itnext discusses the way in which a site that isdivided into operable units should be treated inthe human health evaluation (Section 3.3). RI/FSscoping is discussed in Section 3.4, and Section 3.5addresses the level of effort and detail necessaryfor a human health evaluation.

3.1 GOAL OF THE RI/FS

The goal of the RI/FS is to gatherinformation sufficient to support an informed riskmanagement decision regarding which remedyappears to be most appropriate for a given site.The RI/FS provides the context for all sitecharacterization activity, including the humanhealth evaluation. To attain this goal efficiently,EPA must identify and characterize hazards in away that will contribute directly to the selectionof an appropriate remedy. Program experiencehas shown that Superfund sites are complex, andare characterized by heterogeneous wastes, extremevariability in contamination levels, and a varietyof environmental settings and potential exposurepathways. Consequently, complete characterizationof a site during the RI/FS in the sense ofeliminating uncertainty, is not feasible, cost-effective, or necessary for selection of appropriateremedies. This view has motivated the“streamlined approach” EPA is taking to helpaccomplish the goal of completing an RI/FS in 18months at a cost of $750,000 per operable unit

and $1.1 million per site. The streamlinedapproach recognizes that the elimination of alluncertainties is not possible or necessary andinstead strives only for sufficient data to generallycharacterize a site and support remedy selection.The resulting remedies are flexible and incorporatespecific contingencies to respond to newinformation discovered during remedial action andfollow-up.

3.2 GOAL OF THE RI/FS HUMANHEALTH EVALUATION

As part of the effort to streamline theprocess and reduce the cost and time required toconduct the RI/FS, the Superfund human healthevaluation needs to focus on providinginformation necessary to justify action at a siteand to select the best remedy for the site. Thisshould include characterizing the contaminants,the potential exposures, and the potentiallyexposed population sufficiently to determine whatrisks need to be reduced or eliminated and whatexposures need to be prevented. It is importantto recognize that information should be developedonly to help EPA determine what actions arenecessary to reduce risks, and not to fullycharacterize site risks or eliminate all uncertaintyfrom the analysis.

In a logical extension of this view, EPA hasmade a policy decision to use, whereverappropriate, standardized assumptions, equations,and values in the human health evaluation toachieve the goal of streamlined assessment. Thisapproach has the added benefit of making human

2

health evaluation easier to review, easier tounderstand, and more consistent from site to site.Developing unique exposure assumptions or non-standard methods of risk assessment should not benecessary for most sites. Where justified by site-specific data or by changes in knowledge overtime, however, non-standard methods andassumptions may be used.

3.3 OPERABLE UNITS

Current practice in designing remedies forSuperfund sites often divides sites into operableunits that address discrete aspects of the site (e.g.,source control, ground-water remediation) ordifferent geographic portions of the site. TheNCP defines operable unit as “a discrete actionthat comprises an incremental step toward,comprehensively addressing site problems.” RI/FSs

may be conducted for the entire site and operableunits broken out during or after the feasibilitystudy, or operable units may be treatedindividually from the start, with focused RI/FSsconducted for each operable unit. The best wayto address the risks of the operable unit willdepend on the needs of the site.

The human health evaluation should focus onthe subject of the RI/FS, whether that is anoperable unit or the site as a whole. The baselinerisk assessment and other risk informationgathered will provide the justification for takingthe action for the operable unit. At the sametime, personnel involved in conducting the humanhealth evaluation for a focused RI/FS must bemindful of other potential exposure pathways, andother actions that are being contemplated for thesite to address other potential exposures. Riskassessors should foresee that exposure pathwaysoutside the scope of the focused RI/FS mayultimaately be combined with exposure pathwaysthat are directly addressed by the focused RI/FS.Considering risks from all related operable unitsshould prevent the unexpected discovery of highmultiple pathway risks during the human healthevaluation for the last operable unit. Consider, for example, a site that will be addressed in twooperable units: a surface soil cleanup at thecontamination source and a separate ground-watercleanup. Risks associated with residuals from thesoil cleanup and the ground-water cleanup mayneed to be considered as a cumulative total if

there is the potential for exposure to both mediaat the same time.

3.4 RI/FS SCOPING

Planning the human health evaluation priorto beginning the detailed analysis is an essentialstep in the process. The RPM must make up-front decisions about, for example, the scope ofthe baseline risk assessment, the appropriate levelof detail and documentation, trade-offs betweendepth and breadth in the analysis, and the staffand monetary resources to commit.

Scoping is the initial planning phase of theRI/FS process, and many of the planning stepsbegun here are continued and refined in laterphases. Scoping activities typically begin with thecollection of existing site data, including data fromprevious investigations such as the preliminaryassessment and site inspection. On the basis ofthis information, site management planning isundertaken to identify probable boundaries of thestudy area, to identify likely remedial actionobjectives and whether interim actions may benecessary or appropriate, and to establish whetherthe site may best be remedied as one site or asseveral separate operable units. Once an overallmanagement strategy is agreed upon, the RI/FSfor a specific project or the site as a whole isplanned.

The development of remedial alternativesusually begins during or soon after scoping, whenlikely response scenarios may first be identified.The development of alternatives requires:

identifying remedial action objectives;

identifying potential treatment, resourcerecovery, and containment technologiesthat will satisfy these objectives; and

screening the technologies based on theireffectiveness, implementability, and cost.

Remedial alternatives may be developed to addressa contaminated medium, a specific area of thesite, or the entire site. Alternative remedialactions for specific media and site areas either canbe carried through the FS process separately orcombined into comprehensive alternatives for the

3

entire site. The approach is flexible to allowalternatives to be considered in combination atvarious points in the process. The RI/FS guidancediscusses planning in greater detail.

3.5 LEVEL OF EFFORT/LEVEL OFDETAIL OF THE HUMANHEALTH EVALUATION

An important part of scoping is determiningappropriate level of effort/level of detailthe

necessay for the human health evaluation.Human health evaluation can be thought of asspanning a continuum of complexity, detail, andlevel of effort, just as sites vary in conditions andcomplexity. Some of the site-specific factorsaffecting level of effort that the RPM mustconsider include the following

number and identity of chemicalspresent;

availability of ARARs and/or applicabletoxicity data;

number and complexity of exposurepathways (including complexity of releasesources and transport media), and theneed for environmental fate andtransport modeling to supplementmonitoring data;

necessity for precision of the results,which in turn depends on site conditionssuch as the extent of contaminantmigration, characteristics of potentiallyexposed populations, and enforcementconsiderations (additional quantificationmay be warranted for some enforcementsites); and

quality and quantity of availablemonitoring data.1

This manual is written to address the mostcomplex sites, and as a result not all of the stepsand procedures of the Superfund human healthevaluation process described in this manual applyto all remedial sites. For example, Section 6.6provides procedures and equations for estimatingchemical intakes through numerous exposureroutes, although for many sites, much of thisinformation will not apply (e.g., the exposureroute does not exist or is determined to berelatively unimportant). This manual establishesa generic framework that is broadly applicableacross sites, and it provides specific proceduresthat cover a range of sites or situations that mayor may not be appropriate for any individual site.As a consequence of attempting to cover the widevariety of Superfund site conditions, some of theprocess components, steps, and techniquesdescribed in the manual do not apply to somesites. In addition, most of the components canvary greatly in level of detail. Obviously,determining which elements of the process arenecessary, which are desirable, and which areextraneous is a key decision for each site. Allcomponents should not be forced into the assess-ment of a site, and the evaluation should belimited to the complexity and level of detailnecessary to adequately assess risks for thepurposes described in Sections 3.1 and 3.2.

Planning related to the collection and analysisof chemical data is perhaps the most importantplanning step. Early coordination among the riskassessors, the remainder of the RI/FS team,representatives of other agencies involved in therisk assessment or related studies (e.g., ATSDR,natural resource trustees such as the Departmentof the Interior, state agencies), and the RPM isessential and preferably should occur during thescoping stage of the RI/FS. Detailed guidance onplanning related to collection and analysis ofchemical data is given in Chapter 4 of thismanual.

4

ENDNOTE FOR CHAPTER 3

1. All site monitoring data must be subjected to appropriate quality assurance/quality control programs. Lack of acceptable data maylimit by necessity the amount of data available for the human health evaluation, and therefore may limit the scope of the evaluation.Acceptability is determined by whether data meet the appropriate data quality objectives (see Section 4.1.2).

CHAPTER 4

DATA COLLECTION

CHAPTER 4

DATA COLLECTION

This chapter discusses procedures foracquiring reliable chemical release and exposuredata for quantitative human health risk assessmentat hazardous waste sites.1 The chapter is intendedto be a limited discussion of important samplingconsiderations with respect to risk assessment itis not intended to be a complete guide on how tocollect data or design sampling plans.

Following a general background section(Section 4.1), this chapter addresses the followingeight important areas:

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

review of available site information(Section 4.2);

consideration of modeling parameterneeds (Section 4.3);

definition of background sampling needs(Section 4.4);

preliminary identification of potentialhuman exposure (Section 4.5);

development of an overall strategy forsample collection (Section 4.6);

definition of required QA/QC measures(Section 4.7);

evaluation of the need for SpecialAnalytical Services (Section 4.8); and

activities during workplan developmentand data collection (Section 4.9).

4.1 BACKGROUND INFORMATIONUSEFUL FOR DATACOLLECTION

This section provides background informationon the types of data needed for risk assessment,overall data needs of the RI/FS, reasons and stepsfor identifying risk assessment data needs early,use of the Data Quality Objectives for RemedialResponse Activities (EPA 1987a,b, hereafterreferred to as the DQO guidance), and other dataconcerns.

4.1.1 TYPES OF DATA

In general, the types of site data needed fora baseline risk assessment include the following

contaminant identities;

2

DEFINITIONS FOR CHAPTER 4

Analytes. The chemicals for which a sample is analyzed.

Anthropogenic Background Levels. Concentrations of chemicals that are present in the environment due to human-made, non-site sources (e,g., industry, automobiles).

Contract Laboratory Program (CLP). Analytical program developed for Superfund waste site samples to fill the need for legallydefensible analytical results supported by a high Ievel of quality assurance and documentation.

Data Qualitv Objectives (DQOs). Qualitative and quantitative statements to ensure that data of known and documented qualityare obtained during an RI/FS to support an Agency decision.

Field Sampling Plan (FSP). Provides guidance for all field work by defining in detail the sampling and data gathering methodsto be used on a project.

Naturally Occurring Background Levels. Ambient concentrations of chemicals that are present in the environment and havenot been influenced by humans (e.g., aluminum, rnanganese). I

Qualitv Assurance Project Plan (QAPjP). Describes the policy, organization, functional activities, and quality assurance andquality control protocols necessary to achieve DQOs dictated by the intended use of the data (RI/FS Guidance).

Routine Analytical Services (RAS). The set of CLP analytical protocols that are used to analyze most Superfund site samples.These protocols are provided in the EPA Statements of Work for the CLP (SOW for Inorganics, SOW for Organics) andmust be followed by every CLP laboratory.

Sampling and Analysis Plan (SAP). Consists of a Quality Assurance Project Plan (QAPjP) and a Field Sampling Plan (FSP).

Sample Management Office (SMO). EPA contractor providing rnanagement, operational, and administrative support to theCLP to facilitate optimal use of the program.

Special Analytical Services (SAS). Non-standardized analyses conducted under the CLP to meet user requirements that cannotbe met using RAS, such as shorter analytical turnaround time, lower detection limits, and analysis of non-standard matricesor non-TCL compounds

Statement of Work (S0 W) for the CLP. A document that specifies the instrumentation, sample handling procedures, analyticalparameters and procedures, required quantitation limits, quality control requirements, and report format to be used by CLPlaboratories. The SOW also contains the TAL and TCL..

Target Analyte List (TAL). Developed by EPA for Superfund site sample analyses. The TAL is a list of 23 metals plus totalcyanide routinely analyzed using RAS.

Target Compound List (TCL). Developed by EPA for Superfund site sample analyses. The TCL, is a list of analytes (34volatile organic chemicals, 65 semivolatile organic chemicals, 19 pesticides, 7 polychlorinated biphenyls. 23 metals, andtotal cyanide) routinely analyzed using RAS.

contaminant concentrations in the keysources and media of interest;2

characteristics of sources, especiallyinformation related to release potential;and

characteristics of the environmentalsetting that may affect the fate, transport,and persistence of the contaminants.

Most of these data are obtained during thecourse of a remedial investigation/feasibility study(RI/FS). Other sources of information, such aspreliminary assessment/site inspection (PA/SI)reports, also may be available.

4.1.2 DATA NEEDS AND THE RI/FS

The RI/FS has four primary data collectioncomponents:

(1) characterization of site conditions;

3

(2) determination of the nature of thewastes;

(3) risk assessment; and

(4) treatability testing.

The site and waste characterization components ofthe RI/FS are intended to determinecharacteristics of the site (e.g., ground-watermovement, surface water and soil characteristics)and the nature and extent of contaminationthrough sampling and analysis of sources andpotentially contaminated media. Quantitative riskassessment, like site characterization, requires dataon concentrations of contaminants in each of thesource areas and media of concern. Riskassessment also requires information on othervariables necessary for evaluating the fate,transport, and persistence of contaminants andestimating current and potential human exposureto these contaminants. Additional data might berequired for environmental risk assessments (seeEPA 1989a).

Data also are collected during the RI/FS tosupport the design of remedial alternatives. Asdiscussed in the DQO guidance (EPA 1987a,b),such data include results of analyses ofcontaminated media "before and after" bench-scaletreatability tests. This information usually is notappropriate for use in a baseline risk assessmentbecause these media typically are assessed only fora few individual parameters potentially affected bythe treatment being tested. Also, initialtreatability testing may involve only a screeninganalysis that generally is not sensitive enough anddoes not have sufficient quality assurance/qualitycontrol (QA/QC) procedures for use inquantitative risk assessment.

4.1.3 EARLY IDENTIFICATION OF DATANEEDS

Because the RI/FS and other site studiesserve a number of different purposes (e.g., site and waste characterization, design of remedialalternatives), only a subset of this informationgenerally is useful for risk assessment. To ensurethat all risk assessment data needs will be met, itis important to identify those needs early in theRI/FS planning for a site. The earlier therequirements are identified, the better the chances

are of developing an RI/FS that meets the riskassessment data collection needs.

One of the earliest stages of the RI/FS atwhich risk assessment data needs can be addressedis the site scoping meeting. As discussed in theGuidance for Conducting Remedial Investigationsand Feasibility Studies Under CERCLA (EPA1988a, hereafter referred to as RI/FS guidance),the scoping meeting is part of the initial planningphase of site remediation. It is at this meetingthat the data needs of each of the RI/FScomponents (e.g., site and waste characterization)are addressed together. Scoping meeting attendeesinclude the RPM, contractors conducting theRI/FS (including the baseline risk assessment),onsite personnel (e.g., for construction), andnatural resource trustees (e.g., Department ofInterior). The scoping meeting allowsdevelopment of a comprehensive sampling andanalysis plan (SAP) that will satisfy the needs ofeach RI/FS component while helping to ensurethat time and budget constraints are met. Thus,in addition to aiding the effort to meet the riskassessment data needs, this meeting can helpintegrate these needs with other objectives of theRI/FS and thereby help make maximum use ofavailable resources and avoid duplication of effort.

During scoping activities, the risk assessorshould identify, at least in preliminary fashion, thetype and duration of possible exposures (e.g.,chronic, intermittent), potential exposure routes(e.g., ingestion of fish, ingestion of drinking water,inhalation of dust), and key exposure points (e.g.,municipal wells, recreation areas) for eachmedium. The relative importance of the potentialexposure routes and exposure points indetermining risks should be discussed, as shouldthe consequences of not studying them adequately.Section 4.5 and Chapter 6 provide guidance foridentifying exposure pathways that may exist athazardous waste sites. If potential exposurepathways are identified early in the RI/FS process,it will be easier to reach a decision on thenumber, type, and location of samples needed toassess exposure.

During the planning stages of the RI/FS, therisk assessor also should determine if non-routine(i.e., lower) quantitation limits are needed toadequately characterize risks at a site. SpecialAnalytical Services (SAS) of the EPA Contract

4

Laboratory Program (CLP) may be needed toachieve such lower quantitation limits. (See

Section 4.8 for additional information concerningquantitation limits.)

4.1.4 USE OF THE DATA QUALITYOBJECTIVES (DQO) GUIDANCE

The DQO guidance (EPA 1987a,b) providesinformation on the review of site data and the

on the active role that the risk assessor mustplay.)

After data have been collected, they shouldbe carefully reviewed to identify reliable, accurate,and verifiable numbers that can be used toquantify risks. All analytical data must beevaluated to identify the chemicals of potentialconcern (i.e., those to be carried through the riskassessment). Chapter 5 discusses the criteria to

determination of data quality needs for sampling be considered in selecting the subset of chemical(see the box below).

OVERVIEW OF DQO GUIDANCE

According to the DQ0 guidance (EPA 1987a andb), DQO are qualitative and quantitative statementsestablished prior to data collection, which specify thequality of the data required to support Agency decisionsduring remedial response activities The DQO for aparticular site vary according to the end use of the data(i.e., whether the data are collected to supportpreliminary assessments/site inspections, remedial investigations/feasibility studies, remedial designs, orremedial actions).

The DQO process consists of three stages. In Stage1 (Identify Decision Types), all available site infomationis compiled and analyzed in order to develop aconceptual model of the site that describes suspectedsources, contaminant pathways, and potential receptors.The outcome of Stage 1 is a definition of the objectivesof the site investigation and an identification of datagaps. Stage 2 (Identify Data Uses/Needs) involvesspecifying the data necessary to meet the objectives setin Stage 1, selecting the sampling approaches and theanalytical options for the site, and evaluating multiple-option approaches to allow more timely or cost-effectivedata collection and evaluation. In Stage 3 (Design DataCollection Program), the methods to be used to obtaindata of acceptable quality are specified in such productsas the SAP or the workplan.

Use of this guidance will help ensure that allenvironmental data collected in support of RI/FSactivities are of known and documented quality.

4.1.5 OTHER DATA CONCERNS

The simple existence of a data collection plandoes. not guarantee usable data. The risk assessorshould plan an active role in oversight of datacollection to ensure that relevant data have beenobtained. (See Section 4.9 for more information

data appropriate for baseline risk assessment.Data that do not meet the criteria are notincluded in the quantitative risk assessment; theycan be discussed qualitatively in the riskassessment report, however, or may be the basisfor further investigation.

4.2 REVIEW OF AVAILABLE SITEINFORMATION

Available site information must be reviewedto (1) determine basic site characteristics, (2)initially identify potential exposure pathways andexposure points, and (3) help determine dataneeds (including modeling needs). All availablesite information (i.e., information existing at thestart of the RI/FS) should be reviewed inaccordance with Stage 1 of the DQO process.Sources of available site information include:

RI/FS scoping information;

PA/SI data and Hazard Ranking System(HRS) documentation;

listing site inspection (LSI) data(formally referred to as expanded siteinspection, or ESI);

photographs (e.g., EPA’s EnvironmentalPhotographic Interpretation Center[EPIC]);

records on removal actions taken at thesite; and

information on amounts of hazardoussubstances disposed (e.g., from siterecords).

5

If available, LSI (or ESI) data are especially usefulbecause they represent fairly extensive site studies.

Based on a review of the existing data, therisk assessor should formulate a conceptual modelof the site that identifies all potential or suspectedsources of contamination, types and concentrationsof contaminants detected at the site, potentiallycontaminated media, and potential exposurepathways, including receptors (see Exhibit 4-l).As discussed previously, identification of potentialexposure pathways, especially the exposure points,is a key element in the determination of dataneeds for the risk assessment. Details concerningdevelopment of a conceptual model for a site areprovided in the DQO guidance (EPA 1987a,b) andthe RI/FS guidance (EPA 1988a).

In most cases, site information available atthe start of the RI/FS is insufficient to fullycharacterize the site and the potential exposurepathways. The conceptual model developed at thisstage should be adequate to determine theremaining data needs. The remainder of thischapter addresses risk assessment data needs indetail.

4.3 ADDRESSING MODELINGPARAMETER NEEDS

As discussed in detail in Chapter 6,contaminant release, transport, and fate modelsare often needed to supplement monitoring datawhen estimating exposure concentrations.Therefore, a preliminary site modeling strategyshould be developed during RI/FS scoping toallow model input data requirements to beincorporated into the data collection requirements.This preliminary identification of models andother related data requirements will ensure thatdata for model calibration and validation arecollected along with other physical and chemicaldata at the site. Exhibit 4-2 lists (by medium)several site-specific parameters often needed toincorporate fate and transport models in risk

assessments.

Although default values for some modelingparameters are available, it is preferable to obtainsite-specific values for as many input parametersas is feasible. If the model is not sensitive to a

particular parameter for which a default value isavailable, then a default value may be used.Similarly, default values may be used if obtainingthe site-specific model parameter would be tootime consuming or expensive. For example,certain airborne dust emission models use adefault value for the average wind speed at thesite; this is done because representativemeasurements of wind speed at the site wouldinvolve significant amounts of time (i.e., sampleswould have to be collected over a large part ofthe year).

Some model parameters are needed only ifthe sampling conducted at a site is sufficient tosupport complex models. Such model parametersmay not be necessary if only simple fate andtransport models are used in the risk assessment.

4.4 DEFINING BACKGROUNDSAMPLING NEEDS

Background sampling is conducted todistinguish site-related contamination fromnaturally occurring or other non-site-related levelsof chemicals, The following subsections define thetypes of background contamination and provideguidance on the appropriate location and numberof background samples.

4.4.1 TYPES OF BACKGROUND

There are two different types of backgroundlevels of chemicals:

(1)

(2)

naturally occurring levels, which areambient concentrations of chemicalspresent in the environment that have notbeen influenced by humans (e.g.,aluminum, manganese); and

anthropogeenic levels, which areconcentrations of chemicals that arepresent in the environment due tohuman-made, non-site sources (e.g.,industry, automobiles).

Background can range from localized toubiquitous. For example, pesticides -- most ofwhich are not naturally occurring (anthropogenic)-- may be ubiquitous in certain areas (e.g.,

6

EXHIBIT 4-1

ELEMENTS OF A CONCEPTUAL EVALUATION MODEL

HYPOTHESES TOVARIABLES BE TESTED

SOURCE: EPA 1987a

7

EXHIBIT 4-2

EXAMPLES OF MODELING PARAMETERS FOR WHICHINFORMATION MAY NEED TO BE OBTAINED DURING

A SITE SAMPLING INVESTIGATION

Type of Modeling Modeling Parametersa

Source Characteristics Geometry, physical/chemical conditions, emission rate, emissionstrength, geography

Soil Particle size, dry weight, pH, redox potential, mineral class, organiccarbon and clay content, bulk density, soil porosity

Ground-water Head measurements, hydraulic conductivity (pump and slug testresults), saturated thickness of aquifer, hydraulic gradient, pH,redox potential, soil-water partitioning

Ai r

Surface Water

Prevailing wind direction, wind speeds, stability class, topography,depth of waste, contaminant concentration in soil and soil gas,fraction organic content of soils, silt content of soils, percentvegetation, bulk density of soil, soil porosity

Hardness, pH, redox potential, dissolved oxygen, salinity,temperature, conductivity, total suspended solids, flow ratesand depths for rivers/streams, estuary and embaymentparameters such as tidal cycle, saltwater incursion extent,depth and area, lake parameters such as area, volume, depth,depth to thermocline

Sediment Particle size distribution, organic content, pH, benthic oxygenconditions, water content

Biota Dry weight, whole body, specific organ, and/or edible portionchemical concentrations, percent moisture, lipid content,size/age, life history stage

a These parameters are not necessarily limited to the type of modeling with which they areassociated in this exhibit. For example, many of the parameters listed for surface water are alsoappropriate for sediments.

8

agricultural areas); salt runoff from roads duringperiods of snow may contribute high ubiquitouslevels of sodium. Polycyclic aromatichydrocarbons (PAHs) and lead are other examplesof anthropogenic, ubiquitous chemicals, althoughthese chemicals also may be present at naturallyoccurring levels in the environment due to naturalsources (e.g., forest fires may be a source ofPAHs, and lead is a natural component of soils insome areas).

4.4.2 BACKGROUND SAMPLINGLOCATIONS

Background samples are collected at or nearthe hazardous waste site in areas not influencedby site contamination. They are collected fromeach medium of concern in these offsite areas.That is, the locations of background samples mustbe areas that could not have receivedcontamination from the site, but that do have thesame basic characteristics as the medium ofconcern at the site.

Identifying background location requiresknowing which direction is upgradient/upwind/upstream. In general, the direction of water flowtends to be relatively constant, whereas thedirection of air flow is constantly changing.Therefore, the determination of backgroundlocations for air monitoring requires constant andconcurrent monitoring of factors such as winddirection.

4.4.3 BACKGROUND SAMPLE SIZE

In appropriate circumstances, statistics maybe used to evaluate background sample data.Because the number of background samplescollected is important for statistical hypothesistesting, at some sites a statistician should beconsulted when determining background samplesize. At all sites, the RPM should decide thelevel of statistical analysis appliable to aparticular situation.

Often, rigorous statistical analyses areunnecessary because site- and non-site-relatedcontamination clearly differ. For most sites, theissue will not be whether a difference in chemicalconcentrations can be demonstrated betweencontaminated and background areas, but ratherthat of establishing a reliable representation of the

extent (in three dimensions) of a contaminatedarea. However, statistical analyses are requiredat some sites, making a basic understanding ofstatistics necessary. The following discussionoutlines some basic statistical concepts in thecontext of background data evaluation for riskassessment. (A general statistics textbook shouldbe reviewed for additional detail. Also, the boxbelow lists EPA guidance that might be useful.)

STATISTICAL METHODS GUIDANCE

Statistical Methods for Evaluating Ground-water Monitoring Data from Hazardous WasteFacilities (EPA 1988b)

Surface Impoundment Clean ClosureGuidance Manual (EPA 1988c)

Love Canal Emergency Declaration AreaHabitability Study (EPA 1988d)

Soils Sampling Quality Assurance Guide (EPA1989b)

A statistical test of a hypothesis is a ruleused for deciding whether or not a statement (i.e..the null hypothesis) should be rejected in favor ofa specified alternative statement (i.e., thealternative hypothesis). In the context ofbackground contamination at hazardous wastesites, the null hypothesis can be expressed as“there is no difference between contaminantconcentrations in background areas and onsite,"and the alternative hypothesis can be expressed as“concentrations are higher onsite.” This expressionof the alternative hypothesis implies a one-tailedtest of significance.

The number of background samples collectedat a site should be sufficient to accept or rejectthe null hypothesis with a specified likelihood oferror. In statistical hypothesis testing there aretwo types of error. The null hypothesis may berejected when it is true (i.e., a Type I error), ornot rejected when it is false (i.e., a Type II error).An example of a Type I error at a hazardouswaste site would be to conclude that contaminantconcentrations in onsite soil are higher thanbackground soil concentrations when in fact they

9

are not. The corresponding Type II error wouldbe to conclude that onsite contaminantconcentrations are not higher than backgroundconcentrations when in fact they are. A Type Ierror could result in unnecessary remediation,while a Type II error could result in a failure toclean up a site when such an action is necessary.

In customary notations, (alpha) denotes theprobability that a Type I error will occur, and (beta) denotes the probability that a Type II errorwill occur. Most statistical comparisons refer to , also known as the level of significance of thetest. If = 0.05, there is a 5 percent (i.e., 1 in20) chance that we will conclude thatconcentrations of contaminants are higher thanbackground when they actually are not.

Equally critical considerations in determiningthe number of background samples are concept called “power.” The power of a statisticaltest has the value 1 - and is defined as thelikelihood that the test procedure detects a falsenull hypothesis. Power functions for commonlyused statistical tests can be found in most generalstatistical textbooks. Power curves are a functionof a (which normally is freed at 0.05), sample size(i.e., the number of background and/or onsitesamples), and the amount of variability in thedata. Thus, if a 15 percent likelihood of failingto detect a false null hypothesis is desired (i.e., = 0.15), enough background samples must becollected to ensure that the power of the test isat least 0.85.

A small number of background samplesincreases the likelihood of a Type II error. If aninsufficient number of background samples iscollected, fairly large differences between site andbackground concentrations may not be statisticallysignificant, even though concentrations in themany site samples are higher than the fewbackground samples. To guard against thissituation, the statistical power associated with thecomparison of background samples with sitesamples should be evaluated.

In general, when trying to detect smalldifferences as statistically significant, the numberof background samples should be similar to thenumber of onsite samples that will be used for thecomparison(s) (e.g., the number of samples takenfrom one well). (Note that this does not mean

that the background sample size must equal thetotal number of onsite samples.) Due to theinherent variability of air concentrations (seeSection 4.6), background sample size for air needsto be relatively large.

4.4.4 COMPARING BACKGROUNDSAMPLES TO SITE-RELATEDCONTAMINATION

The medium sampled influences the kind ofstatistical comparisons that can be made withbackground data. For example, air monitoringstations and ground-water wells are normallypositioned based on onsite factors and gradientconsiderations. Because of this purposiveplacement (see Section 4.6.1), several wells ormonitors cannot be assumed to be a randomsample from a single population and hence cannotbe evaluated collectively (i.e., the sampling resultscannot be combined). Therefore, the informationfrom each well or air monitor should be comparedindividually with background.

Because there typically are many site-related,media-specific sampling location data to comparewith background, there usually is a "multiplecomparison problem" that must be addressed. Ingeneral, the probability of experiencing a Type Ierror in the entire set of statistical tests increaseswith the number of comparisons being made. Ifa = 0.05, there is a 1 in 20 chance of a Type Ierror in any single test. If 20 comparisons arebeing made, it therefore is likely that at least oneType I error will occur among all 20 tests.Statistical Analysis of Ground-water MonitoringData at RCRA Facilities (EPA 1989c) is usefulfor designing sampling plans for comparinginformation from many freed locations withbackground.

It may be useful at times to look atcomparisons other than onsite versus background.For example, upgradient wells can be comparedwith downgradient wells. Also, there may beseveral areas within the site that should becompared for differences in site-relatedcontaminant concentration. These areas ofconcern should be established before samplingtakes place. If a more complicated comparisonscheme is planned, a statistician should beconsulted frequently to help distribute thesampling effort and design the analysis.

10

A statistically significant difference betweenbackground samples and site-related contaminationshould not, by itself, trigger a cleanup action. Theremainder of this manual still must be applied sothat the toxicological -- rather than simply thestatistical -- significance of the contamination canbe ascertained.

4 . 5 PRELIMINARY IDENTIFI-CATION OF POTENTIALHUMAN EXPOSURE

A preliminary identification of potentialhuman exposure provides much neededinformation for the SAP. This activity involvesthe identification of (1) media of concern, (2)areas of concern (i.e., general locations of themedia to be sampled), (3) types of chemicalsexpected at the site, and (4) potential routes ofcontaminant transport through the environment(e.g., inter-media transfer, food chain). Thissection provides general information on thepreliminary identification of potential humanexposure pathways, as well as specific informationon the various media. (Also, see Chapter 6 fora detailed discussion of exposure assessment.)

4.5.1 GENERAL INFORMATION

Prior to discussing various specific exposuremedia, general information on the following isprovided: media, types of chemicals, areas ofconcern, and routes of contaminant transport isaddressed.

Media of concern (including biota). For riskassessment purposes, media of concern at a siteare:

any currently contaminated media towhich individuals may be exposed orthrough which chemicals may betransported to potential receptors: and

any currentlv uncontaminated media thatmay become contaminated in the futuredue to contaminant transport.

Several medium-specific factors in sampling mayinfluence the risk assessment. For example,limitations in sampling the medium may limit the

detailed evaluation of exposure pathways describedin Chapter 6. To illustrate this, if soil samplesare not collected at the surface of a site, then itmay not be possible to accurately evaluatepotential exposures involving direct contact withsoils or exposures involving the release ofcontaminants from soils via wind erosion (withsubsequent inhalation of airborne contaminants byexposed individuals). Therefore, based on theconceptual model of the site discussed previously,the risk assessor should make sure thatappropriate samples are collected from eachmedium of concern.

Areas of concern. Areas of concern refer tothe general sampling locations at or near the site.For large sites, areas of concern may be treatedin the RI/FS as “operable units,” and may includeseveral media. Areas of concern also can bethought of as the locations of potentially exposedpopulations (e.g., nearest residents) or biota (e.g.,wildlife feeding areas).

Areas of concern should be identified basedon site-specific characteristics. These areas arechosen purposively by the investigators during theinitial scoping meeting. Areas of concern shouldinclude areas of the site that

(1)

(2)

(3)

(4)

(5)

(6)

have different chemical types;

have different anticipated concentrationsor hot spots;

are a release source of concern;

differ from each other in terms of theanticipated spatial or temporal variabilityof contamination;

must be sampled using differentequipment and/or

are more or less costly to sample.

In some instances, the risk assessor may wantto estimate concentrations that are representativeof the site as a whole, in addition to each area ofconcern. In these cases, two conditions generallyshould be met in defining areas of concern: (1)the boundaries of the areas of concern should notoverlap and (2) all of the areas of concern

11

together should account for the entire area of thesite.

Depending on the exposure pathways thatare being evaluated in the risk assessment, it maynot be necessary to determine site-widerepresentative values. In this case, areas ofconcern do not have to account for the entirearea of the site.

Types of chemicals. The types of chemicalsexpected at a hazardous waste site may dictate thesite areas and media sampled. For example,certain chemicals (e.g., dioxins) thatbioconcentrate in aquatic life also are likely to bepresent in the sediments. If such chemicals areexpected at a particular site and humans areexpected to ingest aquatic life, sampling ofsediments and aquatic life for the chemicals maybe particularly important.

Due to differences in the relative toxicities ofdifferent species of the same chemical (e.g., Cr+3

versus Cr+6), the species should be noted whenpossible.

Routes of contaminant transport. In additionto medium-specific concerns, there may be severalpotential current and future routes of contaminanttransport within a medium and between media ata site. For instance, discharge of ground water orsurface runoff to surface water could occur.Therefore, when possible, samples should becollected based on routes of potential transport.For cases in which contamination has not yetreached points of human exposure but may betransported to those areas in the future, samplingbetween the contaminant source and the exposurelocations should be conducted to help evaluatepotential future concentrations to whichindividuals may be exposed (e.g., throughmodeling). (See Chapter 6 for additionaldiscussion on contaminant transport.)

4.5.2 SOIL

Soil represents a medium of direct contactexposure and often is the main source ofcontaminants released into other media. As such,the number, location, and type of samplescollected from soils will have a significant effecton the risk assessment. See the box on this page

for guidance that provides additional detailedinformation concerning soil sampling, includinginformation on sampling locations, general soiland vegetation conditions, and samplingequipment, strategies, and techniques. In additionto the general sampling considerations discussedpreviously, the following specific issues related tosoil sampling are discussed below theheterogeneous nature of soils, designation of hotspots, depth of samples, and fate and transportproperties.

Heterogeneous nature of soils. One of thelargest problems in sampling soil (or other solidmaterials) is that its generally heterogeneousnature makes collection of representative samplesdifficult (and compositing of samples virtuallyimpossible -- see Section 4.6.3). Therefore, alarge number of soil samples may be required toobtain sufficient data to calculate an exposureconcentration. Composite samples sometimes arecollected to obtain a more homogeneous sampleof a particular area; however, as discussed in alater section, compositing samples also serves tomask contaminant hot spots (as well as areas oflow contaminant concentration).

Designation of hot spots. Hot spots (i.e.,areas of very high contaminant concentrations)may have a significant impact on direct contactexposures. The sampling plan should considercharacterization of hot spots through extensivesampling, field screening, visual observations, ora combination of the above.

12

Depth of samples. Sample depth should beapplicable for the exposure pathways andcontaminant transport routes of concern andshould be chosen purposively within that depthinterval. If a depth interval is chosen purposively,a random procedure to select a sampling pointmay be established. Assessment of surfaceexposures will be more certain if samples arecollected from the shallowest depth that can bepractically obtained, rather than, for example, zeroto two feet. Subsurface soil samples areimportant, however, if soil disturbance is likely orif leaching of chemicals to ground water is ofconcern, or if the site has current or potentialagricultural uses.

Fate and transport properties. The samplingplan should consider physical and chemicalcliaracteristics of soil that are important forevaluating fate and transport. For example, soilsamples being collected to identify potentialsources of ground-water contamination must beable to support models that estimate bothquantities of chemicals leaching to ground waterand the time needed for chemicals to leach to andwithin the ground water.

4.5.3 GROUND WATER

Considerable expense and effort normally arerequired for the installation and development ofmonitoring wells and the collection of ground-water samples. Wells must not introduce foreignmaterials and must provide a representativehydraulic connection to the geologic formations ofinterest. In addition, ground-water samples needto be collected using an approach that adequatelydefines the contaminant plume with respect topotential exposure points. Existing potentialexposure points (e.g., existing drinking water wells)should be sampled.

More detailed information concerning ground-water sampling considerations (e.g., samplingequipment, types, and techniques) can be found inthe references in the box on this page. Inaddition to the general sampling considerationsdiscussed previously in Section 4.5.1, those specificfor ground water -- hydrogeologic properties, welllocation and depth, and filtered vs. unfilteredsamples -- are discussed below.

Hydrogeologic properties. The extent towhich the hydrogeologic properties (e.g., hydraulicconductivity, porosity, bulk density, fractionorganic carbon, productivity) of the aquifer(s) arecharacterized may have a significant effect on therisk assessment. The ability to estimate futureexposure concentrations depends on the extent towhich hydrogeologic properties needed to evaluatecontaminant migration are quantified. Repetitivesampling of wells is necessary to obtain samplesthat are unaffected by drilling and welldevelopment and that accurately reflecthydrogeologic properties of the aquifer(s).

Well location and depth. The location Ofwells should be such that both the horizontal andvertical extent of contamination can becharacterized. Separate water-bearing zones mayhave different aquifer classifications and uses andtherefore may need to be evaluated separately inthe risk assessment. In addition, sinking orfloating layers of contamination may be presentat different depths of the wells.

Filtered vs. unfiltered samples. Data fromfiltered and unfiltered ground-water samples areuseful for evaluating chemical migration in groundwater, because comparison of chemical

13

concentrations in unfiltered versus filtered samplescan provide important information on the form inwhich a chemical exists in ground water. Forinstance, if the concentration of a chemical ismuch greater in unfiltered samples compared tofiltered samples, it is likely that the majority ofthe chemical is sorbed onto particulate matter andnot dissolved in the ground water. Thisinformation on the form of chemical (i.e.,dissolved or suspended on particulate matter) isimportant to understanding chemical mobilitywithin the aquifer.

If chemical analysis reveals significantlydifferent concentrations in the filtered andunfiltered samples, try to determine whether thereis a high concentration of suspended particles orif apparently high concentrations are due tosampling or well construction artifacts.Supplementary samples can be collected in amanner that will minimize the influence of theseartifacts. In addition, consider the effects of thefollowing.

Filter size. A 0.45 urn filter may screenout some potentially mobile particulateto which contaminants are absorbed andthus under-represent contaminantconcentrations. (Recent researchsuggests that a 1.0 urn may be a moreappropriate filter size.)

Pumping velocitv. Pumping at too higha rate will entrain particulate (to whichcontaminants are absorbed) that wouldnot normally be mobile; this couldoverestimate contaminant concentrations.

Sample oxidation. After contact with air,many metals oxidize and form insolublecompounds that may be filtered out thismay underestimate inorganic chemicalconcentrations.

Well construction materials. Corrosionmay elevate some metal concentrationseven in stainless steel wells.

If unfiltered water is of potable quality, datafrom unfiltered water samples should be used toestimate exposure (see Chapter 6). The RPMshould ultimately decide the type of samples that

are collected. If only one type of sample iscollected (e.g., unfiltered), justification for notcollecting the other type of sample (e.g., filtered)should be provided in the sampling plan.

4.5.4 SURFACE WATER AND SEDIMENT

Samples need to be collected from any nearbysurface water body potentially receiving dischargefrom the site. Samples are needed at a sufficientnumber of sampling points to characterizeexposure pathways, and at potential dischargepoints to the water body to determine if the site(or some other source) is contributing to surfacewater/sediment contamination. Some importantconsiderations for surface water/sediment samplingthat may affect the risk assessment for varioustypes and portions of water bodies (i.e., loticwaters, lentic waters, estuaries, sediments) arediscussed below. More detailed informationconcerning surface water and sediment sampling,such as selecting sampling locations and samplingequipment, types, and techniques, is provided inthe references given in the box below.

14

Lotic waters. Lotic waters are fast-movingwaters such as rivers and streams. Variations inmixing across the stream channel and downstreamin rivers and streams can make it difficult toobtain representative samples. Although theselection of sampling points will be highlydependent on the exposure pathways of concernfor a particular site, samples generally should betaken both toward the middle of the channelwhere the majority of the flow occurs and alongthe banks where flow is generally lower. Samplinglocations should be downgradient of any possiblecontaminant sources such as tributaries or effluentoutfalls. Any facilities (e.g., dams, wastewatertreatment plants) upstream that affect flow volumeor water quality should be considered during thetiming of sampling. “Background” releasesupstream could confound the interpretation of

sampling results by diluting contaminants or byincreasing contaminant loads. In general,sampling should begin downstream and proceedupstream.

Lentic waters. Lentic waters are slow-movingwaters such as lakes, ponds, and impoundments.In general, lentic waters require more samplesthan lotic waters because of the relatively lowdegree of mixing of lentic waters. Thermalstratification is a major factor to be consideredwhen sampling lakes. If the water body isstratified, samples from each layer should beobtained. Vertical composites of these layers thenmay be made, if appropriate. For small shallowponds, only one or two sample locations (e.g., theintake and the deepest points) may be adequatedepending on the exposure pathways of concernfor the site. Periodic release of water should beconsidered when sampling impoundments, as thismay affect chemical concentrations andstratification.

Estuaries. Contaminant concentrations inestuaries will depend on tidal flow and salinity-stratification, among other factors. To obtain arepresentative sample, sampling should beconducted through a tidal cycle by taking threesets of samples on a given day: (1) at low tide;(2) at high tide; and (3) at “half tide.” Each layerof salinity should be sampled.

Sediments. Sediment samples should becollected in a manner that minimizes disturbanceof the sediments and potential contamination of

subsequent samples. Sampling in flowing watersshould begin downstream and end upstream.Wading should be avoided. Sediments of differentcomposition (i.e., mud, sand, rock) should not becomposite. Again, it is important to obtain datathat will support the evaluation of the potentialexposure pathways of concern. For example, forpathways such as incidental ingestion, sampling ofnear-shore sediments may be important; however,for dermal absorption of sediment contaminantsduring recreational use such as swimming, samplesfrom different points throughout the water bodymay be important. If ingestion of benthic(bottom-dwelling) species or surface water will beassessed during the risk assessment, sedimentshould be sampled so that characteristics neededfor modeling (e.g., fraction of organic carbon,particle size distribution) can be determined (seeSection 4.3).

4.5.5 AIR

Guidance for developing an air sampling planfor Superfund sites is provided in Procedures forDispersion Modeling and Air Monitoring forSuperfund Air Pathway Analysis (EPA 1989e).That document is Volume IV of a series of fourtechnical guidance manuals called Procedures forConducting Air Pathway Analyses for SuperfundApplications (EPA 1989e-h). The other threevolumes of the series include discussions ofpotential air pathways, air emission sources, andprocedures for estimating potential sourceemission rates associated with both the baselinesite evaluation and remedial activities at the site.

Air monitoring information, along withrecommendations for proper selection andapplication of air dispersion models, is includedin Volume IV. The section on air monitoringcontained in this volume presents step-by-stepprocedures to develop, conduct, and evaluate theresults of air concentration monitoring tocharacterize downwind exposure conditions fromSuperfund air emission sources. The first stepaddressed is the process of collecting andreviewing existing air monitoring informationrelevant to the specific site, including source,receptor, and environmental data. The secondstep involves determining the level ofsophistication for the air monitoring program; thelevels range from simple screening procedures torefined techniques. Selection of a given level will

15

depend on technical considerations (e.g., detectionlimits) and available resources. The third step onair monitoring is development of the airmonitoring plan and includes determination of thetype of air monitors, the number and location ofmonitors, the frequency and duration ofmonitoring, sampling and analysis procedures, andQA/QC procedures. Step four details the day-to-day activities related to conducting the airmaintenance and calibration, and documentationof laboratoty results and QA/QC procedures. Thefifth and final step involves the proceduresnecessary to (1) summarize and evaluate the airmonitoring results for validity, (2) summarize thestatistics used, (3) determine site-related airconcentrations (by comparison of upwind anddownwind concentrations), and (4) estimateuncertainties in the results related to themonitoring equipment and program and theanalytical techniques used in the laboratory.

Given the difficulties of collecting sufficientair samples to characterize both temporal andspatial variability of air concentrations, modeling-- along or in conjunction with monitoring -- isoften used in the risk assessment. For the mostefficient sampling program, the section in VolumeIV on modeling should be used in conjunctionwith the section on monitoring.

Volume IV also contains a comprehensivebibliography of other sources of air monitoringand modeling guidance. Note, however, that whilethis volume contains an extensive discussion onplanning and conducting air sampling, it does notprovide details concerning particular monitoringequipment and techniques. The box on this pagelists some sources of detailed information on airsampling. The following paragraphs addressseveral specific aspects of air sampling: temporaland spatial considerations, emission sources,meteorological conditions.

Temporal and spatial considerations. Thegoal of air sampling at a site is to adequatelycharacterize air-related contaminant exposures. Ata minimum, sampling results should be adequatefor predictive short-term and long-term modeling.When evaluating long-term inhalation exposures,sample results should be representative of thelong-term average air concentrations at the long-term exposure points. This requires an airsampling plan of sufficient temporal scale to

encompass the range of meteorological andclimatic conditions potentially affecting emissions,and of sufficient spatial scale to characterizeassociated air concentrations at potential exposurepoints. If acute or subchronic exposures resultingfrom episodes of unusually large emissions are ofinterest, sampling over a much smaller time scalewould be needed.

Emission sources. Selection of theappropriate type of air monitor will depend onthe emission source(s) being investigated as wellas the exposure routes to be evaluated. Forexample, if inhalation of dust is an exposurepathway of concern, then the monitoringequipment must be able to collect respirable dustsamples.

Meteorological conditions. Site-specificmeteorological conditions should be obtained (e.g.,from the National Weather Service) or recordedduring the air sampling program with sufficientdetail and quality assurance to substantiate andexplain the air sampling results. The review ofthese meteorological data can help indicate thesampling locations and frequencies.Meteorological characteristics also will benecessary if air modeling is to be conducted.

4.5.6 BIOTA

Organisms sampled for human health riskassessment purposes should be those that arelikely to be consumed by humans. This mayinclude animals such as commercial and game fish(e.g., salmon, trout, catfish), shellfish (e.g., oysters,clams, crayfish), fowl (e.g., pheasant, duck), and

16

terrestrial mammals (e.g., rabbit, deer), as well asplants such as grains (e.g., wheat, corn), vegetables(e.g., spinach, carrots), and fruit (e.g., melons,strawberries). An effort should be made tosample species that are consumed most frequentlyby humans. Guidance for collecting biota samplesis provided in the references given in the boxbelow. The following paragraphs address thefollowing special aspects of biota samplingportion vs. whole sampling, temporal concerns,food preference, fish sampling, involvement byother agencies.

Portion vs. whole sampling. If only humanexposure is of concern, chemical concentrationsshould be measured only in edible portion(s) ofthe biota. For many fish species, estimates ofconcentrations in fillets (skin on or skin off) arethe most appropriate measures of exposureconcentrations. Whole body measurements maybe needed, however, for certain species of fishand/or for environmental risk assessments. Forexample, for some species, especially small ones(e.g., smelt), whole body concentrations are mostappropriate. (See Risk Assessment Guidance forSuperfund Environmental Evaluation Manual(EPA 1989a) for more information concerningbiota sampling for environmental assessment.)The edible portion of an organism can vary withspecies and with the potentially exposedsubpopulation.

Temporal concerns, Any conditions that mayresult in non-representative sampling, such assampling during a species’ migration or whenplants are not in season, should be avoided.

Food preferences. At some sites, humansubpopulations in the area may have differentfood consumption patterns that need to beevaluated. For example, some people commonlyeat the hepatopancreas of shellfish. In thesecases, o r g a n c o n c e n t r a t i o n s w o u l d b e m o s tappropriate for estimating exposure. Anotherexample of a less common food preference isconsumption of relatively large quantities ofseaweed and other less commonly eaten seafoodsin some Asian communities.

Fish sampling. It is recommended that fishof “catchable” size be sampled instead of young,small fish because extremely young fish are notlikely to be consumed. Older, larger fish alsogenerally are more likely to have been exposed tosite-specific contaminants for a long time,although for some species (e.g., salmon) thereverse is true. Both bottom-dwelling (benthic)and open-water species should be sampled if bothare used as a food source.

Other agencies. Biota sampling may involveother federal agencies such as the Fish andWildlife Service or the Department of Agriculture.The equivalent state agencies also may beinvolved. In such cases, these agencies should beinvolved early in the scoping process.

4.6 DEVELOPING AN OVERALLSTRATEGY FOR SAMPLECOLLECTION

For each medium at a site, there are severalstrategies for collecting samples. The samplingstrategies for a site must be appropriate for usein a quantitative risk assessment; if inappropriate,even the strictest. QA/QC procedures associatedwith the strategy will not ensure the usability ofsample results. Generally, persons actuallyconducting the field investigation will determinethe strategy. As discussed in Section 4.1, riskassessors also should be involved in discussionsconcerning the strategy. The following areas ofmajor concern (from a risk assessment

17

perspective) are discussed in this section: samplesize, sampling location, types of samples, temporaland meteorological factors, field analyses, and cost of sampling. Many of these areas also arediscussed for specific media in Section 4.5. Seethe box in the opposite column and Section 4.5for more detailed guidance on sampling strategy.

4.6.1 DETERMINE SAMPLE SIZE

Typically, sample size and sample location(see Section 4.6.2) are determined at the sametime. Therefore, much of the discussion in thissubsection is also pertinent to determiningsampling location. The discussion on statistics inSection 4.4 is useful for both sample size andlocation determinations.

A number of considerations are associatedwith determining an appropriate number ofsamples for a risk assessment. Theseconsiderations include the following four factors:

(1)

(2)

(3)

(4)

number of areas of concern that will besampled;

statistical methods that are planned;

statistical performance (i.e., variabilitypower, and certainty) of the data thatwill be collected; and

practical considerations of logistics andcost.

In short, many decisions must be made by therisk assessor related to the appropriate samplesize for an investigation. A statistician cannotestimate an appropriate sample size without thesupporting information provided by a risk assessor.The following paragraphs discuss these four factorsas they relate to sample size determinations.

Areas of concern. A major factor thatinfluences how many samples are appropriate isthe number of areas of concern that areestablished prior to sampling. As discussed in thenext subsection, if more areas of concern areidentified, then more samples generally will beneeded to characterize the site. If the totalvariability in chemical concentrations is reducedsubstantially by subdividing the site into areas ofconcern, then the statistical performance should

improve and result in a more accurate assessmentof the site.

Statistical methods. A variety of statisticalmanipulations may need to be performed on thedata used in the risk assessment. For example,there may be comparisons with backgroundconcentrations, estimates of upper confidencelimits on means, and determinations of theprobability of identifying hot spots. Each of theseanalyses requires different calculations fordetermining a sample size that will yield a

18

specified statistical performance. Some of theavailable guidance, such as the Ground-waterMonitoring guidance (EPA 1986c), the RCRADelisting guidance (EPA 1985d), and the SoilsCleanup Attainment guidance (EPA 1988f),address these strategies in detail.

Statistical performance (i.e., variability,power, and certainty). If samples will be takenfrom an area that is anticipated to have a highdegree of variability in chemical concentrations,then many samples may be required to achieve aspecified level of certainty and power. Ifcontaminant concentrations in an area are highlyvariable and only a few samples can be obtained,then the risk assessor should anticipate (1) a greatdeal of uncertainty in estimating meanconcentrations at the site, (2) difficulty in definingthe distribution of the data (e.g., normal), and (3)upper confidence limits much higher than themean. Identification of multiple areas of concern-- each with its own set of samples and descriptivestatistics -- will help reduce the total variability ifthe areas of concern are defined so that they arevery different in their contaminant concentrationprofiles. Risk assessors should discuss in thescoping meeting both the anticipated variability inthe data and the desired power and certainty ofthe statistics that will be estimated from the data.

As discussed in Section 4.4.3, power is thelikelihood of detecting a false null hypothesis.Power is particularly important when comparingsite characteristics with background. For example,if a 10 percent difference in mean concentrationsneeds to be determined with 99 percent likelihood(i.e., power of 0.99), a very large number ofsamples will likely be needed (unless the site andbackground variabilities are extremely low). Onthe other hand, if the investigator is onlyinterested in whether the onsite average conditionsare 100 times larger than background or canaccept a lower chance of detecting the differenceif it exists (i.e., a lower power), then a smallersample size could be accommodated.

The other statistical performance quantitybesides power that may need to be specified isthe certainty of the calculations. One minus thecertainty is the significance level (i.e., or falsepositive rate (see also Section 4.4.3). The higherthe desired certainty level (i.e., the lower thesignificance level), the greater the true difference

must be to observe a statistical difference. In thecase of upper confidence limits on estimates ofmean concentrations, the higher the desiredcertainty level, the higher will be the upperconfidence limit. This follows from the fact thatin general, as certainty increases (i.e., becomessmaller), the size of the confidence interval alsoincreases.

Practical considerations. Finally, questionsof practicality, logistics, sampling equipment,laboratory constraints, quality assurance, and costinfluence the sample size that will be available fordata analysis. After the ideal sample size hasbeen determined using other factors, practicalconsiderations can be introduced to modify thesample size if necessaty.

4.6.2 ESTABLISH SAMPLING LOCATIONS

There are three general strategies forestablishing sample locations: (1) purposive, (2)completely random, and (3) systematic. Variouscombinations of these general strategies arepossible and acceptable.

Much of the discussion on statistics in thepreceding subsection and in Section 4.4 isappropriate here. Typically, a statistician shouldbe consulted when determining sampling location.

Purposive sampling. Although areas ofconcern are established purposively (e.g., with theintention of identifying contamination), thesampling locations within the areas of concerngenerally should not be sampled purposively if thedata are to be used to provide defensibleinformation for a risk assessment. Purposivelyidentified sampling locations are not discouragedif the objective is site characterization, conductinga chemical inventory, or the evaluation of visuallyobvious contamination. The sampling results,however, may overestimate or underestimate thetrue conditions at the site depending on thestrategies of the sampling team. Due to the biasassociated with the samples, data from purposivelyidentified sampling locations generally should notbe averaged, and distributions of these datagenerally should not be modeled and used toestimate other relevant statistics. After areas ofconcern have been established purposively,ground-water monitoring well locations,continuous air monitor locations, and soil sample

19

locations should be determined randomly orsystematically within the areas of concern.

Random sampling. Random samplinginvolves selecting sampling locations in anunbiased manner. Although the investigator mayhave chosen the area of concern purposively, thelocation of random sampling points within thearea should be independent of the investigator(i.e., unbiased). In addition, the sampling pointsshould be independent of each other; that is, itshould not be possible to predict the location ofone sampling point based on the location ofothers. Random sampling points can beestablished by choosing a series of pairs ofrandom numbers that can be mapped onto acoordinate system that has been established foreach area of concern.

Several positive features are associated withdata collected in a random sampling program.First, the data can be averaged and used toestimate average concentrations for the area ofconcern (rather than simply an average of thesamples that were acquired). Second, estimates ofthe uncertainty of the average and thedistributional form of the concentrationmeasurements are informative and simple toestimate when they are determined from data thatwere obtained randomly. Finally, if there is atrend or systematic behavior to the chemicalconcentrations (e.g., sampling is occurring alonga chemical gradient), then random sampling ispreferred because it reduces the likelihood that allof the high concentration locations are sampled tothe exclusion of the low concentration locations.

Systematic sampling. Systematic samplelocations are established across an area of concernby laying out a grid of sampling locations thatfollow a regular pattern. Systematic samplingensures that the sampling effort across the area ofconcern is uniform and that samples are collectedin each area. The sampling location grid shouldbe determined by randomly identifying a singleinitial location from which the grid is constructed.If such a random component is not introduced,the sample is essentially purposive. The grid canbe formed in several patterns including square,rectangular, triangular, or hexagonal, depending onthe shape of the area. A square pattern is oftenthe simplest to establish. Systematic sampling ispreferable to other types of sampling if the

objective is to search for small areas with elevatedconcentrations. Also, geostatisticalcharacterizations -- as described in the DQOguidance (EPA 1987a,b) -- are best done with datacollected from a systematic sample.

Disadvantages of systematic sampling includethe need for special variance calculations in orderto estimate confidence limits on the averageconcentration. The Soils Cleanup Attainmentguidance (EPA 1988f) discusses these calculationsin further detail.

4.6.3 DETERMINE TYPES OF SAMPLES

Another item of concern is the determinationof the types of samples to be collected. Basically,two types of samples may be collected at a sitegrab and composite.

Grab samples. Grab samples represent asingle unique part of a medium collected at aspecific location and time.

Composite samples. Composite samples --sometimes referred to as continuous samples forair -- combine subsamples from different locationsand/or times. As such, composite samples maydilute or otherwise misrepresent concentrationsat specific points and, therefore, should be avoidedas the only inputs to a risk assessment. Formedia such as soil, sediment, and ground water,composite samples generally may be used to assessthe presence or absence of contamination,however, they may be used in risk assessment onlyto represent average concentrations (and thusexposures) at a site. For example, “hot spots”cannot be determined using composite samples.For surface water and air, composite samples maybe useful if concentrations and exposures areexpected to vary over time or space, as will oftenbe the case in a large stream or river.Composites then can be used to estimate daily ormonthly average concentrations, or to account forstratification due to depth or varying flow ratesacross a stream.

4.6.4 CONSIDER TEMPORAL ANDMETEOROLOGICAL FACTORS

Temporal (time) and meteorological(weather) factors also must be considered whendetermining sampling strategies. The sampling

20

design should account for fluctuations in chemicalconcentrations due to these factors because ingeneral, the variability in sampling resultsincreases with increasing complexity of thesefactors. When these factors are complex,specialized and detailed sampling designs areneeded to maintain a constant and certain level ofaccuracy in the results. Countering this need,however, is the cost of the sampling. Thefollowing paragraphs address the interactions ofthe single sampling event, annual/seasonalsampling cycle, variability estimation, and the costof sampling.

Single sampling event. Variability measuresfrom a single sampling event will underestimatethe overall variability of concentrations across anarea of concern, which in turn will result in theunderestimation of the confidence limits on themean. The reason for this underestimation is thattemporal variability is not included in anevaluation of the total environmental variabilityat the site.

Annual/seasonal sampling cycle. The idealsampling strategy incorporates a full annualsampling cycle. If this strategy cannot beaccommodated in the investigation, at least twosampling events should be considered. Thesesampling events should take place during oppositeseasonal extremes. For example, sampling periodsthat may be considered extremes in temporalsampling include (1) high waterflow water, (2)high recharge/low recharge, (3) windy/calm, and(4) high suspended solids/clear water. This typeof sampling requires some prior knowledge ofregional seasonal dynamics. In addition, asampling team that can mobilize rapidly might beneeded if the particular year of sampling is nottypical and the extreme conditions occur at anunusual time. See the box on this page forexamples of seasonal variability.

Variability estimation. The simple varianceestimators that are often used in risk assessmentrequire that the data are independent oruncorrelated. Certain types of repeated samples,however, (e.g., those from ground-water wells orair monitors) actually are time series data thatmight be correlated. In other words, theconcentration of a contaminant in an aquifermeasured at a well on a given day will depend, inpart, on what the concentration in the aquifer was

SEASONAL VARIABLITY

Regardless of the medium sampled, samplecomposition may vary depending on the time of yearand weather conditions when the sample is collected.For example, rain storms may greatly alter soil composition and thus affect the types and concentrationsof chemicals present on solid material; heavyprecipitation and runoff from snowmelt may directlydilute chemical concentrations or change the types ofchemicals present in surface water heavy rain also mayresult in sediment loading to water bodies, which couldincrease contamination or affect the concentrations ofother contaminants through adsorption and settling inthe water column; if ground-water samples are collectedfrom an area heavily dependent on ground water forirrigation, the composition of a sample collected duringthe summer growing season may greatly differ from thecomposition of a sample collected in the winter.

on the previous day. To reduce this dependence(e.g., due to seasonal variability), sampling ofground-water wells and air monitors should beeither separated in time or the data should beevaluated using statistical models with varianceestimators that can accommodate a correlationstructure. Otherwise, if time series data that arecorrelated are treated as a random sample andused to calculate upper confidence limits on themean, the confidence limits will beunderestimated.

Ideally, samples of various media should becollected in a manner that accounts for time andweather factors. If seasonal fluctuations cannot becharacterized in the investigations, detailsconcerning meteorological, seasonal, and climaticconditions during sampling must be documented.

4.6.5 USE FIELD SCREENING ANALYSES

An important component of the overallsampling strategy is the use of field screeninganalyses. These types of analyses utilizeinstruments that range from relatively simple (e.g.,hand-held organic vapor detectors) to moresophisticated (e.g., field gas chromatography).(See Field Screening Methods Catalog [EPA 1987h]for more information.) Typically, field screeningis used to provide threshold indications ofcontamination. For example, on the basis of soilgas screening, the field investigation team maydetermine that contamination of a particular area

21

is indicated and therefore detailed sampling iswarranted. Although field screening resultsusually are not directly used in the risk

assessment, they are useful for streamliningsampling and the overall RI/FS process.

4.6.6 CONSIDER TIME AND COST OFSAMPLING

Two primary constraints in sampling are timeand cost. Time consuming or expensive samplingstrategies for some media may prohibit multiplesampling points. For example, multiple ground-water wells and air monitors on a grid samplingpattern are seldom located within a single area ofconcern. However, multiple surface water and soilsamples within each area of concern are easier toobtain. In the case of ground water and air,several areas of concern may have to be collapsedinto a single area so that multiple samples will beavailable for estimating environmental variabilityor so that the dynamics of these media can beevaluated using accepted models of fate andtransport.

In general, it is important to remember whendeveloping the sampling strategy that detailedsampling must be balanced against the time andat involved. The goal of RI/FS sampling is notexhaustive site characterization, but rather toprovide sufficient information to form the basisfor site remediation.

4.7 QA/QC MEASURES

This section presents an overview of thefollowing qualityassurance/quality control(QA/QC) considerations that are of particularimportance for risk assessment sampling:sampling protocol, sampling devices, QC samples,collection procedures, and sample preservation.Note, however, that the purpose of this discussionis to provide background information; the riskassessor will not be responsible for most QA/QCevaluations.

The Quality Assurance Field OperationsManual (EPA 1987g) should be reviewed. Inaddition, the EPA Environmental MonitoringSupport Laboratory in Las Vegas, Nevada,

(EMSL-LV) currently is writing a guidancedocument concerning the development of qualityassurance sample designs for Superfund siteinvestigations. Regional QA/QC contacts (e.g.,the regional Environmental Semites Division) orEMSL-LV should be consulted if moreinformation concerning QA/QC procedures forsampling is desired.

4.7.1 SAMPLING PROTOCOL

The sampling protocol for a risk assessmentshould include the following:

objectives of the study;

procedures for sample collection,preservation, handling, and transport;and

analytical strategies that will be used.

Presenting the objectives of the RI sampling isparticularly important because these objectivesalso will determine the focus of the riskassessment. There should be instructions ondocumenting conditions present during sampling(e.g., weather conditions, media conditions).Persons collecting samples must be adequatelytrained and experienced in sample collection. Testevaluations of the precision attained by personsinvolved in sample collection should bedocumented (i.e., the individual collecting asample should do so in a manner that ensuresthat a homogeneous, valid sample is reproduciblyobtained). The discussion of analytical strategiesshould specify quantitation limits to be achievedduring analyses of each medium.

4.7.2 SAMPLING DEVICES

The devices used to collect, store, preserve,and transport samples must not alter the samplein any way (i.e., the sampling materials cannot bereactive, sorptive, able to leach analytes, or causeinterferences with the laboratory analysis). Forexample, if the wrong materials are used toconstruct wells for the collection of ground-watersamples, organic chemicals may be adsorbed to thewell materials and not be present in the collectedsample.

22

4.7.3 QC SAMPLES

Field QC samples (e.g., field blanks, tripblanks, duplicates, split samples) must becollected, stored, transported, and analyzed in amanner identical to those for site samples Themeaning and purpose of blank samples arediscussed in detail in Chapter 5. Field duplicatesamples are usually two samples collectedsimultaneously from the same sampling locationand are used as measures of either thehomogeneity of the medium sampled in aparticular location or the precision in sampling.Split samples are usually one sample that isdivided into equal fractions and sent to separateindependent laboratories for analysis. These splitsamples are used to check precision and accuracyof laboratory analyses. Samples may also be splitin the same laboratory, which can provideinformation on precision. The laboratoryanalyzing the samples should not be aware of theidentity of the field QC samples (e.g., labels onQC samples should be identical to those on thesite samples).

4.7.4 COLLECTION PROCEDURES

Collection procedures should not alter themedium sampled. The general environmentsurrounding the location of the sample shouldremain the same so that the collected samplesare representative of the situation due to the siteconditions, not due to conditions posed by thesampling equipment.

4.7.5 SAMPLE PRESERVATION

Until analysis by the laboratory, anychemicals in the samples must be maintained asclose to the same concentrations and identitiesas in the environment from which they came.Therefore, special procedures may be needed topreserve the samples during the period betweencollection and analysis.

4.8 SPECIAL ANALYTICALSERVICES

EPA’s SAS, operated by the CLP, may benecessary for two main reasons: (1) the standardlaboratory methods used by EPA’s Routine

Analytical Services (RAS) may not be appropriate(e.g., lower detection limits may be needed),4 and(2) chemicals other than those on the targetcompound list (TCL, i.e., chemicals usuallyanalyzed under the Superfund program) may besuspected at the site and therefore may need to beanalyzed. A discussion on the RAS detectionlimits is provided in Chapter 5. Additionalinformation on SAS can be found in the User’sGuide to the Contract Laboratory Program (EPA1988i).

In reviewing the historical data at a site, therisk assessor should determine if non-TCLchemicals are expected. A indicated above, non-TCL chemicals may require. special samplecollection and analytical procedures using SAS.Any such needs should be discussed at the scopingmeeting. SAS is addressed in greater detail inChapter 5.

4.9 TAKING AN ACTIVE ROLEDURING WORKPLANDEVELOPMENT AND DATACOLLECTION

The risk assessor should be sure to take anactive role during workplan development and datacollection. This role involves three main steps:

(1)

(2)

(3)

present risk assessment sampling needsat the scoping meeting

contribute to the workplan and reviewthe Sampling and Analysis Plan; and

conduct interim reviews of outputs ofthe field investigation.

See Chapter 9 for information on the role of theRPM during workplan development and datacollection.

4.9.1 PRESENT RISK ASSESSMENTSAMPLING NEEDS AT SCOPINGMEETING

At the scoping meeting, the uses of samplesand data to be collected are identified, strategiesfor sampling and analysis are developed, DQOsare established, and priorities for sample collection

23

are assigned based on the importance of the datain meeting RI/FS objectives. One of the RI/FS

objectives, of course, is the baseline riskassessment. Therefore, the risk assessment dataneeds and their fit with those of other RI/FScomponents are discussed. If certain riskassessment sampling needs are judged infeasibleby the scoping meeting attendees, all personsinvolved with site investigation should be madeaware of the potential effects of exclusion on therisk assessment.

4.9.2 CONTRIBUTE TO WORKPLAN ANDREVIEW SAMPLING AND ANALYSISPLAN

The outcome of the scoping meeting is thedevelopment of a workplan and a SAP. Theworkplan documents the decisions and evaluationsmade during the scoping process and presentsanticipated future tasks, while the SAP specifiesthe sampling strategies, the numbers, types, andlocations of samples, and the level of qualitycontrol. The SAP consists of a quality assuranceproject plan (QAPjP) and a field sampling plan(FSP). Elements of the workplan and the SAPare discussed in detail in Appendix B of the RI/FSguidance (EPA 1988a). Both the workplan andthe SAP generally are written by the personnelwho will be involved in the collection of thesamples; however, these documents should bereviewed by all personnel who will be using theresulting sample data.

Review the workplan. The workplan shoulddescribe the tasks involved in conducting the riskassessment. It also should describe thedevelopment of a preliminary assessment of publichealth and environmental impacts at the site. Therisk assessor should review the completedworkplan to ensure that all feasible riskassessment sampling needs have been addressed asdiscussed in the scoping meeting. In particular,this review should focus on the descriptions oftasks related to:

field investigation (e.g., source testing,media sampling), especially with respectto

-- background concentrations bymedium,

-- quantification of present and futureexposures, e.g.,

- exposure pathways

- present and potential future landuse

- media that are or may becontaminated

- locations of actual and potentialexposure

- present concentrations atappropriate exposure points,

-- data needs for statistical analysis ofthe above, and

-- data needs for fate and transportmodels;

sample analysis/validation, especially withrespect to

-- chemicals of concern, and-- analytical quantification levels;

data evaluation; and

assessment of risks.

In reviewing the above, the precise informationnecessary to satisfy the remainder of this guidanceshould be anticipated.

Review the SAP. The risk assessor shouldcarefully review and evaluate all sections of theSAP to determine if data gaps identified in theworkplan will be addressed adequately by thesampling program. Of particular importance isthe presentation of the objectives. In the QAPjPcomponent of the SAP, the risk assessor shouldpay particular attention to the QA/QC proceduresassociated with sampling (e.g., number of fieldblanks, number of duplicate samples -- see Section4.8). The SAP should document the detailed, site-specific procedures that will be followed to ensurethe quality of the resulting samples. Specialconsiderations in reviewing the SAP are discussedin Section 4.1.3.

24

In reviewing the FSP, pay particular attentionto the information on sample location andfrequency, sampling equipment and procedures,and sample handling and analysis. As discussedin Section 4.5, the sampling procedures shouldaddress:

each medium of concern,

background concentration,

all potential exposure points within eachmedium;

migration to potential exposure points,including data for models;

potential exposures based on possiblefuture land uses;

sufficient data to satisfy concerns aboutdistributions of sampling data andstatistics; and

number and location of samples.

The analytical plans in the FSP should bereviewed to ensure that DQOs set during thescoping meeting will be met.

The SAP may be revised or amended severaltimes during the site investigation. Therefore, areview of all proposed changes to the SAP thatpotentially may affect the data needs for riskassessment is necessary. Prior to any changes inthe SAP during actual sampling, compliance of the

changes with the objectives of the SAP must bechecked. (If risk assessment objectives are notspecified in the original SAP, they will not beconsidered when changes to an SAP areproposed.)

4.9.3 CONDUCT INTERIM REVIEWS OFFIELD INVESTIGATION OUTPUTS

All sampling results should be reviewed assoon as they are available to determine if the riskassessment data needs outlined in the workplanhave been met by the sampling. Compare theactual number, types, and locations of samplescollected with those planned in the SAP.Sampling locations frequently are changed in thefield when access to a planned sampling locationis obstructed. The number of samples collectedmay be altered if, for instance, there is aninsufficient amount of a certain medium to collectthe planned number of samples (e.g., if severalwells are found to be dry).

If certain sampling needs have not been met,then the field investigators should be contacted todetermine why these samples were not collected.If possible, the risk assessor should obtain samplesto fill these data gaps. If time is critical, SpecialAnalytical Services (see Section 4.7) may be usedto shorten the analytical time. If this is notpossible, then the risk assessor should evaluate allsampling results as discussed in Chapter 5,documenting the potential effect that these datagaps will have on the quantitative risk assessment.In general, the risk assessment should not bepostponed due to these data gaps.

25


1. Some information that is appropriate for the assessment of human health risks also may be suitable and necessary for anenvironmental evaluation of the site. Procedures for conducting an environmental evaluation of the hazardous waste site are outlinedin the companion volume of this guidance, the Environmental Evaluation Manual (EPA 1989a), and are not discussed in this chapter.

2. The term “media” refers to both environmental media (e.g., soil) and biota (e.g., fish).

3. “Areas of Concern” within the context of this guidance should be differentiated from the same terminology used by the Great Lakesenvironmental community. This latter use is defined by the International Joint Commission as an area found to be exceeding the GreatLakes Water Quality Agreement objectives.

4. New routine services that provide lower detection limits are currently under development. Contact the headquarters AnalyticalOperation Branch for further information.

26


American Society of Testing and Materials (ASTM). Undated. A Proposed Guide for Sediment Collection, Storage, Characterization,and Manipulation. Draft. Available from G. Allen Burton, Dept of Biological Sciences, Wright State University, Dayton, Ohio45435.

Provides information concerning how to collect contaminated sediments, sediment spiking, dilution procedures,and QA/QC. Will probably be in the annual ASTM manual.

Environmental Protection Agency (EPA). 1981. Procedures for Handling and Chemical Analysis of Sediment and Water Samples.Great Lakes Laboratory.

Environmental Protection Agency (EPA). 1983. Technical Assistance Document for Sampling and Analysis of Toxic OrganicCompounds in Ambient Air. Office of Research and Development.

Provides guidance to persons involved in designing and implementing ambient air monitoring programs for toxicorganic compounds. Includes guidance on selecting sampling/analytical methods, sampling strategy, QAprocedures, and data format. Outlines policy issues.

Environmental Protection Agency (EPA). 1984. Sediment Sampling Qualitv Assurance User's Guide. Environmental MonitoringSupport Laboratory. Las Vegaa, NV. NTIS PB-85-233-542.

Overview of selected sediment models presented as a foundation for stratification of study of regions andselection of locations for sampling sites, methods of sampling, sampling preparation and analysis. Discussionof rivers, lakes, and estuaries.

Environmental Protection Agency (EPA). 1985a. Practical Guide to Ground-water Sampling. Environmental Research Laboratory.Ada, OK EPA 600/2-85/104.

Contains information on laboratory and field testing of sampling materials and procedures. Emphasizesminimizing errors in sampling and analysis.

Environmental Protection Agency (EPA). 1985b. Methods Manual for Bottom Sediment Sample Collection. Great Lakes NationalProgram Office. EPA 905/4-85/004.

Provides guidance on survey planning, sample collection, document preparation, and quality assurance forsediment sampling surveys. Sample site selection, equipment/containers, collection field observation, preservation,handling custody procedures

Environmental Protection Agency (EPA). 1985c. Cooperative Agreement on the Monitoring of Contaminants in Great Lakes SportFish for Human Health Purposes. Region V, Chicago, IL

Discusses sampling protocols and sample composition used for sport fish (chinook salmon, coho salmon, laketrout, and rainbow trout), maximum composite samples (5 fish) and length ranges which would be applicableto hazardous waste sites contaminating lakes or streams used for recreational fishing.

Environmental Protection Agency (EPA). 1985d. Petitions to Delist Hazardous Wastes Guidance Manual. Office of Solid Waste.EPA/530/SW-85/003.

Environmental Protection Agency (EPA). 1986a. Test Methods for Evaluating Solid Waste (SW-846} Physical/Chemical Methods.Office of Solid Waste.

Provides analytical procedures to test solid waste to determine if it is a hazardous waste as defined under RCRA. Containsinformation for collecting solid waste samples and for determining reactivity, corrosivity, ignitability, composition of waste,and mobility of waste compounds.

Environmental Protection Agency (EPA). 1986b. Field Manual for Grid Sampling of PCB Spill Sites to Verify Cleanups. Office ofToxic Substances. EPA/560/5-86/017.

Provides detailed, step-by-step guidance for using hexagonal grid sampling includes sampling design, collection,QA/QC and reporting.

21

Environmental Protection Agency (EPA). 1986c. Resource Conservation and Recovery Act (RCRA) Ground-water MonitoringTechnical Enforcement Guidance Document. Office of Waste Programs Enforcement.

Contains a detailed presentation of the elements and procedures essential to the design and operation of ground-water monitoring systems that meet the goals of RCRA and its regulations. Includes appendices on statisticalanalysis and some geophysical techniques.

Environmental Protection Agency (EPA). 1987a. Data Quality Objectives for Remedial Response Activities: Development Process.Office of Emergency and Remedial Response and Office of Waste Programs Enforcement. EPA/540/G-87/003. (OSWERDirective 9335.O-7B).

Identifies (1) the framework and process by which data quality objectives (DQOs qualitative and quantitativestatements that specify the quality of the data required to support Agency decisions during remedial responseactivities) are developed and (2) the individuals responsible for development of DQOs. Provides proceduresfor determining a quantifiable degree of certainty that can be used in making site-specific decisions. Providesa formal approach to integration of DQO development with sampling and anafysis plan development. Attemptsto improve the overall quality and cost effectiveness of data collection and analysis activities.

Environmental Protection Agency (EPA). 1987b. Data Qualitv Objectives for Remedial Response Activities Example Scenario: RI/FSActivities at a Site with Contaminated Soils and Ground Water. Office of Emergency and Remedial Response and Office ofWaste Programs Enforcement. EPA/540/G-87/004.

Companion to EPA 1987a. Provides detailed examples of the process for development of data quality objectives(DQOs) for RI/FS activities under CERCLA.

Environmental Protection Agency (EPA). 1987c. A Compendium of Superfund Field Operations Methods. Office of Emergency andRemedial Response. EPA/540/P-87/001. (OSWER Directive 9355.0-14).

Environmental Protection Agency (EPA). 1987d. Handbook Ground Water. Office of Research and Development. EPA/625/6-87/016.

Resource document that brings together the available technical information in a form convenient for personnelinvolved in ground-water management. Also addresses minimization of uncertainties in order to make reliablepredictions about contamination response to corrective or preventative measures.

Environmental Protection Agency (EPA). 1987e. An Overview of Sediment Qualitv in the United States. Office of Water Regulationand Standards.

Good primer. Contains many references.

Environmental Protection Agency (EPA). 1987f. Expanded Site Inspection (ESI} Transitional Guidance for FY 1988. Office ofEmergency and Remedial Response. (OSWER Directive 9345.1-.02).

Provides reader with a consolidated ready reference of general methodologies and activities for conductinginspection work on site-s being investigated for the NPL.

Environmental Protection Agency (EPA). 1987g. Quality Assurance Field Operations Manual. Office of Solid Waste and EmergencyResponse.

Provides guidance for the selection and definition of field methods, sampling procedures, and custodyresponsibilities.

Environmental Protection Agency (EPA). 1987h. Field Screening Methods Catalog. Office of Emergency and Remedial Response.

Provides a listing of methods to be used during field screening, and includes method descriptions, theirapplication to particular sites, their limitations and uses, instrumentation requirements, detection limits, andprecision and accuracy information.

Environmental Protection Agency (EPA). 1988a. Guidance for Conducting Remedial Investigations and Feasibility Studies UnderCERCLA. Interim Final. Office of Emergency and Remedial Response. (OSWER Directive 9355.3-01).

Provides the user (e.g., EPA personnel, state agencies, potentially responsible parties (PRPs), federal facilitycordinators, and contractors assisting in RI/FS-related activities) with an overall understanding of the RI/FSprocess Includes general information concerning scoping meetings, the development of conceptual modefs atthe beginning of a site investigation, sampling, and analysis.

28

Environmental Protection Agency (EPA). 1988b. Statistical Methods for Evaluating Ground Water from Hazardous Waste Facilities.Office of Solid Waste.

Specifies five different statistical methods that are appropriate for ground-water monitoring. Outlines aamplingprocdures and performance standards that are designed to help minimize the occurrence of Type I and TypeII errors.

Environmental Protection Agency (EPA). 1988c. Surface Impoundment Clean Closure Guidance Manual. Office of Solid Waste.

Environmental Protection Agency (EPA). 1988d. Love Canal Emergency Declaration Area Habitability Study Report. Prepared byCH2M Hill and Life Systems for EPA Region II.

Provides a formal comparison of samples with background as well as detailed discussions concerning problemsassociated with sampling to evaluate data.

Environmental Protection Agency (EPA). 1988e. Guidance on Remedial Actions for Contaminated Ground Water at Superfund Sites.Interim Final. Office of Emergency and Remedial Response. (OSWER Directive 9283.1-2).

Provides guidance to develop, evaluate, and select ground-water remedial actions at Superfund sitea, focusingon policy issues and establishing cleanup levels. Also includes discussion of data collection activities forcharacterization of contamination.

Environmental Protection Agency (EPA). 1988f. Statistical Methods for Evaluating the Attainment of Superfund Cleanup Standards.Volume I Soils and Solid Media. Draft. Office of Policy, Planning, and Evaluation.

Provides statistical procedures that can be used in conjunction with attainment objectives defined by EPA todetermine, with the desired confidence, whether a site does indeed attain a cleanup standard. It also providesguidance on sampling of soils to obtain baseline information onsite, monitor cleanup operations, and verifyattainment of cleanup objectives

Environmental Protection Agency (EPA). 1988g. Proposed Guidelines for Exposure-related Measurements. 53 Federal Register 48830(December 2, 1988).

Focuses on general principles of chemical measurements in various physical and biologicalthose who must recommend, conduct, or evaluate an exposure assessment.

Environmental Protection Agency (EPA). 1988h. Interim Report on Sampling Design Methodology.Support Laborstory. Las Vegas, NV. EPA/600/X-88/408.

media. Assists

Environmental Monitoring

Provides guidance concerning the statistical determination of the number of samples to be collected.

Environmental Protection Agency (EPA). 1988i. User’s Guide to the Contract Laboratory Program. Office of Emergency andRemedial Response.

Environmental Protection Agency (EPA). 1989a. Risk Assessment Guidance for Superfund Environmental Evaluation Manual.Interim Final. Office of Emergency and Remedial Response. EPA/540/l-89/00lA (OSWER Directive 9285.7-01),

Environmental Protection Agency (EPA). 1989b. Soil Sampling Qualitv Assurance Guide. Review Draft. Environmental MonitoringSupport Laboratory. Las Vegas, NV.

Replaces earlier edition NTIS Pb-84-198-621. Includes DQOs, QAPP, information concerning the purposeof background sampling, selection of numbers of samples and sampling sites, error control, sample design,sample documentation.

Environmental Protection Agency (EPA). 1989c. Statistical Analysis of Ground-water Monitoring Data at RCRA Facilities. Officeof Solid Waste.

Environmental Protection Agency (EPA). 1989d. Ground-water Sampling for Metals Analyses. Office of Solid Waste and EmergencyResponse. EPA/540/4-89-001.

29

Environmental Protection Agency (EPA). 1989e. Air Superfund National Technical Guidance Series. Volume IV Procedures forDispersion Modeling and Air Monitoring for Superfund Air Pathway Analysis. Interim Final. Office of Air Quality Planning andStandards. Research Triangle Park NC. EPA/450/l-89004.

This volume discusses procedures for dispersion modeling and air monitoring for superfund air pathway analyses.Contains recommendations for proper selection and application of air dispersion models and procedures todevelop, conduct, and evaluate the results of air concentration monitoring to characterize downwind exposureconditions from Superfund air emission sources.

Environmental Protection Agency (EPA). 1989f. Air Superfund National Technical Guidance Series. Volume I: Application of AirPathway Analyses for Superfund Activities. Interim Final. Office of Air Quality Planning and Standards. Research Triangle ParkNC. EPA/450/l-89/001.

Provides recommended procedures for the conduct of air pathway analyses (APAs) that meet the needs of theSuperfund program. The procedures are intended for use by EPA remedial project managers, enforcementproject managers, and air experts as well as by EPA Superfund contractors. The emphasis of this volume isto provide a recommended APA procedure relative to the remedial phase of the Superfund process.

Environmental Protection Agency (EPA). 1989g. Air Superfund National Technical Guidance Series. Volume II Estimation ofBaseline Air Emissions at Superfund sites. Interim Final. Office of Air Quality Planning and Standards. Research Triangle Park,NC. EPA/450/l-89/002.

This volume provides information concerning procedures for developing baseline emissions from landfills andlagoons. Describes baseline emissions from both undisturbed sites and sites where media-disturbing activitiesare taking place. The procedures described for landfills may be applied to solid hazardous waste, and thosefor lagoons may be applied to liquid hazardous waste.

Environmental Protection Agency (EPA). 1989h. Air Superfund National Technical Guidance Series. Volume III: Estimation of Air Emissions from Cleanup Activities at Superfund Sites. Interim Final. Office of Air Quality Planning and Standards. Research

Triangle Park, NC. EPA/450/l-89/003.

This volume provides technical guidance for estimating air emissions from remedial activities at NPL sites thatmay impact local air quality for both onsite workers at a site and the surrounding community while the remedialactivities are occurring. Discusses methods to characterize air quality impacts during soil removal, incineration,and air stripping.

Environmental Protection Agency (EPA). 1989i. Guidance Manual for Assessing Human Health Risks from Chemically ContaminatedFish and Shellfish. Office of Marine and Estuarine Protection. EPA/503/8-89/002.

Study designed to measure concentrations of toxic substances in edible tissues of fish and shellfish.

Environmental Protection Agency (EPA) and Army Corps of Engineers (COE). 1981. Procedures for Handling and Chemical Analysisof Sediment and Water Samples. Technical Committee on Dredged and Fill Material. Technical Report EPA/DE-81-l.

Food and Drug Administration (FDA). 1977. Pesticide Analytical Manual. Volume I.

Provides a skin-on fillet (whole fish sampling) protocol used in USEPA monitoring of sportfish in the GreatLakes. Also includes information on compositing.

Food and Drug Administration (FDA). 1986. Pesticides and Industrial Chemicals in Domestic Foods.

Provides guidance for sampling designs for fishery products from the market.

Freeman, H.M. 1989. Standard Handbook of Hazardous Waste Treatment and Disposal. McGraw-Hill. New York.

Provides detailed information concerning sampling and monitoring of hazardous wastes at remedial action sites(Chapters 12 and 13).

Gilbert, R.O. 1987. Statistical Methods for Environmental Pollution Monitoring. Van Nostrand Reinhold. New York.

Provides statistical analysis information by providing sampling plans, statistical tests, parameter estimationprocedure techniques and references to pertinent publications. The statistical techniques discussed are relativelysimple, and examples, exercise, and case studies are provided to illustrate procedures.

CHAPTER 5

DATA EVALUATION

CHAPTER 5

DATA EVALUATION

After a site sampling investigation has beencompleted (see Chapter 4), a large quantity ofanalytical data is usually available. Each samplemay have been analyzed for the presence of overone hundred chemicals, and many of thosechemicals may have been detected. The followingnine steps should be followed to organize the datainto a form appropriate for a baseline riskassessment:

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

gather all data available from the siteinvestigation and sort by medium(Section 5.1);

evaluate the analytical methods used(Section 5.2);

evaluate the quality of data with respectto sample quantitation limits (Section5.3);

evaluate the quality of data with respectto qualifiers and codes (Section 5.4);

evaluate the quality of data with respectto blanks (Section 5.5);

evaluate tentatively identified compounds(Section 5.6);

compare potential site-relatedcontamination with background (Section5.7);

develop a set of data for use in the riskassessment (Section 5.8); and

if appropriate, further limit the numberof chemicals to be carried through therisk assessment (Section 5.9).

Prior to conducting any of these steps, theEPA remedial project manager (RPM) should beconsulted to determine if certain steps should bemodified, added, or deleted as a result of site-specific conditions. Also, some of the steps maybe conducted outside the context of the riskassessment (e.g., for the feasibility study). Therationale for not evaluating certain data based onany of these steps must be fully discussed in thetext of the risk assessment report.

The following sections address each of thedata evaluation steps in detail, and Exhibit 5-1presents a flowchart of the process. The outcomeof this evaluation is (1) the identification of a set

2

of chemicals that are likely to be site-related and(2) reported concentrations that are of acceptablequality for use in the quantitative risk assessment.If the nine data evaluation steps are followed, thenumber of chemicals to be considered in theremainder of the risk assessment usually will beless than the number of chemicals initiallyidentified. Chemicals remaining in thequantitative risk assessment based upon thisevaluation are referred to in this guidance as“chemicals of potential concern.”

5.1 COMBINING DATAAVAILABLE FROM SITEINVESTIGATIONS

Gather data, which may be from severaldifferent sampling periods and based on severaldifferent analytical methods, from all availablesources, including field investigation team (FIT)reports, remedial investigations, preliminary siteassessments, and ongoing site characterization andalternatives screening activities. Sort data by

medium. A useful table format for presentingdata is shown in Exhibit 5-2.

Evaluate data from different time periods todetermine if concentrations are similar or ifchanges have occurred between sampling periods.If the methods used to analyze samples fromdifferent time periods are similar in terms of thetypes of analyses conducted and the QA/QCprocedures followed, and if the concentrationsbetween sampling periods are similar, then thedata may be combined for the purposes ofquantitative risk assessment in order to obtainmore information to characterize the site. Ifconcentrations of chemicals change significantlybetween sampling periods, it may be useful tokeep the data separate and evaluate risksseparately. Alternatively, one could use only themost recent data in the quantitative riskassessment and evaluate older data in a qualitativeanalysis of changes in concentrations over time.The RPM should be consulted on the eliminationof any data sets from the risk assessment, andjustification for such elimination must be fullydescribed in the risk assessment report.

EXHIBIT 5-2

EXAMPLE OF OUTPUT FORMAT FOR VALIDATED DATA

Area X- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

Note: All values other than qualifiers must be entered as numbers, not as labels.

a Contract-required quantitation limit (unless otherwise noted). Values for illustration only.

b Refer to section 5.4 for an explanation of qualifiers.

c Sample quantitation limit.

5

5.2 EVALUATION OF ANALYTICALMETHODS

Group data according to the types of analysesconducted (e.g., field screening analysis,semivolatiles analyzed by EPA methods for waterand wastewater, semivolatiles analyzed by EPA’sSuperfund Contract Laboratory Program [CLP]procedures) to determine which analytical method

results are appropriate for use in quantitative riskassessment. Often, this determination has beenmade already by regional and contractor staff.

An overview of EPA analytical methods isprovided in the box below. Exhibit 5-3 presentsexamples of the types of data that are not usuallyappropriate for use in quantitative risk assessment,even though they may be available from a siteinvestigation.

6

EXHIBIT 5-3

EXAMPLES OF THE TYPES OF DATA POTENTIALLY UNSUITABLEFOR A QUANTITATIVE RISK ASSESSMENT

Analytical Instrumentor Method Purpose of Analysis Analytical Result

HNu Organic Vapor Detector Health and Safety, Total Organic VaporField Screen

Organic Vapor Analyzer Health and Safety, Total Organic VaporField Screen

Combustible Gas Indicator Health and Safety Combustible Vapors,Oxygen-deficientAtmosphere

Field Gas Chromatographya Field Screen/Analytical Specific Volatile andMethod Semi-volatile Organic

Chemicals

a Depending on the detector used, this instrument can be sufficiently sensitive to yield adequate data foruse in a quantitative risk assessment; however, a confirming analysis by GC/MS should be performed ona subset of the samples in a laboratory prior to use.

7

Analytical results that are not specific for aparticular compound (e.g., total organic carbon

[TOC], total organic halogens [TOX]) or resultsof insensitive analytical methods (e.g., analysesusing portable field instruments such as organicvapor analyzers and other field screening methods)may be useful when considering sources ofcontamination or potential fate and transport ofcontaminants. These types of analytical results,however, generally are not appropriate forquantitative risk assessment; therefore, the riskassessor may not want to include them in thesummary of chemicals of potential concern for thequantitative risk assessment. In addition, theresults of analytical methods associated withunknown, few, or no QA/QC procedures shouldbe eliminated from further quantitative use.These types of results, however, may be useful forqualitative discussions of risk in other sections ofthe risk assessment report.

The outcome of this step is a set of site datathat has been developed according to a standardset of sensitive, chemical-specific methods (e.g.,SW-846 Methods [EPA 1986], EPA 600 Methods[EPA 1984], CLP Statements of Work [EPA1988b,c]), with QA/QC procedures that are well-documented and traceable. The data resultingfrom analyses conducted under the CLP, whichgenerally comprise the majority of results availablefrom a Superfund site investigation, fall into thiscategory.

Although the CLP was developed to ensurethat consistent QA/QC methods are used whenanalyzing Superfund site samples, it does notensure that all analytical results are consistentlyof sufficient quality and reliability for use inquantitative risk assessment. Neither the CLP norQA/QC procedures associated with other methodsmake judgments concerning the ultimate “usability”of the data. Do not accept at face value allremaining analytical results, whether from the CLPor from some other set of analyticalmethodologies. Instead, determine -- according tothe steps discussed below -- the limitations anduncertainties associated with the data so that onlydata that are appropriate and reliable for use ina quantitative risk assessment are carried throughthe process.

5.3 EVALUATION OFQUANTITATION LIMITS

This step involves evaluation of quantitationlimits and detection limits (QLs and DLs) for allof the chemicals assessed at the site. Thisevaluation may lead to the re-analysis of somesamples, the use of “proxy” (or estimated)concentrations, and/or the elimination of certainchemicals from further consideration (because theyare believed to be absent from the site). Typesand definitions of QLs and DLs are presented inthe box on the next page.

Before eliminating chemicals because they arenot detected (or conducting any othermanipulation of the data), the following pointsshould be considered:

(1)

(2)

the sample quantitation limit (SQL) ofa chemical may be greater thancorresponding standards, criteria, orconcentrations derived from toxicityreference values (and, therefore, thechemical may be present at levels greaterthan these corresponding referenceconcentrations, which may result inundetected risk); and

a particular SQL may be significantlyhigher than positively detected values inother samples in a data set.

These two points are discussed in detail in thefollowing two subsections. A third subsectionprovides guidance for situations where only someof the samples for a given medium test positivefor a particular chemical. A fourth subsectionaddresses the special situation where SQLs are notavailable. The final subsection addresses thespecific steps involved with elimination ofchemicals from the quantitative risk assessmentbased on their QLs.

5.3.1 SAMPLE QUANTITATION LIMITS(SQLs) THAT ARE GREATER THANREFERENCE CONCENTRATIONS

As discussed in Chapter 4, QLs needed forthe site investigation should be specified in thesampling plan. - For some chemicals, however,SQLs obtained under RAS or SAS may exceed

8

certain reference concentrations (e.g., maximum Three points should be noted whencontaminant levels [MCLS], concentrations considering this example.corresponding to a 10-6 cancer risk). The box onthe next page illustrates this problem. For certain (1)chemicals (e.g., antimony), the CLP contract-required quantitation limits (CRQLs) exceed thecorresponding reference concentrations fornoncarcinogenic effects, based on the EPA-verifiedreference dose and a 2-liter per day ingestion ofwater by a 70-kilogram person.1 Estimation ofcancer risks for several other chemicals (e.g.,arsenic, styrene) at their CRQLs yields cancerrisks exceeding 10-4, based on the same wateringestion factors. Most potential carcinogens with (2)EPA-derived slope factors have CRQLs that yieldcancer risk levels exceeding 10-6 in water, andnone of the carcinogens with EPA-derived slopefactors have CRQL values yielding less than 10-7

cancer risk levels (as of the publication date ofthis manual; data not shown).

Review of site information and apreliminary determination of chemicalsof potential concern at a site prior tosample collection may allow thespecification of lower QLs (i.e., usingSAS) before an investigation begins (seeChapter 4). This is the most efficientway to minimize the problem of QLsexceeding levels of potential concern.

EPA’s Analytical Operations Branchcurrently is working to reduce the CRQLvalues for several chemicals on the TCLand TAL, and to develop an analyticalservice for chemicals with specialstandards (e.g., MCLs).

TYPES AND DEFINITIONS OF DETECTION LIMITS AND QUANTITATION LIMITS

Strictly interpreted, the detection limit (DL) is the lowest amount of a chemical that can be “seen” above the normal, randomnoise of an analytical instrument or method, A chemical present below that level cannot reliably be distinguished from noise.DLs are chemical-specific and instrument-specific and are determined by statistical treatment of multiple analyses in which theratio of the lowest amount observed to the electronic noise level (i.e., the signal-to-noise ratio) is determined. On any givenday in any given sample, the calculated limit may not be attainable; however, a properly calculated limit can be used as an overallgeneral measure of laboratory performance.

Two types of DLs may be described -- instrument DLs (IDLs) and method DLs (MDLs). The IDL is generally the lowestamount of a substance that can be detected by an instrument it is a measure only of the DL for the instrument, and does notconsider any effects that sample matrix, handling, and preparation may have. The MDL, on the other hand, takes into accountthe reagents, sample matrix, and preparation steps applied to a sample in specific analytical methods.

Due to the irregular nature of instrument or method noise, reproducible quantitation of a chemical is not possible at the DL.Generally, a factor of three to five is applied to the DL to obtain a quantitation limit (QL), which is considered to be the lowestlevel at which a chemical may be accurately and reproducibly quantitated. DLs indicate the level at which a small amount wouldbe “seen; whereas QLs indicate the levels at which measurements can be “trusted.”

Two types of QLs may be described -- contract-required QLs (CRQLs) and sample QLs (SQLs). (Contract-required detectionlimits [CRDL] is the term used for inorganic chemicals. For the purposes of this manual, however, CRQL will refer to bothorganic and inorganic chemicals.) In order to participate in the CLP, a laboratory must be able to meet EPA CRQLs. CRQLsare chemical-specific and vary depending on the medium analyzed and the amount of chemical expected to be present in thesample. As the name implies, CRQLs are not necessarily the lowest detectable levels achievable, but rather are levels that aCLP laboratory should routinely and reliably detect and quantitate in a variety of sample matrices. A specific sample mayrequire adjustments to the preparation or analytical method (e.g., dilution, use of a smaller sample aliquot) in order to beanalyzed. In these cases, the reported QL must in turn be adjusted. Therefore, SQLs, not CRQLs, will be the QLs of interestfor most samples. In fact, for the same chemical, a specific SQL may be higher than, lower than, or equal to SQL values forother samples. In addition, preparation or analytical adjustments such as dilution of a sample for quantitation of an extremelyhigh level of only one compound could result in non-detects for all other compounds included as analytes for a particularmethod, even though these compounds may have been present at trace quantities in the undiluted sample. Because SQLs takeinto account sample characteristics, sample preparation, and analytical adjustments, these values are the most relevant QLs forevaluating nondetected chemicals.

9

(3) In several situations, an analytical the chemical is present in the sample at the SQLlaboratory may be able to attain QLs in (see Section 5.3.4 for situations where SQLs areparticular samples that are below or not available). Carry the chemical through theabove the CRQL values screening risk assessment, essentially conducting

the assessment on the SQL for the particularIf SAS was not specified before sampling chemical. In this way, the risks that would be

began and/or if a chemical is not detected in any posed if the chemical is present at the SQL cansample from a particular medium at the QL, then be compared with risks posed by other chemicalsavailable modeling data, as well as professional at the site.judgment, should be used to evaluate whether thechemical may be present above reference Re-analyze the sample. This (preferred)concentrations. If the available information option discourages elimination of questionableindicates the chemical is not present, see Section chemicals (i.e., chemicals that may be present5.3.5 for guidance on eliminating chemicals. If below their QL but above a level of potentialthere is some indication that the chemical is concern) from the risk assessment. If time allowspresent, then either re-analyze selected samples and a sufficient quantity of the sample is available,using SAS, if time allows, or address the chemical submit a SAS request to re-analyze the samplequalitatively. In determining which option is most at QLs that are below reference concentrations.appropriate for a site, a screening-level risk The possible outcome of this option is inclusionassessment should be performed by assuming that of chemicals positively detected at levels above

10

reference concentrations but below the QLs thatwould normally have been attained under routineanalysis of Superfund samples in the CLPprogram.

Address the chemical qualitatively. A secondand less desirable option for a chemical that maybe present below its QL (and possibly above itshealth-based reference concentration) is toeliminate the chemical from the quantitative riskassessment, noting that if the chemical wasdetected at a lower QL, then its presence andconcentration could contribute significantly to theestimated risks.

5.3.2 UNUSUALLY HIGH SQLs

Due to one or more sample-specific problems(e.g., matrix interferences), SQLs for a particularchemical in some samples may be unusually high,sometimes greatly exceeding the positive resultsreported for the same chemical in other samplesfrom the data set. Even if these SQLs do not

exceed health-based standards or criteria, they maystill present problems. If the SQLs cannot bereduced by re-analyzing the sample (e.g., throughthe use of SAS or sample cleaning procedures toremove matrix interferences), exclude the samplesfrom the quantitative risk assessment if they causethe calculated exposure concentration (i.e., theconcentration calculated according to guidance inChapter 6) to exceed the maximum detected con-centration for a particular sample set. The boxon this page presents an example of how toaddress a situation with unusually high QLs.

5.3.3 WHEN ONLY SOME SAMPLES IN AMEDIUM TEST POSITIVE FOR ACHEMICAL

Most analytes at a site are not positivelydetected in each sample collected and analyzed.Instead, for a particular chemical the data setgenerally will contain some samples with positiveresults and others with non-detected results. Thenon-detected results usually are reported as SQLs.These limits indicate that the chemical was notmeasured above certain levels, which may varyfrom sample to sample. The chemical may bepresent at a concentration just below the reportedquantitation limit, or it may not be present in thesample at all (i.e., the concentration in the sampleis zero).

In determining the concentrations mostrepresentative of potential exposures at the site(see Chapter 6), consider the positively detectedresults together with the non-detected results (i.e.,the SQLs). If there is reason to believe that thechemical is present in a sample at a concentrationbelow the SQL, use one-half of the SQL as aproxy concentration. The SQL value itself can beused if there is reason to believe theconcentration is closer to it than to one-half theSQL (See the next subsection for situationswhere SQLs are not available.) Unless site-specific information indicates that a chemical isnot likely to be present in a sample, do notsubstitute the value zero in place of the SQL (i.e.,do not assume that a chemical that is not detectedat the SQL would not be detected in the sampleif the analysis was extremely sensitive). Also, donot simply omit the non-detected results from therisk assessment.

11

5.3.4 WHEN SQLS ARE NOT AVAILABLE

A fourth situation concerning QLs maySometimes be encountered when evaluating sitedata. For some sites, data summaries may notprovide the SQLs. Instead, MDLs, CRQLS, oreven IDLs may have been substituted wherever achemical was not detected. Sometimes, nodetection or quantitation limits may be providedwith the data. As a first step in these situations,alwavs attempt to obtain the SQLs, because theseare the most appropriate limits to consider whenevaluating non-detected chemicals (i.e., theyaccount for sample characteristics, samplepreparation, or analytical adjustments that maydiffer from sample to sample).

If SQLs cannot be obtained, then, for CLPsample analyses, the CRQL should be used as theQL of interest for each non-detected chemical,with the understanding that these limits mayoverestimate or underestimate the actual SQL.For samples analyzed by methods different fromCLP methods, the MDL may be used as the QL,with the understanding that in most cases this willunderestimate the SQL (because the MDL is ameasure of detection limits only and does notaccount for sample characteristics or matrixinterferences). Note that the IDL should rarelybe used for non-detected chemicals since it is ameasure only of the detection limit for aparticular instrument and does not consider theeffect of sample handling and preparation orsample characteristics.

5.3.5 WHEN CHEMICALS ARE NOTDETECTED IN ANY SAMPLES IN A

MEDIUM

After considering the discussion provided inthe above subsections, generally eliminate thosechemicals that have not been detected in anysamples of a particular medium. On CLP datareports, these chemicals will be designated in eachsample with a U qualifier preceded by the SQL orCRQL (e.g., 10 U). If information exists toindicate that the chemicals are present, theyshould not be eliminated. For example, ifchemicals with similar transport and fatecharacteriatics are detected frequently in soil at asite, and some of these chemicals also are detectedfrequently in ground water while the others arenot detected, then the undetected chemicals are

probably present in the ground water andtherefore may need to be included in the riskassessment as ground-water contaminants.

The outcome of this step is a data set thatonly contains chemicals for which positive data(i.e., analytical results for which measurableconcentrations are reported) are available in atleast one sample from each medium. Unlessotherwise indicated, assume at this point in theevaluation of data that positive data to which nouncertainties are attached concerning either theassigned identity of the chemical or the reportedconcentration (i.e., data that are not “tentative,”“uncertain,” or “qualitative”) are appropriate foruse in the quantitative risk assessment.

5.4 EVALUATION OF QUALIFIEDAND CODED DATA

For CLP analytical results, various qualifiersand codes (hereafter referred to as qualifiers) areattached to certain data by either the laboratoriesconducting the analyses or by persons performingdata validation. These qualifiers often pertain toQA/QC problems and generally indicate questionsconcerning chemical identity, chemicalconcentration, or both. All qualifiers must beaddressed before the chemical can be used inquantitative risk assessment. Qualifiers used bythe laboratory may differ from those used by datavalidation personnel in either identity or meaning.

5.4.1 TYPES OF QUALIFIERS

A list of the qualifiers that laboratories arepermitted to use under the CLP -- and theirpotential use in risk assessment -- is presented inExhibit 5-4. A similar list addressing datavalidation qualifiers is provided in Exhibit 5-5.In general, because the data validation process isintended to assess the effect of QC issues on datausability, validation data qualifiers are attached tothe data after the laboratory qualifiers andsupersede the laboratory qualifiers. If data haveboth laboratory and validation qualifiers and theyappear contradictory, ignore the laboratoryqualifier and consider only the validation qualifier.If qualifiers have been attached to certain data bythe laboratory and have not been removed,revised, or superseded during data validation, then

12

EXHIBIT 5-4

CLP LABORATORY DATA QUALIFIERS AND THEIR POTENTIAL USEIN QUANTITATIVE RISK ASSESSMENT

Indicates:Uncertain Uncertain Include Data in Quantitative

Qualifier Definition Identity? Concentration? Risk Assessment?

Inorganic Chemical Data:a

B

U

E

M

N

S

W

*

+

Reported value is< CRDL, but > IDL.

Compound was analyzed for,but not detected.

Value is estimated due tomatrix interferences.

Duplicate injection precisioncriteria not met.

Spiked sample recovery notwithin control limits.

Reported value was determinedby the Method of StandardAdditions (MSA).

Post-digestion spike for furnaceAA analysis is out of controllimits, while sample absorbanceis <50% of spike absorbance.

Duplicate analysis was notwithin control limits.

Correlation coefficient forMSA was <O.995.

Organic Chemical Data:b

U Compound was analyzed for,but not detected.

No ? Yes

Yes Yes ?

No Yes Yes

No Yes Yes

No Yes Yes

No No Yes

No

No

No

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

?

(continued)

13

EXHIBIT 5-4 (continued)

CLP LABORATORY DATA QUALIFIERS AND THEIR POTENTIAL USEIN QUANTITATIVE RISK ASSESSMENT

Indicates:Uncertain Uncertain Include Data in Quantitative


J Value is estimated,either for a tentativelyidentified compound (TIC)or when a compound is present(spectral identificationcriteria are met, but thevalue is <CRQL).

C Pesticide results wereconfirmed by GC/MS.

B Analyte found in associatedblank as well as in sample.c

E Concentration exceedscalibration range ofGC/MS instrument.

D Compound identified in ananalysis at a secondarydilution factor.

A The TIC is a suspected aldol-condensation product.

X Additional flags definedseparately.

No, for Yes ?TCL chem-icals;

Yes, forTICs

No

No

No

No No Yes

Yes Yes No

-- -- --

No

Yes

Yes

Yes

Yes

Yes

-- = Data will vary with laboratory conducting analyses.

a Source EPA 1988b.

b Source EPA 1988C.

C See Section 5.5 for guidance concerning blank contamination.

14

EXHIBIT 5-5

VALIDATION DATA QUALIFIERS AND THEIRPOTENTIAL USE IN QUANTITATIVE RISK ASSESSMENT

Indicates:uncertain Uncertain Include Data in Quantitative


Inorganic and Organic Chemical Data:a

U The material was analyzedfor, but not detected. Theassociated numerical valueis the SQL.

Yes

J The associated numericalvalue is an estimated quantity.

R Quality control indicates thatthe data are unusable (compoundmay or may not be present).Re-sampling and/or re-analysis isnecessary for verification.

Z No analytical result (inorganicdata only).

--

Q No analytical result (organicdata only).

N Presumptive evidence ofpresence of material (tentativeidentification). b

Yes

No

Yes

Yes

Yes

?

Yes

No

-- --

-- -- --

Yes Yes ?

-- = Not applicable

a Source EPA 1988d,e.

b Organic chemical data only.

15

evaluate the laboratory qualifier itself. If it isunclear whether the data have been validated,contact the appropriate data validation and/orlaboratory personnel.

The type of qualifier and other site-specificfactors determine how qualified data are to beused in a risk assessment. As seen in Exhibits5-4 and 5-5, the type of qualifier attached tocertain data often indicates how that data shouldbe used in a risk assessment. For example, mostof the laboratory qualifiers for both inorganicchemical data and organic chemical data (e.g., J,E, N) indicate uncertainty in the reportedconcentration of the chemical, but not in itsassigned identity. Therefore, these data can beused just as positive data with no qualifiers orcodes. In general, include data with qualifiers thatindicate uncertainties in concentrations but not inidentification.

Examples showing the use of certain qualifieddata are presented in the next two boxes. Thefirst box addresses the J qualifier, the mostcommonly encountered data qualifier in Superfunddata packages. Basically, the guidance here is touse J-qualified concentrations the same way as

positive data that do not have this qualifier. Ifpossible, note potential uncertainties associatedwith the qualifier, so that if data qualified with aJ contribute significantly to the risk, thenappropriate caveats can be attached.

EXAMPLE OF VALIDATED DATA CONTAINING R QUALIFIERS

In this example, concentrations of inorganic chemicalsin ground water have been determined using the CLP'sRAS.

Concentration (µg/L)Chemical Sample 1 Sample 2 Sample 3 Sample 4

Manganese 310 500Ra 30URb 500

a R = Quality control indicates that the data areunusable (compound may or may not be present),

b u = compund was analyzed for, but not detected,Value presented (e.g., 30 U) is the SQL.

These data have been validated, and therefore the Rqnatifiers indicate that the person conducting the datavalidation rejected the data for manganese in Samples2 and 3. The “UR” qualifier means that manganese wasnot detected in Sample 3; however, the data validatorrejected the non-detected result. Eliminate these twosamptes so that the data set now consists of only twosamples (Samples 1 and 4).

An illustration of the use of R-qualified datais presented in the box in this column. Thedefinition, and therefore the use of the Rqualifier, differs depending on whether the datahave been validated or not. (Note that the CLPformerly used R as a laboratory qualifier toindicate low spike recovery for inorganic. Thishas been changed, but older data may still havebeen qualified by the laboratory with an R.) If itis known that the R data qualifier indicates thatthe sample result was rejected by the datavalidation personnel, then this result should beeliminated from the risk assessment; if the R dataqualifier was placed on the data to indicateestimated data due to low spike recovery (i.e., theR was placed on the data by the laboratory and

16

not by the validator), then use the R-qualifieddata in a manner similar to the use of J-qualifieddata (i.e., use the R-qualified concentrations thesame way as positive data that do not have thisqualifier). If possible, note whether the R-Requalified data are overestimates or underestimatesof actual expected chemical concentrations so thatappropriate caveats may be attached if dataqualified with an R contribute significantly to therisk.

5.4.2 USING THE APPROPRIATEQUALIFIERS

The information presented in Exhibits 5-4and 5-5 is based on the most recent EPAguidance documents concerning qualifiers: the

SOW for Inorganic and the SOW for Organics(EPA 1988b,c) for laboratory qualifiers, and theFunctional Guidelines for Inorganic and theFunctional Guidelines for Organics (EPA 1988d,e)for validation qualifiers. The types and definitionsof qualifiers, however, may be periodically updatedwithin the CLP program. In addition, certainEPA regions may have their own data qualifiersand associated definitions. These regionalqualifiers are generally consistent with theFunctional Guidelines, but are designed to conveyadditional information to data users.

In general, the risk assessor should checkwhether the information presented in this sectionis current by contacting the appropriate regionalCLP or headquarters Analytical OperationsBranch staff. Also, if definitions are not reportedwith the data, regional contacts should beconsulted prior to evaluating qualified data.These variations may affect how data with certainqualifiers should be used in a risk assessment.Make sure that definitions of data qualifiers used

in the data set for the site have been reportedwith the data and are current. Never guess aboutthe definition of qualifiers.

5.5 COMPARISON OFCONCENTRATIONSDETECTED IN BLANKS WITHCONCENTRATIONSDETECTED IN SAMPLES

Blank samples provide a measure ofcontamination that has been introduced into asample set either (1) in the field while thesamples were being collected or transported to thelaboratory or (2) in the laboratory during samplepreparation or analysis. To prevent the inclusionof non-site-related contaminants in the riskassessment, the concentrations of chemicalsdetected in blanks must be compared withconcentrations of the same chemicals detected insite samples. Detailed definitions of differenttypes of blanks are provided in the box on thenext page.

Blank data should be compared with resultsfrom samples with which the blanks are associated.It is often impossible, however, to determine theassociation between certain blanks and data. Inthis case, compare the blank data with resultsfrom the entire sample data set. Use theguidelines in the following paragraphs whencomparing sample concentrations with blankconcentrations.

Blanks containing common laboratorycontaminants. As discussed in the CLP SOW forOrganics (EPA 1988c) and the FunctionalGuidelines for Organics (EPA 1988e), acetone, 2-butanone (or methyl ethyl ketone), methylenechloride, toluene, and the phthalate esters areconsidered by EPA to be common laboratorycontaminants. In accordance with the FunctionalGuidelines for Organics (EPA 1988e) and theFunctional Guidelines for Inorganic (EPA 1988d),if the blank contains detectable levels of commonlaboratory contaminants, then the sample resultsshould be considered as positive results only if theconcentrations in the sample exceed ten times themaximum amount detected in anv blank. If theconcentration of a common laboratorycontaminant is less than ten times the blankconcentration, then conclude that the chemicalwas not detected in the particular sample and, inaccordance with EPA guidance, consider theblank-related concentrations of the chemical to be

17

TYPES OF BLANKS

Blanks are analytical quality control samples analyzed in the same manner as site samples. They are used in the measurementof contamination that has been introduced into a sample either (1) in the field while the samples were being collected ortransported to the laboratory or (2) in the laboratory during sample preparation or analysis, Four types of blanks -- trip, field,laboratory calibration, and laboratory reagent (or method) -- are described below. A discussion on the water used for the blankalso is provided.

Trip Blank. This type of blank is used to indicate potential contamination due to migration of volatile organic chemicals(VOCs) from the air on the site or in sample shipping containers, through the septum or around the lid of sampling vials, andinto the sample. A trip blank consists of laboratory distiIled, deionized water in a 40-ml glass vial sealed with a teflon septumThe blank accompanies the empty sample bottles to the field as well as the samples returning to the laboratory for analysis; itis not opened until it is analyzed in the lab with the actual site samples. The containers and labels for trip blanks should bethe same as the containers and labets for actual samples, thus making the laboratory "lind" to the identity of the blanks.

Field Blank. A field blank is used to determine if certain field sampling or cleaning procedures (e.g., insufficient cleaningof sampling equipment) result in cross-contamination of site samples. Like the trip blank, the field blank is a sample of distilled,deionized water taken to the field with empty sample bottles and is analyzed in the laboratory along with the actual samples.Unlike the trip blank, however, the field blank sample is opened in the field and used as a sample would be (e.g., it is pouredthrough cleaned sampling equipment or it is poured from container to container in the vicinity of a gas-powered pump). Aswith trip blanks, the field blanks’ containers and labels should be the same as for actual samples,

Laboratory Calibration Blank. This type of blank is distilled, deionized water injected directly into an instrument withouthaving been treated with reagents appropriate to the analytical method used to analyze actual site samples. This type of blankis used to indicate contamination in the instrument itself, or possibly in the distilled, deionized water.

Laboratory Reagent Or Method Blank. This blank results from the treatment of distilled, deionized water with all of thereagents and manipulations (e.g., digestions or extractions) to which site samples will be subjected. Positive results in thereagent blank may indicate either contamination of the chemical reagents or the glassware and implements used to store orprepare the sample and resulting solutions. Although a laboratory following good laboratoty practices will have its analyticalprocesses under control, in some instances method blank contamination cannot be entirely eliminated.

Water Used for Blanks. For all the blanks described above, results are reliable only if the water comprising the blank wasclean. For example, if the laborstory water comprising the trip blank was contaminated with VOCs prior to being taken to thefield, then the source of VOC contamination in the trip blank cannot be isolated (see laboratory calibration blank).

the quantitation limit for the chemical in thatsample. Note that if all samples contain levels ofa common laboratory contaminant that are lessthan ten times the level of contamination notedin the blank, then completely eliminate thatchemical from the set of sample results.

Blanks containing chemicals that are notcommon laboratory contaminants. As discussedin the previously referenced guidance, if the blankcontains detectable levels of one or more organicor inorganic chemicals that are not considered byEPA to be common laboratory contaminants (e.g.,all other chemicals on the TCL), then considersite sample results as positive only if theconcentration of the chemical in the site sampleexceeds five times the maximum amount detectedin any blank. Treat samples containing less thanfive times the amount in any blank as non-detectsand, in accordance with EPA guidance, consider

the blank-related chemical concentration to be thequantitation limit for the chemical in that sample.Again, note that if all samples contain levels of aTCL chemical that are less than five times thelevel of contamination noted in the blank, thencompletely eliminate that chemical from the set ofsample results.

5.6 EVALUATION OFTENTATIVELY IDENTIFIEDCOMPOUNDS

Both the identity and reported concentrationof a tentatively identified compound (TIC) isquestionable (see the box on the next page forbackground on TICs). Two options for addressingTICS exist, depending on the relative number ofTICs compared to non-TICs.

18

5.6.1 WHEN FEW TICs ARE PRESENT

When only a few TICs are present comparedto the TAL and TCL chemicals, and no historicalor other site information indicates that either aparticular TIC may indeed be present at the site(e.g., because it may be a by-product of a chemicaloperation conducted when the site was active) orthat the estimated concentration may be very high(i.e., the risk would be dominated by the TIC),then generally do not include the TICS in the riskassessment. Otherwise, follow the guidanceprovided in the next subsection. Consult with theRPM about omitting TICS from the quantitative

risk assessment, and document reasons forexcluding TICS in the risk assessment report.

5.6.2 WHEN MANY TICs ARE PRESENT

If many TICs are present relative to the TALand TCL compounds identified, or if TICconcentrations appear high or site informationindicates that TICs are indeed present, thenfurther evaluation of TICs is necessary. Ifsufficient time is available, use SAS to confirmthe identity and to positively and reliably measurethe concentrations of TICs prior to their use inthe risk assessment. If SAS methods to identifyand measure TICs are unavailable, or if there isinsufficient time to use SAS, then the TICS shouldbe included as chemicals of potential concern inthe risk assessment and the uncertainty in bothidentity and concentration should be noted (unlessinformation exists to indicate that the TICs arenot present).

5.7 COMPARISON OF SAMPLESWITH BACKGROUND

In some cases, a comparison of sampleconcentrations with background concentrations(e.g., using the geometric mean concentrations ofthe two data sets) is useful for identifying thenon-site-related chemicals that are found at ornear the site. If background risk might be aconcern, it should be calculated separately fromsite-related risk. Often, however, the comparisonof samples with background is unnecessary becauseof the low risk usually posed by the backgroundchemicals compared to site-related chemicals.

As discussed in Chapter 4, informationcollected during the RI can provide informationon two types of background chemicals: (1)naturally occurring chemicals that have not beeninfluenced by humans and (2) chemicals that arepresent due to anthropogenic sources. Either typeof background chemical can be either localized orubiquitous.

Information on background chemicals mayhave been obtained by the collection of site-specific background samples and/or from othersources (e.g., County Soil Conservation Servicesurveys, United States Geological Survey [USGS]

19

reports). As discussed in Chapter 4, backgroundconcentrations should be from the site or thevicinity of the site.

5.7.1 USE APPROPRIATE BACKGROUNDDATA

Background samples collected during the siteinvestigation should not be used if they wereobtained from areas influenced or potentiallyinfluenced by the site. Instead, the literaturesources mentioned in the previous paragraph maybe consulted to determine background levels ofchemicals in the vicinity of the site. Care must betaken in using literature sources, because the datacontained therein might represent nationwidevariation in a particular parameter rather thanvariation typical of the geographic region orgeological setting in which the site is located. Forexample, a literature source providingconcentrations of chemicals in ground water on anational scale may show a wide range ofconcentrations that is not representative of thevariation in concentrations that would be expectedat a particular site.

5.7.2 IDENTIFY STATISTICAL METHODS

In cases where background comparisons willbe made, any statistical methods that will be usedshould be identified prior to the collection ofsamples (see Chapter 4). Guidance documentsand reports that are available to aid inbackground comparison are listed in Section 4.4.3.Prior to conducting the steps discussed in the nexttwo subsections, the RPM should be consulted todetermine the type of comparison to be made, ifany. Both a justification for eliminating chemicalsbased on a background comparison and a briefoverview of the type of comparison conductedshould be included in the risk assessment report.

5.7.3 COMPARE CHEMICALCONCENTRATIONS WITHNATURALLY OCCURRING LEVELS

As defined previously, naturally occurringlevels are levels of chemicals that are presentunder ambient conditions and that have not beenincreased by anthropogenic sources. If inorganicchemicals are present at the site at naturalIyoccurring levels, they may be eliminated from thequantitative risk assessment. In some cases,

however, background concentrations may presenta significant risk, and, while cleanup may or maynot elimimte this risk, the background risk maybe an important site characteristic to thoseexposed. The RPM will always have the optionto consider the risk posed by naturally occurringbackground chemicals separately.

In general, comparison with naturallyoccurring levels is applicable only to inorganicchemicals, because the majority of organicchemicals found at Superfund sites are notnaturally occurring (even though they may beubiquitous). The presence of organic chemicalsin background samples collected during a siteinvestigation actually may indicate that the samplewas collected in an area influenced by sitecontamination and therefore does not qualify asa true background sample. Such samples shouldinstead be included with other site samples in therisk assessment. Unless a very strong case can bemade for the natural occurrence of an organicchemical, do not eliminate it from the quantitativerisk assessment for this reason.

5.7.4 COMPARE CHEMICALCONCENTRATIONS WITHANTHROPOGENIC LEVELS

Anthropogenic levels are ambientconcentrations resulting from human (non-site)sources. Localized anthropogenic background isoften caused by a point source such as a nearbyfactory. Ubiquitous anthropogenic background isoften from nonpoint sources such as automobiles.In general, do not eliminate anthropogenicchemicals because, at many sites, it is extremelydifficult to conclusively show at this stage of thesite investigation that such chemicals are presentat the site due to operations not related to thesite or the surrounding area.

Often, anthropogenic background chemicalscan be identified and considered separately duringor at the end of the risk assessment. Thesechemicals also can be omitted entirely from therisk assessment, but, as discussed for naturalbackground, they may present a significant risk.Omitting anthropogenic background chemicalsfrom the risk assessment could result in the lossof important information for those potentiallyexposed.

20

5.8 DEVELOPMENT OF A SET OFCHEMICAL DATA ANDINFORMATION FOR USE INTHE RISK ASSESSMENT

After the evaluation of data is complete asspecified in previous sections, a list of the samples(by medium) is made that will be used to estimateexposure concentrations, as discussed in Chapter6 of this guidance. In addition, as shown in theflowchart in Exhibit 5-1, a list of chemicals ofpotential concern (also by medium) will be neededfor the quantitative risk assessment. This listshould include chemicals that were

(1)

(2)

(3)

(4)

(5)

positively detected in at least one CLPsample (RAS or SAS) in a givenmedium, including (a) chemicals with noqualifiers attached (excluding sampleswith unusually high detection limits), and(b) chemicals with qualifiers attachedthat indicate known identities butunknown concentrations (e.g., J-qualifieddata);

detected at levels significantly elevatedabove levels of the same chemicalsdetected in associated blank samples;

detected at levels significantly elevatedabove naturally occurring levels of thesame chemicals;

only tentatively identified but either maybe associated with the site based onhistorical information or have beenconfirmed by SAS; and/or

transformation products of chemicalsdemonstrated to be present.

Chemicals that were not detected in samplesfrom a given medium (i.e., non-detects) but thatmay be present at the site also may be includedin the risk assessment if an evaluation of the riskspotentially present at the detection limit isdesired.

5.9 FURTHER REDUCTION INTHE NUMBER OFCHEMICALS (OPTIONAL)

For certain sites, the list of potentially site-related chemicals remaining after quantitationlimits, qualifiers, blank contamination, andbackground have been evaluated may be lengthy.Carrying a large number of chemicals through aquantitative risk assessment may be complex andit may consume significant amounts of time andresources. The resulting risk assessment report,with its large, unwieldy tables and text, may bedifficult to read and understand, and it maydistract from the dominant risks presented by thesite. In these cases, the procedures discussed inthis section -- using chemical classes, frequency ofdetection, essential nutrient information, and aconcentration-toxicity screen -- may be used tofurther reduce the number of chemicals ofpotential concern in each medium.

If conducting a risk assessment on a largenumber of chemicals is feasible (e.g., because ofadequate computer capability), then theprocedures presented in this section should not beused. Rather, the most important chemicals (e.g.,those presenting 99 percent of the risk) --identified after the risk assessment -- could bepresented in the main text of the report, and theremaining chemicals could be presented in theappendices.

5.9.1 CONDUCT INITIAL ACTIVITIES

Several activities must be conducted beforeimplementing any of the procedures described inthis section: (1) consult with the RPM; (2)consider how the rationale for the procedureshould be documented; (3) examine historicalinformation on the site; (4) consider concentrationand toxicity of the chemicals; (5) examine themobility, persistence, and bioaccumulationpotential of the chemicals; (6) consider specialexposure routes; (7) consider the treatability ofthe chemicals; (8) examine applicable or relevantand appropriate requirements (ARARs); and (9)examine the need for the procedures. Theseactivities are described below.

Consultation with the RPM. If a largenumber of chemicals are of potential concern at

21

a particular site, the RPM should be consulted.Approval by the RPM must be obtained prior tothe elimination of chemicals based on any of theseprocedures. The concentration-toxicity screen inparticular may be needed only in rare instances.

Documentation of rationale. The rationalefor eliminating chemicals from the quantitativerisk assessment based on the procedures discussedbelow must be clearly stated in the risk assessmentreport. This documentation, and its possibledefense at a later date, could be fairly resource-intensive. If a continuing need to justify this stepis expected, then any plans to eliminate chemicalsshould be reconsidered.

Historical information. Chemicals reliablyassociated with site activities based on historicalinformation generally should not be eliminatedfrom the quantitative risk assessment, even if theresults of the procedures given in this sectionindicate that such an elimination is possible.

Concentration and toxicity. Certain aspectsof concentration and toxicity of the chemicals alsomust be considered prior to eliminating chemicalsbased on the results of these procedures. Forexample, before eliminating potentiallycarcinogenic chemicals, the weight-of-evidenceclassification should be considered in conjunctionwith the concentrations detected at the site. Itmay be practical and conservative to retain achemical that was detected at low concentrationsif that chemical is a Group A carcinogen. (Asdiscussed in detail in Chapter 7, the weight-of-evidence classification is an indication of thequality and quantity of data underlying achemical’s designation as a potential humancarcinogen.)

Mobility, persistence, and bioaccumulation.Three factors that must be considered whenimplementing these procedures are the mobility,persistence, and bioaccumulation of the chemicals.For example, a highly volatile (i.e., mobile)chemical such as benzene, a long-lived (i.e.,persistent) chemical such as dioxin, or a readilytaken-up and concentrated (i.e., bioaccumulated)chemical such as DDT, probably should remain inthe risk assessment. These procedures do notexplicitly include a mobility, persistence, orbioaccumulation component, and therefore the

risk assessor must pay special attention to thesefactors.

Special exposure routes. For some chemicals,certain exposure routes need to be consideredcarefully before using these procedures. Forexample, some chemicals are highly volatile andmay pose a significant inhalation risk due to thehome use of contaminated water, particularly forshowering. The procedures described in thissection may not account for exposure routes suchas this.

Treatability. Some chemicals are moredifficult to treat than others and as a result shouldremain as chemicals of potential concern becauseof their importance during the selection ofremedial alternatives.

ARARs. Chemicals with ARARs (includingthose relevant to land ban compliance) usually arenot appropriate for exclusion from the quantitativerisk assessment based on the procedures in thissection. This may, however, depend in part onhow the chemicals’ site concentrations in specificmedia compare with their ARAR concentrationsfor these media.

Need for procedures. Quantitative evaluationof all chemicals of potential concern is the mostthorough approach in a risk assessment. Inaddition, the time required to implement anddefend the selection procedures discussed in thissection may exceed the time needed to simplycarry all the chemicals of potential concernthrough the risk assessment. Usually, carrying allchemicals of potential concern through the riskassessment will not be a difficult task, particularlygiven the widespread use of computer spreadsheetsto calculate exposure concentrations of chemicalsand their associated risks. Although the tablesthat result may indeed be large, computerspreadsheets significantly increase the ability toevaluate a number of chemicals in a relativelyshort period of time. For these reasons, theprocedures discussed here may be needed only inrare instances. As previously stated, the approvalof these procedures by the RPM must be obtainedprior to implementing any of these optionalscreening procedures at a particular site.

22

5.9.2 GROUP CHEMICALS BY CLASS

At times, toxicity values to be used incharacterizing risks are available only for certainchemicals within a chemical class. For example,of the polycyclic aromatic hydrocarbons (PAHs)considered to be potential carcinogens, a slopefactor currently is available (i.e., as this manualwent to press) for benz(a)pyrene only. In thesecases, rather than eliminating the other chemicalswithin the class from quantitative evaluationbecause of a lack of toxicity values, it may beuseful to group data for such a class of chemicals(e.g., according to structure-activity relationshipsor other similarities) for consideration in latersections of the risk assessment. For example, theconcentrations of only one group of chemicals

(e.g., carcinogenic PAHs) would be consideredrather than concentrations of each of the sevencarcinogenic PAHs currently on the TCL.

To group chemicals by class, concentrationsof chemicals within each class are summedaccording to procedures discussed in Chapter 6 ofthis guidance. Later in the risk assessment, thischemical class concentration would be used tocharacterize risk using toxicity values (i.e., RfDsor slope factors) associated with one of thechemicals in the particular class.

Three notes of caution when groupingchemicals should be considered: (1) do not groupsolely by toxicity characteristics; (2) do not groupall carcinogenic chemicals or all noncarcinogenicchemicals without regard to structure-activity orother chemical similarities; and (3) discuss in therisk assessment report that grouping can produceeither over- or under-estimates of the true risk.

5.9.3 EVALUATE FREQUENCY OFDETECTION

Chemicals that are infrequently detected maybe artifacts in the data due to sampling, analytical,or other problems, and therefore may not berelated to site operations or disposal practices.Consider the chemical as a candidate forelimination from the quantitative risk assessmentif: (1) it is detected infrequently in one orperhaps two environmental media, (2) it is notdetected in any other sampled media or at highconcentrations, and (3) there is no reason tobelieve that the chemical may be present.

Available modeling results may indicate whethermonitoring data that show infrequently detectedchemicals are representative of only their samplinglocations or of broader areas. Because chemicalconcentrations at a site are spatially variable, therisk assessor can use modeling results to projectinfrequently detected chemical concentrations overbroader areas when determining whether thesubject chemicals are relevant to the overall riskassessment. Judicious use of modeling tosupplement available monitoring data often canminimize the need for the RPM to resort toarbitrarily setting limits on inclusion ofinfrequently detected chemicals in the riskassessment. Any detection frequency limit to beused (e.g., five percent) should be approved by theRPM prior to using this screen. If, for example,a frequency of detection limit of five percent isused, then at least 20 samples of a medium wouldbe needed (i.e., one detect in 20 samples equalsa five percent frequency of detection).

In addition to available monitoring data andmodeling results, the risk assessor will need toconsider other relevant factors (e.g., presence ofsensitive subpopulations) in recommendingappropriate site-specific limits on inclusion ofinfrequently detected chemicals in the quantitativerisk assessment. For example, the risk assessorshould consider whether the chemical is expectedto be present based on historical data or anyother relevant information (e.g., knowndegradation products of chemicals present at thesite, modeling results). Chemicals expected to bepresent should not be eliminated. (See theexample of chemicals with similar transport andfate characteristics in Section 5.3.5.)

The reported or modeled concentrations andlocations of chemicals should be examined tocheck for hotspots, which may be especiallyimportant for short-term exposures and whichtherefore should not be eliminated from the riskassessment. Always consider detection ofparticular chemicals in all sampled media becausesome media may be sources of contamination forother media. For example, a chemical that isinfrequently detected in soil (a potential ground-water contamination source) probably should notbe eliminated as a site contaminant if the samechemical is frequently detected in ground water.In addition, infrequently detected chemicals with

23

concentrations that greatly exceed referenceconcentrations should not be eliminated.

5.9.4 EVALUATE ESSENTIAL NUTRIENTS

Chemicals that are (1) essential humannutrients, (2) present at low concentrations (i.e.,only slightly elevated above naturally occurringlevels), and (3) toxic only at very high doses (i.e.,much higher than those that could be associatedwith contact at the site) need not be consideredfurther in the quantitative risk assessment.Examples of such chemicals are iron, magnesium,calcium, potassium, and sodium.

Prior to eliminating such chemicals from therisk assessment, they must be shown to be presentat levels that are not associated with adversehealth effects. The determination of acceptabledietary levels for essential nutrients, however,often is very difficult. Literature valuesconcerning acceptable dietary levels may conflictand may change fairly often as new studies areconducted. For example, arsenic -- a potentialcarcinogen -- is considered by some scientists tobe an essential nutrient based on animalexperiments; however, acceptable dietary levels arenot well known (EPA 1988f). Therefore, arsenicshould be retained in the risk assessment, eventhough it may be an essential nutrient atundefined dietary levels. Another example of anutrient that is difficult to characterize is sodium.Although an essential element in the diet, certainlevels of sodium may be associated with bloodpressure effects in some sensitive individuals(although data indicating an association betweensodium in drinking water and hypertension areinadequate [EPA 1987]).

Another problem with determining acceptabledietary levels for essential nutrients is thatnutrient levels often are presented in the literatureas concentrations within the human body (e.g.,blood levels). To identify an essential nutrientconcentration to be used for comparison withconcentrations in a particular medium at a site,blood (or other tissue) levels of the chemical fromthe literature must be converted to concentrationsin the media of concern for the site (e.g., soil,drinking water).

For these reasons, it may not be possible tocompare essential nutrient concentrations with siteconcentrations in order to eliminate essentialnutrient chemicals. In general, only essentialnutrients present at low concentrations (i.e., onlyslightly elevated above background) should beeliminated to help ensure that chemicals presentat potentially toxic concentrations are evaluated inthe quantitative risk assessment.

5.9.5 USE A CONCENTRATION-TOXICITYSCREEN

The objective of this screening procedure isto identify the chemicals in a particular mediumthat -- based on concentration and toxicity -- aremost likely to contribute significantly to riskscalculated for exposure scenarios involving thatmedium, so that the risk assessment is focused onthe “most significant” chemicals.

Calculate individual chemical scores. Twoof the most important factors when determiningthe potential effect of including a chemical in therisk assessment are its measured concentrations atthe site and its toxicity. Therefore, in thisscreening procedure, each chemical in a mediumis first scored according to its concentration andtoxicity to obtain a risk factor (see the box below).Separate scores are calculated for each mediumbeing evaluated.

24

The units for the risk factor Rij depend onthe medium being screened. In general, theabsolute units do not matter, as long as unitsamong chemicals in a medium are the same. Theconcentration used in the above equation shouldbe the maximum detected concentration, andtoxicity values should be obtained in accordancewith the procedures discussed in Chapter 7.

Chemicals without toxicity values cannot bescreened using this procedure. Such chemicalsshould always be discussed in the risk assessmentas chemicals of potential concern; they should notbe eliminated from the risk assessment. Guidanceconcerning chemicals without toxicity values isprovided in Chapter 7.

For some chemicals, both oral and inhalationtoxicity values are available. In these cases, themore conservative toxicity values (i.e., onesyielding the larger risk factor when used in theabove equation) usually should be used. If onlyone exposure route is likely for the medium beingevaluated, then the toxicity values correspondingto that exposure route should be used.

Calculate total chemical scores (per medium).Chemical-specific risk factors are summed toobtain the total risk factor for all chemicals ofpotential concern in a medium (see the box onthis page). A separate Rj will be calculated forcarcinogenic and noncarcinogenic effects. Theratio of the risk factor for each chemical to thetotal risk factor (i.e.,Rij/Rj) approximates therelative risk for each chemical in medium j.

Eliminate chemicals. After carefullyconsidering the factors discussed previously in thissubsection, eliminate from the risk assessmentchemicals with Rij/Rj ratios that are very lowcompared with the ratios of other chemicals in themedium. The RPM may wish to specify a limitfor this ratio (e.g., 0.01; a lower fraction would beneeded if site risks are expected to be high). Achemical that contributes less than the specifiedfraction of the total risk factor for each mediumwould not be considered further in the riskassessment for that medium. Chemicals exceedingthe limit would be considered likely to contributesignificantly to risks, as calculated in subsequent

stages of the risk assessment. This screeningprocedure could greatly reduce the number ofchemicals carried through a risk assessment,because in many cases only a few chemicalscontribute significantly to the total risk for aparticular medium.

The risk factors developed in this screeningprocedure are to be used only for potentialreduction of the number of chemicals carriedthrough the risk assessment and have no meaningoutside of the context of the screening procedure.They should not be considered as a quantitativemeasure of a chemical’s toxicity or risk or as asubstitute for the risk assessment proceduresdiscussed in Chapters 6, 7, and

5.10 SUMMARY ANDPRESENTATION

The section of the risk

8 of this guidance.

OF DATA

assessment reportsummarizing the results of the data collection-andevaluation should be titled “Identification ofChemicals of Potential Concern” (see Chapter 9).Information in this section should be presented inways that readily support the calculation ofexposure concentrations in the exposureassessment portion of the risk assessment.Exhibits 5-6 and 5-7 present examples of tables tobe included in this section of the risk assessmentreport.

25

EXHIBIT 5-6

EXAMPLE OF TABLE FORMAT FOR PRESENTINGCHEMICALS SAMPLED IN SPECIFIC MEDIA

Table XChemicals Sampled in Medium Y

(and in Operable Unit Z, if appropriate)Name of Site, Location of Site

Range Rangeof Sample of Detected

Frequency of Quantitation Concentrations BackgroundChemical Detection a Limits (units) (units) Levels

Chemical A 3/25 5 - 5 0 320-4600 100-140* Chemical B 25/25 1-32 16-72 - -

-- = Not available.

* Identified as a chemical of potential concern based on evaluation of data according to proceduresdescribed in text of report.

a Number of samples in which the chemical was positively detected over the number of samplesavailable.

26

EXHIBIT 5-7

EXAMPLE OF TABLE FORMAT FOR SUMMARIZINGCHEMICALS OF POTENTIAL CONCERN IN

ALL MEDIA SAMPLED

Table WSummary of chemicals of

Potential Concern at Site X, Location Y(and in Operable Unit Z, if appropriate)

Concentration

Chemical Soils Ground Water Surface Water S e d i m e n t s A i r(mg/kg) (ug/L) (ug/L) (ug/kg) (ug/m3)

Chemical A 5-1,100 -- 2 - 3 0 -- --Chemical B 0.5-64 5 - 9 2 -- 100-45,000 --Chemical C -- 15-890 50-11,000 -- --Chemical D 2 - 1 2 -- -- -- 0.1-940

-- = Not available.

27

5.10.1 SUMMARIZE DATA COLLECTIONAND EVALUATION RESULTS IN TEXT

In the introduction for this section of the riskassessment report, clearly discuss in bullet formthe steps involved in data evaluation. If theoptional screening procedure described in Section5.9 was used in determining chemicals of potentialconcern, these steps should be included in theintroduction. If both historical data and currentdata were used in the data evaluation, state thisin the introduction. Any special site-specificconsiderations in collecting and evaluating thedata should be mentioned. General uncertaintiesconcerning the quality associated with either thecollection or the analysis of samples should bediscussed so that the potential effects of theseuncertainties on later sections of the riskassessment can be determined.

In the next part of the report, discuss thesamples from each medium selected for use inquantitative risk assessment. Provide informationconcerning the sample collection methods used(e.g., grab, composite) as well as the number andlocation of samples. If this information isprovided in the RI report, simply refer to theappropriate sections. If any samples (e.g., fieldscreening/analytical samples) were excludedspecifically from the quantitative risk assessmentprior to evaluating the data, document this alongwith reasons for the exclusion. Again, rememberthat such samples, while not used in thequantitative risk assessment, may be useful forqualitative discussions and therefore should not beentirely excluded from the risk assessment.

Discuss the data evacuation either by medium,by medium within each operable unit (if the siteis sufficiently large to be divided into specificoperable units), or by discrete areas within eachmedium in an operable unit. For each medium,if several source areas with different types andconcentrations of chemicals exist, then themedium-specific discussion for each source areamay be separate. Begin the discussion with thosemedia (e.g., wastes, soils) that are potential sources of contamination for other media (e.g.,ground water, surface water/sediments). If nosamples or data were available for a particularmedium, discuss this in the text. For soils data,discuss surface soil results separately from thoseof subsurface soils. Present ground-water results

by aquifer if more than one aquifer was sampled.Discuss surface water/sediment results by thespecific surface water body sampled.

For each medium, identify in the report thechemicals for which samples were analyzed, andlist the analytes that were detected in at least onesample. If any detected chemicals were eliminatedfrom the quantitative risk assessment based onevaluation of data (i.e., based on evaluation ofdata quality, background comparisons, and theoptional screening procedures, if used), providereasons for the elimination in the text (e.g.,chemical was detected in blanks at similarconcentrations to those detected in samples orchemical was infrequently detected).

The final subsection of the text is adiscussion of general trends in the data results.For example, the text may mention (1) whetherconcentrations of chemicals of potential concernin most media were close to the detection limitsor (2) trends concerning chemicals detected inmore than one medium or in more than oneoperable unit at the site. In addition, the locationof hot spots should be discussed, as well as anynoticeable trends apparent from sampling resultsat different times.

5.10.2 SUMMARIZE DATA COLLECTIONAND EVALUATION RESULTS INTABLES AND GRAPHICS

As shown in Exhibit 5-6, a separate table thatincludes all chemicals detected in a medium canbe provided for each medium sampled at ahazardous waste site or for each medium withinan operable unit at a site. Chemicals that havebeen determined to be of potential concern basedon the data evaluation should be designated in thetable with an asterisk to the left of the chemicalname.

For each chemical, present the frequency ofdetection in a certain medium (i.e., the number oftimes a chemical was detected over the totalnumber of samples considered) and the range ofdetected or quantified values in the samples. Donot present the QL or similar indicator of aminimum level (e.g., <10 mg/L, ND) as the lowerend of the range; instead, the lower and upperbound of the range should be the minimum andmaximum detected values, respectively. The range

28

of reported QLs obtained for each chemical invarious samples should be provided in a separatecolumn. Note that these QLs should be sample-specific; CRQLs, MDLs, or other types of non-sample-specific values should be provided onlywhen SQLS are not available. Note that the rangeof QLs would not include any limit values (e.g.,unusually high QLs) eliminated based on theguidance in Section 5.3. Finally, naturallyoccurring concentrations of chemicals used incomparing sample concentrations may be providedin a separate column. The source of thesenaturally occurring levels should be provided in afootnote. List the identity of the samples used in

determining concentrations presented in the tablein an appropriate footnote.

The final table in this section is a list of thechemicals of potential concern presented bymedium at the site or by medium within eachoperable unit at the site. A sample table formatis presented in Exhibit 5-7.

Another useful type of presentation ofchemical concentration data is the isopleth (notshown). This graphic characterizes the monitoredor modeled concentrations of chemicals at a siteand illustrates the spatial pattern ofcontamination.

29


1. Note that the values in this example are for illustration purposes only. Many CRQLs and CRDLs are in the process of beinglowered, and the RfDs and slope factors may have changed.

30


Environmental Protection Agency (EPA). 1984. Methods for Organic Chemical Analysis of Municipal and Industrial Wastewater (EPA600 Methods) as presented in 40 CFR Part 136, Guidelines Establishing Teat Procedures for the Analysis of Pollutants Underthe Clean Water Act.

Used to determine chemicals present in municipal and industrial wastewater as provided under the Clean Water Act.Analytical methods for priority pollutants, including sample preparation, reagents, calibration procedures, QA/QC analyticalprocedures, and calculation.

Environmental Protection Agency (EPA). 1986. Test Methods for Evaluating Solid Waste (SW-846} Physical/Chimcal Methods.Office of Solid Waste.

Provides analytical procedures to test solid waste to determine if it is a hazardous waste as defined under RCRA. containsinformation for collecting solid waste samples and for determining reactivity, corrosivity, ignitability, composition of waste,and mobility of waste components.

Environmental Protection Agency (EPA). 1987. Drinking Water, Proposed Substitution of Contaminants and Proposed List ofAdditional Substances Which May Require Regulation Under the Safe Drinking Water Act. 52 Fedetal Register 25720 (July 8,1987).

Environmental Protection Agency (EPA). 1988a. User’s Guide to the Contract Laboratory Program. Office of Emergency andRemedial Response.

Provides requirements, and analytical procedures of the CLP protocols developed from technical caucus recommendationsfor both organic and inorganic analysis. Contains information on CLP objectives and orientation, CLP structure, descriptionof analytical services, utilization of analytical services, auxiliary support services, and program quality assurance.

Environmental Protection Agency (EPA). 1988b. contract Laboratory Program Statement of Work for Inorganics Analysis Multi-media, Multi-concentration. Office of Emergency and Remedial Response. SOW No. 788.

Provides procedures required by EPA for analyzing hazardous waste disposal site samples (aqueous and solid) for inorganicchemicals (25 elements plus cyanide). Contains analytical, document control, and quality assurance/quality controlprocedures.

Environmental Protection Agency (EPA). 1988c. Contract Laboratory Program Statement of Work for Organics Analysis Multi-media, Multi-concentration. Office of Emergency and Remedial Response. SOW No. 288.

Provides procedure required by EPA for analyzing aqueous and solid hazardous waste samples for 126 volatile, semi-volatile, pesticide, and PCB chemicals. Contains analytical, document control, and quality assurance/quality controlprocedures.

Environmental Protection Agency (EPA). 1988d. Laboratory Data Validation Functional Guidelines for Evaluating Inorganics Analysis.Office of Emergency and Remedial Response.

Provides guidance in laboratory data evaluation and validation for hazardous waste site samples analyzed under the EPACLP program. Aids in determining data problems and shortcomings and potential actions to be taken.

Environmental Protection Agency (EPA). 1988e. LaboratoryData Validation Functional Guidelines for Evaluating Organics Analysis(Functional Guidelines for Organics). Office of Emergency and Remedial Response.

Provides guidance in laboratory data evaluation and validation for hazardous waste site samples analyzed under the EPACLP program. Aids in determining data problems and shortcomings and potential actions to be taken.

Environmental Protection Agency (EPA). 1988f. Special Report on Ingested Inorganic Arsenic; Skin Cancer, Nutritional Essentiality.Risk Assessment Forum. EPA 625/3-87/013.

Technical report concerning the health effects of exposure to ingested arsenic. Includes epidemiologic studies suitable fordose-response evaluation from Taiwan, Mexico, and Germany. Also includes discussions on pathological characteristics andsignificance of arsenic-induced skin lesions, genotoxicity of arsenic, metabolism and distribution, dose-response estimatesfor aresnic ingestion and arsenic as an essential nutrient.

CHAPTER 6

EXPOSURE ASSESSMENT

CHAPTER 6

EXPOSURE ASSESSMENT

This chapter describes the procedures forconducting an exposure assessment as part of thebaseline risk assessment process at Superfundsites. The objective of the exposure assessment isto estimate the type and magnitude of exposuresto the chemicals of potential concern that arepresent at or migrating from a site. The resultsof the exposure assessment are combined withchemical-specific toxicity information tocharacterize potential risks.

The procedures and information presentedin this chapter represent some new approaches toexposure assessment as well as a synthesis ofcurrently available exposure assessment guidanceand information published by EPA. Throughoutthis chapter, relevant exposure assessmentdocuments are referenced as sources of moredetailed information supporting the exposureassessment process.

6.1 BACKGROUND

Exposure is defined as the contact of anorganism (humans in the case of health riskassessment) with a chemical or physical agent(EPA 1988a). The magnitude of exposure isdetermined by measuring or estimating theamount of an agent available at the exchangeboundaries (i.e., the lungs, gut, skin) during aspecified time period. Exposure assessment is thedetermination or estimation (qualitative orquantitative) of the magnitude, frequency,duration, and route of exposure. Exposureassessments may consider past, present, and futureexposures, using varying assessment techniques foreach phase. Estimates of current exposures canbe based on measurements or models of existingconditions, those of future exposures can be basedon models of future conditions, and those of pastexposures can be based on measured or modeledpast concentrations or measured chemical

concentrations in tissues. Generally, Superfundexposure assessments are concerned with currentand future exposures. If human monitoring isplanned to assess current or past exposures, theAgency for Toxic Substances and Disease Registry(ATSDR) should be consulted to take the lead inconducting these studies and in assessing thecurrent health status of the people near the sitebased on the monitoring results.

6.1.1 COMPONENTS OF ANEXPOSURE ASSESSMENT

The general procedure for conducting anexposure assessment is illustrated in Exhibit 6-1.This procedure is based on EPA’s publishedGuidelines for Exposure Assessment (EPA 1986a)and on other related guidance (EPA 1988a,1988b). It is an adaptation of the generalizedexposure assessment process to the particularneeds of Superfund site risk assessments.Although some exposure assessment activities mayhave been started earlier (e.g., during RI/FSscoping or even before the RI/FS process began),the detailed exposure assessment process beginsafter the chemical data have been collected and

2

validated and the chemicals of potential concernhave been selected (see Chapter 5, Section 5.3.3).The exposure assessment proceeds with thefollowing steps.

Step 1-- Characterization of exposure setting(Section 6.2). In this step, the assessorcharacterizes the exposure setting with respectto the general physical characteristics of thesite and the characteristics of the populationson and near the site. Basic site

characteristics such as climate, vegetation,ground-water hydrology, and the presence andlocation of surface water are identified in thisstep. Populations also are identified and aredescribed with respect to those characteristicthat influence exposure, such as locationrelative to the site, activity patterns, and thepresence of sensitive subpopulations. Thisstep considers the characteristics of thecurrent population, as well as those of any

3

EXHIBIT 6-1

THE EXPOSURE ASSESSMENT PROCESS

4

potential future populations that may differunder an alternate land use.

Step 2 -- Identification of exposure pathways(Section 6.3). In this step, the exposureassessor identifies those pathways by whichthe previously identified populations may beexposed. Each exposure pathway describesa unique meehanism by which a populationmay be exposed to the chemicals at ororiginating from the site. Exposure pathwaysare identified based on consideration of thesources, releases, types, and locations ofchemicals at the site; the likely environmentalfate (including persistence, partitioning,transport, and intermedia transfer) of thesechemicals; and the location and activities ofthe potentially exposed populations.Exposure points (points of potential contactwith the chemical) and routes of exposure(e.g., ingestion, inhalation) are identified foreach exposure pathway.

Step 3 -- Quantification of exposure (Section6.4). In this step, the assessor quantifies themagnitude, frequeney and duration ofexposure for each pathway identified in Step2. This step is most often conducted in twostages: estimation of exposure concentrationsand calculation of intakes.

Estimation of exposure concentrations(Section 6.5). In this part of step 3, theexposure assessor determines theconcentration of chemicals that will becontacted over the exposure period.Exposure concentrations are estimated usingmonitoring data and/or chemical transportand environmental fate models. Modelingmay be used to estimate future chemicalconcentrations in media that are currentlycontaminated or that may becomecontaminated, and current concentrations inmedia and/or at locations for which there areno monitoring data.

Calculation of intakes (Section 6.6). In thispart of step 3, the exposure assessorcalculates chemical-specific exposures for eachexposure pathway identified in Step 2.Exposure estimates are expressed in termsof the mass of substance in contact with thebody per unit body weight per unit time (e.g.,

mg chemical per kg body weight per day, alsoexpressed as mg/kg-day). These exposureestimates are termed “intakes” (for thepurposes of this manual) and represent thenormalized exposure rate. Several termscommon in other EPA documents and theliterature are equivalent or related to intake(see box on this page and definitions box onpage 6-2). Chemical intakes are calculatedusing equations that include variables forexposure concentration, contact rate, exposurefrequency, exposure duration, body weight,and exposure averaging time. The values ofsome of these variables depend on siteconditions and the characteristics of thepotentially exposed population.

After intakes have been estimated, they areorganized by population, as appropriate (Section6.7). Then, the sources of uncertainty (e.g.,variability in analytical data, modeling results,parameter assumptions) and their effect on theexposure estimates are evaluated and summarized(Section 6.8). This information on uncertainty isimportant to site decision-makers who mustevaluate the results of the exposure and riskassessment and make decisions regarding thedegree of remediation required at a site. Theexposure assessment concludes with a summary ofthe estimated intakes for each pathway evaluated(Section 6.9).

6.1.2 REASONABLE MAXIMUM EXPOSURE

Actions at Superfund sites should be basedon an estimate of the reasonable maximumexoosure (RME) expected to occur under bothcurrent and future land-use conditions. Thereasonable maximum exposure is defined here as

5

the highest exposure that is reasonably expectedto occur at a site. RMEs are estimated forindividual pathways. If a population is exposedvia more than one pathway, the combination ofexposures across pathways also must represent anRME.

Estimates of the reasonable maximumexposure necessarily involve the use ofprofessional judgment. This chapter providesguidance for determining the RME at a site andidentifies some exposure variable valuesappropriate for use in this determination. Thespecific values identified should be regarded asgeneral recommendations, and could change basedon site-specific information and the particularneeds of the EPA remedial project manager(RPM). Therefore, these recommendations shouldbe used in conjunction with input from the RPMresponsible for the site.

In the past, exposures generally wereestimated for an average and an upper-boundexposure case, instead of a single exposure case(for both current and future land use) asrecommended here. The advantage of the twocase approach is that the resulting range ofexposures provides some measure of theuncertainty surrounding these estimates. Thedisadvantage of this approach is that the upper-bound estimate of exposure may be above therange of possible exposures, whereas the averageestimate is lower than exposures potentiallyexperienced by much of the population. Theintent of the RME is to estimate a conservativeexposure case (i.e., well above the average case)that is still within the range of possible exposures.Uncertainty is still evaluated under this approach.However, instead of combining many sources ofuncertainty into average and upper-boundexposure estimates, the variation in individualexposure variables is used to evaluate uncertainty(See Section 6.8). In this way, the variablescontributing most to uncertainty in the exposureestimate are more easily identified.

6.2 STEP 1: CHARACTERI-ZATION OF EXPOSURESETTING

The first step in evaluating exposure atSuperfund sites is to characterize the site withrespect to its physical characteristics as well asthose of the human populations on and near thesite. The output of this step is a qualitativeevaluation of the site and surrounding populationswith respect to those characteristics that influenceexposure. All information gathered during thisstep will support the identification of exposurepathways in Step 2. In addition, the informationon the potentially exposed populations will beused in Step 3 to determine the values of someintake variables.

6.2.1 CHARACTERIZE PHYSICALSETTING

Characterize the exposure setting with respectto the general physical characteristics of the site.Important site characteristics include thefollowing:

climate (e.g., temperature,precipitation);

meteorology (e.g., wind speed anddirection);

geologic setting (e.g., location andcharacterization of underlying strata);

vegetation (e.g., unvegetated, forested,grassy);

soil type (e.g., sandy, organic, acid,basic);

ground-water hydrology (e.g., depth,direction and type of flow); and

location and description of surface water(e.g., type, flow rates, salinity).

Sources of this information include sitedescriptions and data from the preliminaryassessment (PA), site inspection (SI), and remedialinvestigation (RI) reports. Other sources includecounty soil surveys, wetlands maps, aerial

6

photographs, and reports by the NationalOceanographic and Atmospheric Association(NOAA) and the U.S. Geological Survey (USGS).The assessor also should consult with appropriatetechnical experts (e.g., hydrogeologists, airmodelers) as needed to characterize the site.

6.2.2 CHARACTERIZE POTENTIALLYEXPOSED POPULATIONS

Characterize the populations on or near thesite with respect to location relative to the site,activity patterns, and the presence of sensitivesubgroups.

Determine location of current populationsrelative to the site. Determine the distance anddirection of potentially exposed populations fromthe site. Identify those populations that areclosest to or actually living on the site and that,therefore, may have the greatest potential forexposure. Be sure to include potentially exposeddistant populations, such as public water supplyconsumers and distant consumers of fish orshellfish or agricultural products from the sitearea. Also include populations that could beexposed in the future to chemicals that havemigrated from the site. Potential sources of thisinformation include:

site visit;

other information gathered as part ofthe SI or during the initial stages of theR I

population surveys conducted near thesite;

topographic, land use, housing or othermaps; and

recreational and commercial fisheriesdata.

Determine current land use. Characterizethe activities and activity patterns of thepotentially exposed population. The followingland use categories will be applicable most oftenat Superfund sites:

residential;commercial/industrial; and

recreational.

Determine the current land use or uses ofthe site and surrounding area. The best sourceof this information is a site visit. Look forhomes, playgrounds, parks, businesses, industries,or other land uses on or in the vicinity of the site.Other sources on local land use include

zoning maps;

state or local zoning or other land use-related laws and regulations;

data from the U.S. Bureau of theCensus;

topographic, land use, housing or othermaps and

aerial photographs.

Some land uses at a site may not fit neatlyinto one of the three land use categories andother land use classifications may be moreappropriate (e.g., agricultural land use). At somesites it may be most appropriate to have morethan one land use category.

After defining the land use(s) for a site,identify human activities and activity patternsassociated with each land use. This is basicallya “common sense” evaluation and is not based onany specific data sources, but rather on a generalunderstanding of what activities occur inresidential, business, or recreational areas.

Characterize activity patterns by doing thefollowing.

Determine the percent of time that thepotentially exposed population(s) spendin the potentially contaminated area.For example, if the potentially exposedpopulation is commercial or industrial,a reasonable maximum daily exposureperiod is likely to be 8 hours (a typicalwork day). Conversely, if the populationis residential, a maximum daily exposureperiod of 24 hours is possible.

Determine if activities occur primarilyindoors, outdoors, or both. For example,

7

office workers may spend all their timeindoors, whereas construction workersmay spend all their time outdoors.

Determine how activities change withthe seasons. For example, someoutdoor, summertime recreationalactivities (e.g., swimming, fishing) willoccur less frequently or not at all duringthe winter months. Similarly, childrenare likely to play outdoors leas frequentlyand with more clothing during the wintermonths.

Determine if the site itself may be usedby local populations, particularly if accessto the site is not restricted or otherwiselimited (e.g., by distance). For example,children living in the area could playonsite, and local residents could hunt orhike onsite.

Identify any site-specific populationcharacteristics that might influenceexposure. For example, if the site islocated near major commercial orrecreational fisheries or shellfisheries,the potentially exposed population islikely to eat more locally-caught fish andshellfish than populations located inland.

Determine future land use. Determine if anyactivities associated with a current land use arelikely to be different under an alternate futureland use. For example, if ground water is notcurrently used in the area of the site as a sourceof drinking water but is of potable quality, futureuse of ground water as drinking water would bepossible. Also determine if land use of the siteitself could change in the future. For example, ifa site is currently classified as industrial,determine if it could possibly be used forresidential or recreational purposes in the future.

Because residential land use is most oftenassociated with the greatest exposures, it isgenerally the most conservative choice to makewhen deciding what type of alternate land usemay occur in the future. However, an assumptionof future residential land use may not bejustifiable if the probability that the site willsupport residential use in the future is exceedinglysmall.

Therefore, determine possible alternate futureland uses baaed on available information andprofessional judgment. Evaluate pertinentinformation sources, including (as available):

master plans (city or county projectionsof future land use);

Bureau of the Census projections; and

established land use trends in the generalarea and the area immediatelysurrounding the site (use Census Bureauor state or local reports, or use generalhistorical accounts of the area).

Note that while these sources provide potentiallyuseful information, they should not be interpretedas providing proof that a certain land use will orwill not occur.

Assume future residential land use if it seemspossible based on the evaluation of the availableinformation. For example, if the site is currentlyindustrial but is located near residential areas inan urban area, future residential land use may bea reasonable possibility. If the site is industrialand is located in a very rural area with a lowpopulation density and projected low growth,future residential use would probably be unlikely.In this case, a more likely alternate future landuse may be recreational. At some sites, it may bemost reasonable to assume that the land use willnot change in the future.

There are no hard-and-fast rules by which todetermine alternate future land use. The use ofprofessional judgment in this step is critical. Besure to consult with the RPM about any decisionregarding alternate future land use. Support theselection of any alternate land use with a logical,reasonable argument in the exposure assessmentchapter of the risk assessment report. Alsoinclude a qualitative statement of the likelihoodof the future land use occurring.

Identify subpopulations of potential concern.Review information on the site area to determineif any subpopulations may be at increased riskfrom chemical exposures due to increasedsensitivity, behavior patterns that may result inhigh exposure, and/or current or past exposuresfrom other sources. Subpopulations that may be

8

more sensitive to chemical exposures includeinfants and children, elderly people, pregnant andnursing women, and people with chronic illnesses.Those potentially at higher risk due to behaviorpatterns include children, who are more likely tocontact soil, and persons who may eat largeamounts of locally caught fish or locally grownproduce (e.g., home-grown vegetables).Subpopulations at higher risk due to exposuresfrom other sources include individuals exposed tochemicals during occupational activities andindividuals living in industrial areas.

To identify subpopulations of potentialconcern in the site area, determine locations ofschools, day care centers, hospitals, nursing homes,retirement communities, residential areas withchildren, important commercial or recreationalfisheries near the site, and major industriespotentially involving chemical exposures. Uselocal census data and information from localpublic health officials for this determination.

6.3 STEP 2: IDENTIFICATIONOF EXPOSURE PATHWAYS

This section describes an approach foridentifying potential human exposure pathways ata Superfund site. An exposure pathway describesthe course a chemical or physical agent takes fromthe source to the exposed individual. An exposurepathway analysis links the sources, locations, andtypes of environmental releases with populationlocations and activity patterns to determine thesignificant pathways of human exposure.

An exposure pathway generally consists offour elements: (1) a source and mechanism ofchemical release, (2) a retention or transportmedium (or media in cases involving mediatransfer of chemicals), (3) a point of potentialhuman contact with the contaminated medium(referred to as the exposure point), and (4) anexposure route (e.g., ingestion) at the contactpoint. A medium contaminated as a result of apast release can be a contaminant source for othermedia (e.g., soil contaminated from a previousspill could be a contaminant source for groundwater or surface water). In some cases, the sourceitself (i.e., a tank, contaminated soil) is theexposure point, without a release to any other

medium. In these latter cases, an exposurepathway consists of (1) a source, (2) an exposurepoint, and (3) an exposure route. Exhibit 6-2illustrates the basic elements of each type ofexposure pathway.

The following sections describe the basicanalytical process for identifying exposurepathways at Superfund sites and for selectingpathways for quantitative analysis. The pathwayanalysis described below is meant to be aqualitative evaluation of pertinent site andchemical information, and not a rigorousquantitative evaluation of factors such as sourcestrength, release rates, and chemical fate andtransport. Such factors are considered later inthe exposure assessment during the quantitativedetermination of exposure concentrations (Section6.5).

6.3.1 IDENTIFY SOURCES ANDRECEIVNG MEDIA

To determine possible release sources for asite in the absence of remedial action, use allavailable site descriptions and data from the PA,SI, and RI reports. Identify potential releasemechanisms and receiving media for past, current,and future releases. Exhibit 6-3 lists some typicalrelease sources, release mechanisms, and receivingmedia at Superfund sites. Use monitoring data inconjunction with information on source locationsto support the analysis of past, continuing, orthreatened releases. For example, soilcontamination near an old tank would suggest thetank (source) ruptured or leaked (releasemechanism) to the ground (receiving media). Besure to note any source that could be an exposurepoint in addition to a release source (e.g., openbarrels or tanks, surface waste piles or lagoons,contaminated soil).

Map the suspected source areas and theextent of contamination using the availableinformation and monitoring data. As an aid inevaluating air sources and releases, Volumes I andII of the National Technical Guidance Studies(NTGS; EPA 1989a,b) should be consulted.

9

EXHIBIT 6-2

ILLUSTRATION OF EXPOSUREPATHWAYS

10

EXHIBIT 6-3

COMMON CHEMICAL RELEASE SOURCES ATSITES IN THE ABSENCE OF REMEDIAL ACTION

Receiving ReleaseMedium Mechanism Release Source

Air Volatilization Surface wastes -- lagoons,ponds, pits, spills

Contaminated surface waterContaminated surface soilContaminated wetlandsLeaking drums

Fugitive dustgeneration

Contaminated surface soilWaste piles

Contaminated surface soilSurface water Surface runoff

Episodic overland Lagoon overflowSpills, leaking containers

Contaminated ground waterGround-waterseepage

Surface or buried wastesContaminated soil

Ground water Leaching

Soil Leaching Surface or buried wastes

Contaminated surface soilSurface runoff

Lagoon overflowSpills, leaking containers

Episodic overlandflow

Contaminated surface soilWaste piles

Fugitive dustgenerationdeposition

Contaminated surface soilTracking

Surface wastes -- lagoons,ponds, pits, spills

Contaminated surface soil

Sediment Surface runoff,Episodic overlandflow

Ground-waterseepage

Contaminated ground water

Leaching Surface or buried wastesContaminated soil

Contaminated soil, surfacewater, sediment, groundwater or air

Other biota

Biota Uptake(direct contact,ingestion, inhalation)

11

6.3.2 EVALUATE FATE AND TRANSPORTIN RELEASE MEDIA

Evaluate the fate and transport of thechemicals to predict future exposures and to helplink sources with currently contaminated media.The fate and transport analysis conducted at thisstage of the exposure assessment is not meant toresult in a quantitative evaluation of media-specific chemical concentrations. Rather, theintent is to identify media that are receiving ormay receive site-related chemicals. At this stage,the assessor should answer the questions: Whatchemicals occur in the sources at the site and inthe environment? In what media (onsite andoffsite) do they occur now? In what media andat what location may they occur in the future?Screening-level analyses using available data andsimplified calculations or analytical models mayassist in this qualitative evaluation.

After a chemical is released to theenvironment it may be:

transported (e.g., convected downstreamin water or on suspended sediment orthrough the atmosphere);

physically transformed (e.g., volatilization,precipitation);

chemically transformed (e.g., photolysis,hydrolysis, oxidation, reduction, etc.);

biologically transformed (e.g,biodegradation); and/or

accumulated in one or more media(including the receiving medium).

To determine the fate of the chemicals ofpotential concern at a particular site, obtaininformation on their physical/chemical andenvironmental fate properties. Use computer databases (e.g., SRC’s Environmental FateCHEMFATE, and BIODEG data bases; BIOSIS;AQUIRE) and the open literature as necessaryas sources for up-to-date information on thephysical/chemical and fate properties of thechemicals of potential concern. Exhibit 6-4 listssome important chemical-specific fate parametersand briefly describes how these can be used toevaluate a chemical’s environmental fate.

Also consider site-specific characteristics(identified in Section 6.2.1) that may influencefate and transport. For example, soilcharacteristics such as moisture content, organiccarbon content, and cation exchange capacity cangreatly influence the movement of many chemicals.A high water table may increase the probability ofleaching of chemicals in soil to ground water.

Use all applicable chemical and site-specificinformation to evaluate transport within andbetween media and retention or accumulationwithin a single medium. Use monitoring data toidentify media that are contaminated now and thefate pathway analysis to identify media that maybe contaminated now (for media not sampled) orin the future. Exhibit 6-5 presents someimportant questions to consider when developingthese pathways. Exhibit 6-6 presents a series offlow charts useful when evaluating the fate andtransport of chemicals at a site.

6.3.3 IDENTIFY EXPOSURE POINTS ANDEXPOSURE ROUTES

After contaminated or potentiallycontaminated media have been identified, identifyexposure points by determining if and where anyof the potentially exposed populations (identifiedin Step 1) can contact these media. Considerpopulation locations and activity patterns in thearea, including those of subgroups that may be ofparticular concern. Any point of potential contactwith a contaminated medium is an exposure point.Try to identify those exposure points where theconcentration that will be contacted is thegreatest. Therefore, consider including anycontaminated media or sources onsite as apotential exposure point if the site is currentlyused, if access to the site under current conditionsis not restricted or otherwise limited (e.g., bydistance), or if contact is possible under analternate future land use. For potential offsiteexposures, the highest exposure concentrationsoften will be at the points closest to anddowngradient or downwind of the site. In somecases, highest concentrations may be encounteredat points distant from the site. For example, site-related chemicals may be transported anddeposited in a distant water body where they maybe subsequently bioconcentrated by aquaticorganisms.

12

EXHIBIT 6-4

IMPORTANT PHYSICAL/CHEMICAL ANDENVIRONMENTAL FATE PARAMETERS

K o c provides a measure of the extent of chemical partitioning between organic carbon and water atequilibrium. The higher the K, the more likely a chemical is to bind to soil or sediment than toremain in water.

K dprovides a soil or sediment-specific measure of the extent of chemical partitioning between soilor sediment and water, unadjusted for dependence upon organic carbon. To adjust for thefraction of organic carbon present in soil or sediment (foc) use Kd = Kocxfoc. The higher the Kd,the more likely a chemical is to bind to soil or sediment than to remain in water.

K o w provides a measure of the extent of chemical partitioning between water and octanol atequilibrium. The greater the Kow the more likely a chemical is to partition to octanol than toremain in water. Octanol is used as a surrogate for lipids (fat), and Kow, can be used to predictbioconcentration in aquatic organisms.

Solubility is an upper limit on a chemical’s dissolved concentration in water at a specified temperature.Aqueous concentrations in excess of volubility may indicate sorption onto sediments, thepresence of solubilizing chemicals such as solvents, or the presence of a non-aqueous phaseliquid.

Henry’s Law Constant provides a measure of the extent of chemical partitioning between air and water atequilibrium. The higher the Henry’s Law constant, the more likely a chemical is to volatilizethan to remain in the water.

Vapor Pressure is the pressure exerted by a chemical vapor in equilibrium with its solid or liquid form atany given temperature. It is used to calculate the rate of volatilization of a pure substance from asurface or in estimating a Henry’s Law constant for chemicals with low water solubility. Thehigher the vapor pressure, the more likely a chemical is to exist in a gaseous state.

Diffusivity describes the movement of a molecule in a liquid or gas medium as a result of differences inconcentration. It is used to calculate the dispersive component of chemical transport. Thehigher the diffusivity, the more likely a chemical is to move in response to concentrationgradients.

Bioconcentration Factor (BCF) provides a measure of the extent of chemical partitioning at equilibriumbetween a biological medium such as fish tissue or plant tissue and an external medium such aswater. The higher the BCF, the greater the accumulation in living tissue is likely to be.

Media-specific Half-life provides a relative measure of the persistence of a chemical in a given medium,although actual values can vary greatly depending on site-specific conditions. The greater thehalf-life, the more persistent a chemical is likely to be.

13

EXHIBIT 6-5

IMPORTANT CONSIDERATIONS FOR DETERMININGTHE ENVIRONMENTAL FATE AND TRANSPORTOF THE CHEMICALS OF POTENTIAL CONCERN

AT A SUPERFUND SITE

. What are the principal mechanisms for change or removal in each of the environmentalmedia?

. How does the chemical behave in air, water, soil, and biological media? Does itbioaccumulate or biodegrade? Is it absorbed or taken up by plants?

. Does the agent react with other compounds in the environment?

. Is there intermedia transfer? What are the mechanisms for intermedia transfer? Whatare the rates of tbe intermedia transfer or reaction mechanism?

. How long might the chemical remain in each environmental medium? How does itsconcentration change with time in each medium?

. What are the products into which the agent might degrade or change in the environment?Are these products potentially of concern?

. Is a steady-state concentration distribution in the environment or in specific segments ofthe environment achieved?

1

14

EXHIBIT 6-6

FLOW CHART FORFATE AND TRANSPORT ASSESSMENTS

Environmental fate and transport assessment atmosphere

Source: Adapted from EPA 1988b.

(continued)

15



Environmental fate and transport assessment: surface water and sediment


(continued)

16



Environmental fate and transport assessment: soils and ground water


17

After determining exposure points, identifyprobable exposure routes (i.e., ingestion,inhalation, dermal contact) based on the mediacontaminated and the anticipated activities at theexposure points. In some instances, an exposurepoint may exist but an exposure route may not(e.g., a person touches contaminated soil but iswearing gloves). Exhibit 6-7 presents apopulation/exposure route matrix that can be usedin determining potential exposure routes at a site.

6.3.4 INTEGRATE INFORMATION ONSOURCES, RELEASES, FATE ANDTRANSPORT, EXPOSURE POINTS,AND EXPOSURE ROUTES INTOEXPOSURE PATHWAYS

Assemble the information developed in theprevious three steps and determine the completeexposure pathways that exist for the site. Apathway is complete if there is (1) a source orchemical release from a source, (2) an exposurepoint where contact can occur, and (3) anexposure route by which contact can occur.Otherwise, the pathway is incomplete, such as thesituation where there is a source releasing to airbut there are no nearby people. If available fromATSDR, human monitoring data indicatingchemical accumulation or chemical-related effectsin the site area can be used as evidence tosupport conclusions about which exposurepathways are complete; however, negative datafrom such studies should not be used to concludethat a pathway is incomplete.

From all complete exposure pathways at asite, select those pathways that will be evaluatedfurther in the exposure assessment. If exposureto a sensitive subpopulation is possible, select thatpathway for quantitative evaluation. All pathwaysshould be selected for further evaluation unlessthere is sound justification (e.g., based on theresults of a screening analysis) to eliminate apathway from detailed analysis. Such ajustification could be based on one of thefollowing:

the exposure resulting from the pathwayis much less than that from anotherpathway involving the same medium atthe same exposure point;

the potential magnitude of exposurefrom a pathway is low or

the probability of the exposure occurringis very low and the risks associated withthe occurrence are not high (if apathway has catastrophic consequences,it should be selected for evaluation evenif its probability of occurrence is verylow).

Use professional judgment and experience tomake these decisions. Before deciding to excludea pathwav from quantitative analysis, consult withthe RPM. If a pathway is excluded from furtheranalysis, clearly document the reasons for thedecision in the exposure assessment section of therisk assessment report.

For some complete pathways it may not bepossible to quantify exposures in the subsequentsteps of the analysis because of a lack of data onwhich to base estimates of chemical release,environmental concentration, or human intake.Available modeling results should complement andsupplement the available monitoring data tominimize such problems. However, uncertaintiesassociated with the modeling results may be toolarge to justify quantitative exposure assessmentin the absence of monitoring data to validate themodeling results. These pathways shouldnevertheless be carried through the exposureassessment so that risks can be qualitativelyevaluated or so that this information can beconsidered during the uncertainty analysis of theresults of the exposure assessment (see Section6.8) and the risk assessment (see Chapter 8).

6.3.5 SUMMARIZE INFORMATION ONALL COMPLETE EXPOSUREPATHWAYS

Summarize pertinent information on allcomplete exposure pathways at the site byidentifying potentially exposed populations,exposure media, exposure points, and exposureroutes. Also note if the pathway has beenselected for quantitative evaluation; summarize thejustification if a pathway has been excluded.Summarize pathways for current land use and anyalternate future land use separately. Thissummary information is useful for defining thescope of the next step (quantification of exposure)

18

EXHIBIT 6-7

MATRIX OF POTENTIAL EXPOSURE ROUTES

Exposure Medium/ Residential Commercial/Industrial RecreationalExposure Route Population Population Population

Ingestion L ADermal Contact L A —

Ingestion L A L, CDermal Contact L A L, C

SedimentIncidental Ingestion C A CDermal Contact C A L, C

AirInhalation of VaporPhase Chemicals

Indoors L —outdoors L L

Inhalation ofParticulate

Indoors L —Outdoors L L

Soil/DustIncidental Ingestion L, C L, CDermal Contact L, C L, C

IngestionFish and Shellfish L — LMeat and Game L — LDairy L, C — LEggs L LVegetables L L

AA

AA

AA

— = Exposure of this population via this route is not likely to occur.

19

and also is useful as documentation of theexposure pathway analysis. Exhibit 6-8 providesa sample format for presenting this information.

6.4 STEP 3: QUANTIFICATIONOF EXPOSURE GENERALCONSIDERATIONS

The next step in the exposure assessmentprocess is to quantify the magnitude, frequencyand duration of exposure for the populations andexposure pathways selected for quantitativeevaluation. This step is most often conducted intwo stages: first, exposure concentrations areestimated, then, pathway-specific intakes arequantified. The specific methodology forcalculating exposure concentrations and pathway-specific exposures are presented in Sections 6.5and 6.6, respectively. This section describes someof the basic concepts behind these processes.

6.4.1 QUANTIFYING THE REASONABLEMAXIMUM EXPOSURE

Exposure is defined as the contact of anorganism with a chemical or physical agent. Ifexposure occurs over time, the total exposure canbe divided by a time period of interest to obtainan average exposure rate per unit time. Thisaverage exposure rate also can be expressed as afunction of body weight. For the purposes of thismanual, exposure normalized for time and bodyweight is termed “intake”, and is expressed in unitsof mg chemical/kg body weight-day.

Exhibit 6-9 presents a generic equation forcalculating chemical intakes and defines the intakevariables. There are three categories of variablesthat are

(1)

(2)

(3)

used to estimate intake

chemical-related variable -- exposureconcentration,

variables that describe the exposedpopulation -- contact rate, exposurefrequency and duration, and body weightand

assessment-determined variable --averaging time.

Each intake variable in the equation has arange of values. For Superfund exposureassessments, intake variable values for a givenpathwav should be selected so that thecombination of all intake variables results in anestimate of the reasonable maximum exposure forthat pathway. As defined previously, thereasonable maximum exposure (RME) is themaximum exposure that is reasonably expected tooccur at a site. Under this approach, some intakevariables may not be at their individual maximumvalues but when in combination with othervariables will result in estimates of the RME.Some recommendations for determining the valuesof the individual intake variables are discussedbelow. These recommendations are based onEPA’s determination of what would result in anestimate of the RME. As discussed previously, adetermination of “reasonable” cannot be basedsolely on quantitative information, but alsorequires the use of professional judgment.Accordingly, the recommendations below are basedon a combination of quantitative information andprofessional judgment. These are generalrecommendations, however, and could changebased on site-specific information or the particularneeds of the risk manager. Consult with the RPMbefore varying from these recommendations.

Exposure concentration. The concentrationterm in the intake equation is the arithmeticaverage of the concentration that is contacted overthe exposure period. Although this concentrationdoes not reflect the maximum concentration thatcould be contacted at any one time, it is regardedas a reasonable estimate of the concentrationlikely to be contacted over time. This is becausein most situations, assuming long-term contactwith the maximum concentration is notreasonable. (For exceptions to this generalization,see discussion of hot spots in Section 6.5.3.)

Because of the uncertainty associated withany estimate of exposure concentration. the upperconfidence limit (i.e., the 95 percent upperconfidence limit) on the arithmetic average will beused for this variable. There are standardstatistical methods which can be used to calculatethe upper confidence limit on the arithmeticmean. Gilbert (1987, particularly sections 11.6and 13.2) discusses methods that can be appliedto data that are distributed normally or lognormally. Kriging is another method that

20

EXHIBIT 6-8

EXAMPLE OF TABLE FORMAT FOR SUMMARIZINGCOMPLETE EXPOSURE PATHWAYS AT A SITE

Potentially Exposed Exposure Route, Medium Pathway Selected Reason for SelectionPopulation and Exposure Point for Evaluation? or Exclusion

Residents

Residents

IndustrialWorkers

Residents

Residents

Ingestion of ground waterfrom local wells down-gradient of the site

Inhalation of chemicalsvolatilized from groundwater during home use

Direct contact withchemicals of potentialconcern in soil on thesite

Direct contact with chemi-cals of potential concernin soil on the site

Ingestion of chemicals

Yes

Yes

Yes

Yes

No

Residents use groundwater from local wellsas drinking water.

Some of the chemicalsof potential concern inground water are volatile,and ground water is usedby local residents.

Contaminated soil is inan area potentially usedby outside maintenanceworkers.

Area could be developedin the future as aresidential area.

The potential for signifi-that have accumulated in cant exposure via thisfish located in onsite pathway is low becauseponds none of the chemicals of

potential concern accumulateextensively in fish.

21

EXHIBIT 6-9

GENERIC EQUATION FOR CALCULATINGCHEMICAL INTAKES

22

potentially can be used (Clark 1979 is one ofseveral reference books on kriging). A statisticianshould be consulted for more details or forassistance with specific methods.

If there is great variability in measured ormodeled concentration values (such as when toofew samples are taken or when model inputs areuncertain), the upper confidence limit on theaverage concentration will be high, andconceivably could be above the maximum detectedor modeled value. In these cases, the maximumdetected or modeled value should be used toestimate exposure concentrations. This could beregarded by some as too conservative an estimate,but given the uncertainty in the data in thesesituations, this approach is regarded as reasonable.

For some sites, where a screening levelanalysis is regarded as sufficient to characterizepotential exposures, calculation of the upperconfidence limit on the arithmetic average is notrequired. In these cases, the maximum detectedor modeled concentration should be used as theexposure concentration.

Contact rate. Contact rate reflects theamount of contaminated medium contacted perunit time or event. If statistical data are availablefor a contact rate, use the 95th percentile valuefor this variable. (In this case and throughout thischapter, the 90th percentile value can be used ifthe 95th percentile value is not available.) Ifstatistical data are not available, professionaljudgment should be used to estimate a valuewhich approximates the 95th percentile value. (Itis recognized that such estimates will not beprecise. They should, however, reflect areasonable estimate of an upper-bound value.)

Sometimes several separate terms are used toderive an estimate of contact rate. For example,for dermal contact with chemicals in water,contact rate is estimated by combining informationon exposed skin surface area, dermal permeabilityof a chemical, and exposure time. In suchinstances, the combination of variables used toestimate intake should result in an estimateapproximating the 95th percentile value.Professional judgment will be needed to determinethe appropriate combinations of variables. (Morespecific guidance for determining contact rate forvarious pathways is given in Section 6.6.)

Exposure frequency and duration. Exposurefrequency and duration are used to estimate thetotal time of exposure. These terms aredetermined on a site-specific basis. If statisticaldata are available, use the 95th percentile valuefor exposure time. In the absence of statisticaldata (which is usually the case), use reasonableconservative estimates of exposure time. Nationalstatistics are available on the upper-bound (90thpercentile) and average (50th percentile) numberof years spent by individuals at one residence(EPA 1989d). Because of the data on which theyare based, these values may underestimate theactual time that someone might live in oneresidence. Nevertheless, the upper-bound value of30 years can be used for exposure duration whencalculating reasonable maximum residentialexposures. In some cases, however, lifetimeexposure (70 years by convention) may be a moreappropriate assumption. Consult with the RPMregarding the appropriate exposure duration forresidential exposures. The exposure frequency andduration selected must be appropriate for thecontact rate selected. If a long-term averagecontact rate (e.g., daily fish ingestion rate averagedover a year) is used, then a daily exposurefrequency (i.e., 365 days/year) should be assumed.

Body weight. The value for body weight isthe average body weight over the exposure period.If exposure occurs only during childhood years,the average child body weight during the exposureperiod should be used to estimate intake. Forsome pathways, such as soil ingestion, exposurecan occur throughout the lifetime but the majorityof exposure occurs during childhood (because ofhigher contact rates). In these cases, exposuresshould be calculated separately for age groupswith similar contact rate to body weight ratios; thebody weight used in the intake calculation foreach age group is the average body weight for thatage group. Lifetime exposure is, then calculatedby taking the time-weighted average of exposureestimates over all age groups. For pathwayswhere contact rate to body weight ratios are fairlyconstant over a lifetime (e.g., drinking wateringestion), a body weight of 70 kg is used.

A constant body weight over the period ofexposure is used primarily by convention, but alsobecause body weight is not always independent ofthe other variables in the exposure equation (mostnotably, intake). By keeping body weight

23

constant, error from this dependence is minimized.The average body weight is used because, whencombined with the other variable values in theintake equation, it is believed to result in the beatestimate of the RME. For example, combining a95th percentile contact rate with a 5th percentilebody weight is not considered reasonable becauseit is unlikely that smallest person would have thehighest intake. Alternatively, combining a 95thpercentile intake with a 95th percentile bodyweight is not considered a maximum because asmaller person could have a higher contact rate tobody weight ratio.

Averaging time, The averaging time selecteddepends on the type of toxic effect being assessed.When evaluating exposures to developmentaltoxicants, intakes are calculated by averaging overthe exposure event (e.g., a day or a singleexposure incident). For acute toxicants, intakesare calculated by averaging over the shortestexposure period that could produce an effect,usually an exposure event or a day. Whenevacuating longer-term exposure tononcarcinogenic toxicants, intakes are calculatedby averaging intakes over the period of exposure(i.e., subchronic or chronic daily intakes). Forcarcinogens, intakes are calculated by proratingthe total cumulative dose over a lifetime (i.e.,chronic daily intakes, also called lifetime averagedaily intake). This distinction relates to thecurrently held scientific opinion that themechanism of action for each category is different(see Chapter 7 for a discussion). The approachfor carcinogens is based on the assumption thata high dose received over a short period of timeis equivalent to a corresponding low dose spreadover a lifetime (EPA 1986b). This approachbecomes problematic as the exposures in questionbecome more intense but less frequent, especiallywhen there is evidence that the agent has showndose-rate related carcinogenic effects. In somecases, therefore, it may be necessary to consult atoxicologist to assess the level of uncertaintyassociated with the exposure assessment forcarcinogens. The discussion of uncertainty shouldbe included in both the exposure assessment andrisk characterization chapters of the riskassessment report.

6.4.2 TIMING CONSIDERATIONS

At many Superfund sites, long-term exposureto relatively low chemical concentrations (i.e.,chronic daily intakes) are of greatest concern. Insome situations, however, shorter-term exposures(e.g., subchronic daily intakes) also may beimportant. When deciding whether to evaluateshort-term exposure, the following factors shouldbe considered:

the toxicological characteristics of thechemicals of potential concern;

the occurrence of high chemicalconcentrations or the potential for alarge release;

persistence of the chemical in theenvironment; and

the characteristics of the population thatinfluence the duration of exposure.

Toxicity considerations. Some chemicals canproduce an effect after a single or very short-termexposure to relatively low concentrations. Thesechemicals include acute toxicants such as skinirritants and neurological poisons, anddevelopmental toxicants. At sites where thesetypes of chemicals are present, it is important toassess exposure for the shortest time period thatcould result in an effect. For acute toxicants thisis usually a single exposure event or a day,although multiple exposures over several days alsocould result in an effect. For developmentaltoxicants, the time period of concern is theexposure event. This is based on the assumptionthat a single exposure at the critical time indevelopment is sufficient to produce an adverseeffect. It should be noted that the critical timereferred to can occur in almost any segment ofthe human population (i.e., fertile men andwomen, the conceptus, and the child up to the ageof sexual maturation [EPA 1989e]).

Concentration considerations. Manychemicals can produce an effect after a single orvery short-term exposure, but only if exposure isto a relatively high concentration. Therefore, itis important that the assessor identify possiblesituations where a short-term exposure to a highconcentration could occur. Examples of such a

24

situation include sites where contact with a small,but highly contaminated area is possible (e.g., asource or a hot spot), or sites where there is apotential for a large chemical release (e.g.,explosions, ruptured drums, breached lagoondikes). Exposure should be determined for theshortest period of time that could produce aneffect.

Persistence considerations. Some chemicalsmay degrade rapidly in the environment. In thesecases, exposures should be assessed only for thatperiod of time in which the chemical will bepresent at the site. Exposure assessments in thesesituations may need to include evaluations ofexposure to the breakdown products, if they arepersistent or toxic at the levels predicted to occurat the site.

Population considerations. At some sites,population activities are such that exposure wouldoccur only for a short time period (a few weeksor months), infrequently, or intermittently.Examples of this would be seasonal exposuressuch as during vacations or other recreationalactivities. The period of time over whichexposures are averaged in these instances dependson the type of toxic effect being assessed (seeprevious discussion on averaging time, Section6.4.1).

6.5 QUANTIFICATION OFEXPOSURE: DETERMINA-TION OF EXPOSURECONCENTRATIONS

This section describes the basic approachesand methodology for determining exposureconcentrations of the chemicals of potentialconcern in different environmental media usingavailable monitoring data and appropriate models.As discussed in Section 6.4.1, the concentrationterm in the exposure equation is the averageconcentration contacted at the exposure point orpoints over the exposure period. When estimatingexposure concentrations, the objective is toprovide a conservative estimate of this averageconcentration (e.g., the 95 percent upperconfidence limit on the arithmetic mean chemicalconcentration).

This section provides an overview of the basicconcepts and approaches for estimating exposureconcentrations. It identifies what type ofinformation is needed to estimate concentrations,where to find it, and how to interpret and use it.This section is not designed to provide all theinformation necessary to derive exposureconcentrations and, therefore, does not detail thespecifics of potentially applicable models norprovide the data necessary to run the models orsupport concentration estimates. However,sources of such information, including theSuperfund Exposure Assessment Manual (SEAM;EPA 1988b) are referenced throughout thediscussion.

6.5.1 GENERAL CONSIDERATIONS FORESTIMATING EXPOSURECONCENTRATIONS

In general, a great deal of professionaljudgment is required to estimate exposureconcentrations. Exposure concentrations may beestimated by (1) using monitoring data alone, or(2) using a combination of monitoring data. andenvironmental fate and transport models. In mostexposure assessments, some combination ofmonitoring data and environmental modeling willbe required to estimate exposure concentrations.

Direct use of monitoring data. Use ofmonitoring data to estimate exposureconcentrations is normally applicable whereexposure involves direct contact with themonitored medium (e.g., direct contact withchemicals in soil or sediment), or in cases wheremonitoring has occurred directly at an exposurepoint (e.g., a residential drinking water well orpublic water supply). For these exposurepathways, monitoring data generally provide thebest estimate of current exposure concentrations.

As the first step in estimating exposureconcentrations, summarize available monitoringdata. The manner in which the data aresummarized depends upon the site characteristicand the pathways being evaluated. It may benecessary to divide chemical data from a particularmedium into subgroups based on the location ofsample points and the potential exposurepathways. In other instances, as when thesampling point is an exposure point (e.g., whenthe sample is from an existing drinking water well)

25

it may not be appropriate to group samples at all,but may be most appropriate to treat the sample

data separately when estimating intakes. Still, inother instances, the assessor may wish to use themaximum concentration from a medium as theexposure concentration for a given pathway as ascreening approach to place an upper bound onexposure. In these cases it is important toremember that if a screening level approachsuggests a potential health concern. the estimatesof exposure should be modified to reflect moreprobable exposure conditions.

In those instances where it is appropriate togroup sampling data from a particular medium,calculate for each exposure medium and eachchemical the 95 percent upper confidence limit onthe arithmetic average chemical concentration.See Chapter 5 for guidance on how to treatsample concentrations below the quantitationlimit.

Modeling approaches. In some instances, itmay not be appropriate to use monitoring dataalone, and fate and transport models may berequired to estimate exposure concentrations.Specific instances where monitoring data alonemay not be adequate are as follows.

Where exposure points are spatiallyseparate from monitoring points.Models may be required when exposurepoints are remote from sources ofcontamination if mechanisms for releaseand transport to exposure points exist(e.g., ground-water transport, airdispersion),

Where temporal distribution of data islacking. Typically, data from Superfundinvestigations are collected over arelatively short period of time. Thisgenerally will give a clear indication ofcurrent site conditions, but both long-term and short-term exposure estimatesusually are required in Superfundexposure assessments. Although theremay be situations where it is reasonableto assume that concentrations willremain constant over a long period oftime, in many cases the time span of themonitoring data is not adequate topredict future exposure concentrations.

Environmental models may be requiredto make these predictions.

Where monitoring data are restricted bythe limit of quantitation. Environmentalmodels may be needed to predictconcentrations of contaminants that maybe present at concentrations that arebelow the quantitation limit but that maystill cause toxic effects (even at such lowconcentrations). For example, in thecase of a ground-water plume discharginginto a river, the dilution afforded by theriver may be sufficient to reduce theconcentration of the chemical to a levelthat could not be detected by directmonitoring. However, as discussed inSection 5.3.1, the chemical may besufficiently toxic or bioaccumulative thatit could present a health risk atconcentrations below the limit ofquantitation. Models may be requiredto make exposure estimates in thesetypes of situations.

A wide variety of models are available foruse in exposure assessments. SEAM (EPA 19S8b)and the Exposure Assessment Methods Handbook(EPA 1989f) describe some of the modelsavailable and provide guidance in selectingappropriate modeling techniques. Also, theCenter for Exposure Assessment Modeling(CEAM -- Environmental Research Laboratory(ERL) Athens), the Source Receptor AnalysisBranch (Office of Air Quality Planning andStandards, or OAQPS), and modelers in EPAregional offices can provide assistance in selectingappropriate models. Finally, Volume IV of theNTGS (EPA 1989c) provides guidance for air andatmospheric dispersion modeling for Superfundsites. Be sure to discuss the fate and transportmodels to be used in the exposure assessment withthe RPM.

The level of effort to be expended inestimating exposure concentrations will depend onthe type and quantity of data available, the levelof detail required in the assessment, and theresources available for the assessment. In general,estimating exposure concentrations will involveanalysis of site monitoring data and application ofsimple, screening-level analytical models. Themost important factor in determining the level of

26

effort will be the quantity and quality of theavailable data. In general, larger data sets willsupport the use of more sophisticated models.

Other considerations. When evaluatingchemical contamination at a site, it is importantto review the spatial distribution of the data andevaluate it in ways that have the most relevanceto the pathway being assessed. In short, considerwhere the contamination is with respect to knownor anticipated population activity patterns. Mapsof both concentration distribution and activitypatterns will be useful for the exposureassessment. It is the intersection of activitypatterns and contamination that defines anexposure area. Data from random sampling orfrom systematic grid pattern sampling may bemore representative of a given exposure pathwaythan data collected only from hot spots.

Generally, verified GC/MS laboratory datawith adequate quality control will be required tosupport quantitative exposure assessment. Fieldscreening data generally cannot be incorporatedwhen estimating exposure concentrations becausethey are derived using less sensitive analyticalmethods and are subject to less stringent qualitycontrol.

Other areas to be considered in estimatingexposure concentrations are as follows.

Steady-state vs. non-steadv-stateconditions. Frequently, it may benecessary to assume steady-stateconditions because the informationrequired to estimate non-steady-stateconditions (such as source depletionrate) is not readily available. This islikely to overestimate long-term exposureconcentrations for certain pathways.

Number and type of exposure parametersthat must be assumed. In developingexposure models, values for site-specificparameters such as hydraulicconductivity, organic carbon content ofsoil, wind speed and direction, and soiltype may be required. These values maybe generated as part of the RI. In easeswhere these values are not available,literature values may be substituted. Inthe absence of applicable literature

values, the assessor must consider if areliable exposure concentration estimatecan be made.

Number and type of fate processes tobe considered. In some cases, exposuremodeling may be limited toconsiderations of mass balance, dilution,dispersion, and equilibrium partitioning.In other cases, models of more complexfate processes, such as chemical reaction,biodegradation, and photolysis may beneeded. However, prediction of suchfate processes requires significantly largerquantities of model calibration andvalidation data than required for lesscomplex fate processes. For those siteswhere these more complex fate processesneed to be modeled, be sure to consultwith the RPM regarding the added datarequirements.

6.5.2 ESTIMATE EXPOSURECONCENTRATIONS IN GROUNDWATER

Exposure concentrations in ground water canbe based on monitoring data alone or on acombination of monitoring and modeling. Insome cases, the exposure assessor may favor theuse of monitoring data over the use of complexmodels to develop exposure concentrations. It ismost appropriate to use ground-water samplingdata as estimates of exposure concentrations whenthe sampling points correspond to exposurepoints, such as samples taken from a drinkingwater tap. However, samples taken directly froma domestic well or drinking water tap should beinterpreted cautiously. For example, where thewater is acidic, inorganic chemicals such as leador copper may leach from the distribution system.Organic chemicals such as phthalates may migrateinto water from plastic piping. Therefore,interpretations of these data should consider thetype and operation of the pumping, storage, anddistribution system involved.

Most of the time, data from monitoring wellswill be used to estimate chemical concentrationsat the exposure point. Several issues should beconsidered when using monitoring well data toestimate these concentrations. First, determine ifthe aquifer has sufficient production capacity and

21

is of sufficient quality to support drinking wateror other uses. If so, it generally should beassumed that water could be drawn from anywherein the aquifer, regardless of the location ofexisting wells relative to the contaminant plume.In a few situations, however, it may not bereasonable to assume that water will be drawnfrom directly beneath a specific source (e.g., awaste management unit such as a landfill) in thefuture. In these cases, it should be assumed thatwater could be drawn from directly adjacent to thesource. Selection of the location(s) used toevaluate future ground-water exposures should bemade in consultation with the RPM. Second,compare the construction of wells (e.g., drinkingwater wells) in the area with the construction ofthe monitoring wells. For example, drinking waterwells may draw water from more than one aquifer,whereas individual monitoring wells are usuallyscreened in a specific aquifer. In some cases itmay be appropriate to separate data from twoaquifers that have very limited hydraulicconnection if drinking water wells in the areadraw) water from only one of them. Consult ahydrogeologist for assistance in the aboveconsiderations.

Another issue to consider is filtration ofwater samples. While filtration of ground-watersamples provides useful information forunderstanding chemical transport within an aquifer(see Section 4.5.3 for more details), the use offiltered samples for estimating exposure is verycontroversial because these data mayunderestimate chemical concentrations in waterfrom an unfiltered tap. Therefore, data fromunfiltered samples should be used to estimateexposure concentrations. Consult with the RPMbefore using data from filtered samples.

Ground-water monitoring data are often oflimited use for evaluating long-term exposureconcentrations because they are generallyrepresentative of current site conditions and notlong-term trends. Therefore, ground-water modelsmay be needed to estimate exposureconcentrations. Monitoring data should be usedwhen possible to calibrate the models.

Estimating exposure concentrations in groundwater using models can be a complex task becauseof the many physical and chemical processes thatmay affect transport and transformation in ground

water. Among the important mechanisms thatshould be considered when estimating exposureconcentrations in ground water are leaching fromthe surface, advection (including infiltration, flowthrough the unsaturated zone, and flow withground water), dispersion, sorption (includingadsorption, resorption, and ion exchange), andtransformation (including biological degradation,hydrolysis, oxidation, reduction, complexation,dissolution, and precipitation). Anotherconsideration is that not all chemicals may bedissolved in water, but may be present instead innonaqueous phases that float on top of groundwater or sink to the bottom of the aquifer.

The proper selection and application of soiland ground-water models requires a thoroughunderstanding of the physical, chemical, andhydrogeologic characteristics of the site. SEAM(EPA 1988b) provides a discussion of the factorscontrolling soil and ground-water contaminantmigration as well as descriptions of various soiland ground-water models. For more in-depthguidance on the selection and application ofappropriate ground-water models, consultSelection Criteria for Mathematical Models Used inExposure Assessments Ground-water Models (EPA1988c). As with all modeling, the assessor shouldcarefully evaluate the applicability of the model tothe site being evaluated, and should consult witha hydrogeologist as necessary.

If ground-water modeling is not used, currentconcentrations can be used to represent futureconcentrations in ground water assuming steady-state conditions. This assumption should be notedin the exposure assessment Chapter and in theuncertainties and conclusions’assessment.

6.5.3 ESTIMATE EXPOSURECONCENTRATIONS IN

of the risk

SOIL

Estimates of current exposure concentrationsin soil can be based directly on summarizedmonitoring data if it is assumed thatconcentrations remain constant over time. Suchan assumption may not be appropriate for somechemicals and some sites where leaching,volatilization, photolysis, biodegradation, winderosion, and surface runoff will reduce chemicalconcentrations over time. Soil monitoring dataand site conditions should be carefully screened to

28

identify situations where source depletion is likelyto be important. SEAM (EPA 1988b) givessteady-state equations for estimating many of theseprocesses. However, incorporating these processesinto the calculation of exposure concentrations forsoil involves considerable effort. If a modelingapproach is not adopted in these situations,assume a constant concentration over time andbase exposure concentrations on monitoring data.This assumption should be clearly documented,

In evaluating monitoring data for theassessment of soil contact exposures, the spatialdistribution of the data is a critical factor. Thespatial distribution of soil contamination can beused as a basis for estimating the averageconcentrations contacted over time if it is assumedthat contact with soil is spatially random (i.e., ifcontact with soil in all areas of the site is equallyprobable). Data from random sampling programsor samples from evenly spaced grid networksgenerally can be considered as representative ofconcentrations across the site. At many siteshowever, sampling programs are designed tocharacterize only obviously contaminated soils orhot spot areas. Care must be taken in evaluatingsuch data sets for estimating exposureconcentrations. Samples from areas where directcontact is not realistic (such as where a steepslope or thick vegetation prevents current access)should not be considered when estimating currentexposure concentrations for direct contactpathways. Similarly, the depth of the sampleshould be considered, surface soil samples shouldbe evaluated separately from subsurface samplesif direct contact with surface soil or inhalation ofwind blown dust are potential exposure pathwaysat the site.

In some cases, contamination may beunevenly distributed across a site, resulting in hotspots (areas of high contamination relative toother areas of the site). If a hot spot is locatednear an area which, because of site or populationcharacteristics, is visited or used more frequently,exposure to the hot spot should be assessedseparately. The area over which the activity isexpected to occur should be considered whenaveraging the monitoring data for a hot spot. Forexample, averaging soil data over an area the sizeof a residential backyard (e.g., an eighth of anacre) may be most appropriate for evaluatingresidential soil pathways.

6.5.4 ESTIMATE EXPOSURECONCENTRATIONS IN AIR

There are three general approaches toestimating exposure concentrations in air: (1)ambient air monitoring, (2) emissionmeasurements coupled with dispersion modeling,and (3) emission modeling coupled with dispersionmodeling. Whichever approach is used, theresulting exposure concentrations should be asrepresentative as possible of the specific exposurepathways being evaluated. If long-term exposuresare being evaluated, the exposure concentrationsshould be representative of long-term averages.If short-term exposures are of interest, measuredor modeled peak concentrations may be mostrepresentative.

If monitoring data have been collected at asite, their adequacy for use in a risk assessmentshould be evaluated by considering howappropriate they are for the exposures beingaddressed. Volume II of the NTGS (EPA 1989b)provides guidance for measuring emissions andshould be consulted when evaluating theappropriateness of emission data. See Chapter 4(Section 4.5.5) for factors to consider whenevaluating the appropriateness of ambient airmonitoring data. As long as there are nosignificant analytical problems affecting airsampling data, background levels are notsignificantly higher than potential site-relatedlevels, and site-related levels are not below theinstrument detection limit, air monitoring data canbe used to derive exposure concentrations. Therestill will be uncertainties inherent in using thesedata because they usually are not representativeof actual long-term average air concentrations.This may be because there were only a few samplecollection periods, samples were collected duringonly one type of meteorological or climaticcondition, or because the source of the chemicalswill change over time. These uncertainties shouldbe mentioned in the risk assessment.

In the absence of monitoring data, exposureconcentrations often can be estimated usingmodels. Two kinds of models are used toestimate air concentrations: emission models thatpredict the rate at which chemicals may bereleased into the air from a source, and dispersionmodels that predict associated concentrations inair at potential receptor points.

29

Outdoor air modeling. Emissions may occuras a result of the volatilization of chemicals fromcontaminated media or as a result of thesuspension of onsite soils. Models that predictemission rates for volatile chemicals or dustrequire numerous input parameters, many ofwhich are site-specific. For volatile chemicals,emission models for surface water and soil areavailable in SEAM (EPA 1988b). Volume IV ofthe NTGS (EPA 1989c) also provides guidance forevaluating volatile emissions at Superfund sites.Emissions due to suspension of soils may resultfrom wind erosion of exposed soil particles andfrom vehicular disturbances of the soil. Topredict soil or dust emissions, EPA’s fugitive dustmodels provided in AP42 (EPA 1985b) or modelsdescribed in SEAM (1988b) may be used.Volume IV of the NTGS (EPA 1989c) also willbe useful in evaluating fugitive dust emissions atSuperfund sites. Be sure to critically review allmodels before use to determine their applicabilityto the situation and site being evaluated. Ifnecessary, consult with air modelers in EPAregional offices, the Exposure Assessment Groupin EPA headquarters or the Source ReceptorAnalysis Branch in OAQPS.

After emissions have been estimated ormeasured, air dispersion models can be applied toestimate air concentrations at receptor points. Inchoosing a dispersion model, factors that must beconsidered include the type of source and thelocation of the receptor relative to the source.For area or point sources, EPA’s Industrial SourceComplex model (EPA 1987a) or the simpleGaussian dispersion models discussed in SEAM(EPA 1988b) can provide air concentrationsaround the source Other models can be foundin Volume IV of the NTGS (EPA 1989c). TheSource Receptor Analysis Branch of OAQPS alsocan be contacted for assistance. Again, criticallyreview all models for their applicability.

Indoor air modeling. Indoor emissions mayoccur as a result of transport of outdoor-generateddust or vapors indoors, or as a result ofvolatilization of chemicals indoors during use ofcontaminated water (e.g., during showering,cooking, washing). Few models are available forestimating indoor air concentrations from outsidesources. For dust transport indoors, it cangenerally be assumed that indoor concentrationsare less than those outdoors. For vapor transport

indoors, concentrations indoors and outdoors canbe assumed to be equivalent in most cases.However, at sites where subsurface soil gas orground-water seepage are entering indoors, vaporconcentrations inside could exceed those outdoors.Vapor concentrations resulting from indoor use ofwater may be greater than those outdoors,depending on the emission source characteristics,dispersion indoors, and indoor-outdoor sirexchange rates. Use models discussed in theExposure Assessment Methods Handbook (EPA1989f) to evaluate volatilization of chemicals fromindoor use of water.

6.5.5 ESTIMATE EXPOSURECONCENTRATIONS IN SURFACEWATER

Data from surface water sampling andanalysis may be used alone or in conjunction withfate and transport models to estimate exposureconcentrations. Where the sampling pointscorrespond to exposure points, such as atlocations where fishing or recreational activitiestake place, or at the intake to a drinking watersupply, the monitoring data can be used alone toestimate exposure concentrations. However, thedata must be carefully screened. The complexityof surface water processes may lead to certainlimitations in monitoring data. Among these arethe following.

Temporal representativeness. Surfacewater bodies are subject to seasonalchanges in flow, temperature, and depththat may significantly affect the fate andtransport of contaminants. Releases tosurface water bodies often depend onstorm conditions to produce surfacerunoff and soil erosion. Lakes aresubject to seasonal stratification andchanges in biological activity. Unless thesurface water monitoring program hasbeen designed to account for thesephenomena, the data may not representlong-term average concentrations orshort-term concentrations that may occurafter storm events.

Spatial representativeness. Considerablevariation in concentration can occur withrespect to depth and lateral location insurface water bodies. sample locations

30

should be examined relative to surfacewater mixing zones. Concentrationswithin the mixing zone may besignificantly higher than at downstreampoints where complete mixing has takenplace.

Quantitation limit limitations. Wherelarge surface water bodies are involved,contaminants that enter as a result ofground-water discharge or runoff fromrelatively small areas may be significantlydiluted. Although standard analyticalmethods may not be able to detectchemicals at these levels, the toxic effectsof the chemicals and/or their potentialto bioaccumulate may neverthelessrequire that such concentrations beassessed.

Contributions from other sources.Surface water bodies are normally subjectto contamination from many sources(e.g., pesticide runoff, stormwater,wastewater discharges, acid minedrainage). Many of the chemicalsassociated with these sources may bedifficult to distinguish from site-relatedchemicals. In many cases backgroundsamples will be useful in assessing site-related contaminants from othercontaminants (see Section 4.4).However, there may be other caseswhere a release and transport model maybe required to make the distinction.

Many analytical and numerical models areavailable to estimate the release of contaminantsto surface water and to predict the fate ofcontaminants once released. The models rangefrom simple mass balance relationships tonumerical codes that contain terms for chemicaland biological reactions and interactions withsediments. In general, the level of informationcollected during the RI will tend to limit the useof the more complex models.

There are several documents that can beconsulted when selecting models to estimatesurface water exposure concentrations, includingSEAM (EPA 1988b), the Exposure AssessmentMethods Handbook (EPA 1989f), and Selection

Criteria for Mathematical Models Used in ExposureAssessments: Surface Water Mode1s (EPA 1987b).SEAM lists equations for surface water runoff andsoil erosion and presents the basic mass balancerelationships for estimating the effects of dilution.A list of available numerical codes for morecomplex modeling also is provided. The selectioncriteria document (EPA 1987b) provides a morein-depth discussion of numerical codes and othermodels. In addition, it provides guidelines andprocedures for evaluating the appropriate level ofcomplexity required for various applications. Thedocument lists criteria to consider when selectinga surface water model, including (1) type of waterbody, (2) presence of steady-state or transientconditions, (3) point versus non-point sources ofcontamination, (4) whether 1, 2, or 3 spatialdimensions should be considered, (5) the degreeof mixing, (6) sediment interactions, and (7)chemical processes. Each of the referenceddocuments should be consulted prior to anysurface water modeling.

6.5.6 ESTIMATE EXPOSURECONCENTRATIONS IN SEDIMENTS

In general, use sediment monitoring data toestimate exposure concentrations. Sedimentmonitoring data can be expected to provide bettertemporal representativeness than surface waterconcentrations. This will especially be true in thecase of contaminants such as PCBs, PAHs, andsome inorganic chemicals, which are likely toremain bound to the sediments. When usingmonitoring data to represent exposureconcentrations for direct contact exposures, datafrom surficial, near-shore sediments should beused.

If modeling is needed to estimate sedimentexposure concentrations, consult SEAM (EPA1988b). SEAM treats surface water and sedimenttogether for the purpose of listing availablemodels for the release and transport ofcontaminants. Models for soil erosion releasesare equally applicable for estimating exposureconcentrations for surface water and sediment.Many of the numerical models listed in SEAMand the surface water selection criteria document(EPA 1987b) contain sections devoted to sedimentfate and transport.

31

6.5.7 ESTIMATE CHEMICALCONCENTRATIONS IN FOOD

Fish and shellfish. Chemical concentrationsin fish and shellfish may be measured orestimated. Site-specific measured values arepreferable to estimated values, but before usingsuch values, evaluate the sampling plan todetermine if it was adequate to characterize thepopulation and species of concern (see Section4.5.6 for some sampling considerations). Alsoexamine analytical procedures to determine if thequantitation limits were low enough to detect thelowest concentration potentially harmful tohumans. Inadequate sampling or high levels ofquantitation may lead to erroneous conclusions.

In the absence of adequate tissuemeasurements, first consider whether the chemicalbioconcentrates (i.e., is taken up from water) orbioaccumulates (i.e., is taken up from food,sediment, and water). For example, low molecularweight volatile organic chemicals do notbioaccumulate in aquatic organisms to a greatextent. Other chemicals accumulate in somespecies but not in others. For example, PAHstend to accumulate in mollusk species but not infish, which rapidly metabolize the chemicals. Forthose chemicals that bioconcentrate in aquaticspecies of concern, use the organism/waterpartition coefficient (i.e., bioconcentration factor,or BCF) approach to estimate steady-stateconcentrations. BCFs that estimate concentrationsin edible tissue (muscle) are generally moreappropriate for assessing human exposures fromfish or shellfish ingestion than those that estimateconcentrations in the whole body, although this isnot true for all aquatic species or applicable to allhuman populations consuming fish or shellfish.When data from multiple experiments areavailable, select the BCF from a test that used aspecies most similar to the species of concern atthe site, and multiply the BCF directly by thedissolved chemical concentration in water toobtain estimates of tissue concentrations. Beaware that the study from which the BCF isobtained should reflect a steady state or

equilibrium condition, generally achieved overlong-term exposures (although some chemicalsmay reach steady state rapidly in certain species).For some chemicals, BCFs may overestimate tissuelevels in fish that may be exposed only for a shortperiod of time.

When no BCF is available, estimate the BCFwith a regression equation based on octanol/waterpartition coefficients (Kow). Several equations areavailable in the literature. Those developed forchemicals with structural similarities to thechemical of concern should be used in preferenceto general equations because of better statisticalcorrelations.

The regression equation approach toestimating BCFs can overestimate orunderestimate concentrations in fish tissuedepending upon the chemical of concern and thestudies used to develop the regression equations.For example, high molecular weight PAHs (suchas benz(a)pyrene) with high Kow values lead tothe prediction of high fish tissue residues.However, PAHs are rapidly metabolized in theliver, and do not appear to accumulatesignificantly in fish. Regression equations usingK ow cannot take into account suchpharmacokinetics, and thus may overestimatebioconcentration. On the other hand, studies usedto develop regression equations which were notrepresentative of steady-state conditions will tendto underestimate BCFs.

Typical methods for estimating fish tissueconcentrations are based on dissolved chemicalconcentrations in water. While chemicals presentin sediment and biota may also bioaccumulate infish, there are only limited data available toestimate contributions to fish from these sources.However, chemicals that readily adsorb tosediments, such as PCBs, can be present in surfacewater at concentrations below detection limits andstill significantly bioaccumulate. Some models areavailable to assess the contribution of chemicalconcentrations in sediment to chemicalconcentrations in aquatic biota. CEAM (ERLAthens) may be of assistance in choosing andapplying an appropriate model.

Plants. Site-related chemicals may be presentin plants as a result of direct deposition ontoplant surfaces, uptake from the soil, and uptakefrom the air. When possible, samples of plants orplant products should be used to estimateexposure concentrations. In the absence ofmonitoring data, several modeling approaches areavailable for estimating exposure concentrations inplants. Use of these models, however, can

32

introduce substantial uncertainty into an exposureassessment.

If deposition onto plants is the source of thechemical, air deposition modeling can be used inconjunction with plant interception fractions toestimate uptake. The plant interception fractioncan be estimated by methods published in theliterature or can be developed for a specific cropby considering crop yield and the area of the plantavailable for deposition.

If soil contamination is the source of thechemical, calculate the concentration in plants bymultiplying soil to plant partition coefficients bysoil concentrations. Use the open literature orcomputerized data bases to obtain thesecoefficients from field, microcosm, or laboratoryexperiments that are applicable to the type ofvegetation or crop of concern (see EPA 1985csludge documents for some). In the absence ofmore specific information, use general BCFspublished in the literature that are not crop-specific (see Baes et al. 1984 for some). Whenusing these parameters, it is important to considerthat many site-specific factors affect the extent ofuptake. These factors include pH, the amount oforganic material present in soil, and the presenceof other chemicals.

When literature values are not available,consider equations published in the literature forestimating uptake into the whole plant, into theroot, and translocation from the root into aboveground parts (see Calamari et al. 1987). Suchmethods require physical/chemical parameters suchas Kow or molecular weight and were developedusing a limited data base. Scientific judgmentmust always be applied in the development andapplication of any partition coefficient, andcaution must be applied in using these values inrisk assessment.

Terrestrial animals, Use tissue monitoringdata when available and appropriate for estimatinghuman exposure to chemicals in the terrestrialfood chain. In the absence of tissue monitoring

data, use transfer coefficients together with thetotal chemical mass ingested by an animal per dayto estimate contaminant concentrations in meat,eggs, or milk. Data to support modeling ofuptake by terrestrial animals generally are notavailable for birds, but are available for some

mammalian species. Terrestrial mammals such ascattle are simultaneously exposed to chemicalsfrom several sources such as water, soil, cornsilage, pasture grass, and hay. Cattle ingestvarying amounts of these sources per day, each ofwhich will contain a different contaminantconcentration. Because all sources can beimportant with regard to total body burden, anapproach based upon the daily mass of chemicalingested per day is recommended because it canbe applied to input from many sources.

Obtain transfer coefficients from theliterature (see Ng et al. 1977, 1979, 1982 Baes etal. 1984 for some), or calculate them directly fromfeeding studies (see Jensen et al. 1981; Jensen andHummel 1982; Fries et al. 1973; Van Bruwaeneet al. 1984). In the absence of this information,use regression equations in the literature for theestimation of transfer coefficients (see Travis andArms 1988). It is important to be aware thatregression equations that use feeding study resultsfrom short-term exposures may underestimatemeat or milk concentrations. In addition,regression equations which rely on Kow values mayoverestimate exposures for chemicals such asbenz(a)pyrene that are rapidly metabolized.Information on the amount of feed, soil and wateringested by dairy and beef cows is available in theliterature and should be combined with chemicalconcentrations in these media to estimate a dailydose to the animal.

6.5.8 SUMMARIZE EXPOSURECONCENTRATIONS FOR EACHPATHWAY

Summarize the exposure concentrationsderived for each pathway. Exhibit 6-10 presentsa sample format.

6.6 QUANTIFICATION OFEXPOSURE: ESTIMATIONOF CHEMICAL INTAKE

This section describes the methodology forcalculating chemical-specific intakes for thepopulations and exposure pathways selected forquantitative evaluation. The general equation forestimating intake was shown in Exhibit 6-9.Remember that the intakes calculated in this step

33

EXHIBIT 6-10

EXAMPLE OF TABLE FORMAT FOR SUMMARIZINGEXPOSURE CONCENTRATIONS

34

are expressed as the amount of chemical at theexchange boundary (e.g., skin, lungs, gut) andavailable for absorption. Intake, therefore, is notequivalent to absorbed dose, which is the amountof a chemical absorbed into the blood stream.

The sections that follow give standardequations for estimating human intakes for allpossible exposure routes at a site. Values forequation variables are presented for use inevaluating residential exposures. Considerationsfor deriving pathway-specific variable values forpopulations other than residential (i.e.,commercial/industrial or recreational) also aregiven. In general, both upper-bound (e.g., 95thpercentile or maximum values) and average (meanor median) values are presented. These valuescan be used to calculate the RME or to evaluateuncertainty. A general discussion of whichvariable values should be used to calculate theRME was provided in Section 6.4.1; more specificguidance follows. A discussion of the uncertaintyanalysis is presented in Section 6.8.

The information presented below is organizedby exposure medium and exposure route.

6.6.1 CALCULATE GROUND-WATER ANDSURFACE WATER INTAKES

Individuals may be exposed to chemicals ofpotential concern in ground water and surfacewater by the following routes:

(1) ingestion of ground water or surfacewater used as drinking water;

(2) incidental ingestion of surface waterwhile swimming and

(3) dermal contact with ground water orsurface water.

Inhalation exposures to chemicals that havevolatilized from surface or ground water arecovered in Section 6.6.3.

Intake from drinking water. Calculateresidential intakes from ingestion of ground wateror surface water used as drinking water, using theequation and variable values presented. in Exhibit6-11. As discussed in section 6.5.3, chemicalconcentration in water (CW) should be based on

data from unfiltered samples. Develop pathway-specific variable values as necessary. Ingestionrates (IR) could be lower for residents who spenda portion of their day outside the home (e.g., atwork). Also, exposure frequency (EF) may varywith land use. Recreational users and workersgenerally would be exposed less frequently thanresidents.

Intake from ingestion of surface water whileswimming. Calculate intakes from incidentalingestion of surface water while swimming. Usethe equation and variable values presented inExhibit 6-12. Chemical concentration in water(CW) should represent unfiltered concentrations.Incidental ingestion rates (IR) while swimminghave not been found in the available literature.SEAM (EPA 1988b) recommends using anincidental ingestion rate of 50 ml/hour ofswimming. Exposure duration (ED) will generallybe less for recreational users of a surface watercompared to residents living near the surfacewater. Workers are not expected to be exposedvia this pathway.

Intake from dermal contact. Calculateintakes from dermal contact with water whileswimming, wading, etc., or during household use(e.g., bathing).

Use the equation and variable valuespresented in Exhibit 6-13. In this case, thecalculated exposure is actuallv the absorbed dose,not the amount of chemical that comes in contactwith the skin (i.e., intake). This is becausepermeability constants (PC) reflect the movementof the chemical across the skin to the stratumcorneum and into the bloodstream. Be sure torecord this information in the summary ofexposure assessment results so that the calculatedintake is compared to an appropriate toxicityreference value in the risk characterizationchapter. Note that PC are based on anequilibrium partitioning and likely result in anover-estimation of absorbed dose over shortexposure periods (e.g., < 1 hr). The openliterature should be consulted for chemical-specificPC values. The values in SEAM (EPA 1988b) arecurrently being reviewed and should not be usedat this time. If chemical-specific PC values arenot available, the permeability of water can beused to derive a default value. (See Blank et al.[1984] for some values [e.g., 8.4x10-4cm/hr].) Note

35

EXHIBIT 6-11

RESIDENTIAL EXPOSURE: INGESTION OFCHEMICALS IN DRINKING WATER a

(AND BEVERAGES MADE USING DRINKING WATER)

aSee Section 6.4.1 and 6.6.1 for a discussion of which variable values should be used to calculate thereasonable maximum exposure. In general, combine 95th or 90th percentile values for contact rateand exposure frequency and duration variables.

36

EXHIBIT 6-12

RESIDENTIAL EXPOSURE:INGESTION OF CHEMICALS IN SURFACE WATER

WHILE SWIMMING a

a See Section 6.4.1 and 6.6.1 for a discussion of which variable values should be used to calculate thereasonable maximum exposure. In general, combine 95th or 90th percentile values for contact rateand exposure frequency and duration variables.

37

EXHIBIT 6-13

RESIDENTIAL EXPOSURE:DERMAL CONTACT WITH CHEMICALS IN WATERa

a See Section 6.4.1 and 6.6.1 for a discussion of which variable values should be used to calculate thereasonable maximum exposure. In general, combine 95th or 90th percentile values for contact rate andexposure frequency and duration variables. Use 50th percentile values for SA; see text for rationale.

(continued)

38


RESIDENTIAL EXPOSUREDERMAL CONTACT WITH CHEMICALS IN WATERa

NOTE: Values for children were calculated using age-specific body surface areas and the averagepercentage of total body surface area represented by particular body parts in children,presented in EPA 1985a. Values for adults presented in EPA 1989d or calculated frominformation presented in EPA 1985a. Information on surface area of other body parts (e.g.,head, feet) and for female children and adults also is presented in EPA 1985a, 1989dDifferences in body part surface areas between sexes is negligible.

PC: Consult open literature for values [Note that use of PC values results inan estimate of absorbed dose.]

ET: Pathway-specific value (consider local activity patterns if informationis available)

2.6 hrs/day (national average for swimming; USDOI inEPA 1988b, EPA 1989d)

EF: Pathway-specific value (should consider local climatic conditions[e.g., number of days above a given temperature] and age of potentiallyexposed population)

7 days/year (national average for swimming USDOI in EPA 1988b,EPA 1989d)

ED: 70 years (lifetime; by convention)30 years (national upper-bound time (90th percentile) at one residence;

EPA 1989d)9 years (national median time (50th percentile) at one residence;

EPA 1989d)

CF: 1 liter/1000 cm3

BW: 70 kg (adult, average; EPA 1989d)Age-specific values (EPA 1985a, 1989d)

AT: Pathway-specific period of exposure for noncarcinogenic effects(i.e., ED x 365 days/year), and 70 year lifetime for carcinogenic effects(i.e., 70 years x 365 days/year).

a See Section 6.4.1 and 6.6.1 for a discussion of which variable values should be used to calculatethe reasonable maximum exposure. In general, combine 95th or 90th percentile values forcontact rate and exposure frequency and duration variables.

39

that this approach may underestimate dermalpermeability for some organic chemicals.

To calculate the reasonable maximumexposure for this pathway, 50th percentile values,instead of 95th percentile values, are used for thearea of exposed skin (SA). This is becausesurface area and body weight are stronglycorrelated and 50th percentile values are mostrepresentative of the surface area of individuals ofaverage weight (e.g., 70 kg) which is assumed forthis and all other exposure pathways. Estimatesof exposure for this pathway are still regarded asconservative because generally conservativeassumptions are used to estimate dermalabsorption (PC) and exposure frequency andduration.

Consider pathway-specific variations for theintake variables. SA will vary with activity andthe extent of clothing worn. For example, agreater skin surface area would be in contact withwater during bathing or swimming than whenwading. Worker exposure via this pathway willdepend on the type of work performed at the site,protective clothing worn, and the extent of wateruse and contact.

6.6.2 CALCULATE SOIL SEDIMENT, ORDUST INTAKES

Individuals may be exposed to chemicals ofpotential concern in soil, sediment, or dust by thefollowing routes:

(1) incidental ingestion; and(2) dermal contact.

Inhalation exposures to airborne soil or dust arediscussed in Section 6.6.3.

Incidental ingestion. Calculate intakes fromincidental ingestion of chemicals in soil byresidents using the equation and variable valuespresented in Exhibit 6-14. Consider populationcharacteristics that might influence variable values.Exposure duration (ED) may be less for workersand recreational users.

The value suggested for ingestion rate (IR)for children 6 years old and younger are basedprimarily on fecal tracer studies and account foringestion of indoor dust as well as outdoor soil.

These values should be viewed as representativeof long-term average daily ingestion rates forchildren and should be used in conjunction withan exposure frequency of 365 days/year. A termcan be used to account for the fraction of soil ordust contacted that is presumed to becontaminated (FI). In some cases, concentrationsin indoor dust can be equal to those in outdoorsoil. Conceivably, in these cases, FI could beequal to 1.0.

For ingestion of chemicals in sediment, usethe same equation as that used for ingestion ofsoil. Unless more pathway-specific values can befound in the open literature, use as defaultvariable values the same values as those used foringestion of soil. In most instances, contact andingestion of sediments is not a relevant pathwayfor industrial/commercial land use (a notableexception to this could be workers repairingdocks).

Dermal contact. Calculate exposure fromdermal contact with chemicals in soil by residentsusing the equation and variable values presentedin Exhibit 6-15. As was the case with exposure tochemicals in water, calculation of exposure for thispathway results in an estimate of the absorbeddose, not the amount of chemical in contact withthe skin (i.e., intake). Absorption factors (ABS)are used to reflect the resorption of the chemicalfrom soil and the absorption of the chemicalacross the skin and into the blood stream.Consult the open literature for information onchemical-specific absorption factors. In theabsence of chemical-specific information, useconservative assumptions to estimate ABS.

Again, as with dermal exposure to water, 50thpercentile body surface area (SA) values are usedto estimate contact rates. These values are usedalong with average body weight because of thestrong correlation between surface area and bodyweight. Contact rates may vary with time of yearand may be greater for individuals contacting soilsin the warmer months of the year when lessclothing is worn (and hence, more skin is availablefor contact). Adherence factors (AF) are availablefor few soil types and body parts. The literatureshould be reviewed to derive AF values for othersoil types and other body parts. Exposurefrequency (EF) is generally determined using site-specific information and professional judgment.

40

EXHIBIT 6-14

RESIDENTIAL EXPOSURE:

INGESTION OF CHEMICALS IN SOILa

Equation:

Where:

CS = Chemical Concentration in Soil (mg/kg)IR = Ingestion Rate (mg soil/day)CF = Conversion Factor (10 -6 kg/mg)FI = Fraction Ingested from Contaminated Source (unitless)EF = Exposure Frequency (days/years)ED = Exposure Duration (years)BW = Body Weight (kg)AT = Averaging Time (period over which exposure is averaged -- days)

Variable Values:

CS: Site-specific measured value

IR: 200 mg/day (children, 1 through 6 years old; EPA 1989g)100 mg/day (age groups greater than 6 years old; EPA 1989g)

NOTE: IR values are default values and could change basedon site-specific or other information. Research is currently ongoingto better define ingestion rates. IR values do not apply to individualswith abnormally high soil ingestion rates (i.e., pica).

CF: 10 -6 kg/mg

FI: Pathway-specific value (should consider contaminant location andpopulation activity patterns)

EF: 365 days/year

ED: 70 years (lifetime; by convention)30 years (national upper-bound time (90th percentile) at one

residence; EPA 1989d)9 years (national median time (50th percentile) at one residence;

EPA 1989d)

BW: 70 kg (adult, average; EPA 1989d)16 kg (children 1 through 6 years old, 50th percentile; EPA 1985a)


a See Section 6.4.1 and 6.6.2 for a discussion of which variable values should be used to calculatethe reasonable maximum exposure. In general, use 95th or 90th percentile values for contact rateand exposure frequency and duration variables.

41,

EXHIBIT 6-15

RESIDENTIAL EXPOSURE:DERMAL CONTACT WITH CHEMICALS IN SOILa

a See Section 6.4.1 and 6.6.2 for a discussion of which variable values should be used to calculate the reason-able maximum exposure. In general, combine 95th or 90th percentile values for contact rate and exposurefrequency variables. Use 50th percentile values for SA; see text for rationale.

(continued)

42


RESIDENTIAL EXPOSURE:DERMAL CONTACT WITH CHEMICALS IN SOILa

NOTE (continued): Information on surface area of other body parts (e.g., head, feet) and forfemalechilden and adults also is presented in EPA 1985a, 1989d. Differences in body part surfaceareas between sexes is negligible.

AF: 1.45 mg/cm 2 — commercial potting soil (for hands; EPA 1989d, EPA1988b)

2.77 mg/cm2 -- kaolin clay (for hands; EPA 1989d, EPA 1988b)

ABS: Chemical-specific value (this value accounts for resorption ofchemical from the soil matrix and absorption of chemical acrossthe skin; generally, information to support a determination of ABS islimited — see text)

EF: Pathway-specific value (should consider local weather conditions[e.g.,number of rain, snow and host-free days] and age of potentiallyexposed population)



EPA 1989d)



a See Section 6.4.1 and 6.6.2 for a discussion of which variable values should be used to calculate thereasonable maximum exposure. In general, combine 95th or 90th percentile values for contact rateand exposure frequency and duration variables.

43

“Best guess” values for children potentially usefulin risk assessments are 3 times/week for fall andspring days (> 32°F) and 5 times/week for summer

days when children are not attending school. Asdiscussed previously, in some cases, concentrationsin indoor dust could be equal to that in outdoorenvironments. Therefore, at some sites, EF couldbe 365 days/year. Worker and recreational usercontact rates are dependent on the type of activityat the site. Exposure duration (ED) and exposurefrequency (EF) may be lower for workers andrecreational users.

For dermal contact with sediment or dust,use the same equation as that for dermal contactwith soil. As default values, also use the variablevalues given for dermal contact with soil unlessmore pathway-specific values can be found in theopen literature. Adherence factors for somesediments (particularly sandy sediments) are likelyto be much less than for soils because contactwith water may wash the sediment off the skin.Exposure frequency for sediments also is probablylower than that for soils at many sites.

6.6.3 CALCULATE AIR INTAKES

Individuals may be exposed to chemicals ofpotential concern in air by inhalation of chemicalsin the vapor phase or adsorbed to particulate.Dermal absorption of vapor phase chemicals isconsidered to be lower than inhalation intakes inmany instances and generally is not considered inSuperfund exposure assessments.

As with other pathways, the inhalationintakes are expressed in units of mg/kg-day. Thecombination of inhalation intakes with inhalationRfDs (expressed in concentration units of mg/m3)will be discussed in Chapters 7 and 8.

Inhalation of vapor-phase chemicals.Calculate intakes from inhalation of vapor phasechemicals using the equation and variable valuespresented in Exhibit 6-16. Consider variationswith land use. Exposure time (ET) will generallybe less for workers and recreational users. Forexposure times less than 24 hours per day, anhourly inhalation rate (IR) based on activity, age,and sex should be used instead of the daily IRvalues. Exposure duration (ED) may also be lessfor workers and recreational users.

Inhalation of particulate phase chemicals.Calculate intakes from inhalation of particulatephase chemicals by modifying the equations andvariable values presented in Exhibit 6-16 forvapor-phase exposures. Derive inhalationestimates using the particulate concentration inair, the fraction of the particulate that isrespirable (i.e., particles 10 urn or less in size)and the concentration of the chemical in therespirable fraction. Note that it may be necessaryto adjust intakes of particulate phase chemicals ifthey are to be combined with toxicity values thatare based cn exposure to the chemical in thevapor phase. This adjustment is done in the riskcharacterization step.

6.6.4 CALCULATE FOOD INTAKES

Individuals may be exposed by ingestion ofchemicals of potential concern that haveaccumulated in food. The primary food items ofconcern are:

(1) fish and shellfish;

(2) vegetables and other produce; and

(3) meat, eggs, and dairy products (domesticand game species).

Ingestion of fish and shellfish. Calculateintakes from ingestion of fish and shellfish usingthe equation and variable values given in Exhibit6-17. Exposure will depend in part on theavailability of suitable fishing areas. The chemicalconcentration in fish or shellfish (CF) should bethe concentration in the edible tissues (whenavailable). The edible tissues will vary withaquatic species and with population eating habits.Residents near major commercial or recreationalfisheries or shell fisheries are likely to ingestlarger quantities of locally caught fish and shellfishthan inland residents. In most instances, workersare not likely to be exposed via this pathway,although at some sites this may be possible.

Ingestion of vegetables or other produce.Calculate intakes from ingestion of contaminatedvegetables or other produce using the equationand variable values given in Exhibit 6-18. Thispathway will be most significant for farmers andfor rural and urban residents consuminghomegrown fruits and vegetables. For

44

EXHIBIT 6-16

RESIDENTIAL EXPOSUREINHALATION OF AIRBORNE (VAPOR PHASE) CHEMICALSab

Equation:

Where:

CA = Contaminant Concentration in Air (mg/m3) IR = Inhalation Rate (m3/hour)ET = Exposure Time (hours/day)EF = Exposure Frequency (days/year)ED = Exposure Duration (years)BW = Body Weight (kg)AT = Averaging Time (period over which exposure is averaged — days)

Variable Values:

CA: Site-specific measured or modeled value

IR: 30 m3/day (adult, suggested upper bound value; EPA 1989d)20 m3/day (adult, average; EPA 1989d)Hourly rates (EPA 1989d)Age-specific values (EPA 1985a)Age, sex, and activity based values (EPA 1985a)0.6 m3/hr — showering (all age groups; EPA 1989d)

ET Pathway-specific values (dependent on duration of exposure-relatedactivities)

12 minutes — showering (90th percentile; EPA 1989d)7 minutes — showering (50th percentile; EPA 1989d)

EF: Pathway-specific value (dependent on frequency of showering or otherexposure-related activities)



EPA 1989d)

BW: 70 kg (adult, average EPA 1989d)Age-specific values (EPA 1985a, 1989d)


“See Section 6.4.1 and 6.6.3 for a discussion of which variable values should be used to calculate thereasonable maximum exposure. In general, use 95th or 90th percentile values for contact rate andexposure frequency and duration variables.

bThe equation and variable values for vapor phase exposure can be used with modification to calculateparticulate exposure. See text.

45,

Equation:

Where:

CF = Contaminant Concentration in Fish (mg/kg)IR = Ingestion Rate (kg/meal)FI = Fraction Ingested from Contaminated Source (unitless)EF = Exposure Frequency (meals/year)ED = Exposure Duration (years)BW = Body Weight (kg)AT = Averaging Time (period over which exposure is averaged — days)

Variable Values:

CF: Site-specific measured or modeled value

IR: 0.284 kg/meal (95th percentile for fin fish; Pao et al. 1982)0.113 kg/meal (50th percentile for fin fish; Pao et al. 1982)

132 g/day (95th percentile daily intakes averaged over three daysfor consumers of fin fish Pao et al. 1982)

38 g/day (50th percentile daily intake, averaged over three daysfor consumers of fin fish; Pao et al. 1982)

6.5 g/day (daily intake averaged over a year; EPA 1989d.NOTE: Daily intake values should be used in conjunction withan exposure frequency of 365 days/year.)

Specific values for age, sex, race, region and fish species areavailable (EPA 1989d, 1989h)

FI: Pathway-specific value (should consider local usage patterns)

EF: Pathway-specific value (should consider local population patternsif information is available)

48 days/year (average per capita for fish and shellfish; EPA ToleranceAssessment System in EPA 1989h)

ED: 70 years (Iifetime; by convention)30 years (national upper-bound time (90th percentile) at one residence;


EPA 1989d)



EXHIBIT 6-17

RESIDENTIAL EXPOSURE: FOOD PATHWAY --INGESTION OF CONTAMINATED FISH AND SHELLFISHa

aSee Section 6.4.1 and 6.6.4 for a discussion of which variable values should be used to calculate thereasonable maximum exposure. In general, use 95th or 90th percentile values for intake rate andexposure frequency and duration variables.

46

EXHIBIT 6-18

RESIDENTIAL EXPOSURE: FOOD PATHWAY --INGESTION OF CONTAMINATED FRUITS AND VEGETABLES a

Equation:

Where:

CF = Contaminant Concentration in Food (mg/kg)IR = Ingestion Rate (kg/meal)FI = Fraction Ingested from Contaminated Source (unitless)EF = Exposure Frequency (meals/year)ED = Exposure Duration (years)BW = Body Weight (kg)AT = Averaging Time (period over which exposure is averaged — days)

VariabIe Values:

CF: Site-specific measured value or modeled value based on soilconcentration and plant:soil accumulation factor or deposition factors

lR: Specific values for a wide variety of fruits and vegetables are available(Pao et al. 1982)

FI: Pathway-specific value (should consider location and size ofcontaminated area relative to that of residential areas, as well asanticipated usage patterns)

EF: Pathway-specific value (should consider anticipated usage patterns)

ED: 70 years (Iifetime; by convention)30 years (national upper-bound time (90th percentile) at one residence;

EPA 1989d)9 years (national median time (SOth percentile) at one residence;

EPA 1989d)



aSee Section 6.4.1 and 6.6.4 for a discussion of which variable values should be used to calculate thereasonable maximum exposure. In general, use 95th or 90th percentile values for contact rate andexposure frequency and duration variables.

47

contaminated backyard gardens, the fraction offood ingested that is contaminated (FI) can beestimated using information on the fraction offruits or vegetables consumed daily that is homegrown (HF). EPA (1989d) provides HF values forfruit (0.20, average; 0.30 worst-case) andvegetables (0.25, average; 0.40, worst-case),(Worst-case values can be used as estimates of the95th percentile value.) Pao et al. (1982) providesspecific values for a variety of fruits andvegetables.

Workers are not likely to be exposed via thispathway. Recreational users could be exposedfrom consuming wild fruits or vegetables from thesite, although such exposures are likely to benegligible.

Ingestion of meat, eggs, and dairy products.Calculate intakes from ingestion of contaminatedmeat and dairy products using the equation andvariable values given in Exhibit 6-19. Derivepathway-specific values as necessary. Ruralresidents may consume poultry as well as livestockand wild game that have been exposed tocontaminants at the site. The fraction of foodingested daily that is contaminated (FI) can beestimated for beef and dairy products usinginformation provided in EPA (1989d) on thefraction of these foods that is homegrown (HF).HF for beef is estimated to be 0.44 (average) and0.75 (worst-case). HF for dairy products isestimated to be 0.40 (average) and 0.75 (worst-case). (Worst-case values can be used as estimatesof the 95th percentile value.) Consider land-usevariations. Workers are not likely to be exposedvia this pathway. Exposure duration (ED) andexposure frequency (EF) will likely be less forrecreational users (e.g., hunters).

6.7 COMBINING CHEMICALINTAKES ACROSSPATHWAYS

As discussed previously, the RME at a sitereflects the RME for a pathway as well as theRME across pathways. A given population maybe exposed to a chemical from several exposureroutes. For example, residents may be exposed tochemicals in ground water via ingestion ofdrinking water and via inhalation of chemicals that

have volatilized from ground water during its use.They also could be exposed to chemicals in vaporsor dust that have migrated from the site, Tocalculate an exposure that is a reasonablemaximum across pathways, it may be necessary tocombine the RME for one pathway with anestimate of more typical exposure for anotherpathway (see Section 8.3.1). The average variablevalues identified in the previous sections can beused to calculate intakes for these more typicalexposures. At this point in the assessment,estimated intakes are not summed acrosspathways; this is addressed in the riskcharacterization chapter. However, the assessorshould organize the results of the previousexposure analyses (including any estimates oftypical exposure) by grouping all applicableexposure pathway for each exposed population.This organization will allow risks from appropriateexposures to be combined in the riskcharacterization chapter (see Exhibit 6-22 for asample summary format).

6.8 EVALUATINGUNCERTAINTY

The discussion of uncertainty is a veryimportant component of the exposure assessment.Based on the sources and degree of uncertaintyassociated with estimates of exposure, thedecision-maker will evaluate whether the exposureestimates are the maximum exposures that can bereasonably expected to occur. Section 8.4 providesa discussion of how the exposure uncertaintyanalysis is incorporated into the uncertaintyanalysis for the entire risk assessment.

The discussion of uncertainty in the exposureassessment chapter should be separated into twoparts. The first part is a tabular summary of thevalues used to estimate exposure and the range ofthese values. The table should include thevariables that appear in the exposure equation aswell as those used to estimate exposureconcentrations (e.g., model variables). A simpleexample of this table is shown in Exhibit 6-20.For each variable, the table should include therange of possible values, the midpoint of therange (useful values for this part are given inExhibits 6-11 through 6-19), and the value used toestimate exposure. In addition, a brief description

48

EXHIBIT 6-19

RESIDENTIAL EXPOSURE: FOOD PATHWAY --INGESTION OF CONTAMINATED MEAT, EGGS,

AND DAIRY PRODUCTS a

Equation:

Where:

CF = Contaminant Concentration in Food (mg/kg)IR = Ingestion Rate (kg/meal)FI = Fraction Ingested from Contaminated Source (unitless)EF = Exposure Frequency (meals/year)ED = Exposure Duration (years)BW = Body Weight (kg)AT = Averaging Time (period over which exposure is averaged — days)

Variable Values:

CF; Site-specific measured or modeled value. Based on soilconcentrations, plant (feed) accumulation factors, and feed-to-meator feed-to-dairy product transfer coefficients

IR: 0.28 kg/meal — beef (95th percentile Pao et al. 1982)0.112 kg/meal — beef (50th percentile Pao et al. 1982)Specific values for other meats are available (Pao et al. 1982)

0.150 kg/meal - eggs (95th percentile; Pao et al. 1982)0.064 kg/meal -- eggs (50th percentile; Pao et al. 1982)

Specific values for milk, cheese and other dairy products are available(Pao et al. 1982)

FI: Pathway-specific value (should consider location and size of contaminatedarea relative to that of residential areas, as well as anticipated usagepatterns)

EF: Pathway-specific value (should consider anticipated usage patterns)

ED: 70 years (lifetime; by convention)30 years (national upper-bound time (90tb percentile) at one residence


EPA1989d)



a See Section 6.4.1 and 6.6.4 for a discussion of which variable values should be used to calculatethe reasonabk ma.ximum exposure. In general, use 95th or 90th percentile vaIues for contact rateand exposure frequency and duration.

49i

EXHIBIT 6-20

EXAMPLE OF TABLE FORMAT FOR SUMMARIZINGVALUES USED TO ESTIMATE EXPOSURE

Variable Range Midpoint Value Used Brief Rationale

PCB concentration ND -3,500 250in soil (mg/kg) (arithmetic mean)

Chronic exposure 1,400 95th percentile upperbound(mg/kg) estimate of mean concentration

Acute exposure 3,500 Maximum detected concentration(mg/kg)

Adult soil ingestion 0 - 1 7 0 17 100 Range based on assumptionsrate (mg/d) (arithmetic mean) regarding soil adherence and

percent ingestion. Value usedis from EPA 1989g.

Exposure frequency 1 - 7 3 5 Best professional judgment.(days/wk)

Exposure duration 1 - 2 0 10 20 Best professional judgment.(years)

50

of the selection rationale should be included. Thediscussion that accompanies the table in theexposure assessment chapter should identify whichvariables have the greatest range and provideadditional justification for the use of values thatmay be less certain.

The second part of the uncertainty discussionis to summarize the major assumptions of theexposure assessment, to discuss the uncertaintyassociated with each, and to describe how thisuncertainty is expected to affect the estimate ofexposure. Sources of uncertainty that should beaddressed include 1) the monitoring data, whichmay or may not be representative of actualconditions at the site; 2) the exposure models,assumptions and input variables used to estimateexposure concentrations; and 3) the values of theintake variables used to calculate intakes. Eachof these sources should be discussed in thesummary section of the exposure assessment. Atable may be useful in summarizing thisinformation. Exhibit 6-21 presents a sampleformat.

A supplemental approach to uncertaintyanalysis is to use analytical methods (e.g., first-order uncertainty analysis) or numerical methods(e.g., Monte Carlo analysis). These methods and

their limitations are described in greater detail inSection 8.4 It is recommended that these analysesbe used only after approval of the EPA projectmanager, and then, only as a part of theuncertainty analysis (and not as a basis for thereasonable maximum exposure).

6.9 SUMMARIZING ANDPRESENTING THE EXPOSUREASSESSMENT RESULTS

At this point, the exposure assessor shouldsummarize the results of the exposure assessment.The summary information should be presented intable format and should list the estimatedchemical-specific intakes for each pathway. Thepathways should be grouped by population so thatrisks can be combined across pathways asappropriate. The summary information should befurther grouped by current and future usecategories. Within these categories, subchronicand chronic daily intakes should be summarizedseparately. Exhibit 6-22 presents a sample formatfor this summary information. In addition to thesummary table, provide sample calculations foreach pathway, to aid in the review of thecalculations.

51

EXHIBIT 6-21

EXAMPLE OF AN UNCERTAINTY TABLE FOREXPOSURE ASSESSMENT

EFFECT ON EXPOSURE a

PotentialPotential Potential MagnitudeMagnitude Magnitude for Over-for Over- for Under- or UnderEstimation Estimation Estimation

ASSUMPTION of Exposure of Exposure of Exposure

Environmental Sampling and AnalysisSufficient samples may not have Moderatebeen taken to characterize the mediabeing evaluated, especially withrespect to currently available soil data.

Systematic or random errors in thechemical analyses may yield erroneousdata.

Chemicals in fish will be at Lowequilibrium with chemicalconcentrations in water.

Use of a Gaussian dispersion modelto estimate air concentrations offsite.

Use of a box model to estimate Lowair concentrations onsite.

Use of Cowherd’s model to estimatevehicle emission factors.

The standard assumptions regardingbody weight, period exposed, lifeexpectancy, population characteristics,and lifestyle may not be representativeof any actual exposure situation.

The amount of media intake is assumed Moderateto be constant and representativeof the exposed population.

Assumption of daily lifetime Moderate toexposure for residents. High

Use of “hot spot” soil data for Moderate toupper-bound lifetime exposure High

Low

Moderate

a As a general guideline, assumptions marked as “low”, may affect estimates of exposure by less than oneorder of magnitude; assumptions marked “moderate” may affect estimates of exposure by between one andtwo orders of magnitude; and assumptions marked “high” may attect estimates of exposure by more thantwo orders of magnitude.

52

EXHIBIT 6-22

EXAMPLE OF TABLE FORMAT FOR SUMMARIZINGTHE RESULTS OF THE EXPOSURE ASSESSMENT–

CURRENT LAND USEa

Population Exposure Pathway Chemical Chronic Daily Intake (CDI) (mg/kg-day)Carcinogenic Noncarcinogenic

Effects Effects

Residents Ingestion of ground water Benzene 0.00025 — b

that has migrated from Chlordane 0.00015 0.00035the site to downgradient Phenol c— 0.1

clocal wells Cyanide — 0.0003

Nitrobenzene c— O.0001

bInhalation of chemicals Benzene 0.000013 —that have volatilized fromground water during use

Ingestion of fish Chlordane 0.00008 0.00019that have accumulated MEK — 0.005chemicals in nearby Phenol c— 0.08lake

a Similar tables should be prepared for all subchronic daily intake (SDI) estimates as well as for all CDIand SDI estimates under future land use conditions.

b CDI for noncarcinogenic effects not calculated for benzene because it doeS not have an EPA-verifi~chronic reference dose (as of the publication date of this manual).

c CDI for carcinogenic effects not calculated for chemicals not considered by EPA to be potential humancarcinogens (as of the publicatation date of this manual).

53


Baes, C.F., III, Sharp, R.D., Sjoreen, AL., and Shore, R. W. 1984. A Review and Analysis of Parameters for Assessing Transport ofEnvironmentally Released Radionuclides through Agriculture. Oak Ridge National Laboratory. Prepared for U.S. Departmentof Energy. ORNL-5786.

Blank, I.H., Moloney, J., Alfred, B.S., Simon, I., and Apt, C. 1984. The Diffusion of Water Across the Stratum Corneum as a Functionof its Water Content. J. Invest. Derm. 82188-194.

Calamari, D., Vighi, M., and Bacci, E. 1987. The Use of Terrestrial Plant Biomass as a Parameter in the Fugacity Model.Chemosphere. 16:2539-2364.

Clark, I. 1979. Practical Geostatistics. Applied Science Publishers, Ltd. London.

Environmental Protection Agency (EPA). 1985a. Development of Statistical Distributions or Ranges of Standard Factors Used inExposure Assessments. Office of Health and Environmental Assessment.

Environmental Protection Agency (EPA). 1985b. Compilation of Air Pollutant Emission Factors. Volume 1. Stationary Point andArea Sources. Fourth Edition. Office of Research and Development. Research Triangle Park, NC

Environmental Protection Agency (EPA). 1985c. Environmental Profiles and Hazard Indices for Constituents of Municipal Sludge.Office of Water. (Individual documents are available for a number of substances).

Environmental Protection Agency (EPA). 1986a. Guidelines for Exposure Assessment. 51 Federal Regi ster 34042 (September 24,1986).

Environmental Protection Agency (EPA). 1986b. Guidelines for Carcinogen Risk Assessment. 51 Federal Register 33992 (September24, 1986).

Environmental Protection Agency (EPA). 1987a. Industrial Source Complex (ISC) Dispersion Model User’s Guide. Volume I. Officeof Air Quality Planning and Standards. Research Triangle Park NC. EPA/450/4-88/002a.

Environmental Protection Agency (EPA). 1987b. Selection Criteria for Mathematical Models Used in Exposure Assessments SurfaceWater Models. Office of Health and Environmental Assessment. EPA/600/8-87/042.

Environmental Protection Agency (EPA). 1988s. Proposed Guidelines for Exposure-related Measurements. 53 Federal Register 48830(December 2, 1988).

Environmental Protection Agency (EPA). 1988b. Superfund Exposure Assessment Manual. Office of Emergency and RemedialResponse. EPA/540/l-88/001. (OSWER Directive 9285.5-1).

Environmental Protection Agency (EPA). 1988c. Selection Criteria for Mathematical Models Used in Exposure Assessment Ground-water Models. Office of Health and Environmental Assessment. EPA/600/8-88/075.

Environmental Protection Agency (EPA). 1989a. Air Superfund National Technical Guidance Series. Volume I: Application of AirPathway Analyses for Superfund Activities. Interim Final. Office of Air Quality Planning and Standards. Research Triangle Park,NC. EPA/450/l-89/001.

Environmental Protection Agency (EPA). 1989b. Air Superfund National Technical Guidance Series. Volume II: Estimation ofBaseline Air Emissions at Superfund Sites. Interim Final. Office of Air Quality Planning and Standards. Research Triangle Park,NC. EPA/450/l-89/002.

Environmental Protection Agency (EPA). 1989c. Air Superfund National Technical Guidance Series. Volume IV: Procedures forDispersion Modeling and Air Monitoring for Superfund Air Pathway Analysis. Interim Fina;. Office of Air Quality Planning andStandards. Research Triangle Park, NC. EPA/450/l-89/004.

Environmental Protection Agency (EPA). 1989d. Exposure Factors Handbook. Office of Health and Environmental Assessment.EPA/600/8-89/043.

Environmental Protection Agency (EPA). 1989e. Proposed Amendments to the Guidelines for the Health Assessment of SuspectDevelopmental Toxicants. 54 Federal Register 9386 (March 6, 1989).

54

Environmental Protection Agency (EPA). 1989f. Exposure Assessment Methods Handbook. Draft. Office of Health andEnvironmental Assessment.

Environmental Protection Agency (EPA). 1989g. Interim Final Guidance for Soil Ingestion Rates. Office of Solid Waste andEmergency Response. (OSWER Directive 9850.4).

Environmental Protection Agency (EPA). 1989h. Guidance Manual for Assessing Human Health Risks From Chemicallv ContaminatedFish and Shellfish. Office of Marine and Estuarine Protection. EPA/503/8-89/002

Fries, G.F., Marrow, G.S., and Gordon, CH. 1973. Long-term Studies of Residue Retention and Excretion by Cows Fed aPolychlorinated Biphenyl (Ardor 1254). J. Agric. Food Chem. 21:117-121.

Gilbert, R.O. 1987. Statistical Methods for Environmental Pollution Monitoring. Van Nostrand Reinhold. New York.

Jensen, D.J., Hummel, RA, Mahle, N.H., Kocher, CW., and Higgins, H.S. 1981. A Residue Study on Beef Cattle Consuming 2,3,7,8-Tetrachlorodibenzo-p-dioxin. J. Agric. Food Chem. 29:265-268.

Jensen, D.J. and Hummel, R.A. 1982 Secretion of TCDD in Milk and Cream Following the Feeding of TCDD to Lactating DairyCoWS. Bull. Env. Contain. Toxicol. 29:440-446.

Ng, Y.C, Colsher, C.S., Quinn, D.J. and Thompson, S.E. 1977. Transfer Coefficients for the Prediction of the Dose to Man Via theForage-Cow-Miik Pathwav from Radionuclides Released to the Biosphere. Lawrence Livermore National Laboratory, Univ.Califomia. Prepared for U.S. Dept. of Energy. UCRL-5139.

Ng, Y.C., Colsher, C.., and Thompson, S.E. 1979. Transfer Factors for Assessing the Dose from Radionuclides in AgriculturalProducts. Biological Implications of Radionuctides Released From Nuclear Induatries. In: Proceedings of an InternationalSymposium on Biological Implications of Radionuclides Released from Nuclear Induatries. Vienna. March 26-30, 1979. IAEA-SM-237/54. Vol. II.

Ng, Y.C, Colsher, C.., and Thompson, S.E. 1982 Transfer Coefficients for Assessing the Dose from Radionucfides in Meat and Egg.Lawrence Livermore National Laboratory NUREG/CR-2976.

Pao, E.M., Fleming, KH., Gueuther, P.M., and Mickle, SJ. 1982. Food Commonly Eaten by Individuals: Amount Per Day and PerEating Occassion. U.S. Department of Agriculture.

Schaum, J.L 1984. Risk Analysis of TCDD Cpntaminated Soil. Office of Health and Environmental Assessment, U.S. EnvironmentalProtection Agency. EPA/600/8-84/031.

Travis, CC and Arms, AD. 1988. Bioconcentration of Organics in Beef, Milk and Vegetation. Environ. Sci. Teehnol. 22:271-274.

Van Bruwaene. R.. Gerber. G.B., Kerchmann. R.. Colard. J. and Van Kerkom. J. 1984. Metabolism of 51Cr, 54Mn, 59Fe and 60Coin Lactating Dairy Cows Health Physics 46:1069-1082.

CHAPTER 7

TOXICITY ASSESSMENT

CHAPTER 7

TOXICITY ASSESSMENT

The purpose of the toxicity assessment is toweigh available evidence regarding the potentialfor particular contaminants to cause adverseeffects in exposed individuals and to provide,where possible, an estimate of the relationshipbetween the extent of exposure to a contaminantand the increased likelihood and/or severity ofadverse effects.

Toxicity assessment for contaminants foundat Superfund sites is generally accomplished intwo steps: hazard identification and dose-responseassessment. These two steps were first discussedin the National Academy of Sciences’ publicationentitled Risk Assessment in the Federal Government- Managing the Process and more recently inEPA’s Guidelines for Carcinogen Risk Assessment(NAS 1983, EPA 1986). The first step, hazardidentification, is the process of determiningwhether exposure to an agent can cause anincrease in the incidence of a particular adversehealth effect (e.g., cancer, birth defect) andwhether the adverse health effect is likely to occurin humans. Hazard identification involvescharacterizing the nature and strength of theevidence of causation. The second step, dose-response evaluation is the process ofquantitatively evaluating the toxicity informationand characterizing the relationship between thedose of the contaminant administered or receivedand the incidence. of adverse health effects in theexposed population. From this quantitative dose-response relationship, toxicity values (e.g.,reference doses and slope factors) are derived thatcan be used to estimate the incidence or potentialfor adverse effects as a function of human

exposure to the agent. These toxicity values areused in the risk characterization step to estimatethe likelihood of adverse effects occurring inhumans at different exposure levels.

Toxicity assessment is an integral part of theoverall Superfund site risk assessment. Although

toxicity information is critical to the riskassessment, the amount of new toxicologicalevaluation of primary data required to completethis step is limited in most cases. EPA hasperformed the toxicity assessment step fornumerous chemicals and has made available theresulting toxicity information and toxicity values,which have undergone extensive peer review. Atsome sites, however, there will be significant dataanalysis and interpretation issues that should beaddressed by an experienced toxicologist. Thischapter provides step-by-step guidance for locatingEPA toxicity assessments and accompanyingvalues, and advises how to determine which valuesare most appropriate when multiple values exist.Prior to this procedural discussion, background

3

information regarding EPA’s methods for toxicityassessment is provided to assist the risk assessorin understanding the basis of the toxicity valuesand the limitations of their use. The steps of thetoxicity assessment are illustrated in Exhibit 7-1.

Derivation and interpretation of toxicityvalues requires toxicological expertise and shouldnot be undertaken by those without training andexperience. Detailed guidance for deriving toxicityvalues is beyond the scope of this document. Forthose persons interested in obtaining additionalinformation about EPA’s methods for toxicityassessment, references to appropriate guidancedocuments are given throughout this chapter.

7.1 TYPES OF TOXICOLOGICALINFORMATION CONSIDEREDIN TOXICITY ASSESSMENT

This section summarizes information fromseveral EPA documents (especially EPA 1989a,f)on the basic types of data used in toxicityassessment. As part of the hazard identificationstep of the toxicity assessment, EPA gathersevidence from a variety of sources regarding thepotential for a substance to cause adverse healtheffects (carcinogenic and noncardnogenic) inhumans. These sources may include controlledepidemiologic inveatigations, clinical studies, and

experimental animal studies. Supportinginformation may be obtained from sources such asin vitro test results and comparisons of structure-activity relationships.

7.1.1 HUMAN DATA

Well-conducted epidemiologic studies thatshow a positive association between an agent anda disease are accepted as the most convincingevidence about human risk. At present, however,human data adequate to serve as the sole basis ofa dose-response assessment are available for onlya few chemicals. Humans are generally exposedin the workplace or by accident, and because thesetypes of exposures are not intentional, thecircumstances of the exposures (concentration andtime) may not be well known. Often theincidence of effects is low, the number of exposedindividuals is small, the latent period betweenexposure and disease is long, and exposures are tomixed and multiple substances. Exposedpopulations may be heterogeneous, varying in age,sex, genetic constitution, diet, occupational andhome environment, activity patterns, and othercultural factors affecting susceptibility. For thesereasons, epidemiologic data require carefulinterpretation. If adequate human studies(confirmed for validity and applicability) existthese studies are given first priority in the dose-response assessment, and animal toxicity studiesare used as supportive evidence.

4

EXHIBIT 7-1

STEPS IN TOXICITY ASSESSMENT

Step 1: Gather Toxicity Information--Qualitative and Quantitative-for Substances Being Evaluated

Step 2: Identify Exposure Periods forWhich Toxicity Values Are Necessary

Step 3: Determine Toxicity Values forNoncarcinogenic Effects

Step 4: Determine Toxicity Values forCarcinogenic Effects

Step 5: Summarize Toxicity Information

5

Human studies having inadequate exposure-response information for a quantitative assessmentare often used as supporting data. Such studiesmay establish a qualitative relationship betweenenvironmental exposures and the presence of anadverse effect in exposed human populations. Forexample, case reports of exposures resulting ineffects similar to the types of effects observed inanimals provide support for the conclusions drawnfrom the animal data.

7.1.2 ANIMAL DATA

The toxicity data base for most chemicalslacks sufficient information on toxic effects inhumans. In such cases, EPA may infer thepotential for the substance to cause an adverseeffect in humans from toxicity information drawnfrom experiments conducted on non-humanmammals, such as the rat, mouse, rabbit, guineapig, hamster, dog, or monkey. The inference thathumans and animals (mammals) are similar, onaverage, in intrinsic susceptibility to toxiccliemicals and that data from animals can in manycases be used as a surrogate for data from humansis the basic premise of modern toxicology. Thisconcept is particularly important in the regulationof toxic chemicals. There are occasions, however,in which observations in animals may be ofuncertain relevance to humans. EPA considersthe likelihood that the agent will have adverseeffects in humans to increase as similar results areobserved across sexes, strains, species, and routesof exposure in animal studies.

7.1.3 SUPPORTING DATA

Several other types of studies used to supportconclusions about the likelihood of occurrence ofadverse health effects in humans are describedbelow. At the present time, EPA considers all ofthese types of data to be supportive, notdefinitive, in assessing the potential for adversehealth effects in humans.

Metabolic and other pharmacokinetic studiesmay be used to provide insights into themechanism of action- of a particular compound.By comparing the metabolism of a compoundexhibiting a toxic effect in an animal with thecorresponding metabolism in humans, evidence forthe potential of the compound to have toxiceffects in humans may be obtained.

Studies using cell cultures or microorganismsmay be used to provide insights into a compound’spotential for biological activity. For example, testsfor point mutations, numerical and structuralchromosome aberrations, DNA damage/repair, andcell transformation may provide supportiveevidence of carcinogenicity and may giveinformation on potential mechanisms ofcarcinogenicity. It should be noted, however, thatlack of positive results in short-term tests forgenotoxicity is not considered a basis fordiscounting positive results in long-termcarcinogenicity studies in animals.

Structure-activity studies (i.e., predictions oftoxicologic activity based on analysis of chemicalstructure) are another potential source ofsupporting data. Under certain circumstances, theknown activity of one compound may be used toestimate the activity of another structurally relatedcompound for which specific data are lacking.

7.2 TOXICITY ASSESSMENT FORNONCARCINOGENIC EFFECTS

This section summarizes how the types oftoxicity information presented in Section 7.1 areconsidered in the toxicity assessment fornoncarcinogenic effects. A reference dose, orRfD, is the toxicity value used most often inevaluating noncarcinogenic effects resulting fromexposures at Superfund sites. Additionally, One-day or Ten-day Health Advisories (HAs) may beused to evaluate short-term oral exposures. Themethods EPA uses for developing RfDs and HAsare described below. Various types of RfDs areavailable depending on the exposure route (oralor inhalation), the critical effect (developmentalor other), and the length of exposure beingevaluated (chronic, subchronic, or single event).This section is intended to be a summarydescription only for additional details, refer to theappropriate guidelines and other sources listed asreferences for this chapter (especially EPA 1986b,EPA 1989b-f).

A chronic RfD is defined as an estimate(with uncertainty spanning perhaps an order ofmagnitude or greater) of a daily exposure level forthe human population, including sensitivesubpopulations, that is likely to be without an

6

appreciable risk of deleterious effects during alifetime. Chronic RfDs are specifically developedto be protective for long-term exposure to acompound. As a guideline for Superfund programrisk assessments, chronic RfDs generally should beused to evaluate the potential noncarcinogeniceffects associated with exposure periods between7 years (approximately 10 percent of a humanlifetime) and a lifetime. Many chronic RfDs havebeen reviewed and verified by an intra-AgencyRfD Workgroup and entered into the Agency’sIntegrated Risk Information System (IRIS).

Prior to the development of RfDs, noncareinogeniceffects of chronic exposures were evaluated using valuescalled acceptable daily intakes (ADIs] or acceptableintakes for chronic exposure (AICs). While ADIs andAICs are similar in concept to RfDs, RfDs have beenderived using a more strictly defined methodology andrepresent the Agency’s preferred toxicity values.Furthermore, many chronic RfDs have been reviewedand verified by an intra-Agency RfD Workgroup; theseverified RfDs represent an Agency consensus and arepreferred over other RfDs that have not undergone suchreview (see Section 7.2.7, Verification of RfDs).Similarly, acceptable intakes for subchronic exposures(AISs) have been superseded by the more strictlydefined subchronic RfD values. Therefore, the formerterminology (ADI, AIC, AIS) should no longer be usedin Superfund program risk assessments.

More recently, EPA has begun developingsubchronic RfDs (RfDs), which are useful- forcharacterizing potential noncarcinogenic effectsassociated with shorter-term exposures, anddevelopmental RfDs (RfDdts), which are usefulspecifically for assessing potential developmentaleffects resulting from exposure to a compound.As a guideline for Superfund program riskassessments, subchronic RfDs should be used toevaluate the potential noncarcinogenic effects ofexposure periods between two weeks and sevenyears. Such short-term exposures can result whena particular activity is performed for a limitednumber of years or when a chemical with a shorthalf-life degrades to negligible concentrationswithin several months. Developmental RfDs areused to evaluate the potential effects on adeveloping organism following a single exposureevent.

7.2.1 CONCEPT OF THRESHOLD

For many noncarcinogenic effects, protectivemechanisms are believed to exist that must beovercome before an adverse effect is manifestedFor example, where a large number of cellsperform the same or similar function, the cellpopulation may have to be significantly depletedbefore an effect is seen. As a result, a range ofexposures exists from zero to some finite valuethat can be tolerated by the organism withessentially no chance of expression of adverseeffects. In developing a toxicity value forevaluating noncarcinogenic effects (i.e., an RfD),the approach is to identify the upper bound ofthis tolerance range (i.e., the maximumsubthreshold level). Because variability exists inthe human population, attempts are made toidentify a subthreshold level protective of sensitiveindividuals in the population. For most chemicals,this level can only be estimated; the RfDincorporates uncertainty factors indicating thedegree or extrapolation used to derive theestimated value. RfD summaries in IRIS alsocontain a statement expressing the overallconfidence that the evaluators have in the RtD(high, medium, or low). The RfD is generallyconsidered to have uncertainty spanning an orderof magnitude or more, and therefore the RfDshould not be viewed as a strict scientificdemarcation between what level is toxic andnontoxic.

7.2.2 DERIVATION OF AN ORAL RfD (RfDo)

Identifying the critical study and determiningthe NOAEL. In the development of oral RfDs, allavailable studies examining the toxicity of achemical following exposure by the oral route aregathered and judged for scientific merit.Occasionally, studies based on other exposureroutes (e.g., inhalation) are considered, and thedata are adjusted for application to the oral route.Any differences between studies are reconciled andan overall evaluation is reached. If adequatehuman data are available, this information is usedas the basis of the RfD. Otherwise, animal studydata are used; in these cases, a series ofprofessional judgments are made that involve,among other considerations, an assessment of therelevance and scientific quality of the experimentalstudies. If data from several animal studies arebeing evaluated, EPA first seeks to identify the

7

animal model that is most relevant to humansbased on a defensible biological rationale, forinstance, using comparative metabolic andpharmacokinetic data. In the absence of a speciesthat is clearly the most relevant, EPA assumesthat humans are at least as sensitive to thesubstance as the most sensitive animal speciestested. Therefore, as a matter of science policy,the study on the most sensitive species (thespecies showing a toxic effect at the lowestadministered dose) is selected as the critical studyfor the basis of the RfD. The effect characterizedby the “lowest-observed-adverse-effect-level”(LOAEL) after dosimetric conversions to adjustfor species differences is referred to as the criticaltoxic effect.

After the critical study and toxic effect havebeen selected, EPA identifies the experimentalexposure level representing the highest level testedat which no adverse effects (including the criticaltoxic effect) were demonstrated. This highest “no-observed-adverse-effect level” (NOAEL) is the keydatum obtained from the study of the dose-response relationship. A NOAEL observed in ananimal study in which the exposure wasintermittent (such as five days per week) isadjusted to reflect continuous exposure.

The NOAEL is selected based in part on theassumption that if the critical toxic effect isprevented, then all toxic effects are prevented.The NOAEL for the critical toxic effect shouldnot be confused with the “no-observed-effect level”(NOEL). The NOEL corresponds to the exposurelevel at which no effect at all has been observed;frequently, effects are observed that are notconsidered to be of toxicological significance. Insome studies, only a LOAEL rather than aNOAEL is available. The use of a LOAEL

MULTIPLE TOXIC EFFECTS AND RfDS

The RfD is developed from a NOAEL for the mostsensitive, or critical, toxic effect based in part on the assumption that if the critical toxic effect is preventad,then all toxic effects are prevented. It should beremembered during the risk characterization step of the risk assessment that if exposure levels exceed the RfD,then adverse effects in addition to the critical toxiceffect may begin to appear.

however, requires the use of an additionaluncertainty factor (see below).

Applying uncertainty factors. The RfD isderived from the NOAEL (or LOAEL) for thecritical toxic effect by consistent application ofuncertainty factors (UFs) and a modifying factor(MF). The uncertainty factors generally consist ofmultiples of 10 (although values less than 10 aresometimes used), with each factor representing aspecific area of uncertainty inherent in theextrapolation from the available data. The basesfor application of different uncertainty factors areexplained below.

A UF of 10 is used to account forvariation in the general population andis intended to protect sensitivesubpopulations (e.g., elderly, children).

A UF of 10 is used when extrapolatingfrom animals to humans. This factor isintended to account for the interspeciesvariability between humans and othermammals.

A UF of 10 is used when a NOAELderived from a subchronic instead of achronic study is used as the basis for achronic RfD.

A UF of 10 is used when a LOAEL isused instead of a NOAEL. This factoris intended to account for theuncertainty associated with extrapolatingfrom LOAELs to NOAELs.

In addition to the UFS listed above, a modifyingfactor (MF) is applied.

An MF ranging from >0 to 10 isincluded to reflect a qualitativeprofessional assessment of additionaluncertainties in the critical study and inthe entire data base for the chemical notexplicitly addressed by the precedinguncertainty factors. The default valuefor the MF is 1.1

To calculate the RfD, the appropriate NOAEL(or the LOAEL if a suitable NOAEL is notavailable) is divided by the product of all of the

8

applicable uncertainty factors and the modifyingfactor. That is:

RfD = NOAEL or LOAEL/(UFl x UF2.. xM F

Oral RfDs typically are expressed as onesignificant figure in units of mg/kg-day. Theseconcepts are shown graphically in EPA (1989g).To date, most RfDs developed by EPA andincluded in the sources listed in Section 7.4 arebased on administered doses, not absorbed doses(see box on page 7-10).

7.2.3 DERIVATION OF AN INHALATIONRfD (RfDi)

The methods EPA uses in the derivation ofinhalation RfDs are similar in concept to thoseused for oral RfDs; however, the actual analysisof inhalation exposures is more complex than oralexposures due to (1) the dynamics of therespiratory system and its diversity across speciesand (2) differences in the physiochemicalproperties of contaminants. Additionalinformation can be found in EPA’s InterimMethods for Development of Inhalation ReferenceDoses (EPA 1989d).

Identifying the critical study and determiningthe NOAEL. Although in theory the identificationof the critical study and the determination of theNOAEL is similar for oral and inhalationexposures, several important differences should benoted. In selecting the most appropriate study,EPA, considers differences in respiratory anatomyand physiology, as well as differences in thephysiochemical characteristics of the contaminant.Differences in respiratory anatomy and physiologymay affect the pattern of contaminant depositionin the respiratory tract, and the clearance andredistribution of the agent. Consequently, thedifferent species may not receive the same dose ofthe contaminant at the same locations within therespiratory tract even though both species wereexposed to the same particle or gas concentration.Differences in the physiochemical characteristics

of the contaminants, such as the size and shape ofa particle or whether the contaminant is anaerosol or a gas, also influence deposition,clearance, and redistribution.

In inhalation exposures, the target tissue maybe a portion of the respiratort tract or, if thecontaminant can be absorbed and distributedthrough the body, some extra respiratory organ.Because the pattern of deposition may influenceconcentrations at the alveolar exchange boundaryor different tissues of the lung, the toxic healtheffect observed may be more directly related tothe pattern of deposition than to the exposureconcentration. Consequently, EPA considers thedeposition, clearance mechanisms, and thephysiochemical properties of the inhaled agent indetermining the effective dose delivered to thetarget organ.

Doses calculated in animals are converted toequivalent doses in humans on the basis ofcomparative physiological considerations (e.g.,ventilator parameters, regional lung surfaceareas). Additionally, if the exposure period wasdiscontinuous, it is adjusted to reflect continuousexposure.

Applying uncertainty factors. The inhalationRfD is derived from the NOAEL by applyinguncertainty factors similar to those listed abovefor oral RfDs. The UF of 10 is used whenextrapolating from animals to humans, in additionto calculation of the human equivalent dose, toaccount for interspecific variability in sensitivity tothe toxicant. The resulting RfD value forinhalation exposure is generally reported as aconcentration in air (in mg/m3 for continuous, 24hour/day exposure), although it may be reportedas a corresponding inhaled intake (in mg/kg-day).A human body weight of 70 kg and an inhalationrate of 20 m3/day are used to convert between aninhaled intake expressed in units of mg/kg-day anda concentration in air expressed in mg/m3.

7.2.4 DERIVATION OF A SUBCHRONIC RfD(RfDs)

The chronic RfDs described above pertain tolifetime or other long-term exposures and may beoverly protective if used to evaluate the potentialfor adverse health effects resulting fromsubstantially less-than-lifetime exposures. Forsuch situations, EPA has begun calculating toxicityvalues specifically for subchronic exposuredurations, using a method similar to that outlinedabove for chronic RfDs, EPA’s EnvironmentalCriteria and Assessment Office develops

9

subchronic RfDs and, although they have beenpeer-reviewed by Agency and outside reviewers,RfDs values have not undergone verification by anintra-Agency workgroup (see Section 7.2.7). Asa result, subchronic RfDs are considered interimrather than verified toxicity values and are notplaced in IRIS.

Development of subchronic reference dosesparallels the development of chronic referencedoses in concept; the distinction is one ofexposure duration. Appropriate studies areevaluated and a subchronic NOAEL is identified.The RfDs is derived from the NOAEL by theapplication of UFs and MF as outlined above.When experimental data are available only forshorter exposure durations than desired, anadditional uncertainty factor is applied. This issimilar to the application of the uncertainty factorfor duration differences when a chronic RfD isestimated from subchronic animal data. On theother hand, if subchronic data are missing and achronic oral RfD derived from chronic data exists,the chronic oral RfD is adopted as the subchronicoral RfD. There is no application of anuncertainty factor to account for differences inexposure duration in this instance.

7.2.5 DERIVATION OF A DEVELOPMENTALTOXICANT RfD (RfDdt)

In developing an RfDdt, evidence is gatheredregarding the potential of a substance to causeadverse effects in a developing organism as aresult of exposure prior to conception (eitherparent), during prenatal development, orpostnatally to the time of sexual maturation.Adverse effects can include death, structuralabnormality, altered growth, and functionaldeficiencies. Maternal toxicity also is considered.The evidence is assessed, and the substance isassigned a weight-of-evidence designationaccording to the scheme outlined below andsummarized in the box in the opposite column.In this scheme, three levels are used to indicatethe assessor’s degree of confidence in the data:definitive evidence, adequate evidence, andinadequate evidence. The definitive and adequateevidence categories are subdivided as to whetherthe evidence demonstrates the occurrence or theabsence of adverse effects.

WEIGHT-OF-EVIDENCE SCHEME FORDEVELOPMENTAL TOXICITY

Definitive Evidence for:

- Human Developmental Toxicity

- No Apparent Human Developmental Toxicity

Adequate Evidence for

- Potential Human Developmental Toxicity

- No Apparent Potential Human DevelopmentalToxicity

Inadequate Evidence for Determining PotentialHuman Developmental Toxicity

After the weight-of-evidence designation isassigned, a study is selected for the identificationof a NOAEL. The NOAEL is converted to anequivalent human dose, if necessary, and dividedby uncertainty factors similar to those used in thedevelopment of an oral RfD. It should beremembered that the RfDdt is based on a shortduration of exposure because even a singleexposure at a critical time (e.g., during gestation)may be sufficient to produce adversedevelopmental effects and that chronic exposureis not a prerequisite for developmental toxicity tobe manifested. Therefore, RfDdt values areappropriate for evaluating single event exposures,which usually are not adjusted based on theduration of exposure. Additional information onthe derivation of RfDdt values is available inEPA’s Proposed Amendments to the Guidelines forthe Health Assessment of Suspect DevelopmentalToxicants (EPA 1989e).

7.2.6 ONE-DAY AND TEN-DAY HEALTHADVISORIES

Reference values that may be useful forevaluating potential adverse effects associated withoral exposures of shorter duration have beendeveloped by the Office of Drinking Water.These values are known as One-day and Ten-dayHealth Advisories, which are issued asnonregulatory guidance. Health Advisory valuesare concentrations of contaminants in drinkingwater at which adverse health effects would not beexpected to occur for an exposure of the specified

10

duration. The Health Advisory values are basedon data describing noncarcinogenic effects and arederived by dividing a NOAEL or LOAEL by theappropriate uncertainty and modifying factors.They are based on a 10-kg child assumed to drink1 liter of water per day, and a margin of safety isincluded to protect sensitive members of thepopulation. One-day and Ten-day HealthAdvisories do not consider any carcinogenic riskassociated with the exposure even if the compoundis a potential carcinogen. For additionalinformation on the derivation of Health Advisoryvalues, refer to the Agency’s guidance document(EPA 1989c).

7.2.7 VERIFICATION OF RfDS

EPA has formed an RfD Workgroupcomposed of members from many EPA offices toverify existing Agency RfDs and to resolveconflicting toxicity assessments and toxicity valueswithin the Agency. The Workgroup reviews theinformation regarding the derivation of an RfDfor a substance and summarizes its evaluations,conclusions, and reservations regarding the RfDin a standardized summary form from one toseveral pages in length. This form containsinformation regarding the development of theRfD, such as the chosen effect levels anduncertainty factors, as well as a statement on theconfidence that the evaluators have in the RfDitself, the critical study, and the overall data base(high, medium, or low). Once verified, these data

ABSORBED VERSUSADMINISTERED DOSE

Toxicity values -- for both noncarcinogenic andcarcinogenic effects -- are generally calculated fromcritical effect levels baaed on administered rather thanabsorbed doses. It is important, therefore, to comparesuch toxicity values to exposure estimates expressed asintakes (corresponding to administered doses), not asabsorbed doses. For the few toxicity values that have been based on absorbed doses, either the exposureestimate or the toxicity value should be adjusted tomake the values comparable (i.e., compare exposuresestimated as absorbed doses to toxicity values repressedas absorbed doses, and exposures estimated as intakesto toxicity values expressed as administered doses) SeeAppendix A for guidance on making adjustments forabsorption efficiency.

evaluation summaries are entered into IRIS andare available for public access.

Workgroup-approved RfDs are referred to asverified RfDs. Those RfDs awaiting workgroupapproval are referred to as interim RfDs. At thetime of this manual’s publication, only chronicRfDs are being verified. No workgroup has beenestablished to verify subchronic RfDs ordevelopmental RfDs.

7.3 TOXICITY ASSESSMENT FORCARCINOGENIC EFFECTS

This section describes how the types oftoxicity information presented in Section 7.1 areconsidered in the toxicity assessment forcarcinogenic effects. A slope factor and theaccompanying weight-of-evidence determinationare the toxicity data most commonly used toevaluate potential human carcinogenic risks. Themethods EPA uses to derive these values areoutlined below. Additional information can beobtained by consulting EPA’s Guidelines forCarcinogen Risk Assessment (EPA 1986a) andAppendix B to IRIS (EPA 1989a).

7.3.1 CONCEPT OF NONTHRESHOLDEFFECTS

Carcinogenesis, unlike many noncarcinogenichealth effects, is generally thought to be aphenomenon for which risk evaluation based onpresumption of a threshold is inappropriate. Forcarcinogens, EPA assumes that a small number ofmolecular events can evoke changes in a singlecell that can lead to uncontrolled cellularproliferation and eventually to a clinical state ofdisease. This hypothesized mechanism forcarcinogenesis is referred to as “nonthreshold”because there is believed to be essentially no levelof exposure to such a chemical that does not posea finite probability, however small, of generatinga carcinogenic response. That is, no dose isthought to be risk-free. Therefore, in evaluatingcancer risks, an effect threshold cannot beestimated. For carcinogenic effects, EPA uses atwo-part evaluation in which the substance first isassigned a weight-of-evidence classification, andthen a slope factor is calculated.

11

7.3.2 ASSIGNING A WEIGHT OF EVIDENCE

In the first step of the evaluation, theavailable data are evaluated to determine the likelihood that the agent is a human carcinogen.The evidence is characterized separately for humanstudies and animal studies as sufficient, limited,inadequate, no data, or evidence of no effect. Thecharacterizations of these two types of data arecombined, and based on the extent to which theagent has been shown to be a carcinogen inexperimental animals or humans, or both, theagent is given a provisional weipht-of-evidenceclassification. EPA scientists then adjust theprovisional classification upward or downward,based on other supporting evidence ofcarcinogenicity (see Section 7.1.3). For a furtherdescription of the role of supporting evidence, seethe EPA guidelines (EPA 1986a).

The EPA classification system for weight ofevidence is shown in the box in the oppositecolumn. This system is adapted from theapproach taken by the International Agency forResearch on Cancer (IARC 1982).

7.3.3 GENERATING A SLOPE FACTOR2

In the second part of the evaluation, basedon the evaluation that the chemical is a known orprobable human carcinogen, a toxicity value thatdefines quantitatively the relationship betweendose and response (i.e., the slope factor) iscalculated. Slope factors are typically calculatedfor potential carcinogens in classes A, Bl, and B2.Quantitative estimation of slope factors for thechemicals in class C proceeds on a case-by-casebasis.

Generally, the slope factor is a plausibleupper-bound estimate of the probability of aresponse per unit intake of a chemical over alifetime. The slope factor is used in riskassessments to estimate an upper-bound lifetimeprobability of an individual developing cancer asa result of exposure to a particular level of apotential carcinogen. Slope factors should alwavsbe accompanied by the weight-of-evidenceclassification to indicate the strength of theevidence that the agent is a human carcinogen.

Identifying the appropriate data set. Inderiving slope factors, the available information

about a chemical is evaluated and an appropriatedata set is selected. In choosing appropriate datasets, human data of high quality are preferable toanimal data. If animal data are used, the speciesthat responds most similarly to humans (withrespect to factors such as metabolism, physiology,and pharmacokinetics) is preferred. When noclear choice is possible, the most sensitive speciesis given the greatest emphasis. Occasionally, insituations where no single study is judged mostappropriate, yet several studies collectively supportthe estimate, the geometric mean of estimatesfrom all studies may be adopted as the slope.This practice ensures the inclusion of all relevantdata.

Extrapolating to lower doses. Because riskat low exposure levels is difficult to measuredirectly either by animal experiments or byepidemiologic studies, the development of a slopefactor generally entails applying a model to theavailable data set and using the model toextrapolate from the relatively high dosesadministered to experimental animals (or theexposures noted in epidemiologic studies) to thelower exposure levels expected for human contactin the environment.

12

A number of mathematical models andprocedures have been developed to extrapolatefrom carcinogenic responses observed at highdoses to responses expected at low doses.Different extrapolation methods may provide areasonable fit to the observed data but may leadto large differences in the projected risk at lowdoses. In keeping with EPA’s Guidelines forCarcinogen Risk Assessment (EPA 1986a) and theprinciples outlined in Chemical Carcinogens: AReview of the Science and Its Associated Principles(OSTP 1985), the choice of a low-doseextrapolation model is governed by consistentwith current understanding of the mechanism ofcarcinogenesis, and not solely on goodness-of-fitto the observed tumor data. When data arelimited and when uncertainty exists regarding themechanisms of carcinogenic action, the EPAguidelines and OSTP principles suggest thatmodels or procedures that incorporate low-doselinearity are preferred when compatible with thelimited information available. EPA’s guidelinesrecommend that the linearized multistage modelbe employed in the absence of adequateinformation to the contrary. Among the othermodels available are the Weibull, probit, logit,one-hit, and gamma multihit models, as well asvarious time-to-tumor models. Most of thesemodels are less conservative (i.e., predict lowercancer potency) than the linearized multistagemodel. These concepts and models are showngraphically in EPA (1989g) and OTA (1981).

In general, after the data are fit to theappropriate model, the upper 95th percentconfidence limit of the slope of the resulting dose-response curve is calculated. This value is knownas the slope factor and represents an upper 95thpercent confidence limit on the probability of aresponse per unit intake of a chemical over alifetime (i.e., there is only a 5 percent chance thatthe probability of a response could be greater thanthe estimated value on the basis of theexperimental data and model used). In somecases, slope factors based on human dose-responsedata are based on the “best” estimate instead ofthe upper 95th percent confidence limits. Becausethe dose-response curve generally is linear only inthe low-dose region, the slope factor estimate onlyholds true for low doses. Information concerningthe limitations on use of slope factors can befound in IRIS.

Determining equivalent human doses. Whenanimal data are used as a basis for extrapolation,the human dose that is equivalent to the dose inthe animal study is calculated using theassumption that different species are equallysensitive to the effects of a toxicant if they absorbthe same amount of the agent (in milligrams) perunit of body surface area. This assumption ismade only in the absence of specific informationabout the equivalent doses for the chemical inquestion. Because surface area is approximatelyproportional to the 2/3 power of body weight, theequivalent human dose (in mg/day, or other unitsof mass per unit time) is calculated by multiplyingthe animal dose (in identical units) by the ratio ofhuman to animal body weights raised to the 2/3power. (For animal doses expressed as mg/kgday,the equivalent human dose, in the same units, iscalculated by multiplying the animal dose by theratio of animal to human body weights raised tothe 1/3 power.)

When using animal inhalation experiments toestimate lifetime human risks for partially solublevapors or gases, the air concentration (ppm) isgenerally considered to be the equivalent dosebetween species based on equivalent exposuretimes (measured as fractions of a lifetime). Forinhalation of particulate or completely absorbedgases, the amount absorbed per unit of bodysurface area is considered to be the equivalentdose between species.

Summary of dose-response parameters.Toxicity values for carcinogenic effects can beexpressed in several ways. The slope factor isusually, but not always, the upper 95th percentconfidence limit of the slope of the dose-responsecurve and is expressed as (mg/kg-day)-l. If theextrapolation model selected is the linearizedmultistage model, this value is also known as theq1*. That is:

Slope factor = risk per unit dose= risk per mg/kg-day

Where data permit, slope factors listed in IRISare based on absorbed doses, although to datemany of them have been based on administereddoses. (The qualifiers related to absorbed versusadministered dose given in the box on page 7-10apply to assessment of cancer risk as well as toassessment of potential noncarcinogenic effects.)

13

Toxicity values for carcinogenic effects alsocan be expressed in terms of risk per unitconcentration of the substance in the mediumwhere human contact occurs. These measures,called unit risks, are calculated by dividing theslope factor by 70 kg and multiplying by theinhalation rate (20 rn3/day) or the waterconsumption rate (2 liters/day), respectively, forrisk associated with unit concentration in air orwater. Where an absorption fraction less than 1.0has been applied in deriving the slope factor, anadditional conversion factor is necessary in thecalculation of unit risk so that the unit risk willbe on an administered dose basis. Thestandardized duration assumption for unit risks isunderstood to be continuous lifetime exposure.Hence, when there is no absorption conversionrequired:

air unit risk = risk per ug/m3

= slope factor x l/70 kg x20 m3/day x 10-3

water unit risk = risk per ug/L= slope factor x 1/70 kg x

2 L/day x 10-3

The multiplication by 10-3 is necessary to convertfrom mg (the slope factor, or q1*, is given in(mg/k-day) -1) to ug (the unit risk is given in(ug/m3 )-1 or (ug/L)-1).

7.3.4 VERIFICATION OF SLOPE FACTORS

EPA formed the Carcinogen Risk AssessmentVerification Endeavor (CRAVE) Workgroup tovalidate Agency carcinogen risk assessments andresolve conflicting toxicity values developed byvarious program offices. Workgroup membersrepresent many different EPA offices and arescientists experienced in issues related to both thequalitative and quantitative risk assessment ofcarcinogenic agents. Slope factors verified byCRAVE have undergone extensive peer reviewand represent an Agency consensus. CRAVE-verified review summaries (similar to RfDWorkgroup summaries) are entered into the IRISdata base.

7.4 IDENTIFYING APPROPRIATETOXICITY VALUES FOR SITERISK ASSESSMENT

Using the methods outlined above, EPA hasperformed toxicity assessments for many chemicalsfound at Superfund sites and has made the resultsavailable for use. This section provides step-by-step methods for locating appropriate toxicityinformation, including numerical toxicity values, tobe used in Superfund risk assessments. Becauseone’s confidence in toxicity values depends heavilyon the data base and the methods of extrapolationused in their development, guidance is alsoincluded for identifying the important informationon which these values are based.

7.4.1 GATHER TOXICITY INFORMATIONFOR CHEMICALS BEING EVALUATED

In the first step of the toxicity assessment,information is collected regarding the toxic effectsthat occur following exposure to the chemicalbeing evaluated. Particular attention. should bepaid to the route of exposure, the frequency andlength of exposure, and the doses at which theadverse effects are expected to occur. Chemicalshaving potential reproductive or developmentaleffects should be flagged. Later in the evaluation,special reference doses for developmental effectscan be sought for these chemicals.

Several sources may provide useful toxicityinformation and references to primary literature,although only some of them should be used assources for slope factors and reference doses (asexplained below).

Integrated Risk Information System (IRIS).3

IRIS is an EPA data base containing up-to-datehealth risk and EPA regulatory information fornumerous chemicals. IRIS contains only thoseRfDs and slope factors that have been verified bythe RfD or CRAVE Workgroups andconsequently, is considered to be the preferredsource of toxicity information. Information inIRIS supersedes all other sources. Only ifinformation is not available in IRIS for thechemical being evaluated should the sources belowbe consulted. IRIS consists of a collection ofcomputer files on individual chemicals. Existinginformation on the chemicals is updated as new

14

scientific data are reviewed. New files and new chemicals are added as information becomes available. These chemical files contain descriptiveand quantitative information in the followingcategories:

oral and inhalation chronic referencedoses;

oral and inhalation slope factors andunit risks for chronic exposure tocarcinogens;

Health Advisories from EPA’s Office ofDrinking Water;

EPA regulatory action summaries; and

supplemental data on acute healthhazards and physical/chemical properties.

To ensure access to the most up-to-datechemical information, IRIS is only available on-line. For information on how to access this database, call IRIS User Support at 513-569-7254 orsee the Federal Register notice regarding theavailability of IRIS (EPA 1988a).

Should EPA regional staff have specifictechnical or scientific questions about anyverification workgroup’s analysis of particular datacited in IRIS, the Agency contact for a particularchemical (identified at the end of each IRIS file)should be consulted. If new data are identifiedsuggesting that existing IRIS information may beoutdated, or if there is concern or disagreementabout the overall findings of particular files, theAgency IRIS coordinator should be consulted.The IRIS coordinator can assist in makingarrangements should discussions with a verificationworkgroup be needed.

Health Effects Assessment Summary Tables(HEAST). Formerly “The Quarterly” andassociated references, HEAST is a tabularpresentation of toxicity information and values forchemicals for which Health Effects Assessments(HEAs), Health and Environmental Effects Documents (HEEDs), Health and EnvironmentalEffects Profiles (HEEPs), Health AssessmentDocuments (HADs), or Ambient’ Air QualityCriteria Documents (AAQCDs) have beenprepared. HEAST summarizes interim (and some

verified) RfDs and slope factors as well as othertoxicity information for specific chemicals. Inaddition, HEAST directs readers to the mostcurrent sources of supporting toxicity informationthrough an extensive reference section. Therefore,HEAST is especially helpful when verifiedinformation for a chemical is not in IRIS.HEAST, which is updated quarterly, also providesa valuable pointer system for identifying currentreferences on chemicals that are not in IRIS.

HEAST can be obtained upon request fromthe Superfund Docket (FTS or 202-382-3046).The Docket will mail copies of HEAST to callersand place requesters on a mailing list to receivean updated version quarterly. HEAs, HEEDs,HEEPs, HADs, and AAQCDS referenced inHEAST are available through EPA’s Center forEnvironmental Research Information (CERI) inCincinnati, OH (513-569-7562 or FTS 684-7562)or the National Technical Information Service(NTIS), 5285 Port Royal Road, Springfield, VA22161 (703-487-4650 or 800-336-4700).

EPA criteria documents. These documentsinclude drinking water criteria documents, drinkingwater Health Advisory summaries, ambient waterquality criteria documents, and air quality criteriadocuments, and contain general toxicityinformation that can be used if information for achemical is not available through IRIS or theHEAST references. Criteria documents areavailable through NTIS at the address given above.Information on drinking water criteria documentscan be obtained through the Safe Drinking WaterHotline (800-426-4791).

Agency for Toxic Substances and DiseaseRegistry (ATSDR) toxicological profiles. ATSDRis developing toxicological profiles for 275hazardous substances found at Superfund sites.The first 200 substances to be addressed havebeen identified in Federal Register notices (EPA1987, 1988b). These profiles contain generaltoxicity information and levels of exposureassociated with lethality, cancer, genotoxicity,neurotoxicity, developmental and reproductivetoxicity, immunotoxicity, and systemic toxicity (i.e.,hepatic, renal, respiratory, cardiovascular,gastrointestinal, hematological, musculoskeletal,and dermal/ocular effects). Health effects inhumans and animals are discussed by exposureroute (i.e., oral, inhalation, and dermal) and

15

HIERARCHY OF TOXICITY INFORMATION

Because toxcity information may change rapidly and quickly become outdated, care should be taken to find the most recentinformation available. IRIS is updated monthly, provides verified RfDs and slope factors, and is the Agency’s preferred sourceof toxicity information. Only if values unavailable in IRIS should other information sources be consulted.

HEAST is the second most current source of toxicity information of irnportance to Superfund. Unlike IRIS, HEAST providesinformation regarding interim as well as verified RfDs and slope factors. Readers are directed to supporting toxicity informationfor interim and verified vatues in an extensive reference section of HEAST. HEAST informatiom should only be sought for thosechemicals not listed in IRIS..

Toxicity information, RfDs, and slope factors also can be found in other EPA documents. Although these values were developed by offices within the Agency, they have not necessarily been verified by the RfD or CRAVE Workgroups. The useof up-to-date verified information is preferred to the use of interim information and, therefore, toxicity information should beobtained from other EPA references only if informatkm could not be found in IRIS or HEAST. Before using references otherthan those cited in IRIS or HEAST, check with ECAO at 513-569-7300 (FTS 684-7300) to see if more current information isavailable.

duration (i.e., acute, intermediate, and chronic).Also included in the profiles are chapters onphysiochemical properties, environmental fate,potential for human exposure, analytical methods,and regulatory and advisory status. Contact NTISat the address given on the previous page forfurther information on the status or availability ofa particular profile.

EPA’s Environmental Criteria andAssessment Office (ECAO). ECAO may becontacted at 513-569-7300 (FTS 684-7300) forgeneral toxicological information as well as fortechnical guidance concerning route-to-routeextrapolations, toxicity values for dermalexposures, and the evaluation of chemicals withouttoxicity values. The requestor should identify theirneed for a “rapid response request” (within 48hours) for interim guidance on Superfund health-related issues. Contractors must give the nameand address of their RPM or regional riskassessment contact before ECAO will respond.RPMs and regional contacts will be sent a copyof ECAO’s response to the contractor.

Open literature. A primary literature searchmay be valuable for determining whether new dataare available that may affect IRIS information.

7.4.2 DETERMINE TOXICITY VALUES FORNONCARCINOGENIC EFFECTS (RfDs)

After general toxicity information for thechemicals of concern has been located, the nextstep is to identify the appropriate toxicity values

to be used in evaluating noncarcinogenic effectsassociated with the specific exposures beingassessed. First, by referring to the exposureinformation generated in Chapter 6, the exposureperiods for which toxicity values are necessary andthe exposure route for each chemical beingevaluated should be determined. The appropriatetoxicity values for the chemical for each exposureduration and route of exposure can then beidentified using the sources listed above.

For Superfund risk assessments, chronic RfDsshould be identified for evaluating exposureperiods between seven years and a lifetime,subchronic RfDs for exposure periods between twoweeks and seven years, and One- or Ten-dayHealth Advisories for oral exposure periods of lessthan two weeks. According to EPA (1988c), One-day Health Advisories are applicable to exposureperiods as long as five days and Ten-day HealthAdvisories are applicable to exposure periods aslong as two weeks. Developmental RfDs shouldbe identified for evaluating single exposure eventsand other very short exposures (e.g., one day).Note that for some substances and some exposuresituations, more than one of the toxicity valueslisted above may be needed to adequately assesspotential noncarcinogenic effects.

Because carcinogens also commonly evokenoncarcinogenic effects, RfDs should be sought for..all chemicals being carried through the riskassessment, including carcinogens. The RfDsderived for carcinogens, however, are based onnoncancer effects and should not be assumed to

16

be protective against carcinogenicity. A sampleformat for summarizing RfDs and other toxicity values is shown in Exhibit 7-2. This informationwill be needed in the risk characterization step(see Exhibits 8-3 and 8-4).

7.4.3 DETERMINE TOXICITY VALUES FORCARCINOGENIC EFFECTS (SLOPEFACTORS)

In this step of the toxicity assessment,appropriate toxicity values for evaluating thecarcinogenic risks associated with exposure areidentified. First, by referring to the exposureinformation generated in Chapter 6, the route ofexposure for the potential carcinogens beingevaluated should be identified. Slope factors for

these chemicals can then be identified using thehierarchy of sources listed in the box on page7-15. Slope factors for all potential carcinogenshaving a weight-of-evidence classification of A, B,or C should be sought. A notation of the EPAweight-of-evidence classification should always beincluded with the slope factor. A sample formatfor summarizing the required toxicity values isshown in Exhibit 7-3. This information will beneeded in the risk characterization stepExhibit 8-2).

7.5 EVALUATING CHEMICALSFOR WHICH NO TOXICITYVALUES ARE AVAILABLE

(see

If EPA-derived RfDs and slope factors areavailable for the chemicals being examined, thesevalues should always be used in the riskassessment. Use of EPA-derived toxicity valuesprevents duplication of effort and ensuresconsistency among risk assessments. If EPA-derived toxicity values are not available, thefollowing measures are recommended.

7.5.1 ROUTE-TO-ROUTE EXTRAPOLATION

For cases in which EPA-derived toxicity values are not available for the route of exposure

being considered but are available for anotherroute, EPA recommends contacting ECAO forguidance on route-to-route extrapolation. Iftoxicity information is not available from ECAO,a qualitative rather than quantitative evaluation of

the chemical is recommended. The implicationsof the absence of this chemical from the riskestimate should be discussed in the uncertaintysection.

7.5.2 DERMAL EXPOSURE

No RfDs or slope factors are available forthe dermal route of exposure. In some cases,however, noncarcinogenic or carcinogenic risksassociated with dermal exposure can be evaluatedusing an oral RfD or oral slope factor,respectively. EPA recommends contacting ECAOfor guidance on appropriate methods forevaluating dermal exposure for specific chemicals;some general guidance for calculating intakes viathe dermal route and making appropriatecomparisons with oral RfD values is given inAppendix A. In brief, exposures via the dermalroute generally are calculated and expressed asabsorbed doses. These absorbed doses arecompared to an oral toxicity value that has beenadjusted, if necessary, so that it too is expressedas an absorbed dose.

It is inappropriate to use the oral slopefactor to evaluate the risks associated with dermalexposure to carcinogens such as benz(a)pyrene,which cause skin cancer through a direct action atthe point of application. These types of skincarcinogens and other locally active compoundsmust be evaluated separately from the abovemethod; consult ECAO for guidance. Generallyonly a qualitative assessment of risks from dermalexposure to these chemicals is possible. This doesnot apply to carcinogens such as arsenic, whichare believed to cause skin cancer through asystemic rather than local action.

If information is not available from ECAO,the assessor should describe the effects of thechemical qualitatively and discuss the implicationsof the absence of the chemical from the riskestimate in the uncertainty section of the riskassessment.

7.5.3 GENERATION OF TOXICITY VALUES

If EPA-derived toxicity values are unavailablebut adequate toxicity studies are available, onemay derive toxicity values using Agencymethodology. Any such derivation should be done

EXHIBIT 7-2

EXAMPLE OF TABLE FORMAT FORTOXICITY VALUES: POTENTIAL NONCARCINOGENIC EFFECTS

18

EXHIBIT 7-3

EXAMPLE OF TABLE FORMAT FORTOXICITY VALUES: POTENTIAL CARCINOGENIC EFFECTS

Slope Factor (SF) Weight-of-Evidence Type of SF Basis/Chemical (mg/kg-day) -1 Classification Cancera SF Source

Oral Route

Benzene

Chlordane

Inhalation

. . . . .

0.029*

1.3*

A*

B2*

Leukemia

--

Waterb/IRIS

Waterb/IRIS

Route

. . . . . . . . . . . . . . . . . . . .

* Values for illustration only.

a Identify type(s) of cancer in this table for Class A carcinogens only.

b Slope factor based on administered dose in drinking water and assumed absorption fraction of 1.0.

19

in conjunction with the regional risk assessmentcontact, who will submit the derivation to ECAOfor approval. Contact with ECAO should beestablished early in the process to eliminate anyduplication of effort because ECAO may haveinformation on the chemical being evaluated.

7.6 UNCERTAINTIES RELATEDTO TOXICITY INFORMATION

Toxicity information for many of thechemicals found at Superfund sites is oftenlimited. Consequently, there are varying degreesof uncertainty associated with the toxicity valuescalculated. Sources of uncertainty associated withtoxicity values may include

using dose-response information fromeffects observed at high doses to predictthe adverse health effects that may occurfollowing exposure to the low levelsexpected from human contact with theagent in the environment

using dose-response information fromshort-term exposure studies to predictthe effects of long-term exposures, andvice-versa;

using dose-response information fromanimal studies to predict effects inhumans; and

using dose-response information fromhomogeneous animal populations orhealthy human populations to predict theeffects likely to be observed in thegeneral population consisting ofindividuals with a wide range ofsensitivities.

An understanding of the degree ofuncertainty associated with toxicity values is animportant part of interpreting and using thosevalues. Therefore, as part of the toxicityassessment for Superfund sites, a discussion of thestrength of the evidence of the entire range ofprincipal and supporting studies should beincluded. The degree of confidence ascribed toa toxicity value is a function of both the qualityof the individual study from which it was derived

and the completeness of the supporting database. EPA-verified RfDs found in IRIS areaccompanied by a statement of the confidence thatthe evaluators have in the RfD itself, the criticalstudy, and the overall data base. All EPA-verifiedslope factors are accompanied by a weight-of-evidence classification, which indicates thelikelihood that the agent is a human carcinogen.The weight-of-evidence classification is based onthe completeness of the evidence that the agentcauses cancer in experimental animals andhumans. These designations should be used asone basis for the discussion of uncertainty.

The discussion of uncertainty also shouldinclude an indication of the extent to which ananalysis of the results from different studies givea consistent, plausible picture of toxicity. Thegreater the strength of the evidence, the greaterone’s confidence in the conclusions drawn. Thefollowing factors add to the strength of theevidence that the chemical poses a hazard tohumans and should be considered:

similar effects across species, strains, sex,and routes of exposure;

clear evidence of a dose-responserelationship;

a plausible relationship among data onmetabolism, postulated mechanism ofaction, and the effect of concern (seeSection 7.1.3);

similar toxicity exhibited by structurallyrelated compounds (see Section 7.1.3]and

some link between the chemical andevidence of the effect of concern inhumans (see Section 7.1.1).

High uncertainty (low confidence; lowstrength of evidence) indicates that the toxicityvalue might change if additional chronic toxicitydata become available. Low uncertainty (highconfidence) is an indication that a value is lesslikely to change as more data become available,because there is consistent among the toxicresponses observed in different species, sexes,study designs, or in dose-response relationships.The lower the uncertainty about toxicity values,

20

the more confidence a decision-maker can have inthe risk assessment results. Often, highconfidence is associated with values that are basedon human data for the exposure route of concern.

7.7 SUMMARIZATION ANDPRESENTATION OF THETOXICITY INFORMATION

This section discusses methods for presentingtoxicity information in the risk assessmentdocument for the chemicals being evaluated.

7.7.1 TOXICITY INFORMATION FOR THEMAIN BODY OF THE TEXT

A short description of the toxic effects ofeach chemical carried through the assessment innon-technical language should be prepared forinclusion in the main body of the risk assessment.Included in this description should be informationon the effects associated with exposure to thechemical and the concentrations at which theadverse effects are expected to occur in humans.Toxicity values should be accompanied by a briefdescription of the overall data base and theparticular study from which the value was derived.In addition, a notation should be made of thecritical effect and any uncertainty factors used inthe calculation. For any RfD value obtained fromIRIS, a notation of the degree of confidenceassociated with the determination should also beincluded. To aid in the risk characterization, itshould be indicated if absorption efficiency was

considered and also what exposure averagingperiods are appropriate for comparison with thevalue.

Summary tables of toxicity values for allchemicals should be prepared for inclusion in themain body of the risk assessment report. RfDs inthe table should be accompanied with theuncertainty factors used in their derivation, theconfidence rating given in IRIS (if applicable). anda notation of the critical effect. Slope factorsshould always be accompanied by EPA’s weight-of-evidence classification.

7.7.2 TOXICITY INFORMATION FORINCLUSION IN AN APPENDIX

If toxicity values were derived in conjunctionwith the regional risk assessment contact andECAO for chemicals lacking EPA-derived values,a technical documentation/justification of themethod of derivation should be prepared andincluded in the appendix of the risk assessmentreport. Included in this explanation should be adescription of the toxic effects of the chemicalsuch as information regarding the noncarcinogenic,carcinogenic, mutagenic, reproductive, anddevelopmental effects of the compound. Alsopresented should be brief descriptions (species,route of administration, dosages, frequency ofexposure, length of exposure, and critical effect)of the studies from which the values were derivedas well as the actual method of derivation.References for the studies cited in the discussionshould be included.

21


1. The MF is set less than one for a and number of subatances to account for nutritional essentiality.

2. The slope factor is occasionally referred to as a cancer potency factor; however, use of this terminology is not recommended.

3. The quantitative risk values and supporting information found in IRIS representa consensus judgement of EPA’s Reference DoseWorkgroup or Carcinogen Risk Assessment Verification Endeavor (CRAVE) Workgroup. These workgroups are composed of scientistsfrom EPA’s program offices and the Office of Research and Development. The concept of Agency-wide consensus is one of the moatvaluable aspects of IRIS.

22



Environmental Protection Agency (EPA). 1986b. Guidelines for the Health Assessment of Suspect Developmental Toxicants. 51Federal Regis ter 34028 (September 24, 1986).

Environmental Protection Agency (EPA). 1987. First Priotity List of Hazardous Substances That Will Be the Subject of ToxicologicalProfiles. 52 Federal Register 12866 (April 17, 1987).

Environmental Protection Agency (EPA). 1988a. Availability of the Integrated Risk Information System (IRIS). 53 Federal Register20162 (June 2, 1988).

Environmental Protection Agency (EPA). 1988b. Hazardous Substances List, Toxicological List, Toxico1ogical Profilcs; Second List. 53 FederalRegister 41280 (October 20, 1988)

Environmental Protection Agency (EPA). 1988c. Office of Drinking Water Health Advisories. Reviews of EnvironmentalContamination and Toxicology 104.

Environmental Protection Agency (EPA). 1989a. EPA Approach for Assessing the Risk Associated with Exposure to EnvironmentalCarcinogens. Appendix B to the Integrated Risk Information System (IRIS).

Environmental Protection Agency (EPA). 1989b. General Quantitative Risk Assessment Guidelines for NonCancer Health Effects.External Review Draft. Risk Assessment Forum Technical Panel on Risk Assessment Guidelines for Noncancer Health Effects.ECAO-CIN-538.

Environmental Protection Agency (EPA). 1989c. Guidelinces for Authors of EPA Office of Water Health Advisories for Drinking WaterContaminants. Office of Drinking Water.

Environmental Protection Agency (EPA). 1989d. Interim Methods for Development of Inhalation Reference Doses. EnvironmentalCriteria and Assessment Office. EPA/600/8-88/066F.

Environmental Protection Agency (EPA). 1989e. Proposed Amendments to the Guidelines for the Health Assessment of SuspectDevelopmental Toxicants. 54 Federal Register 9386 (March 6, 1989).

Envionmental Protection Agency (EPA). 1989f. Reference Dose (RfD) : Description and Use in Health Risk Assessments. AppendixA to the Integrated Risk Information System (IRIS).

Environmental Protection Agency (EPA). 1989g. Guidance Manual for Assessing Human Health Risks from Chemically ContaminatedFish and Shellfish. Office of Marine and Estuarine Protection. EPA/503/8-89/002

International Agency for Research on Cancer (IARC). 1982. IARC Monographs on the Evaluation of the Carcinogenic Risk ofChemicals to Humans. Supplement 4. Lyon, France.

National Academy of Sciences (NAS). 1983. Risk Assessment in the Federal Government: Managing the Process. National AcademyPress. Washington, D.C

Office of Science and Technology Policy (OSTP). 1985. Chemical Carcinogens A Preview of the Science and Its Associated Principles.50 Federal Register 10372 (March 14, 1985).

Office of Technology Assessment (OTA). 1981. Assessment of Technologies for Determining Cancer Risks from the Environment.Congress of the United States. Washington, D.C

CHAPTER 8

RISK CHARACTERIZATION

CHAPTER 8


This chapter describes the final step of thebaseline health risk assessment process, riskcharacterization. In this step, the toxicity andexposure assessments are summarized andintegrated into quantitative and qualitativeexpressions of risk. To characterize potentialnoncarcinogenic effects, comparisons are madebetween projected intakes of substances andtoxicity values; to characterize potentialcarcinogenic effects, probabilities that anindividual will develop cancer over a lifetime ofexposure are estimated from projected intakes andchemical-specific dose-response information.Major assumptions, scientific judgments, and tothe extent possible, estimates of the uncertaintiesembodied in the assessment are also presented.

Risk characterization also serves as the bridgebetween risk assessment and risk management andis therefore a key step in the ultimate sitedecision-making process. This step assimilates riskassessment information for the risk manager(RPM or regional upper management involved insite decision-making) to be considered alongsideother factors important for decision-making suchas economics, technical feasibility, and regulatorycontext. The risk characterization methodsdescribed in this chapter are consistent with EPA’spublished risk assessment guidelines. Exhibit 8-1is an overview of risk characterization, andillustrates how it relates to the preceding toxicityand exposure assessments and to the refinementof preliminary remediation goals.

In the following sections, the riskcharacterization methodology is described. Thereare separate discussions for carcinogenic andnoncarcinogenic effects because the methodologydiffers for these two modes of chemical toxicity.In addition to giving instructions for calculatingnumerical estimates of risk, this chapter providesguidance for interpreting, presenting, andqualifying the results. A risk characterization

cannot be considered complete unless thenumerical expressions of risk are accompanied byexplanatory text interpreting and qualifying theresults— .

8.1 REVIEW OF OUTPUTS FROMTHE TOXICITY ANDEXPOSURE ASSESSMENTS

Most sites being assessed will involve theevaluation of more than one chemical of concernand might include both carcinogenic andnoncarcinogenic substances. The first step in riskcharacterization is to gather, review, compare, andorganize the results of the exposure assessment(e.g., intakes for all exposure pathways and land-uses and for all relevant substances) and toxicityassessment (e.g., toxicity values for all exposure

3

EXHIBIT 8-1

STEPS IN RISK CHARACTERIZATION

4

routes and relevant substance). The followingtwo subsections describe how to organize theoutputs from the exposure and toxicity assessmentsand how to check for the consistency and validityof the information from the preceding exposureand toxicity assessments.

8.1.1 GATHER AND ORGANIZEINFORMATION

For each exposure pathway and land useevaluated in the exposure assessment, check thatall information needed to characterize risk isavailable. The necessary exposure information isoutlined in the box below.

For each chemical or substance evaluated inthe toxicity assessment, use the checklist providedin the next box to ensure that all informationneeded to characterize risk is available.

8.1.2 MAKE FINAL CONSISTENCY ANDVALIDITY CHECK

Check the consistency and validity of keyassumptions common to the exposure outputs andthe toxicity outputs for each contaminant andexposure pathway of concern. These assumptionsinclude the averaging period for exposure, theexposure route, and the absorption adjustments.The basic principle is to ensure that the exposure

estimates correspond as closely as possible withthe assumptions used in developing the toxicityvalues.

Averaging period for exposure. If the toxicityvalue is based on average lifetime exposure (e.g.,slope factors), then the exposure duration mustalso be expressed in those terms. For estimatingcancer risks, always use average lifetime exposure;i.e., convert less-than-lifetime exposures toequivalent lifetime values (see EPA 1986a,Guidelines for Carcinogen Risk Assessment). Onthe other hand, for evaluating potentialnoncarcinogenic effects of less-than-lifetimeexposures, do not compare chronic RfDs to short-term exposure estimates, and do not convertshort-term exposures to equivalent lifetime valuesto compare with the chronic RfDs. Instead, usesubchronic or shorter-term toxicity values toevaluate short-term exposures. Check that theestimated exposure duration is sufficiently similarto the duration of the exposure in the study usedto identify the toxicity value to be protective ofhuman health (particularly for subchronic and

5

shorter-term effects). A toxicologist should reviewthe comparisons. In the absence of short-termtoxicity values, the chronic RfD may be used as aninitial screening value; i.e., if the ratio of theshort-term exposure value to the chronic RtD isless than one, concern for potential adverse healtheffects is low. If this ratio exceeds unity, however,more appropriate short-term toxicity values areneeded to confirm the existence of a significanthealth threat. ECAO may be consulted forassistance in finding short-term toxicity values.

Exposure route. Check that all toxicity valuesused for each exposure pathway being evaluatedat the site are consistent with the route ofexposure (e.g., oral to oral, inhalation toinhalation). It is not possible to extrapolatebetween exposure routes for some substances thatproduce localized effects dependent upon theroute of exposure. For example, a toxicity valuebased on localized lung tumors that result onlyfrom inhalation exposure to a substance would notbe appropriate for estimating risk associated withdermal exposure to the substance. At this time,EPA considers it appropriate only to extrapolatedermal toxicity values from values derived for oralexposure. It is not recommended that oral toxicityreference values be extrapolated casually frominhalation toxicity values, although thisextrapolation may be performed on a case-by-casebasis in consultation with ECAO. In general,inhalation values should not be extrapolated fromoral values. (Also, see Section 7.5.1.)

Inhalation RfDi values obtained from IRISwill usually be expressed as ambient airconcentrations (i.e., mg/m3), instead of asadministered doses (i.e., mg/kg-day). It may benecessary, therefore, to calculate the RfDi in unitsof mg/kg-day for comparison with the intakeestimated in the exposure assessment. The RfDi

expressed in mg/kg-day would be equal to the

RDi in mg/m3 multiplied by 20 m3 air inhaledper person per day divided by 70 kg per person.

Absorption adjustment. Check that theexposure estimates and the toxicity values areeither both expressed as absorbed doses or bothexpressed as intakes (i.e., administered doses).Except for the dermal route of exposure, theexposure estimates developed using the methodsprovided in Chapter 6 should be in the form ofintakes, with no adjustments made for absorption.However, there are three types of absorptionadjustments that might be necessary orappropriate depending on the available toxicityinformation. These are described below. Samplecalculations for these absorption adjustments areprovided in Appendix A

(1) Dermal exposures. The output of theexposure assessment for dermal exposureis expressed as the amount of substanceabsorbed per kg body weight per day. Ittherefore may be necessary to derive anabsorbed dose toxicity value from anadntinistered dose toxicity value to comparewith the exposure estimate. See AppendixA for sample calculations.

(2) Absorbed-dose toxicitv value. For thesubstances for which the toxicity value isexpressed as an absorbed rather thanadministered dose (e.g., inhalation slopefactor in IRIS for trichloroethylene andseveral other substances), one shouldexpress exposure as an absorbed doserather than as an intake. See Appendix A-

(3) Adjustment for medium of exposure.Adjusting for different relative absorptionefficiencies based on the medium ofexposure (e.g., food, soil, or water for oralexposure, vapor or particulate forinhalation exposure) is occasionallyappropriate, but not generallyrecommended unless there are strongarguments for doing so. Many oral RfDand slope factor values assume ingestion inwater even when based on studies thatemployed administration in corn oil bygavage or in feed. Thus, in most cases, theunadjusted toxicity value will provide areasonable or conservative estimate of risk.See Appendix A.

6

8.2 QUANTIFYING RISKS

This section deacribes steps for quantifying riskor hazard indices for both carcinogenic andnoncarcinogenic effects to be applied to eachexposure pathway analyzed. The first subsectioncovers promdures for individual substances, andis followed by a subsection on procedures forquantifying risks associated with simultaneousexposures to several substances. Sample tableformats for recording the results of thesecalculations as well as recording associatedinformation related to uncertainty and absorptionadjustments are provided in Exhibits 8-2 through8-4.

8.2.1 CALCULATE RISKS FOR INDIVIDUALSUBSTANCES

Carcinogenic effects. For carcinogens, risksare estimated as the incremental probability of anindividual developing cancer over a lifetime as aresult of exposure to the potential carcinogen(i.e., incremental or excess individual lifetimecancer risk). The guidelines provided in thissection are consistent with EPA’s (1986a)Guidelines for Caarcinogen Risk Assessment. Forsome carcinogens, there may be sufficientinformation on mechanism of action that amodification of the approach outlined below iswarranted. Alternative approaches may beconsidered in consultation with ECAO on a case-by-case basis.

The slope factor (SF) converts estimated dailyintakes averaged over a lifetime of exposuredirectly to incremental risk of an individualdeveloping cancer. Because relatively low intakes(compared to those experienced by teat animals)are most likely from environmental exposures atSuperfund sites, it generally can be assumed thatthe dose-response relationship will be linear in thelow-dose portion of the multistage model dose-response curve. (See the Background Document2 of IRIS for a discussion of the multistagemodel.) Under this assumption, the slope factoris a constant, and risk will be directly related to

intake. Thus, the linear form of the carcinogenicrisk equation is usually applicable for estimatingSuperfund site risks. This linear low-doseequation is described in the next box.

However, this linear equation is valid only atlow risk levels (i.e., below estimated risks of 0.01).For sites where chemical intakes might be high(i.e., risk above 0.01), an alternate calculationequation should be used, The one-hit equation,which is consistent with the linear low-dose modelgiven above and described in the box on page8-11, should be used instead.

Because the slope factor is often an upper95th percentile confidence limit of the probabilityof response based on experimental animal dataused in the multistage model the carcinogenic riskestimate will generally be an upper-boundestimate. This means that EPA is reasonablyconfident that the “true risk” will not exceed therisk estimate derived through use of this modeland is likely to be less than that predicted

Noncarcinogenic effects. The measure used todescribe the potential for noncarcinogenic toxicityto occur in an individual is not expressed as theprobability of an individual suffering an adverseeffect. EPA does not at the present time use aprobabilistic approach to estimating the potentialfor noncarcinogenic health effects. Instead, the

EXHIBIT 3-2

EXAMPLE OF TABLE FORMAT FOR CANCER RISK ESTIMATES

Nearby Residential Population in Area Y -- Total Cancer Risk (weight of evidence predominantly B2)d

* Values for illustration only. SF = Slope Factora Identify type of cancer in this table for Class A carcinogens only. CDI = Chronic Daily Intakeb All cancer risks should be expressed as one significant figure only.C Slope factor based on dose administered in drinking water and resumed absorption fraction of 1.0.d Summarize weight of evidence for carcinogens contributing most to the total cancer risk estimate.

EXHIBIT 8-3

EXAMPLE OF TABLE FORMAT FOR CHRONIC HAZARD INDEX ESTIMATES

* Valucs for illustration only.

a All hazard indices and hazard quotients shouldbe expressed as one significant figure only.

b If the hazard index is greater than 1.0, seesection 8.2.2 for guidance on possiblesegregation of hazard index by endpoint.

c RfD expressed as administered dose.dUncertainty adjustment of 1,000 used to

represent combined H, A, S, & L extrapolationa.

Abbreviations for Uncertainty Adjustments: MF = Modifying factor for EPA verifiedFactor of 10 used for each adjustment RfDs. This factor represents profes-unless indicated otherwise. sional judgment on overall data base

not specifically addressed byH = variation in human sensitivity uncertainty adjustments.A = animal to human extrapolationS = extrapolation from subchronic to chronic NOAEL CDI = Chronic Daily IntakeL = extrapolation from LOAEL to NOAEL RfD = Chronic Reference Dose

Confidence Level: L = low, M = medium, H = high.

EXHIBIT 8-4

EXAMPLE OF TABLE FORMAT FOR SUBCHRONIC HAZARD INDEX ESTIMATES

* Values for illustration only. Abbreviations for Uncertainty Adjustments: MF = Modifying factor for EPA RtDss.Factor of 10 used for each adjustment,

a All hazard indices and hazard quotients shouldThis factor represents professional

unless indicated otherwise.be expressed as one significant figure only.

judgment on overall data base notspecifically addressed by uncertainty

b If hazard index is greater than 1.0, see H = variation in human sensitivity adjustments.Section 8.22 for guidance on possible A = animal to human extrapolationsegregation of hazard index by endpoint. L = extrapolation from LOAEL to NOAEL SDI = Subchronic Daily Intake

c RfDs expressed as administered dose. RfDs = Subchronic Reference Dose

11

potential for noncarcinogenic effects is evaluatedbycomparing an exposure level over a specifiedtime period (e.g., lifetime) with a reference dosederived for a similar exposure period. This ratioof exposure to toxicity is called a hazard quotientand is described in the box in the oppositecolumn.

The noncancer hazard quotient assumes thatthere is a level of exposure (i.e., RfD) belowwhich it is unlikely for even sensitive populationsto experience adverse health effects. If theexposure level (E) exceeds this threshold (i.e., ifE/RfD exceeds unity), there may be concern forpotential noncancer effects. As a rule, the greaterthe value of E/RfD above unity, the greater thelevel of concern. Be sure, however. not tointerpret ratios of E/RfD as statisticalprobabilities; a ratio of 0.001 does not mean thatthere is a one in one thousand chance of theeffect occurring. Further, it is important toemphasize that the level of concern does notincrease linearly as the RfD is approached orexceeded because RfDs do not have equalaccuracy or precision and are not based on thesame severity of toxic effects. Thus, the slopes ofthe dose-response curve in excess of the RfD canrange widely depending on the substance.

Three exposure durations that will needseparate consideration for the possibility ofadverse noncarcinogenic health effects are chronic,

subchronic, and shorter-term exposures. Asguidance for Superfund, chronic exposures forhumans range in duration from seven years to alifetime; such long-term exposures are almostalways of concern for Superfund sites (e.g.,inhabitants of nearby residences, year-round usersof specified drinking water sources). Subchronichuman exposures range in duration from twoweeks to seven years (as a Superfund programguideline) and are often of concern at Superfundsites. For example, children might attend a juniorhigh school near the site for no more than two orthree years. Exposures less than two weeks induration are occasionally of concern at Superfundsites. For example, if chemicals known to bedevelopmental toxicants are present at a site,short-term exposures of only a day or two can beof concern.

8.2.2 AGGREGATE RISKS FOR MULTIPLESUBSTANCES

At most Superfund sites, one must assesspotential health effects of more than one chemical(both carcinogens and other toxicants).Estimating risk or hazard potential by consideringone chemical at a time might significantlyunderestimate the risks associated withsimultaneous exposures to several substances. Toassess the overall potential for cancer andnoncancer effects posed by multiple chemicals,EPA (1986b) has developed Guidelines for theHealth Risk Assessment of Chemical Mixtures thatcan also be applied to the case of simultaneousexposures to several chemicals from a variety of

12

sources by more than one exposure pathway.Although the calculation procedures differ forcarcinogenic and noncarcinogenic effects, both setsof procedures assume dose additivity in theabsence of information on specific mixtures.

Information on specific mixtures found atSuperfund sites is rarely available. Even if suchdata exist, they are often difficult to use.Monitoring for “mixtures” or modeling themovement of mixtures across space and timepresent technical problems given the likelihoodthat individual components will behave differentlyin the environment (i.e., fate and transport). Ifdata are available on the mixtures present at thesite, but are not adequate to support aquantitative evaluation, note the information in

the “assumptions” documentation.

Carcinogenic effects. The cancer risk equationdescribed in the box below estimates theincremental individual lifetime cancer risk forsimultaneous exposure to several carcinogens andis based on EPA’s (1986a,b) risk assessmentguidelines. This equation represents anapproximation of the precise equation forcombining risks which accounts for the jointprobabilities of the same individual developingcancer as a consequence of exposure to two ormore carcinogens. The difference between theprecise equation and the approximation describedin the box is negligible for total cancer risks lessthan 0.1. Thus, the simple additive equation isappropriate for most Superfund risk assessments.

The risk summation techniques described inthe box on this page and in the footnote assumethat intakes of individual substances are small.They also assume independence of action by thecompounds involved (i.e., that there are nosynergistic or antagonistic chemical interactionsand that all chemicals produce the same effect,i.e., cancer). If these assumptions are incorrect,over- or under-estimation of the actual multiple-substance risk could result.

Calculate a separate total cancer risk for eachexposure pathway by summing the substance-specific cancer risks. Resulting cancer riskestimates should be expressed using one significantfigure only. Obviously, the total cancer risk foreach pathway should not exceed 1. Exhibit 8-2provides a sample table format for presentingestimated cancer risks for specified exposurepathways in the “Total Pathway Risk” column.

There are several limitations to this approachthat must be acknowledged. First, because eachslope factor is an upper 95th percentile estimateof potency, and because upper 95th percentiles ofprobability distributions are not strictly additive,the total cancer risk estimate might becomeartificially more conservative as risks from anumber of different carcinogens are summed. Ifone or two carcinogens drive the risk, however,this problem is not of concern. Second, it oftenwill be the case that substances with differentweights of evidence for human carcinogenicity areincluded. The cancer risk equation for multiplesubstances sums all carcinogens equally, giving asmuch weight to class B or C as to class Acarcinogens. In addition, slope factors derivedfrom animal data will be given the same weight asslope factors derived from human data. Finally,the action of two different carcinogens might notbe independent. New tools for assessingcarcinogen interactions are becoming available(e.g., Arcos et al. 1988), and should be consideredin consultation with the RPM. The significanceof these concerns given the circumstances at aparticular site should be discussed and presentedwith the other information described in Section8.6.

Noncarcinogenic effects. To assess the overallpotential for noncarcinogenic effects posed bymore than one chemical, a hazard index (HI)approach has been developed based on EPA’s

13

(1986b) Guidelines for Health Risk Assessment ofChemical Mixtures. This approach assumes that

simultaneous subthreshold exposures to severalchemicals could result in an adverse health effect.It also assumes that the magnitude of the adverseeffect will be proportional to the sum of the ratiosof the subthreshold exposures to acceptableexposures. The hazard index is equal to the sumof the hazard quotients, as described in the boxbelow, where E and the RfD represent the sameexposure period (e.g., subchronic, chronic, orshorter-term). When the hazard index exceedsunity, there may be concern for potential healtheffects. While any single chemical with anexposure level greater than the toxicity value willcause the hazard index to exceed unity, formultiple chemical exposures, the hazard index canalso exceed unity even if no single chemicalexposure exceeds its RfD.

It is important to calculate the hazard indexseparately for chronic, subchronic, and shorter-term exposure periods as described below. It isalso important to remember to include RfDs forthe noncancer effects of carcinogenic substances.

(1) Noncarcinogenic effects -- chronicexposures. For each chronic exposurepathway (i.e., seven year to lifetimeexposure), calculate a separate chronichazard index from the ratios of the chronicdaily intake (CDI) to the chronic reference

dose (RfD) for individual chemicals asdescribed in the box below. Exhibit 8-3provides a sample table format forrecording these results in the “PathwayHazard Index” column.

(2) Noncarcinogenic effects -- subchronicexposures. For each subchronic exposurepathway (i.e., two week to seven yearexposure), calculate a separate subchronichazard index from the ratios of thesubchronic daily intake (SDI) to thesubchronic reference dose (RfDs) forindividual chemicals as described in the boxon the next page. Exhibit 8-4 provides asample table format for recording theseresults in the “Pathway Hazard Index”column. Add only those ratioscorresponding to subchronic exposures thatwill be occurring simultaneously.

(3) Noncarcinogenic effects -- less than twoweek exposures. The same procedure maybe applied for simultaneous shorter-termexposures to several chemicals. Fordrinking water exposures, 1- and 10-dayHealth Advisories can be used as referencetoxicity values. Depending on availabledata, a separate hazard index might also becalculated for developmental toxicants(using RfDdts), which might cause adverse

14

effects following exposures of only a fewdays. See Guidelines for the HealthAssessment of Suspect DevelopmentalToxicants (EPA 1986G EPA 1989) forfurther guidance.

There are several limitations to this approachthat must be acknowledged. As mentioned earlier,the !evel of concern does not increase linearly asthe reference dose is approached or exceededbecause the RfDs do not have equal accuracy orprecision and are not based on the same severityof effect. Moreover, hazard quotients arecombined for substances with RfDs based oncritical effects of varying toxicological significance.Also, it will often be the case that RfDs ofvarying levels of confidence that include differentuncertainty adjustments and modifying factors willbe combined (e.g., extrapolation from animals tohumans, from LOAELS to NOAELS, from oneexposure duration to another).

Another limitation with the hazard indexapproach is that the assumption of dose additivityis most properly applied to compounds thatinduce the same effect by the same mechanism ofaction. Consequently, application of the hazardindex equation to a number of compounds thatare not expected to induce the same type ofeffects or that do not act by the same mechanism,although appropriate as a screening-levelapproach, could overestimate the potential for

effects. This possibility is generally not of concernif only one or two substances are responsible fordriving the HI above unity. If the HI is greaterthan unity as a consequence of summing severalhazard quotients of similar value, it would beappropriate to segregate the compounds by effectand by mechanism of action and to deriveseparate hazard indices for each group.

Segregation of hazard indices. Segregation ofhazard indices by effect and mechanism of actioncan be complex and time-consuming because it isnecessary to identify all of the major effects andtarget organs for each chemical and then toclassify the chemicals according to target organ(s)or mechanism of action. This analysis is notsimple and should be performed by a toxicologist.If the segregation is not carefully done, anunderestimate of true hazard could result. Agencyreview of particularly complex or controversialcases can be requested of ECAO through theregional risk assessment support staff.

The procedure for recalculating the hazardindex by effect and by mechanism of action isbriefly described in the box on the next page. Ifone of the effect-specific hazard indices exceedsunity, consideration of the mechanism of actionmight be warranted. A strong case is required,however, to indicate that two compounds whichproduce adverse effects on the same organ system(e.g., liver), although by different mechanisms,should not be treated as dose additive. Any suchdetermination should be reviewed by ECAO.

If there are specific data germane to theassumption of dose-additivity (e.g., if twocompounds are present at the same site and it isknown that the combination is five times moretoxic than the sum of toxicities for the twocompounds), then modify the development of thehazard index accordingly. Refer to the EPA(1986b) mixtures guidelines for discussion of ahazard index equation that incorporatesquantitative interaction data. If data on chemicalinteractions are available, but are not adequate tosupport a quantitative assessment, note theinformation in the “assumptions” beingdocumented for the site risk assessment.

15

PROCEDURE FOR SEGREGATION OFHAZARD INDICES BY EFFECT

Segregation of hazard indices requires identificationof the major effects of each chemical, including thoseseen at higher doses than the critical effect (e.g., thechemical may cause liver damage at a dose of 100mg/kg-day and neurotoxicity at a dose of 250 mg/kg-day). Major effect categories include neurotoxicity,developmental toxicity, reproductive toxicity,immunotoxicity, and adverse effects by target organ (i.e.,hepatic, renal, respiratory, cardiovascular,gastrointestinal, hematological, musculoskekal, anddermal/ocular effects). Although higher exposure levelsmay be required to produce adverse health effects otherthan the critical effect, the RfD can be used as thetoxicity value for each effect category as a conservativeand simplifying step.

INFORMATION SOURCES FORSEGREGATION OF HAZARD INDICES

Of the available information sources, the ATSDRToxicological Profiles are well suited in format andcontent to allow a rapid determination of additionalhealth effects that may occur at exposure levels higherthan those that produce the critical effect. Readersshould be aware that the ATSDR definitions ofexposure durations are somewhat different than EPA'sand are indepedent of species; acute - up to 14 days;intermediate -- more than 14 days to 1 year; chronic-- greater than one year. IRIS contains only limitedinformation on health effects beyond the critical effect,and EPA criteria documents and HEAs, HEEPs, andHEEDs may not systematically cover all health effectsobserved at doses higher those associated with the mostsensitive effects.

8.3 COMBINING RISKS ACROSSEXPOSURE PATHWAYS

This section gives directions for combining themulti-chemical risk estimates across exposurepathways and provides guidance for determiningwhen such aggregation is appropriate.

In some Superfund site situations, anindividual might be exposed to a substance orcombination of substances through severalpathways. For example, a single individual mightbe exposed to substance(s) from a hazardous waste

site by consuming contaminated drinking waterfrom a well, eating contaminated fish caught nearthe site, and through inhalation of dust originatingfrom the site. The total exposure to variouschemicals will equal the sum of the exposures byall pathways. One should not automatically sumrisks from all exposure pathways evaluated for asite, however. The following subsections describehow to identify exposure pathways that should becombined and, for these, how to sum cancer risksand noncancer hazard indices across multipleexposure pathways.

8.3.1 IDENTIFY REASONABLE EXPOSUREPATHWAY COMBINATIONS

There are two steps required to determinewhether risks or hazard indices for two or morepathways should he combined for a single exposedindividual or group of individuals. The first is toidentify reasonable exposure pathwaycombinations. The second is to examine whetherit is likely that the same individuals wouldconsistently face the “reasonable maximumexposure” (RME) by more than one pathway.

Identify exposure pathways that have thepotential to expose the same individual orsubpopulation at the key exposure areas evaluatedin the exposure assessment, making sure toconsider areas of highest exposure for eachpathway for both current and future land uses(e.g., nearest downgradient well, nearest downwindreceptor). For each pathway, the risk estimatesand hazard indices have been developed for aparticular exposure area and time period; they donot necessarily apply to other locations or timeperiods. Hence, if two pathways do not affect thesame individual or subpopulation, neitherpathway’s individual risk estimate or hazard indexaffects the other, and risks should not becombined.

Once reasonable exposure pathwaycombinations have been identified, it is necessaryto examine whether it is likely that the sameindividuals would consistently face the RME asestimated by the methods described in Chapter 6.Remember that the RME estimate for eachexposure pathway includes many conservative andupper-bound parameter values and assumptions(e.g., upper 95th confidence limit on amount ofwater ingested, upper-bound duration of occupancy

16

of a single residence). Also, some of the exposureparameters are not predictable in either space ortime (e.g., maximum downwind concentration mayshift compass direction, maximum ground-waterplume concentration may move past a well). Forreal world situations in which contaminantconcentrations vary over time and space, the sameindividual may or may not experience the RMEfor more than one pathway over the same periodof time. One individual might face the RMEthrough one pathway, and a different individualface the RME through a different pathway. Onlyif you can explain why the key RME assumptionsfor more than one pathway apply to the sameindividual or subpopulation should the RME risks

for more than one pathway be combined.

In some situations, it may be appropriate tocombine one pathway’s RME risks with otherpathways’ risk estimates that have been derivedborn more typical exposure parameter values. Inthis way, resulting estimates of combined pathwayrisks may better relate to RME conditions.

If it is deemed appropriate to sum risks andhazard indices across pathways, the risk assessorshould clearly identify those exposure pathwaycombinations for which a total risk estimate orhazard index is being developed. The rationalesupporting such combinations should also beclearly stated. Then, using the methods describedin Sections 8.3.2 and 8.3.3, total cancer riskestimates and hazard indices should be developedfor the relevant exposure areas and individuals (orsubpopulations). For example, Exhibits 8-2 and8-3 illustrate the combination of cancer riskestimates and chronic noncancer hazard indices,respectively, for a hypothetical nearby residentialpopulation exposed to contaminants from a siteby two exposure pathways: drinking contaminatedground water from private wells and ingestion ofcontaminated fish caught in the local river. Inthis hypothetical example, it is “known” that thefew families living next to the site consume morelocally caught fish than the remaining communityand have the most highly contaminated wells ofthe area.

The following two subsections describe how tosum risks and hazard indices for multiple exposurepathways for carcinogenic and noncarcinogenicsubstances, respectively.

8.3.2 SUM CANCER RISKS

First, sum the cancer risks for each exposurepathway contributing to exposure of the sameindividual or subpopulation. For Superfund riskassessments, cancer risks from various exposurepathways are assumed to be additive, as long asthe risks are for the same individuals and timeperiod (i.e., less-than-lifetime exposures have allbeen converted to equivalent lifetime exposures).This summation is described in the box below.The sample table format given in Exhibit 8-2provides a place to record the total cancer riskestimate.

As described in Section 8.2.2, although theexact equation for combining risk probabilitiesincludes terms for joint risks, the differencebetween the exact equation and the approximationdescribed above is negligible for total cancer risksof less than 0.1.

8.3.3 SUM NONCANCER HAZARD INDICES

To assess the overall potential fornoncarcinogenic effects posed by several exposurepathways, the total hazard index for each exposureduration (i.e., chronic, subchronic. and shorter-term) should be calculated separately. Thisequation is described in the box on the next page.The sample table format given in Exhibit 8-3provides a place to record the total exposurehazard index for chronic exposure durations.

When the total hazard index for an exposedindividual or group of individuals exceeds unity,there may be concern for potential noncancerhealth effects. For multiple exposure pathways,the hazard index can exceed unity even if nosingle exposure pathway hazard index exceedsunity. If the total hazard index exceeds unity and

’ Page 8-17

if combining exposure pathways has resulted incombining hazard indices based on differentchemicals, one may need to consider segregatingthe contributions of the different chemicalsaccording to major effect (see Section 8.2.2.).

8.4 ASSESSMENT ANDPRESENTATION OFUNCERTAINTY

This section discusses practical approaches toassessing uncertainty in Superfund site riskassessments and describes ways to present keyinformation bearing on the level of confidence inquantitative risk estimates for a site. The riskmeasures used in Superfund site risk assessmentsusually are not fully probabilistic estimates of riskbut conditional estimates given a considerablenumber of assumptions about exposure andtoxicity (e.g., risk given a particular future landuse). Thus, it is important to fully specify theassumptions and uncertainties inherent in the riskassessment to place the risk estimates in properperspective. Another use of uncertaintycharacterization can be to identify areas where amoderate amount of additional data collectionmight significantly improve the basis for selectionof a remedial alternative.

Highly quantitative statistical uncertaintyanalysis is usually not practical or necessary forSuperfund site risk assessments for a number of

reasons, not the least of which are the resourcerequirements to collect and analyze site data insuch a way that the results can be presented asvalid probability distributions. As in allenvironmental risk assessments, it already isknown that uncertainty about the numericalresults is generally large (i.e., on the range of atleast an order of magnitude or greater).Consequently, it is more important to identify thekey site-related variables and assumptions thatcontribute most to the uncertainty than toprecisely quantify the degree of uncertainty in therisk assessment. Thus, the focus of this section ison qualitative/semi-quantitative approaches thatcan yield useful information to decision-makers fora limited resource investment.

There are several categories of uncertaintiesassociated with site risk assessments. One is theinitial selection of substances used to characterizeexposures and risk on the basis of the samplingdata and available toxicity information. Othersources of uncertainty are inherent in the toxicityvalues for each substance used to characterize risk.Additional uncertainties are inherent in theexposure assessment for individual substances andindividual exposures. These uncertainties areusually driven by uncertainty in the chemicalmonitoring data and the models used to estimateexposure concentrations in the absence ofmonitoring data, but can also be driven bypopulation intake parameters. Finally, additionaluncertainties are incorporated in the riskassessment when exposures to several substancesacross multiple pathways are summed.

The following subsections describe how tosummarize and discuss important site-specificexposure uncertainties and the more generaltoxicity assessment uncertainties.

8.4.1 IDENTIFY AND EVALUATElMPORTANT SITE-SPECIFICUNCERTAINTY FACTORS

Uncertainties in the exposure assessmenttypically include most of the site-specificuncertainties inherent in risk characterization, andthus are particularly important to summarize foreach site. In risk assessments in general, and inthe exposure assessment in particular, severalsources of uncertainty need to be addressed: (1)definition of the physical setting, (2) model

18

applicability and assumptions, (3) transport, fate,and exposure parameter values, and (4) trackinguncertainty, or how uncertainties are magnifiedthrough the various steps of the assessment.Some of these sources of uncertainty can bequantified while others are best addressedqualitatively.

Definition of the physical setting. The initialcharacterization of the physical setting that definesthe risk assessment for a Superfund site involvesmany professional judgments and assumptions.These include definition of the current and futureland uses, identification of possible exposurepathways now and in the future, and selection ofsubstances detected at the site to include in thequantitative risk assessment. In Superfund riskassessments, particular attention should be givento the following aspects of the definition of thephysical setting.

Likelihood of exposure pathwavs and landuses actually occurring. A large part of therisk assessment is the estimation of cancerrisks or hazard indices that are conditionalon the existence of the exposure conditionsanalyzed; e.g., if a residential developmentis built on the site 10 years from now, thehealth risks associated with contaminantsfrom the site would be X. It is importantto provide the RPM or other risk managerwith information related to the likelihoodthat the assumed conditions will occur toallow interpretation of a conditional riskestimate in the proper context. Forexample, if the probability that a residentialdevelopment would be built on the site 10or 50 years from now is very small,different risk management decisions mightbe made than if the probability is high.Present the information collected duringscoping and for the exposure assessmentthat will help the RPM to identify therelative likelihood of occurrence of eachexposure pathway and land use, at leastqualitatively (e.g., institutional land-usecontrols, zoning, regional developmentplans).

The chemicals not included in thequantitative risk estimate as a consequenceof missing information on health effects orlack of quantitation in the chemical

analysis may represent a significant sourceof uncertainty in the final risk estimates.If chemicals with known health effects wereeliminated from the risk assessment on thebasis of concentration or frequency ofdetection, one should now review andconfirm whether or not any of thechemicals previously eliminated shouldactually be included. For substancesdetected at the site, but not included in thequantitative risk assessment because of datalimitations, discuss possible consequencesof the exclusion on the risk assessment.

A checklist of uncertainty factors related to thedefinition of the physical setting is described inthe box below.

Model applicability and assumptions. Thereis always some doubt as to how well an exposuremodel or its mathematical expression (e.g.,ground-water transport model) approximates thetrue relationships between site-specificenvironmental conditions. Ideally, one would liketo use a fully validated model that accounts for allthe known complexities in the parameter

19

interrelationships for each assessment. At present,however, only simpIe, partially validated modelsare available and commonly used. As aconsequence, it is important to identify key modelassumptions (e.g., linearity, homogeneity, steady-state conditions, equilibrium) and their potentialimpact on the risk estimates. In the absence offield data for model validation, one could performa limited sensitivity analysis (i.e., vary assumptionsabout functional relationships) to indicate themagnitude of uncertainty that might be associatedwith model form. At a minimum, one should listkey model assumptions and indicate the potentialimpact of each on risk with respect to bothdirection and magnitude, as shown in the boxbelow. A sample table format is presented inExhibit 6-21 of Chapter 6.

Parameter value uncertainty. During thecourse of a risk assessment, numerous parametervalues are included in the calculations of chemicalfate and transport and human intake. A first stepin characterizing parameter value uncertainty inthe baseline risk assessment is to identify the keyparameters influencing risk. This usually can beaccomplished by expert opinion or by an explicitsensitivity analysis. In a sensitivity analysis, thevalues of parameters suspected of driving the risksare varied and the degree to which changes in theinput variables result in changes in the riskestimates are summarized and compared (e.g., theratio of the change in output to the change ininput). It is important to summarize theuncertainty associated with key parameters, asdescribed below.

Significant site data gaps might haverequired that certain parameter values beassumed for the risk assessment. For

example, no information on the frequencywith which individuals swim in a nearbystream might be available for a site, and anassumed frequency and duration ofswimming events based on a nationalaverage could have driven the exposureestimate for this pathway.

Significant data uncertainties might existfor other parameters, for example, whetheror not the available soil concentrationmeasurements are representative of thetrue distribution of soil contaminantconcentrations.

Tracking uncertainty. Ideally, one would liketo carry through the risk assessment theuncertainty associated with each parameter inorder to characterize the uncertainty associatedwith the final risk estimates. A more practicalapproach for Superfund risk assessments is todescribe qualitatively how the uncertainties mightbe rnagnified or biased through the risk modelsused. General quantitative, semi-quantitative, andqualitative approaches to uncertainty analysis aredescribed below.

Quantitative approach. Only on the rareoccasions that an RPM may indicate the need fora quantitative uncertainty analysis should one beundertaken. As mentioned earlier, a highlyquantitative statistical uncertainty analysis isusually not practical or necessary for Superfundsites.

If a quantitative analysis is undertaken for asite, it is necessary to involve a statistician in thedesign and interpretation of that analysis. Aquantitative approach to characterizing uncertaintymight be appropriate if the exposure models aresimple and the values for the key inputparameters are well known. In this case, the firststep would be to characterize the probabilitydistributions for key input parameter values(either using measured or assumed distributions).The second step would be to propagate parametervalue uncertainties through the analysis usinganalytic (e.g., first-order Taylor seriesapproximation) or numerical (e.g., Monte Carlosimulation) methods, as appropriate. Analyticmethods might be feasible if there are a fewparameters with known distributions and linearrelationships. Numerical methods (e.g., Monte

20

Carlo simulation) can be suitable for morecomplex relationships, but must be done on acomputer and can be resource intensive even withtime-saving techniques (e.g., Latin Hypercubesampling).

Two common techniques of propagatinguncertainty are first-order analyses and MonteCarlo simulations. First-order analysis is based onthe assumption that the total variance of a modeloutput variable is a function of the variances ofthe individual model input variables and thesensitivity of the output variable to changes ininput variables. The sensitivity of the outputvariable is defined by the first derivative of thefunction or model, which can be generatedanalytically or numerically. A Monte Carlosimulation estimates a distribution of exposures orrisk by repeatedly solving the model equation(s).The probability distribution for each variable inthe model must be defined. The computer selectsrandomly from each distribution every time theequation is solved. From the resulting outputdistribution of exposures or risk, the assessor canidentify the value corresponding to any specifiedpercentile (e.g., the 95th percentile in theexposure distribution).

These quantitative techniques requiredefinition of the distribution of all inputparameters and knowledge of the degree ofdependence (i.e., covariance) among parameters.The value of first-order analyses or Monte Carlosimulations in estimating exposure or riskprobability distributions diminishes sharply if oneor more parameter value distributions are poorlydefined or must be assumed. These techniquesalso become difficult to document and to reviewas the number of model parameters increases.Moreover, estimating a probability distribution forexposures and risks can lead one into a false senseof certainty about the analysis. Even in the mostcomprehensive analyses, it will generally be truethat not all of the sources of uncertainty can beaccounted for or all of the parametercodependencies recognized. Therefore, in additionto documenting all input distributions and

covariances, it is very important to identify all ofthe assumptions and incomplete information thathave not been accounted for in the quantitativeuncertainty analysis (e.g., likelihood that aparticular land use will occur) when presenting theresults.

References describing numerical methods ofpropagating uncertainty through a risk analysisinclude Burmaster and von Stackelberg (1988),Hoffman and Gardner (1983), Iman and Helton(1988), and NRC (1983). References describinganalytic methods of tracking uncertainty includeHoffman and Gardner (1983), NRC (1983),Downing et al. (1985), and Benjamin and Cornell(1970).

Semi-quantitative approach. Often availabledata are insufficient to fully describe parameterdistributions, but are sufficient to describe thepotential range of values the parameters mightassume. In this situation, sensitivity analyses canbe used to identify influential model inputvariables and to develop bounds on thedistribution of exposure or risk. A sensitivityanalysis can estimate the range of exposures orrisk that result from combinations of minimumand maximum values for some parameters andmid-range values for others. The uncertainty foran assessment of this type could be characterizedby presenting the ranges of exposure or riskgenerated by the sensitivity analysis and bydescribing the limitations of the data used toestimate plausible ranges of model input variables(EPA 1985).

Qualitative approach. Sometimes, a qualitativeapproach is the most’ practical approach todescribing uncertainty in Superfund site riskassessments given the use of the information (e.g.,identifying areas where the results may bemisleading). Often the most practical approachto characterizing parameter uncertainty will be todevelop a quantitative or qualitative description ofthe uncertainty for each parameter and to simplyindicate the possible influence of theseuncertainties on the final risk estimates givenknowledge of the models used (e.g., a specificground-water transport model). A checklist ofuncertainty factors related to the definition ofparameters is described in the box on page 8-22.A sample table format is provided in Exhibit6-21 of Chapter 6.

Consider presentation of information on keyparameter uncertainties in graphic form toillustrate clearly to the RPM or other riskmanagers the significance of various assumptions.For example, Exhibit 8-5 plots assumptionsregarding contaminated fish ingestion and resulting

21

EXHIBIT 8-5

EXAMPLE OF PRESENTATION OF IMPACT OF EXPOSURE ASSUMPTIONSON CANCER RISK ESTIMATE

Ingestion of Fish Contaminated with Chemical X

(30 mg X/Kg Fish Wet Weight)

aThe risk of developing cancer is plotted on a log scale. A risk of 10 -4 indicates a probability

of 1 chance in 10,000 and a risk of 10-sindicates a probability of 1 chance in 100,000 of an

individual developing cancer.

22

impacts on the cancer risk estimate for thisexposure pathway. Exhibit 8-6 illustrates thesignificance of these same assumptions for thehazard index estimates for contaminated fishconsumption. Additionally, maps showingisopleths of risks resulting from modeled airexposures such as emissions near the site mayassist the RPM or risk manager in visualizing thesignificance of current or future site risks for acommunity.

8.4.2 IDENTIFY AND EVALUATE TOXICITYASSESSMENT UNCERTAINTYFACTORS

For substances that contribute most to theestimates of cancer risk and noncancer hazardindices, summarize the uncertainty inherent in thetoxicity values for the durations of exposureassessed. Some of the information (e.g., weight ofevidence for potential human carcinogens,uncertainty adjustments for noncancer toxicity

values) has already been recorded in the sampletable formats provided in Exhibits 8-2 through8-4. Other information will be developed duringthe toxicity assessment itself (see Chapter 7).The box on page 8-24 provides a checklist ofuncertainties that apply to most toxicityassessments.

Multiple substance exposure uncertainties.Uncertainties associated with summing risks orhazard indices for several substances are ofparticular concern in the risk characterization step.The assumption of dose additivity ignores possiblesynergisms or antagonisms among chemicals, andassumes similarity in mechanisms of action andmetabolism. Unfortunately, data to assessinteractions quantitatively are generally lacking.In the absence of adequate information, EPAguidelines indicate that carcinogenic risks shouldbe treated as additive and that noncancer hazardindices should also be treated as additive. Theseassumptions are made to help prevent anunderestimation of cancer risk or potentialnoncancer health effects at a site.

Be sure to discuss the availability ofinformation concerning potential antagonistic orsynergistic effects of chemicals for which cancerrisks or hazard indices have been summed for thesame exposed individual or subpopulations. Onthe basis of available information concerningtarget organ specificity and mechanism of action,indicate the degree to which treating the cancerrisks as additive may over- or under-estimate risk.If only qualitative information is availableconcerning potential interactions or dose-additivityfor the noncarcinogenic substances, discusswhether the information indicates that hazardindices may have been over- or under-estimated.This discussion is particularly important if thetotal hazard index for an exposure point is slightlybelow or slightly above unity, or if the totalhazard index exceeds unity and the effect-specifichazard indices are less than unity, and if theuncertainty is likely to significantly influence therisk management decision at the site.

8.5 CONSIDERATION OF SITE-SPECIFIC HUMAN STUDIES

This section describes how to compare theresults of the risk characterization step with

23

EXHIBIT 8-6

EXAMPLE OF PRESENTATION OF IMPACT OF EXPOSURE ASSUMPTIONSON HAZARD INDEX ESTIMATE

Ingestion of Fish Contaminated with Chemical Y(10 mg Y/Kg Fish Wet Weight)

---- Fillet with Skin — F i l l e t O n l y

24

ATSDR health assessments and other site-specifichuman studies that might be available. The firstsubsection outlines how to compare an ATSDRhealth assessment for the site with the risk resultssummarized in the previous sections (Sections 8.2,8.3, and 8.4). The second subsection discusseswhen epidemiological or health studies mightprovide useful information for assessing exposuresand health risks associated with contaminantsfrom a site.

8.5.1 COMPARE WITH ATSDR HEALTHASSESSMENT

ATSDR health assessments were defined andcompared to the RI/FS risk assessment in Section2.2.2. As of 1989, preliminary ATSDR healthassessments should be completed before the RI/FSrisk assessment is initiated and therefore shouldbe available to the risk assessor as early as“scoping.” The steps for comparing the

preliminary ATSDR health assessment with thebaseline risk assessment are outlined below.

Review again the ATSDR health assessmentfindings and conclusions. These will be largelyqualitative in nature. If the ATSDR healthassessment identifies exposure pathways orchemicals of concern that have not been includedin the RI/FS baseline risk assessment, describe theinformation supporting the decision not to includethese parameters. If there are differences in thequalitative conclusions of the health assessmentand the quantitative conclusions of the baselinerisk assessment, explain the differences, if possible,and discuss their implications.

8.5.2 COMPARE WITH OTHER AVAILABLESITE-SPECIFIC EPIDEMIOLOGICALOR HEALTH STUDIES

For most Superfund sites, studies of humanexposure or health effects in the surroundingpopulation will not be available. However, ifcontrolled epidemiological or other health studieshave been conducted, perhaps as a consequenceof the preliminary ATSDR health assessment orother community involvement, it is important toinclude this information in the baseline riskassessment as appropriate. However, not all suchstudies provide meaningful information in thecontext of Superfund risk assessments.

One can determine the availability of otherepidemiological or health studies for populationspotentially exposed to contaminants from the siteby contacting the ATSDR RegionalRepresentative, the Centers for Disease Controlin Atlanta, Georgia, and state and local healthagencies as early in the risk assessment process aspossible. It is important to avoid use of anecdotalinformation or data from studies that mightinclude a significant bias or confounding factor,however. Isolated reports of high body levels ofsubstances that are known to be present at thesite in a few individuals living near the site arenot sufficient evidence to confirm the hypothesisthat these individuals have received significantexposures from the site. Nor can isolated reportsof disease or symptoms in a few individuals livingnear the site be used to confirm the hypothesisthat the cause of the health effects in theseindividuals was exposure to contamination fromthe site. A trained epidemiologist should review

25

anv available studies in order to identifv possiblestudy limitations and implications for site riskfindings. The small populations and variableexposures predominating at most Superfund siteswill make it extremely difficult to detect site-related effects using epidemiological techniques.

If site-specific health or exposure studies havebeen identified and evaluated as adequate, oneshould incorporate the study findings into theoverall risk characterization to strengthen theconclusions of the risk assessment (e.g., the riskassessment predicts elevated blood lead levels andthe human exposure study documented elevatedblood lead levels only among those exposed toground water contaminated by the site). Becauseof the generally large and different types ofuncertainties associated with the risk assessmentand actual health studies, a qualitative, notquantitative, comparison between the two types ofstudies is generally warranted. Areas ofagreement and disagreement between the healthstudy(ies) and the risk assessment should bedescribed and factors that might contribute to anydisagreement discussed.

8.6 SUMMARIZATION ANDPRESENTATION OF THEBASELINE RISKCHARACTERIZATIONRESULTS

This section provides guidance on interpretingand presenting the risk characterization results.The results of the baseline evaluation should notbe taken as a characterization of absolute risk.An important use of the risk and hazard indexestimates is to highlight potential sources of riskat a site so that they may be dealt with effectivelyin the remedial process. It is the responsibility ofthe risk assessment team to develop conclusionsabout the magnitude and kinds of risk at the siteand the major uncertainties affecting the riskestimates. It is not the responsibility of the riskassessment team to evaluate the significance of therisk in a program context, or whether and howthe risk should be addressed, which are riskmanagement decisions.

The ultimate user of the risk characterizationresults will be the RPM or other risk manager for

the site. This section therefore outlines apresentation of material that is designed to assistthe risk manager in using risk information toreach site-specific decisions.

8.6.1 SUMMARIZE RISK INFORMATION INTEXT

The final discussion of the risk characterizationresults is a key component of the riskcharacterization. The discussion provides a meansof placing the numerical estimates of risk andhazard in the context of what is known and whatis not known about the site and in the context ofdecisions to be made about selection of remedies.At a minimum, the discussion should include:

confidence that the key site-relatedcontaminants were identified and discussionof contaminant concentrations relative tobackground concentration ranges;

a description of the various types of cancerand other health risks present at the site(e.g., liver toxicity, neurotoxicity),distinguishing between known effects inhumans and those that are predicted tooccur based on animal experiments;

level of confidence in the quantitativetoxicity information used to estimate risksand presentation of qualitative informationon the toxicity of substances not includedin the quantitative assessment

level of confidence in the exposureestimates for key exposure pathways andrelated exposure parameter assumptions;

the magnitude of the cancer risks andnoncancer hazard indices relative to theSuperfund site remediation goals in theNCP (e.g., the cancer risk range of 10-4 to10-7 and noncancer hazard index of 1.0);

the major factors driving the site risks (e.g.,substances, pathways, and pathwaycombinations);

the major factors reducing the certainty inthe results and the significance of theseuncertainties (e.g., adding risks over severalsubstances and pathways);

26

exposed population characteristics, and

comparison with site-specific health studies,when available.

In addition, if the size of the potentiallyexposed population is large, the presentation ofpopulation numbers may be of, assistance to theRPM, especially in evaluating risks in the contextof current land use. Individual risk estimatesbaaed on the reasonable maximum exposure(RME) should not be presented as representativeof a broadly defined population, however.

8.6.2 SUMMARIZE RISK INFORMATION INTABLES

A tabular summary of the cancer risks andnoncancer hazard indices should be prepared forall exposure pathways and land uses analyzed andfor all substances carried through the riskassessment. These tables must be accompanied byexplanatory text, as described in the previoussection, and should not be allowed to stand aloneas the entire risk characterization. The sampletable formats presented in Chapter 6 and inExhibits 8-2 to 8-6 provide basic summary formats.Exhibits 8-7 and 8-8 provide examples of optionalpresentations that might assist in visualization ofthe risk assessment results. These bar graphspresent the baseline cancer risk estimates and

noncancer hazard indices, respectively, by pathwayfor an identified subpopulation near the site. Thestacked bars in Exhibit 8-8 allow the reader toimmediately identify the pathway(s) contributingmost to the total hazard index as well as identifythe substances driving the indices in each pathway.Reference levels are also provided (e.g., hazardindex of 1.0). Exhibits 8-5 and 8-6 introduced inSection 8.4.1 provide examples of figures thatcould help the RPM or other risk managervisualize the impact of various assumptions anduncertainties on the final risk or hazard indexestimate. In addition, graphics relating risk level(or magnitude of hazard index) to concentrationsof substances in environmental media and cost of“treatment” could allow the RPM or other riskmanager to weigh the benefits of various remedialalternatives more easily. Examples of the last typeof graphics are presented in Part C of thismanual.

In a few succinct concluding paragraphs,summarize the results of the risk characterizationstep. It is the responsibility of the risk assessmentteam members, who are familiar with all steps inthe site risk assessment, to highlight the majorconclusions of the risk assessment. The discussionshould summarize both the qualitative and thequantitative findings of cancer risks and noncancerhazards, and properly qualify these by mention ofmajor assumptions and uncertainties in theassessment.

27

EXHIBIT 8-7

EXAMPLE OF PRESENTATION OF RELATIVE CONTRIBUTION OF INDIVIDUALCHEMICALS TO EXPOSURE PATHWAY AND TOTAL CANCER RISK ESTIMATES

Nearby Resident Population

Excess Lifetime Cancer Risks 3 x 10 -4

aThe risk of developing cancer is plotted on a log scale. A risk of 10-4 indicates a probabilityof 1 chance in 10,000 of an individual developing cancer. Risks of 10-5and 10-6 correspond toprobabilities of 1 chance in 100,000 and 1 chance in 1,000,000, respectively. Values inparentheses represent EPAs weight-of-evidence classification of the agent as a potentialhuman carcinogen: A = human carcinogen; and B2 = probable human carcinogen(with sufficient evidence in animals and inadequate or no evidence in humans).

28

EXHIBIT 8-8

EXAMPLE OF PRESENTATION OF RELATIVE CONTRIBUTION OF INDIVIDUALCHEMICALS TO EXPOSURE PATHWAY AND TOTAL HAZARD INDEX ESTIMATES

Nearby Resident Population

Chronic Hazard Index = 0.6

aThe hazard index is equal to the sum of the hazard quotients (i.e., exposure

level/RfD) for each chemical. It is not a probability; a hazard index or

quotient of <1.0 indicates that it is unlikely for even sensitive populations to

experience adverse health effects.

29


1. The probability of an individual developing cancer following exposure to more than one carcinogen is the probability of developingcancer from at least one of the carcinogen. For two carcinogens, the precise equation for estimating this probability is riakl + riak2 -probability (riskl, risk2) where the latter term is the joint probability of the two risks occurring in the same individual. If the risk toagent 1 is distributed in the population independently of the risk to agent 2, the latter term would equal (riak l)(risk2). This equationcan be expanded to evaluate risks from more than two substances.

30


Arcos, J., Woo, Y.T., and Lai, D. 1988: Data Base on Binary Combination Effects of Chemical Carcinogens. Environ. Carcino.Revs. [J. Environ. Sci. Health Pt. C] 6:1-150.

Benjamin, J.R. and C.A. Cornell. 1970. Probability, Statistics, and Decision-making for Civil Engineers. McGraw Hill. New York.

Burmaster, D.E. and K von Stackelberg. 1988. A New Method for Uncertainty and Sensitivity Analysis in Public Health RiskAssessments at Hazardous Waste Sites Using Monte Carlo Techniques in a Spreadsheet. Pages 550-556 in Superfund ’88,Proceedings of the 9th National Conference. Washington, D.C. Sponsored by the Hazardous Materials Control Research Institute.

Downing, D. J., Gardner, R. H., and Hoffman, F. O. 1985. Response Surface Methodologies for Uncertainty Analysis in AssessmentModels. Technometrics 27:151-163.

Environmental Protection Agency (EPA). 1985. Methodology for Characterization of Uncertainty in Expos ure Assessments. Preparedby Research Triangle Institute. NTIS: PB85-240455.


EnviromentalProtection Agency (EPA). 1986b. Guidelines for the Health Risk Assessment of Chemical Mixtures. 51 FederalRegister 34014 (September 24, 1986).

Environmental Protection Agency (EPA). 1986c. Guideline for the Health Assessment of Suspect Developmental Toxicants. 51Federal Register 34028 (September 24, 1986).

Environmental Protection Agency (EPA). 1989. Proposed Amendments to the Guidelines for the Health Assessment of SuspectDevelopmental Toxicants. 54 Federal Register 93S6 (March 6, 1989).

Hoffman, F. O. and R. H. Gardner. 1983. Evaluation of Uncertainties in Radiological Assessment Models. In RadiologicalAssessment. A Textbook on Environmental Dose Analysis. Till, J. E., and H.R. Meyer, (eds.). Prepared for Office of NuclearReactor Regulation, U.S. Nuclear Regulatory Commission. Washington, DC. NRC FIN B0766. NUREG/CR-3332.

Iman, R. L and J. C. Helton. 1988. An Investigation of Uncertainty and Sensitivity Analysis Techniques for Computer Models.Risk Analysis 871-90.

IRIS. Integrated Risk Information System (data base). 1989. U.S. Environmental Protection Agency, Office of Research andDevelopment.

Metcalf, D.R. and J.W. Pegram. 1981. Uncertainty Propagation in Probabilistic Risk Assessment: A Comparative Study. Transactionsof the American Nuclear Society 38:483-484.

Nuclear Regulatoty Commission (NRC). 1983. PRA Procedures Guide - A Guide to the Performance of Probabilistic RiskAssessments for Nuclear Power Plants. Office of Nuclear Regulatory Research, Washington, D.C. NUREG/CR-2300. Vol. 2

Vesefy, W. E. and D. M. Rasmuson. 1984. Uncertainties in Nuclear Probabilistic Risk Analysis Risk Analysis 4:313-322.

CHAPTER 9

DOCUMENTATION, REVIEW, ANDMANAGEMENT TOOLS FOR

THE RISK ASSESSOR,REVIEWER, AND MANAGER

This chapter provides tools for thedocumentation, review, and management of thebaseline risk assessment. These tools will helpensure completeness and consistency throughoutthe risk assessment and in the reporting ofassessment results. Section 9.1 providesdocumentation tools (for risk assessors), Section9.2 provides review tools (for risk assessmentreviewers), and Section 9.3 provides managementtools (for remedial project managers [RPMs] andother decision-makers concerned with the site).

9.1 DOCUMENTATION TOOLS

Throughout Chapters 4 to 8 of this manual,guidance is provided to the risk assessor on howto summarize and document many beginning,intermediate, and final steps of the riskassessment. The purpose of this section is toconsolidate that guidance, provide a final check toensure that all appropriate documentation hasbeen completed, and provide additionalinformation that should be helpful. This sectionaddresses (1) basic principles of documenting aSuperfund site risk assessment (e.g., key “dos” anddon’ts”, the rationale for consistency), (2) asuggested outline and guidance for the riskassessment report, and (3) guidance for providingrisk assessment summaries in other key reports.

9.1.1 BASIC PRINCIPLES

There are three basic principles fordocumenting a baseline risk assessment:

(1)

(2)

(3)

address the main objectives of the riskassessment

communicate using clear, concise, andrelevant text, graphics, and tables; and

use a consistent format.

Addressing the objectives. The objectives ofthe baseline risk assessment -- to help determinewhether additional response action is necessary atthe site, to provide a basis for determiningresidual chemical levels that are adequatelyprotective of public health, to provide a basis forcomparing potential health impacts of variousremedial alternatives, and to help supportselection of the “no-action” remedial alternative(where appropriate) -- should be consideredcarefully during the documentation of the, riskassessment. Recognizing these objectives earlyand presenting the results of the risk assessmentwith them in mind will assist the RPM and otherdecision-makers at the site with readily obtainingand using the necessary information to evaluatethe objectives. Failing to recognize theimportance of the objectives could result in a riskassessment report that appears misdirected and/orunnecessary.

Communicating. Clearly and conciselycommunicating the relevant results of the riskassessment can be one of the most importantaspects of the entire RI/FS. If done correctly, auseful instrument for mitigating public healththreats will have been developed. If doneincorrectly, however, risks could beunderemphasized, possibly leading to the

2

occurrence of adverse health effects, or they couldbe overemphasized, possibly leading to theunnecessary expenditure of limited resources. Seethe box below for some helpful hints oncommunicating the baseline risk assessment.

Many skills for communicating the baselinerisk assessment also can be learned by reviewingthe literature on risk communication. Thefollowing box lists just some of the literature thatis available. Courses on the subject also exist.

Using a consistent format. A consistentformat for all Superfund risk assessments is

strongly recommended for four important reasons:

(1) it encourages consistency andcompleteness in the assessment itself;

(2)

(3)

(4)

it allows for easier review of the riskassessments;

it encourages consistent use of theresults by RPMs and other decision-makers; and

it helps demonstrate to the public andothers that risk assessments areconducted using the same framework (ifnot the same specific procedures).

Using other formats can lead to slower reviewtimes, different interpretations of similar results,and the charge that risk assessments areinappropriately being conducted differently fromone site to another. The following subsectionsprovide guidance on the use of consistent formats.

9.1.2 BASELINE RISK ASSESSMENTREPORT

The baseline risk assessment report referencesand supports the RI/FS report. Depending on thesite, the risk assessment report can range from asmall, simple document with no appendices thatcan simply be added to the RI/FS report as achapter, to a large, complex document with manyappendices that can “stand alone.” This subsectionprovides general guidance on how to organize thebaseline risk assessment report and whichinformation should be included in the report.More detailed guidance, however, is found byfollowing the guidance in previous chapters of this

3

manual. Careful use of that guidance will ensurea well-documented baseline risk assessment report.

Exhibit 9-1 provides a suggested outline forthe full baseline risk assessment report. Thisoutline generally follows the flow of the riskassessment and the organization of this manual.The “bulleted” items are not necessarily sectionheadings, but rather are often items that shouldbe considered when writing the report. Note that,as with the manual, not all components of theoutline are applicable to all sites. This isespecially true if the risk assessment report will bea chapter in the RI/FS report. At some sites, andespecially when the risk assessment report will bea stand-alone document, more site-specific itemscould be added to the report.

Examples of tables and graphics that shouldbe included in the report are presented as exhibitsin previous chapters of this manual. Note,however, that additional tables and graphics maybe useful.

This suggested outline may be used as areview guide by risk assessors (and risk assessmentreviewers) to ensure that all appropriatecomponents of the assessment have beenaddressed. Section 9.2 addresses review tools ingreater detail.

9.1.3 OTHER KEY REPORTS

Two important reports that must includesummaries of the baseline risk assessment are (1)the remedial investigation/feasibility study (RI/FS)report and (2) the record of decision (ROD)report.

Summary for the RI/FS report. One of thechapters of the RI/FS typically is devoted to asummary of the baseline risk assessment. Part ofthis summary should address the human healthevaluation (the other part should address theenvironmental evaluation). The human healthsummary should follow the same outline as thefull baseline risk assessment report, with almosteach section of the summary being a distillationof each full report chapter. The riskcharacterization chapter is an exception, however,in that it could be included in the RI/FS reportessentially unchanged. Most tables and graphicsshould be included unchanged as well. For more

information, see Guidance for Conducting RemedialInvestigations and Feasibility Studies UnderCERCLA (EPA 1988b).

Summary for the ROD report. The RODdocuments the remedial action selected for a site.It consists of three basic components: (1) aDeclaration (2) a Decision Summary;, and (3) aResponsiveness Summary. The second component,a Decision Summary, provides an overview of thesite-specific factors and analyses that led to theselection of the remedy. Included in thiscomponent is a summary of site risks. As withthe risk assessment summary for the RI/FS report,the summary for the ROD report should followthe same outline as the full risk assessment. Thissummary, however, should be much moreabbreviated than the RI/FS summary, althoughcare must be taken to address all of the relevantsite-specific results. For more information, seeInterim Final Guidance on Preparing SuperfundDecision Documents: The Proposed Plan, theRecord of Decision, Explanation of SignificantDifferences, and the Record of Decision Amendment(EPA 1989).

9.2 REVIEW TOOLS

This section provides guidelines on reviewinga risk assessment report. A checklist of manyessential criteria that should be adequatelyaddressed in any good risk assessment is provided(Exhibit 9-2). The checklist touches upon issuesthat are often problematic and lead to difficultyand delay in the review of risk assessments.Principal questions are presented in the checklistwith qualifying statements or follow-up questions,as well as references to appropriate chapters andsections of this manual. The checklist is intendedas a guide to assist the preliminary reviewer byensuring that critical issues concerning the qualityand adequacy of information are not overlookedat the screening level review of risk assessments.Experience has shown that reviewers should payparticular attention to the following concerns.

Were all appropriate media sampled?

Were any site-related chemicals (e.g.,human carcinogens) eliminated fromanalysis without appropriate justification?

4

EXHIBIT 9-1

SUGGESTED OUTLINE FOR A BASELINE RISK ASSESSMENT REPORT

1.0 INTRODUCTION

1.1 OveviewGeneral problem at siteSite-specific objectives of risk assessment

1.2 Site Background Site description Map of site General history

-- Ownership-- Operations-- Contamination

Significant site reference pointsGeographic location relative to offsite areas of interestGeneral sampling locations and media

1.3 Scope of Risk AssessmentComplexity of assessment and rationaleOverview of study design

1.4 Organization of Risk Assessment Report

2.0 IDENTIFICATION OF CHEMICALS OF POTENTIAL CONCERN

2.1 General Site-specific Data Collection ConsiderationsDetailed historical information relevant to data collectionPreliminary identification of potential human exposureModeling parameter needsBackground samplingSampling locations and mediaSampling methodsQA/QC methodsSpecial analytical services (SAS)

2.2 General Site-specific Data Evaluation ConsiderationsSteps used (including optional screening procedure steps, if used)QA/QC methods during evaluationGeneral data uncertainty

2.3 Environmental Area or Operable Unit 1 (Complete for All Media)Area- and media-specific sample collection strategy (e.g., sample size, sampling locations)Data from site investigations

(continued)

5



2.4

2.X

Evaluation of analytical methodsEvaluation of quantitation limitsEvaluation of qualified and coded dataChemicals in blanksTentatively identified compoundsComparison of chemical concentrations with backgroundFurther limitation of number of chemicalsUncertainties, limitations, gaps in quality of collection or analysis

Environmental Area or Operable Unit 2 (Repeat for All Areas or Operable Units, AsAppropriate)

Summary of Chemicals of Potential Concern

3.0 EXPOSURE ASSESSMENT

3.1 Characterization of Exposure Setting Physical Setting-- Climate-- Vegetation-- Soil type-- Surface hydrology-- Ground-water hydrologyPotentially Exposed Populations-- Relative locations of populations with respect to site-- Current land use-- Potential alternate future land uses-- Subpopulations of potential concern

3.2 Identification of Exposure PathwaysSources and receiving mediaFate and transport in release mediaExposure points and exposure routesIntegration of sources, releases, fate and transport mechanisms, exposure points, and exposureroutes into complete exposure pathwaysSummary of exposure pathways to be quantified in this assessment

3.3 Quantification of ExposureExposure concentrationsEstimation of chemical intakes for individual pathways

(continued)

6



3.4 Identification of UncertaintiesCurrent and future land-useEnvironmental sampling and analysisExposure pathways evaluatedFate and transport modelingParameter values

3.5 Summary of Exposure Assessment

4.0 TOXIClTY ASSESSMENT

4.1 Toxicity Information for NonCarcinogenic EffectsAppropriate exposure periods for-toxicity valuesUp-to-date RfDs for all chemicalsOne- and ten-day health advisories for shorter-term oral exposuresOverall data base and the critical study on which the toxicity value is based (including thecritical effect and the uncertainty and modifying factors used in the calculation)Effects that may appear at doses higher than those required to elicit the critical effectAbsorption efficiency considered

4.2 Toxicity Information for Carcinogenic EffectsExposure averaged over a lifetimeUp-to-date slope factors for all carcinogensWeight-of-evidence classification for all carcinogensType of cancer for Class A carcinogensConcentration above which the dose-response curve is no longer linear

4.3 Chemicals for Which No EPA Toxicity Values Are AvailableReview by ECAOQualitative evaluationDocumentation/justification of any new toxicity values developed

4.4 Uncertainties Related to Toxicity InformationQuality of the individual studiesCompleteness of the overall data base

4.5 Summary of Toxicity Information

5.0 RISK CHARACTERIZATION

5.1 Current Land-use ConditionsCarcinogenic risk of individual substancesChronic hazard quotient calculation (individual substances)Subchronic hazard quotient calculation (individual substances)

(continued)

Psgc 9-7



Shorter-term hazard quotient calculation (individual substances)Carcinogenic risk (multiple substances)Chronic hazard index (multiple substances)Subchronic hazard index (multiple substances)Shorter-term hazard index calculation (multiple substances)Segregation of hazard indicesJustification for combining risks across pathwaysNoncarcinogenic hazard index (multiple pathways)Carcinogenic risk (multiple pathways)

5.2 Future Land-use ConditionsCarcinogenic risk of individual substancesChronic hazard quotient calculation (individual substances)Subchronic hazard quotient calculation (individual substances)Carcinogenic risk (multiple substances)Chronic hazard index (multiple substances)Subchronic hazard index (multiple substances)Segregation of hazard indicesJustification for combining risks across pathwaysNoncarcinogenic hazard index (multiple pathways)Carcinogenic risk (multiple pathways)

5.3 UncertaintiesSite-specific uncertainty factors-- Definition of physical setting-- Model applicability and assumptions-- Parameter values for fate/transport and exposure calculations

Summary of toxicity assessment uncertainty-- Identification of potential health effects-- Derivation of toxicity value-- Potential for synergistic or antagonistic interactions-- Uncertainty in evaluating less-than-lifetime exposures

5.4 Comparison of Risk Characterization Results to Human StudiesATSDR health assessmentSite-specific health studies (pilot studies or epidemiological studies)Incorporation of studies into the overall risk characterization

5.5 Summary Discussion and Tabulation of the Risk CharacterizationKey site-related contaminants and key exposure pathways identifiedTypes of health risk of concernLevel of confidence in the quantitative information used to estimate riskPresentation of qualitative information on toxicity

(continued)

8

6.0 SUMMARY

SUGGESTED OUTLINE


FOR A BASELINE RISK ASSESSMENT REPORT

Confidence in the key exposure estimates for the key exposure pathwaysMagnitude of the carcinogenic and noncarcinogenic risk estimatesMajor factors driving riskMajor factors contributing to uncertaintyExposed population characteristicsComparison with site-specific health studies

6.16.26.36.4

Chemicals of Potential ConcernExposure AssessmentToxicity AssessmentRisk Characterization

9

EXHIBIT 9-2

REVIEWER CHECKLIST

1.0 GENERAL CONCERNS

Were the site-specific obJective(s) of the risk assessment stated? (HHEM - 1)

Was the scope of the assessment described (e.g., in terms of the complexity of the assessment andrationale, data needs, and overview of the study design)? (HHEM -1.1.1, 3.5)

Was an adequate history of site activities provided, including a chronology of land use (e.g.,specifying agriculture, industry, recreation, waste deposition, and residential development at thesite)? (HHEM -2.1.4, 9.1)

Was an initial qualitative overview of the nature of contamination included (e.g., specifying in ageneral manner the kinds of contaminants, media potentially contaminated)? (HHEM -2.1.4, 9.1)

Was a general map of the site depicting boundaries and surface topography included, whichillustrates site features, such as fences, ponds, structures, as well as geographical relationshipsbetween specific potential receptors and the site? (HHEM -2.1.4, 9.1)

2.0 CONCERNS IN REVIEWING DATA COLLECTI0N AND EVALUATION

2.1 Data Collection

Was an adequate “conceptual model” of the site discussed? (HHEM - 4.2)

-- a qualitative discussion of potential or suspected sources of contamination, types andconcentrations of contaminants detected at the site, potentially contaminated media, as wellas potential exposure pathways and receptors

Was an adequate Data Quality Objectives (DQO) statement provided? (HHEM 4.1.4)

-- a statement specifying both the qualitative and quantitative nature of the sampling damin terms of relative quality and intent for use, issued prior to data collection, which helpsto ensure that the data collected will be appropriate for the intended objectives of the study

Were key site characteristics documented? (HHEM -4.3, 4.5)

--

--

soil/sediment parameters (e.g., particle size, redox potential, mineral class, organic carbonand clay content, bulk density, and porosity)

hydrogeological parameters (e.g., hydraulic gradient, pH/Eh, hydraulic conductivity, location,saturated thickness, direction, and rate of flow of aquifers, relative location of bedrock layer)

(continued)

10


REVIEWER CHECKLIST

--

--

hydrological parameters (e.g., hardness, pH, dissolved oxygen, salinity, temperature, totalsuspended solids, flow rates, and depths of rivers or streams; estuary and embaymentparameters such as tidal cycle, range, and area; as well as lake parameters such as area,volume, depth, and depth to thermocline)

meteorological parameters (e.g., direction of prevailing wind, average wind speed,temperature, humidity, annual average and 24 hour maximum rainfall)

Were all appromiate media sampled?

-- was there adequate justification for

-- were literature estimates employedreferenced properly?

(HHEM -4.4,4.5, 4.6)

any omissions?

for omissions in background sampling and were they

Were all key areas sampled, based on all available information (e.g., preliminary assessment,field screening)? (HHEM -4.4, 4.5, 4.6)

Did sampling include media along potential routes of migration (e.g., between the contaminantsource and potential future exposure points)? (HHEM -4.5, 4.6)

Were sampling locations consistent with nature of contamination (e.g., at the appropriatedepth)? (HHEM -4.5, 4.6)

Were sampling efforts consistent with field screening and visual observations in locating "hotspots" ?(HHEM -4.5, 4.6)

Were detailed sampling maps provided, indicating the location, type (e.g., grab, composite,duplicate), and numerical code of each sample? (HHEM - 5.10)

Did sampling include appropriate QA/QC measures (e.g., replicates, split samples, trip and fieldblanks)? (HHEM -4.7, 5.4)

Were background samples collected from appropriate areas (e.g., areas proximate to the site,free of potential contamination by site chemicals or anthropogenic sources, and similar to thesite in topography, geology, meteorology, and other physical characteristics)? (HHEM -4.4,5.7)

2.2 Data Evaluation

Were any site-related chemicals (e.g. , human carcinogens) eliminated from analysis withoutappropriate justification? (HHEM - 5.9)

(continued)

11


REVIEWER CHECKLIST

--

--

--

--

as infrequently detected chemicals (HHEM - 5.3.3, 5.9.3)

as non-detects in a specific medium without employing a “proxy” concentration (HHEM -5.3)

as common laboratory contaminants even though sample concentrations were significantlyhigher than that found in blanks? (HHEM - 5.5)

as present at a “ubiquitous level”? (HHEM - 5.7)

Were inappropriate “proxy concentrations” assigned to site-related chemicals? (HHEM - 5.3)

-- was a value of zero or the instrument detection limit (IDL) assigned?

-- was an erroneous sample-specific quantitation limit employed?

Were appromiate analytical methods employed for collection of data upon which risk estimatesare based? (HHEM - 5.2)

-- were the methods consistent with the requisite level of sensitivity?

-- were established procedures with adequate QA/QC measures employed?

Did the data meet the Data Quality Objectives (DQO)? (HHEM - 4.1.4)

-- were the sampling methods consistent with the intended uses of data?

Were appropriate data qualifiers employed? (HHEM - 5.4)

Were special analytical services (SAS) employed when appropriate? (HHEM - 5.3)

-- was SAS employed as an adjunct to routine analysis in cases where certain contaminantswere suspected at low levels, as non-TCL chemicals, in non-standard matrices, or insituations requiring a quick turnaround time?

3.0 CONCERNS IN REVIEWING THE EXPOSURE ASSESSMENT

Were “reasonable maximum exposures” considered (i.e., the highest exposures that are reasonablyexpected to occur)? (HHEM -6.1.2, 6.4.1, 6.6)

Were current and future land uses considered? (HHEM -6.1.2, 6.2)

(continued)

12


REVIEWER CHECKLIST

Was residential land use considered as an alternative future land use? (HHEM - 6.2.2)

-- if not, was a valid rationale provided?

Were all potential sensitive subpopulations considered (e.g., elderly people, pregnant or nursingwomen, infants and children, and people with chronic illnesses)? (HHEM - 6.2.2)

Were all significant contaminant sources considered? (HHEM - 6.3.1)

Were all potential contaminant release mechanisms considered, such as volatilization, fugitive dustemission, surface runoff/overland flow, leaching to ground water, tracking by humans/anirnals, andsoil gas generation? (HHEM - 6.3.1)

Were all potential contaminant transport pathways considered, such as direct air transportdownwind, diffusion in surface water, surface water flow, ground-water flow, and soil gas migration?(HHEM - 6.3)

Were all relevant cross-media transfer effects considered, such as volatilization to air, wetdeposition, dry deposition, ground-water discharge to surface, and ground-water recharge fromsurface water? (HHEM - 6.3)

Were all media potentiallv associated with exposure considered? (HHEM - 6.2, 6.3)

Were all relevant site-speific characteristic considered, including topographical, hydrogeological,hydrological, and meteorological parameters? (HHEM - 6.1, 6.3)

Were all possible exposure pathwavs considered? (HHEM - 6.3)

-- was a valid rationale offered for exclusion of any potential pathways from quantitativeevaluation?

Were all “spatial relationships” adequately considered as factors that could affect the level ofexposure (e.g., hot spots in an area that is frequented by children, exposure to ground water fromtwo aquifers that are not hydraulically connected and that differ in the type and extent ofcontamination)? (HHEM - 6.2, 6.3)

Were appropriate approaches employed for calculating average exposure concentrations? (HHEM- 6.4, 6.5)

-- was a valid rationale provided for using geometric or arithmetic means?

Were appromiate or standard default values used in exposure calculations (e.g., age-specific bodyweights, appropriate exposure frequency and duration values)? (HHEM - 6.4, 6.5, 6.6)

(continued)

13


REVIEWER CHECKLIST

4.0 CONCERNS IN REVIEWING THE TOXICITY ASSESSMENT

Was the exclusion of anv carcinogen from analysis adequately justified (e.g., were “weight-of-evidence” classifications and completeness of exposure pathways considered in this decision)?(HHEM - 5.9, 7.3)

Were appropriate “route-to-route” extrapolations performed in cases where a toxicity value wasapplied across differing routes of exposure? (HHEM -7.5.1, 8.1.2)

-- were the extrapolations based on appropriate guidance?

Were appropriate toxicitv values employed based on the nature of exposure? (HHEM - 7.4, 7.5)

-- were subchronic vs. chronic RfDs applied correctly based on the duration of exposure?

-- were all sensitive subpopulations, such as pregnant or nursing women potentially requiringdevelopmental RfDs (RfDdts), considered in the selection of the toxicity values used?

Were the toxicity values that were used consistent with the values contained within the IntegratedRisk Information System (IRIS) or other EPA documents? (HHEM - 7.4, 7.5)

5.0 CONCERNS IN REVIEWING THE RISK CHARACTERIZATION

Were exposure estimates and toxicity values consistently expressed as either intakes or absorbeddoses for each chemical taken through risk characterization? (HHEM - 8.1.2)

-- was a valid rationale given for employing values based on absorbed dose?

Were all site-related chemicals that were analyzed in the exposure assessment considered in riskcharacterization? (HHEM - 8.1.2)

-- were inconsistencies explained?

Were risks appropriately summed only across exposure pathways that affect the same individualor population subgroup, and in which the same individual or population subgroup faces the“reasonable maximum exposure,” based on the assumptions employed in the exposure assessment?(HHEM - 8.3)

Were sources of uncertainty adequately characterized? (HHEM - 8.4)

14

Were current and future land usesconsidered?

Were all significant contaminant sourcesconsidered?

Were appropriate or standard defaultvalues used in exposure calculations?

Were the toxicity values that were usedconsistent with the values containedwithin the Integrated Risk InformationSystem (IRIS) or other EPA documents?

Although the checklist addresses many pertinentissues, it is not a complete listing of all potentialconcerns, since this objective is beyond the scopeof a preliminary review tool. In addition, some ofthe concerns listed are not necessarily appropriatefor all risk assessment reports.

The recommended steps in reviewing a riskassessment report are as follows:

(1) compare the risk assessment reportoutline to the suggested outline inSection 9.1 of this chapter (i.e., Exhibit9-l);

(2) use the checklist in this section (i.e.,Exhibit 9-2); and

(3) conduct a comprehensive review.

The outline (Exhibit 9-1) and the checklist(Exhibit 9-2) are intended only as tools to assistin a preliminary review of a risk assessment, andare not designed to replace the good judgmentneeded during the comprehensive review. Thesetwo tools should provide a framework, however,for the timely screening of risk assessments byreviewers with a moderate level of experience in

the area. If these steps are followed in order,then some of the major problems with a riskassessment report (if any) can be identified beforesignificant resources are expended during thecomprehensive review.

9.3 MANAGEMENT TOOLS

This section provides a concise checklist forthe RPM to use in carrying out their role in therisk assessment process (see Exhibit 9-3). Otherdecision-makers at the site also may find this’checklist useful. Specific points at which themanagers should be involved, or may be calledupon to become involved, during the riskassessment are discussed in Chapters 4 through 8of the manual. This checklist extracts informationfrom those chapters, and also includes pointers onplanning and involvement for the manager. Thepurpose of the checklist is to involve managers inthe direction and development of the riskassessment and thereby avoid serious mistakes orcostly misdirections in focus or level of effort.

Although the checklist is shaped to suggestwhen and how the manager should becomeinvolved in the risk assessment process, it isassumed that part of the manager’s involvementwill require consultation with technical resourcesavailable in the region or state. The checklistadvises consulting the “regional risk assessmentsupport staff” at a number of points in theprocess. This contact may not be one person, butcould be a number of different technical peoplein the region, such as a toxicologist,hydrogeologist, or other technical reviewer. Themanager should become aware of the resourcesavailable to him or her, and use them whenappropriate to ensure that the risk assessmentdeveloped is useful and accurate.

15

EXHIBIT 9-3

CHECKLIST FOR MANAGER INVOLVEMENT

1. GETTING ORGANIZED

Ensure that the workplan for the risk assessment contractor support is in place (if needed).

Identify EPA risk assessment support persomel (to be used throughout the risk assessmentprocess).

Gather relevant information, such as appropriate risk assessment guidances and site-specific data and reports.

Identify available state, county, and other non-EPA resources.

2. BEFORE THE SCOPING MEETING

Make initial contact with risk assessor.

Provide risk asseasor with available guidances and site data.

Determine (or review) data collection needs for risk assessment, considering:-- modeling parameter needs;-- type and location of background sample,

the preliminary identification of potential human exposure;-- strategies for sample collection appropriate to site/risk assessment data needs;-- statistical methods;-- QA/QC measures of particular importance to risk assessment;-- special analytical services (SAS) needs;-- alternate future land use and-- location(s) in ground water that will be used to evaluate future ground-water exposures.

3. AT THE SCOPING MEETING

Present risk assessment data collection needs.

Ensure that the risk assessment data collection needs will be considered in developmentof the sampling and analysis plan.

Where limited resources require that less-than-optimal sampling be conducted, discuss potential impacts on risk assessment results.

4. AFTER THE SCOPING MEETING

Ensure that the risk assessor reviews and approves the sampling and analysis plan.

Consult with ATSDR if human monitoring is planned.

(continued)

16



5. DURING SAMPLING AND ANALYSIS

Ensure that risk assessment needs are being met during sampling.

Provide risk assessor with any preliminary sampling results so that he/she can determineif sampling should be refocused.

Consult with ATSDR to obtain a status report on any human monitoring that is being conducted.Provide any results to risk assessor.

6. DURING DEVELOPMENT OF RISK ASSESSMENT

Meet with risk assessor to discuss basis of excluding chemicals from the risk assessment(and developing the list of chemicals of potential concern). Confirm appropriateness ofexcluding chemicals.

Confirm determination of alternate future kind use.

Confirm location(s) in ground water that will be used to evaluate future ground-water exposures.

Understand basis for selection of pathways and potentially exposed populations.

Facilitate discussions between risk assessor and EPA risk assessment support personnelon the following points:

--

--

--

--

the need for any major exposure, fate, and transport models (e.g., air or ground-waterdispersion models) used;

site-specific exposure assumptions;

non-EPA-derived toxicity values; and

appropriate level of detail for uncertainty analysis, and the degree to which uncertainties willbe quantified.

Discuss and approve combination of pathway risks and hazard indices.

Ensure that end results of risk characterization have been compared with ATSDR healthassessments and other site-specific human studies that might be available.

7. REVIEWING THE RISK ASSESSMENT

Allow sufficient time for review and incorporation of comments.

Ensure that reviewers’ comments are incorporated.

(continued)

17



8. COMMUNICATING THE RISK ASSESSMENT

Plan a briefing among technical staff to discuss significant findings and uncertainties.

Discuss development of graphics, tools, and

Consult with other groups (e.g.,

Brief upper management.

community

presentations to assist risk management decisions

relations staff), as appropriate.

18


Bean, M.C. (CH2M Hill). 1987. Tools for Environmental Professionals Involved in Risk Communication at Hazardous Waste FacilitiesUndergoing Siting. Permitting, or Remediation. Presented at the Air Pollution Control Association Annual Meeting. New York,June 21-26, 1987.

Environmental Protection Agency (EPA). 1986. Explaining Environmental Risk. Office of Toxic Substances.

Environmental Protection Agency (EPA). 1988a. Seven Cardinal Rules of Risk Communication. Office of Policy Analysis.

Environmental Protection Agency (EPA). 1988b. Guidance for Conducting Remedial Investigations and Feasibiliy Studies UnderCERCLA. Office of Emergency and Remedial Response. (OSWER Directive 9355.3-01).

Environmental Protection Agency (EPA). 1989. Interim Final Guidance on Preparing Superfund Decision Documents The ProposedPlan, the Record of Decision, Explanation of Significant Differences, and the Record of Decision Amendment. Office ofEmergency and Remedial Response. (OSWER Directive 9355.3-02).

New Jersey Department of Environmental Protection (NJDEP). 1987. Improving Dialogue with Communities A Short Guide forGovernment Risk Communication. Division of Science and Research.

CHAPTER 10

RADIATION RISK ASSESSMENTGUIDANCE

There are many sites contaminated withradioactive substances that are included on theNational Priorities List (NPL), and additional sitesare expected in future NPL updates. This chapterprovides supplemental baseline risk assessmentguidance for use at these sites. This guidance isintended as an overview of key differences inchemical and radionuclide assessments, and not asa comprehensive, stand-alone approach forassessing the risks posed by radiation.

The reader should be familiar with theguidance provided in Chapters 2 through 9 beforeproceeding further in Chapter 10. Although thediscussions in the previous chapters focusprimarily on chemically contaminated sites, muchof the information presented is also applicable tothe evaluation of radioactively contaminatedSuperfund sites For consistency and completeness,the topics discussed in each section of this chapterparallel the topics covered in each of the previouschapters.

After a brief introduction to some of thebasic principles and concepts of radiationprotection (Section 10.1), seven additional areasare addressed:

(1)

(2)

(3)

(4)

R e g u l a t i o n o f RadioactivelyContaminated sites (Section 10.2);

Data Collection (Section 10.3);

Data Evaluation (Section 10.4);

Exposure and Dose Assessment (Section10.5);

(5)

(6)

(7)

Toxicity Assessment (Section 10.6);

Risk Characterization (Section 10.7); and

Documentation, Review, andManagement Tools for the RiskAssessor, Reviewer, and Manager(Section 10.8).

3

There are special hazards associated withhandling radioactive waste and EPA stronglyrecommends that a health physicist experienced inradiation measurement and protection beconsulted prior to initiating any activities at a site

suspected of being contaminated with radioactivesubstances. EPA also recommends that theremedial project manager (RPM) or on-scenecoordinator (OSC) should designate both achemical risk assessor and a radiation riskassessor. These individuals should work closelywith each other and the RPM to coordinateremedial activities (e.g., site scoping, health andsafety planning, sampling and analysis) andexchange information common to both chemicaland radionuclide assessments, including data onthe physical characteristics of the site, potentiallyimpacted populations, pathways of concern, andfate and transport models used. At the conclusionof the remedial investigation/feasibility study(RI/FS) process, the RPM should issue a singlereport that summarizes and integrates the resultsfrom both the chemical and the radiation riskassessments.

A two-phase evaluation is described for theradiation risk assessment. As discussed in Section10.5, procedures established by the InternationalCommission on Radiological Protection (ICRP1979) and adopted by EPA in Federal GuidanceReport No. 11 (EPA 1988) are used to estimatethe radiation dose equivalent to humans frompotential exposures to radionuclides through allpertinent exposure pathways at a site. Thoseestimates of dose equivalent may be used forcomparison with established radiation protectionstandards and criteria. However, this methodologywas developed for regulation of occupationalradiation exposures for adults and is notcompletely applicable for estimating health risk tothe general population at a Superfund site.Therefore, a separate methodology is presented inSection 10.7.2 for estimating health risk based onthe age-averaged lifetime excess cancer incidenceper unit intake (and per unit extenal exposure)for radionuclides of concern. Radiation riskassessments for Superfund sites should includeestimates of both the dose equivalent computedas described in Section 10.5, and the health riskattributable to radionuclide exposures computedusing the approach described in Section 10.7.

Only summary-level information is presentedin this chapter, and references are provided to anumber of supporting technical documents forfurther information. In particular, the reader isencouraged to consult Volume 1 of theBackground Information Document for the DraftEnvironmental Impact Statement for ProposedNESHAPS for Radionuclides (EPA 1989a) for amore comprehensive discussion of EPA’s currentrisk assessment methodology for radionuclides.

For additional radiation risk assessmentinformation and guidance, RPMs and otherinterested individuals can contact the Office ofRadiation Programs (ORP) within EPAheadquarters at 202-475-9630 (FTS 475-9630).Interested individuals also can contact theRegional Radiation Program Managers withineach of the EPA regional offices for guidance andhealth physics support.

10.1 RADIATION PROTECTIONPRINCIPLES ANDCONCEPTS

Radioactive atoms undergo spontaneousnuclear transformations and release excess energyin the form of ionizing radiation. Suchtransformations are referred to as radioactivedecay. As a result of the radioactive decayprocess, one element is transformed into another;the newly formed element, called a decay product,will possess physical and chemical propertiesdifferent from those of its parent, and may also beradioactive. A radioactive species of a particularelement is referred to as a radionuclide orradioisotope. The exact mode of radioactivetransformation for a particular radionuclidedepends solely upon its nuclear characteristics, andis independent of the nuclide’s chemicalcharacteristics or physical state. A fundamentaland unique characteristic of each radionuclide isits radioactive half-life, defined as the timerequired for one half of the atoms in a givenquantity of the radionuclide to decay. Over 1,600different radionuclides have been identified todate, with half-lives ranging from fractions of asecond to millions of years. Selected radionuclidea

4

of potential importance at Superfund sites arelisted in Exhibit 10-1.

Radiation emitted by radioactive substancescan transfer sufficient localized energy to atomsto remove electrons from the electric field of theirnucleus (ionization). In living tissue this energytransfer can destroy cellular constituents andproduce electrically charged molecules (i.e., freeradicals). Extensive biological damage can lead toadverse health effects. The type of ionizingradiation emitted by a particular radionuclidedepends upon the exact nature of the nucleartransformation, and may include emission of alphaparticles, electrons (beta particles or positrons),and neutrons; each of these transformations maybe accompanied by emission of photons (gammaradiation or x-rays). Each type of radiationdiffers in its physical characteristics and in itsability to inflict damage to biological tissue. Thesecharacteristics and effects are summarized in thebox on this page.

Quantities of radionuclides are typicallyexpressed in terms of activity at a given time t(A(t)). The SI unit of activity is the becquerel(Bq), which is defined as the quantity of a givenradionuclide in which one atom is transformed persecond (i.e., one decay per second). Theconventional unit of activity is the curie (Ci),which is defined as the quantity of a givenradionuclide in which 3.7xl010 atoms undergonuclear transformation each second; one curie isapproximately equivalent to the decay rate of onegram of Ra-226. A more convenient unit ofactivity for expressing environmentalconcentrations of radionuclides is the picoCurie(pCi), which is equal to l0-12 Ci. Occassionally,activity is expressed incorrectly in terms of countsper second (cps) or counts per minute (cpm):these refer to the number of transformations perunit time measured by a particular radiationdetector and do not represent the true decay rateof the radionuclide. To derive activity values,count rate measurements are multiplied byradioisotope-specific detector calibration factors.

5

EXHIBIT 10-1

RADIOLOGICAL CHARACTERISTICS OF SELECTED RADIONUCLIDESFOUND AT SUPERFUND SITESa

Average Radiation Enemies (MeV/decay) b

Nuclide Half-life c Alpha Beta, Electron x, Gamma

a Source: ICRP 1983 (except Ba-137m data from Kocher 1981).b Computed as the sum of the products of the energies and yields of individual radiations.c Half-life expressed in years (y), days (d), and hours (h).

6

The activity per unit mass of a givenradionuclide is called the specific activity, and isusually expressed in units of bequerels per gram(Bq/g) or curies per gram (Ci/g). The shorter thehalf-life of the radionuclide, the greater is itsspecific activity. For example, Co-60 has aradioactive half-life of about 5 years and a specificactivity of 4xl013 Bq/g, whereas Np-237 has ahalf-life of 2 million years and a specific activityof 3X107 Bq/g.

Several terms are used by health physicists todescribe the physical interactions of different typesof radiations with biological tissue, and to definethe effects of these interactions on human health.One of the first terms developed was radiationexposure, which refers to the transfer of energyfrom a radiation field of x- or gamma rays to aunit mass of air. The unit for this definition ofexposure is the roentgen (R), expressed ascoulombs of charge per kilogram of air (1 R =2.58x10 -4 C/kg).

The term exposure is also defined as thephysical contact of the human body with radiation.Internal exposure refers to an exposure that occurswhen human tissues are subjected to radiationsfrom radionuclides that have entered the body viainhalation, ingestion, injection, or other routes.External exposure refers to the’ irradiation ofhuman tissues by radiations emitted byradionuclides located outside the body eitherdispersed in the air or water, on skin surfaces, ordeposited on ground surfaces. All types ofradiation may contribute to internal exposure,whereas only photon, beta, and neutron radiationscontribute significantly to external exposure.

Ionizing radiation can cause deleteriouseffects on biological tissues only when the energyreleased during radioactive decay is absorbed intissue. The absorbed dose (D) is defined as themean energy imparted by ionizing radiation perunit mass of tissue. The SI unit of absorbed doseis the joule per kilogram, also assigned the specialname the gray (1 Gy = 1 joule/kg). Theconventional unit of absorbed dose is the rad (1rad = 100 ergs per gram = 0.01 Gy).

For radiation protection purposes, it isdesirable to compare doses of different types of

radiation. The absorbed dose of any radiationdivided by the absorbed dose of a reference.radiation (traditionally 250 kVp x-rays) thatproduces the same biological endpoint is calledthe Relative Biological Effectiveness or RBE. Forregulatory purposes, an arbitrary consensus RBEestimate called the Quality Factor or Q is oftenused. The dose equivalent (H) was developed tonormalize the unequal biological effects producedfrom equal absorbed doses of different types ofradiation. The dose equivalent is defined as:

H = DQN

where D is the absorbed dose, Q is a qualityfactor that accounts for the RBE of the type ofradiation emitted, and N is the product of anyadditional modifying factors, Quality factorscurrently assigned by the InternationalCommission on Radiological Protection (ICRP)include values of Q=20 for alpha particles, Q=1Ofor neutrons and protons, and Q= 1 for betaparticles, positrons, x-rays, and gamma rays (ICRP1984). These factors may be interpreted asfollows: on average, if an equal amount of energyis absorbed, an alpha particle will inflictapproximately 20 times more damage to biologicaltissue than a beta particle or gamma ray, andtwice as much damage as a neutron. Themodifyng factor is currently assigned a value ofunity (N= 1) for all radiations. The SI unit ofdose equivalent is the sievert (Sv), and theconventional unit is the rem (1 rem = 0.01 Sv).

7

The dose delivered to tissues from radiationsexternal to the body occurs only while theradiation field is present. However, the dosedelivered to body tissues due to radiations fromsystemically incorporated radionuclides maycontinue long after intake of the nuclide hasceased. Therefore, internal doses to specifictissues and organs are typically reported in termsof the committed dose equivalent (HT,50), whichis defined as the integral of the dose equivalent ina particular tissue T for 50 years after intake(corresponding to a working lifetime).

When subjected to equal doses of radiation,organs and tissues in the human body will exhibitdifferent cancer induction rates. To account forthese differences and to normalize radiation dosesand effects on a whole body basis for regulationof occupational exposure, the ICRP developed theconcept of the effective dose equivalent (HE) andcommitted effective dose equivalent (HE,50), whichare defined as weighted sums of the organ-specific

committed dose equivalents (i.e., respectively. Weighting factors, wT, are based onselected stochastic risk factors specified by theICRP and are used to average organ-specific doseequivalents (ICRP 1977, 1979). The effective doseequivalent is equal to that dose equivalent,delivered at a uniform whole-body rate, that

corresponds to the same number (but possibly adissimilar distribution) of fatal stochastic healtheffects as the particular combination of committedorgan dose equivalents (see the box on this page).

A special unit, the working level (WL), isused to describe exposure to the short-livedradioactive decay products of radon (Rn-222).Radon is a naturally occurring radionuclide thatis of particular concern because it is ubiquitous,it is very mobile in the environment, and it decaysthrough a series of short-lived decay products thatcan deliver a significant dose to the lung wheninhaled. The WL is defined as any combinationof short-lived radon decay products in one liter ofair that will result in the ultimate emission of1.3xl0 5 MeV of alpha energy. The working levelmonth (WLM) is defined as the exposure to 1WL for 170 hours (1 working month).

Radiation protection philosophy encouragesthe reduction of all radiation exposures as low asreasonably achievable (ALARA), in considerationof technical, economic, and social factors.Further, no practice involving radiation exposureshould be adopted unless it provides a positive netbenefit. In addition to these general guidelines,specific upper limits on radiation exposures anddoses have been established by regulatoryauthorities as described in the following section.

8

Additional discussion on the measurement ofradioactivity is provided in Sections 10.3 and 10.4,and the evaluation of radiation exposure and doseis discussed further in Section 10.5. Discussion ofpotential health impacts from ionizing radiationis presented in Section 10.6.

10.2 REGULATION OFRADIOACTIVELYCONTAMINATED SITES

Chapter 2 briefly describes the statutes,regulations, guidance, and studies related to thehuman health evaluation process for chemicalcontaminants. The discussion describes CERCLAas amended by SARA and the RI/FS process.Since radionuclides are classified as hazardoussubstances under CERCLA, this information isalso applicable to radioactively contaminated sites.Chapter 2 also introduces the concept ofcompliance with applicable or relevant andappropriate requirements (ARARs) in federaI andstate environmental laws as required by SARA.Guidance on potential ARARs for theremediation of radioactively contaminated sitesunder CERCLA is available in the CERCLACompliance with Other Laws Manual (EPA 1989c).Only a brief summary of regulatory authorities ispresented here.

The primary agencies with regulatoryauthority for the cleanup of radioactivelycontaminated sites include EPA the NuclearRegulatory Commission (NRC), the Departmentof Energy (DOE), and state agencies. Otherfederal agencies, including the Department ofTransportation (DOT) and Department of Defense(DOD), also have regulatory programs (but morelimited) for radioactive materials. Also, nationaland international scientific advisory organizationsprovide recommendations related to radiationprotection and radioactive waste management, buthave no regulatory authority. The following is abrief description of the main functions and areasof jurisdiction of these agencies and organizations.

EPA’s authority to protect public healthand the environment from adverse effectsof radiation exposure is derived fromseveral statutes, including the AtomicEnergy Act, the Clean Air Act, the

Uranium Mill Tailings Radiation ControlAct (UMTRCA), the Nuclear WastePolicy Act, the Resource Conservationand Recovery Act (RCRA), andCERCLA. EPA’s major responsibilitieswith regard to radiation include thedevelopment of federal guidance andstandards, assessment of newtechnologies, and surveillance ofradiation in the environment. EPA alsohas lead responsibility in the federalgovernment for advising all federalagencies on radiation standards. EPA’sradiation standards apply to manydifferent types of activities involving alltypes of radioactive material (i.e., source,byproduct, special nuclear, and naturallyoccurring and accelerator producedradioactive material [NARM]). Forsome of the EPA standards,implementation and enforcementresponsibilities are vested in otheragencies, such as NRC and DOE.

NRC licenses the possession and use ofcertain types of radioactive material atcertain types of facilities. Specifically, theNRC is authorized to license source,byproduct, and special nuclear material.The NRC is not authorized to licenseNARM, although NARM may bepartially subject to NRC regulation whenit is associated with material licensed bythe NRC. Most of DOE’s operationsare exempt from NRC’s licensing andregulatory requirements, as are certainDOD activities involving nuclearweapons and the use of nuclear reactorsfor military purposes.

DOE is responsible for conducting oroverseeing radioactive materialoperations at numerous government-owned/contractor-operated facilities.DOE is also responsible for managingseveral inactive sites that containradioactive waste, such as sites associatedwith the Formerly Utilized SitesRemedial Action Program (FUSRAP),the Uranium Mill Tailings RemedialAction Program (UMTRAP), the GrandJunction Remedial Action Program(GJRAP), and the Surplus Facilities

9

Management Program (SFMP). DOE isauthorized to control all types ofradioactive materials at sites within itsjurisdiction.

Other federal agencies with regulatoryprograms applicable to radioactive wasteinclude DOT and DOD. DOT hasissued regulations that set forthpackaging, labeling, record keeping, andreporting requirements for the transportof radioactive material (see 49 CFRParts 171 through 179). Most of DOD’sradioactive waste management activitiesare regulated by NRC and/or EPAHowever, DOD has its own program forcontrolling wastes generated for certainnuclear weapon and reactor operationsfor military purposes. Other agencies,such as the Federal EmergencyManagement Agency (FEMA) and theDepartment of the Interior (DOI), mayalso play a role in radioactive wastecleanups in certain cases.

States have their own authority andregulations for managing radioactivematerial and waste. In addition, 29states (Agreement States) have enteredinto agreements with the NRC, wherebythe Commission has relinquished to thestates its regulatory authority oversource, byproduct, and small quantitiesof special nuclear material. BothAgreement States and NonagreementStates can also regulate NARM. Suchstate-implemented regulations arepotential ARARs.

The National Council on RadiationProtection and Measurements (NCRP)and the International Commission onRadiological Protection (ICRP) providerecommendations on human radiationprotection. The NCRP was charteredby Congress to collect, analyze, develop,and disseminate information andrecommendations about radiationprotection and measurements. TheICRP’S function is basically the same,but on an international level. Although

10

neither the NCRP nor the ICRP haveregulatory authority, the i rrecommendations serve as the basis formany of the general (i.e., notsource-specific) regulations on radiationprotection developed at state and federallevels.

The standards, advisories, and guidance ofthese various groups are designed primarily to beconsistent with each other, often overlapping inscope and purpose. Nevertheless, there areimportant differences between agencies and

programs in some cases. It is important thatthese differences be well understood so that when

more than one set of standards is potentiallyapplicable to or relevant and appropriate for thesame CERCLA site, RPMs will be able toevaluate which standards to follow. In general,determination of an ARAR for a sitecontaminated with radioactive materials requiresconsideration of the radioactive constituentspresent and the functional operations thatgenerated the site, whose regulatory jurisdictionthe site falls under, and which regulation is mostprotective, or if relevant and appropriate, mostappropriate given site conditions.

For further information on radiationstandards, advisories, and guidance, RPMs shouldconsult the detailed ARARs guidance document(EPA 1989c), as well as EPA’s ORP and/orRegional Radiation Program Managers.

10.3 DATA COLLECTION

Data collection needs and procedures for sitescontaminated with radioactive substances are verysimilar to those described in Chapter 4 forchemically contaminated sites. There are,however, some basic differences that simplify datacollection for radionuclides, including the relativeease and accuracy with which natural backgroundradiation and radionuclide contaminants can bedetected in the environment when compared withchemical contaminants.

The pathways of exposure and themathematical models used to evaluate thepotential health risks associated with radionuclidesin the environment are similar to those used forevaluating chemical contaminants. Many of the

radionuclides found at Superfund sites behave inthe environment like trace metals. Consequently,the types of data needed for a radiation riskassessment are very similar to those required fora chemical contaminant risk assessment. Forexample, the environmental, land use, anddemographic data needed and the procedures usedto gather the data required to model fate andeffect are virtually identical. The primarydifferences lie in the procedures used tocharacterize the radionuclide contaminants. In thesections that follow, emphasis is placed on theprocedures used to characterize the radionuclidecontaminants and not the environmental settingthat affects their fate and effects, since the latterhas been thoroughly covered in Chapter 4.

10.3.1 RADIATION DETECTION METHODS

Field and laboratory methods used to identifyand quantify concentrations of radionuclides in theenvironment are, in many cases, more exact, lesscostly, and more easily implemented than thoseemployed for chemical analyses. Selection of aradiometric method depends upon the number ofradionuclides of interest, their activities and typesof radiations emitted, as well as on the level ofsensitivity required and the sample size available.In some cases, the selection process requires priorknowledge of the nature and extent of radioactivecontamination present onsite. See the referencesprovided in the box on page 10-12 for detailedguidance on sample collection and preparation,radiochemical procedures, and radiation countersand measurement techniques. The followingdiscussion provides an overview of a few of theradiation detection techniques and instrumentscurrently used to characterize sites contaminatedwith radioactive materials.

Field methods utilize instrumental techniquesrather than radiochemical procedures to determinein-situ identities and concentrations ofradionuclides, contamination profiles, and externalbeta/gamma exposure rates. Field instrumentsdesigned for radiation detection (see Exhibit 10-2) are portable, rugged, and relatively insensitiveto wide fluctuations in temperature and humidity.At the same time, they are sensitive enough todiscriminate between variable levels of backgroundradiation from naturally occurring radionuclidesand excess radiation due to radioactive waste.Because of the harsh conditions in which they are

EXHIBIT 10-2

TYPES OF FIELD RADIATION DETECTION INSTRUMENTS

a None of these surface monitors is suitable for tritium detection.

Source NCRP Report No. 57 (NCRP 1978).

12

sometimes operated, and because their detectionefficiency varies with photon energy, all fieldinstruments should be properly calibrated in thelaboratory against National Bureau of Standards(NBS) radionuclide sources prior to use in thefield. Detector response should also be testedperiodically in the field against NBS check-sourcesof known activity.

Commonly used gamma-ray survey metersinclude Geiger-Muller (G-M) probes, sodiumiodide (NaI(T1)) crystals, and solid-stategermanium diodes (Ge(Li)) coupled to ratemeters,scalers, or multichannel analyzers (MCAs). Theseinstruments provide measurements of overallexposure rates in counts per minute, ormicroRoentgens or microrem per hour. However,only NaI and Ge(Li) detectors with MCAs provideenergy spectra of the gamma rays detected andcan therefore verify the identity of specificradionuclides. Thin window G-M detectors andPancake (ionization) probes are used to detectbeta particles. Alpha-particle surface monitorsinclude portable air proportional, gas proportional,and zinc sulfide (ZnS) scintillation detectors,which all have very thin and fragile windows. Thereferences in the box on this page provideadditional information on several other surveytechniques and instruments, such as aerial gamma

surveillance used to map gamma exposure ratecontours over large areas.

Laboratory methods involve both chemicaland instrumental techniques to quantify low-levelsof radionuclides in sample media. Thepreparation of samples prior to counting is animportant consideration, especially for samplescontaining alpha- and beta-emitting radionuclidesthat either do not emit gamma rays or emitgamma rays of low abundance. Samplepreparation is a multistep process that achievesthe following three objectives: (1) the destructionof the sample matrix (primarily organic material)to reduce alpha- and beta-particle self-absorption(2) the separation and concentration ofradionuclides of interest to increase resolution andsensitivity and (3) the preparation of the samplein a suitable form for counting. Appropriateradioactive tracers (i.e., isotopes of theradionuclides of interest that are not present inthe sample initially, but are added to the sampleto serve as yield determinants) must be selectedand added to the sample before a radiochemicalprocedure is initiated.

For alpha counting, samples are prepared asthin-layer (low mass) sources on membrane filtersby coprecipitation with stable carriers or on metaldiscs by electrodeposition. These sample filters anddiscs are then loaded into gas proportionalcounters, scintillation detectors, or alphaspectrometry systems for measurement (see Exhibit10-3). In a proportional counter, the sample isimmersed in a counting gas, usually methane andargon, and subjected to a high voltage field: alphaemissions dissociate the counting gas creating anionization current proportional to the sourcestrength, which is then measured by the systemelectronics. In a scintillation detector, the sampleis placed in contact with a ZnS phosphor againstthe window of a photomultiplier (PM) tube: alphaparticles induce flashes of light in the phosphorthat are converted to an electrical current in thePM tube and measured. Using alpha spectrometry,the sample is placed in a holder in an evacuatedchamber facing a solid-state, surface-barrierdetector: alpha particles strike the detector andcause electrical impulses, which are sorted bystrength into electronic bins and counted. Allthree systems yield results in counts per minute,which are then converted into activity units usingdetector- and radionuclide-specific calibration

TYPES OF LABORATORY

EXHIBIT 10 -3

RADIATION DETECTION INSTRUMENTSa

Typical ActivityType of Instrument Range (mCi) Typical Sample Form Data Acquisition and Display

Gas Proportional Counters 10 -7 to 10-3 Film disc mount, gas Ratemeter or scaler

Liquid-Scintillation Counters 10 -7 to 10-3 Up to 20 ml of liquid gel Accessories for for background subtraction, quench correction,internal standard sample comparison

NaI (T1) Cylindrical or Well Crystals 10 -6 to 10-3 Liquid, solid, or comtained gas, Ratemeter<4 ml

Discriminators for measuring various energy regions

Multichannel analyzer, or computer plus analog-to-digitalconverter

Computational accessories for full-energy-peak identification,quantification, and spectrum stripping

Ionization Chambers 10 -2 to 103 Liquid, solid, or contained gas Ionization-current measurement;(can be large in size) digital (mCi) readout, as in dose calibrators

Solid-state Detectors 10 -2 to 10 Various Multichannel analyzer or computer with various readoutoptions

aSource: NCRP Report No. 58 (NCRP 1985a).

14

values. Alpha spectrometry is the only system,however, that can be used to identify specificalpha-emitting radionuclides.

For beta counting, samples are prepared bothas thin-sources and as solutions mixed withscintillation fluid, similar in function to aphosphor. Beta-emitting sources are counted ingas proportional counters at higher voltages thanthose applied for alpha counting or in scintillationdetectors using phosphors specifically constructedfor beta-particle detection. Beta-emitters mixedwith scintillation fluid are counted in 20 ml vialsin beta-scintillation counters: beta-particleinteractions with the fluid produce detectable lightflashes. Like alpha detectors, beta detectorsprovide measurements in counts per minute, whichare converted to activity units using calibrationfactors. It should be noted, however, that fewdetection systems are available for determining theidentity of individual beta-emitting radionuclides,because beta particles are emitted as a continuousspectrum of energy that is difficult to characterizeqnd ascribe to any specific nuclide.

It is advisable to count all samples intact ina known geometry on a NaI or Ge(Li) detectorsystem prior to radiochemical analysis, becausemany radionuclides that emit gamma rays insufficient abundanw and energy can be detectedand measured by this process. Even complexgamma-ray spectra emitted by multipleradionuclide sources can be resolved using Ge(Li)detectors, MCAs, and software packages, andspecific radionuclide concentrations can. bedetermined. If the sample activity is low or ifgamma rays are feeble, then more rigorous alphaor beta analyses are advised.

10.3.2 REVIEWING AVAILABLE SITEINFORMATION

In Chapter 4, reference is made to reviewingthe site data for chemical contaminants inaccordance with Stage 1 of the Data QualityObjectives (DQO) process (see box on Page 4-4).This process also applies to radionuclides. Forfurther guidance on the applicability of DQOS toradioactively contaminated sites, consult EPA’sOffice of Radiation Programs.

10.3.3 ADDRESSING MODELINGPARAMETER NEEDS

Exhibits 4-1 and 4-2 describe the elements ofa conceptual model and the types of informationthat may be obtained during a site samplinginvestigation. These exhibits apply to radioactivelycontaminated sites with only minor modifications.For example, additional exposure pathways fordirect external exposure from immersion incontaminated air or water or from contaminatedground surfaces may need to be addressed forcertain radionuclides; these exposure pathways arediscussed further in subsequent sections. Inaddition, several of the parameters identified inthese exhibits are not as important or necessaryfor radiological surveys. For example, theparameters that are related primarily to themodeling of organic contaminants, such as thelipid content of organisms, are typically notneeded for radiological assessments.

10.3.4 DEFINING BACKGROUNDRADIATION SAMPLING NEEDS

As is the case with a chemically contaminatedsite, the background characteristics of aradioactively contaminated site must be definedreliably in order to distinguish natural backgroundradiation and fallout from the onsite sources ofradioactive waste. With the possible exception ofindoor sources of Rn-222, it is often possible tomake these distinctions because the radiationdetection equipment and analytical techniquesused are very precise and sensitive. At achemically contaminated site, there can be manypotential and difficult-to-pinpoint offsite sourcesfor the contamination found onsite, confoundingthe interpretation of field measurements. With aradioactively contaminated site, however, this isnot usually a problem because sources ofradionuclides are, in general, easier to isolate andidentify. In fact, some radionuclides are sospecifically associated with particular industriesthat the presence of a certain radioactivecontaminant sometimes acts as a “fingerprint” toidentify its source. Additional information on thesources of natural background and man-maderadiation in the environment may be found in thereferences listed in the box on the next page.

15

10.3.5 PRELIMINARY IDENTIFICATIONOF POTENTIAL EXPOSURE

Identification of environmental media ofconcern, the types of radionuclides expected at asite, areas of concern (sampling locations), andpotential routes of radionuclide transport throughthe environment is an important part of theradiological risk assessment process. Potentialmedia of concern include soil, ground water,surface water, air, and biota, as discussed inChapter 4. Additional considerations forradioactively contaminated sites are listed below.

Usually a very limited number ofradionuclides at a site contributesignificantly to the risk. During the sitescoping meeting, it is appropriate toconsult with a health physicist not onlyto develop a conceptual model of thefacility, but also to identify the

anticipated critical radionuclides andpathways.

In addition to the environmental mediaidentified for chemically contaminatedsites, radioactively contaminated sitesshould be examined for the potentialpresence of external radiation fields.Many radionuclides emit both beta andgamma radiation, which can createsignificant external exposures.

There are other components in theenvironment that may or may not becritical exposure pathways for the public,but that are very useful indicators of theextent and type of contamination at asite. These components includesediment, aquatic plants, and fish, whichmay concentrate and integrate theradionuclide contaminants that may be(or have been) present in the aquaticenvironment at a site. Accordingly,though some components of theenvironment may or may not beimportant direct routes of exposure toman, they can serve as indicators ofcontamination.

10.3.6 DEVELOPING A STRATEGY FORSAMPLE COLLECTION

The discussions in Chapter 4 regardingsample location, size, type, and frequency apply aswell to radioactively contaminated sites with thefollowing additions and qualifications. First, theresolution and sensitivity of radioanalyticaltechniques permit detection in the environment ofmost radionuclides at levels that are well belowthose that are considered potentially harmful.Analytical techniques for nonradioactive chemicalsare usually not this sensitive.

For radionuclides, continuous monitoring ofthe site environment is important, in addition tothe sampling and monitoring programs describedin Chapter 4. Many field devices that measureexternal gamma radiation, such as continuousradon monitors and high pressure ionizationchambers, provide a real time continuous recordof radiation exposure levels and radionuclideconcentrations. Such devices are useful fordetermining the temporal variation of radiation

16

levels at a contaminated site and for comparingthese results to the variability observed atbackground locations. Continuous measure-meritsprovide an added level of resolution forquantifying and characterizing radiological risk.

Additional factors that affect the frequency ofsampling for radionuclides, besides those discussedin Chapter 4, include the half-lives and the decayproducts of the radionuclides. Radionuclides withshort half-lives, such as Fe-59 (half-life = 44.5days), have to be sampled more frequently becauserelatively high levels of contamination can bemissed between longer sampling intervals. Thedecay products of the radionuclides must also beconsidered, because their presence can interferewith the detection of the parent nuclides ofinterest, and because they also may be importantcontributors to risks.

10.3.7 QUALITY ASSURANCE ANDQUALITY CONTROL (QA/QC)MEASURES

The QA/QC concepts described in Chapter4 also apply to sampling and analysis programs forradionuclides, although the procedures differ.Guidance regarding sampling and measurement ofradionuclides and QA/QC protocols for theiranalyses are provided in the publications listed inthe box on this page.

The QA/QC protocols used for radionuclideanalysis were not developed to meet the evidentialneeds of the Superfund program, however, it islikely that many of the current radiologicalQA/QC guidance would meet the intent ofSuperfund requirements. Some areas whereradiological QA/QC guidance may not meet theintent of Superfund are listed below.

The degree of standardization forradiochemical procedures may be lessrigorous in the QA/QC protocols thanthat required for chemical labs underthe Contract Laboratory Program (CLP).In radiochemical laboratories, severaldifferent techniques may be used toanalyze for a specific radionuclide in agiven matrix with comparable results.The CLP requires all participatingchemical laboratories to use standardizedtechniques.

The required number and type of QCblanks are fewer for radionuclidesamples. For example, a “trip” blank isnot generally used because radionuclidesamples are less likely to becontaminated from direct exposure to airthan are samples of volatile organics.

Limited guidance is available that specifiesfield QA/QC procedures (see the box on thispage). These and other issues related to QA/QCguidance for radiological analyses are discussedfurther in the Section 10.4.

10.4 DATA EVALUATION

Chapter 5 describes the procedures fororganizing and evaluating data collected during asite sampling investigation for use in riskassessment. The ten-step process outlined forchemical data evaluation is generally applicable tothe evaluation of radioactive contaminants,although many of the details must be modified toaccommodate differences in sampling andanalytical methods.

17

10.4.1 COMBINING DATA FROMAVAILABLE SITE INVESTIGATIONS

All available data for the site should begathered for evaluation and sorted byenvironmental medium sampled, analytical methods, and sampling periods. Decisions shouldbe made, using the process described in Section5.1, to combine, evaluate individually, or eliminatespecific data for use in the quantitative riskassessment.

10.4.2 EVALUATING ANALYTICALMETHODS

As with chemical data, radiological datashould be grouped according to the types ofanalyses performed to determine which data areappropriate for use in quantitative risk assessment.Analytical methods for measuring radioactivecontaminants differ from those for measuringorganic and inorganic chemicals. Standardlaboratory procedures for radionuclide analyses arepresented in references, such as those listed in thebox on page 10-12. Analytical methods includealpha, beta, and gamma spectrometry, liquidscintillation counting, proportional counting, andchemical separation followed by spectrometry,depending on the specific radionuclides of interest.

Laboratory accreditation procedures for theanalysis of radionuclides also differ. Radionuclideanalyses are not currently conducted as part of theRoutine Analytical Services (RAS) under theSuperfund CLP. However, these analyses may beincluded under Special Analytical Service-s (SAS).The EPA Environmental RadioactivityIntercomparison Program, coordinated by theNuclear Radiation Assessment Division of theEnvironmental Monitoring Systems Laboratory inLas Vegas (EMSL-LV), provides quality assuranceoversight for participating radiation measurementlaboratories (EPA 1989b). Over 300 federal, state,and private laboratories participate in some phaseof the program, which includes analyses for avariety of radionuclides in media (e.g., water, air,milk, and food) with activity concentrations thatapproximate levels that may be encountered in theenvironment. Similar intercomparison programsfor analysis of thermoluminescent dosimeters(TLDs) for external radiation exposure ratemeasurements are conducted by the DOEEnvironmental Measurements Laboratory (EML)

and the DOE Radiological and EnvironmentalServices Laboratory (RESL).

In both cases, these intercomparisonprograms are less comprehensive than the CLP interms of facility requirements other than analysisof performance evaluation samples, such aslaboratory space and procedural requirements,instrumentation, training, and quality control.However, until such time as radiationmeasurements become fully incorporated in theCLP, use of laboratories that successfullyparticipate in these intercomparison studies maybe the best available alternative for ensuringhigh-quality analytical data. Regardless oflaboratory accreditation, all analytical resultsshould be carefully scrutinized and not acceptedat face value.

As discussed in Chapter 5 for chemicalanalyses, radioanalytical results that are notspecific for a particular radionuclide (e.g., grossalpha, gross beta) may have limited usefulness forquantitative risk assessment. They can be usefulas a screening tool, however. External gammaexposure rate data, although thought of as ascreening measurement, can be directly applied asinput data for a quantitative risk assessment.

10.4.3 EVALUATING QUANTITATIONLIMITS

Lower limits of detection (LLDs), orquantitation limits, for standard techniques formost radionuclide analyses are sufficiently low toensure the detection of nuclides at activityconcentrations well below levels of concern.There are exceptions, however: someradionuclides with very low specific activities, longhalf-lives, and/or low-energy decay emissions (e.g.,I-129, C-14) are difficult to detect precisely usingstandard techniques. To achieve lower LLDs, alaboratory may (1) use more sensitivemeasurement techniques and/or chemicalextraction procedures; (2) analyze larger samplesizes; or (3) increase the counting time of thesample. A laboratory may also choose to applyall three options to increase detection capabilities.Exhibit 10-4 presents examples of typical LLDsusing standard analytical techniques.

The same special considerations noted forchemical analyses would also apply for

18

EXHIBIT 10-4

EXAMPLES OF LOWER LIMITS OF DETECTION (LLD)FOR SELECTED RADIONUCLIDES USING STANDARD ANALYTICAL METHODa

LLDIsotope Sample Mediab pCi Bq Methodology

CO-60

Sr-90

Cs-137

Pb-210

Ra-226

Th-232

U-234U-235U-238

-Water-Soil (dry wt.)-Biota (wet wt.)c

-Aird

-Water

-Water

-Soil (dry wt.)

-Biota (wet wt.)

-Air

-Water-Soil (dry wt.)-Biota (wet wt.)-Air

-Water

-Soil (dry wt.)-Biota (wet wt.)-Air



100.10.125

1

100.310.310.330

0.20.20.25

1000.10.10.10.11

0.020.20.020.3

0.020.10.010.2

0.40.0040.0040.9

0.04

0.40.010.040.010.040.011

0.0070.0070.0070.2

40.0040.0040.0040.0040.04

0.00070.0070.00070.01

0.00070.0040.00040.007

Gamma SpectrometryGamma SpectrometryGamma SpectrometryGamma Spectrometry

Radiochemistry

Gamma SpectrometryRadiochemistryGamma SpectrometryRadiochemistryGamma SpectrometryRadiochemistryGamma Spectrometry

RadiochemistryRadiochemistryRadiochemistryRadiochemistry

Gamma SpectrometryRadiochemistryRadon Daughter EmanationRadon Daughter EmanationRadon Daughter EmanationAlpha Spectrometry

Alpha SpectrometryRadiochemistryAlpha SpectrometryAlpha Proportional Counter

Alpha SpectrometryAlpha SpectrometryAlpha SpectrometryAlpha Spectrometry

(continued)

19


EXAMPLES OF LOWER LIMITS OF DETECTION (LLD)FOR SELECTED RADIONUCLIDES USING STANDARD ANALYTICAL METHODSa

LLDIsotope Sample Mediab pCi Bq Methodology

PU-238 -Water 0.02 0.0007 Alpha SpectrometryPU-239 -Soil (dry wt.) 0.1 0.004 Alpha SpectrometryPU-240 -Biota (wet wt.) 0.01 0.0004 Alpha Spectrometry

-Air 0.2 0.007 Alpha Spectrometry

a

b

c

d

Source: U.S. Environmental Protection Agency Eastern Environmental Radiation Facility (EPA-EERF), Department of EnergyEnvironmental Measurements Laboratory (DOE-EML), and commercial laboratories. Note that LLDs are radionudide-, media-,sample size-, and laboratory-specific higher and lower LLDs than those reported above arc possible. The risk assessor shouldrequest and report the LLDs supplied by the laboratory performing the_

Nominal sample sizes water (1 liter), soil (1 kg dry wt.), biota (1 kg wet wt.) and air (1 filter sample).

Biota includes vegetation, fish, and meat.

Air refers to a sample of 300 m3 of air collected on a filter, which is analyzed for the radionudide of interest.

20

radionuclides that are not detected in any samplesfrom a particular medium, but are suspected to bepresent at a site. In these cases, three options maybe applied: (1) re-analyze the sample using moresensitive methods; (2) use the LLD value as a “proxy” concentration to evaluate the potential risks at the detection limit; or (3) evaluate thepossible risk implication of the radionuclidequalitatively. An experienced health physicistshould decide which of these three options wouldbe most appropriate.

When multiple radionuclides are present in a sample, various interferences can occur that may reduce the analytical sensitivity for a particularradionuclide. Also, in some areas of highbackground radioactivity from naturally occurringradionuclides, it may be difficult to differentiatebackground contributions from incremental sitecontamination. It may be possible to eliminatesuch interferences by radiochemical separation orspecial instrumental techniques.

A sample with activity that is nondetectableshould be reported as less than the appropriatesample and radionuclide-specific LLD value.However, particular caution should be exercisedwhen applying this approach to radionuclides thatare difficult to measure and possess unusually highdetection limits, as discussed previously. In mostcases where a potentially important radionuclidecontaminant is suspected, but not detected, in asample, the sample should be reanalyzed usingmore rigorous radiochemical procedures and moresophisticated detection techniques.

If radionuclide sample data for a site arereported without sample-specific radionuclidequantitation limits, the laboratory conducting theanalyses should be contacted to determine theappropriate LLD values for the analyticaltechniques and sample media.

10.4.4 EVALUATING QUALIFIED ANDCODED DATA

Various data qualifiers and codes may be attached to problem data from inorganic andorganic chemical analyses conducted under theCLP as shown in Exhibits 5-4 and 5-5. Theseinclude laboratory qualifiers assigned by the

laboratory conducting the analysis and datavalidation qualifiers assigned by personnel involvedin data validation. These qualifiers pertain toQA/QC problems and generally indicate questionsconcerning chemical identity, chemicalconcentration, or both. No corresponding systemof qualifiers has been developed for radioanalyticaldata, although certain of the CLP data qualifiersmight be adopted for use in reportingradioanalytical data. The health physicist shoulddefine and evaluate any qualifiers attached to datafor radionuclide analyses. Based on the discussionsin Chapter 5, the references on methods listedabove, and professional judgment, the healthphysicist should eliminate inappropriate data fromuse in the risk assessment.

10.4.5 COMPARING CONCENTRATIONSDETECTED IN BLANKS WITHCONCENTRATIONS DETECTED INSAMPLES

The analysis of blank samples (e.g., laboratoryor reagent blanks, field blanks, calibration blanks)is an important component of a properradioanalytical program. Analysis of blanksprovides a measure of contamination introducedinto a sample during sampling or analysisactivities.

The CLP provides guidance for inorganic andorganic chemicals that are not common laboratorycontaminants. According to this guidance, if ablank contains detectable levels of any uncommonlaboratory chemical, site sample results should beconsidered positive only if the measuredconcentration in the sample exceeds five times themaximum amount detected in any blank. Samplescontaining less than five times the blankconcentration should be classified as nondetects,and the maximum blank-related concentrationshould be specified as the quantitation limit forthat chemical in the sample. Though they arenot considered to be common laboratorycontaminants, radionuclides should not beclassified as nondetects using the above CLPguidance. Instead, the health physicist shouldevaluate all active sample preparation andanalytical procedures for possible sources ofcontamination.

21

10.4.6 EVALUATING TENTATIVELYIDENTIFIED RADIONUCLIDES

Because radionuclides are not included on theTarget Compound List (TCL), they may beclassified as tentatively identified compounds(TICs) under CLP protocols. In reality, however,radioanalytical techniques are sufficiently sensitivethat the identity and quantity of radionuclides ofpotential concern at a site can be determined witha high degree of confidence. In some cases,spectral or matrix interferences may introduceuncertainties, but these problems usually can beovercome using special radiochemical and/orinstrumental methods. In cases where aradionuclide’s identity is not sufficientlywell-defined by the available data set: (1) furtheranalyses may be performed using more sensitivemethods, or (2) the tentatively identifiedradionuclide may be included in the riskassessment as a contaminant of potential concernwith notation of the uncertainty in its identity andconcentration.

10.4.7 COMPARING SAMPLES WITHBACKGROUND

It is imperative to select, collect, and analyzean appropriate number of background samples tobe able to distinguish between onsite sources ofradionuclide contaminants from radionuclidesexpected normally in the environment.Background measurements of direct radiation andradionuclide concentrations in all media ofconcern should be determined at samplinglocations geologically similar to the site, butbeyond the influence of the site. Screeningmeasurements (e.g., gross alpha, beta, and gamma)should be used to determine whether moresensitive radionuclide-specific analyses arewarranted. Professional judgment should be usedby the health physicist to select appropriatebackground sampling locations and analyticaltechniques. The health physicist should alsodetermine which naturally occurring radionuclides(e.g., uranium, radium, or thorium) detected onsiteshould be eliminated from the quantitative riskassessment. All man-made radionuclides detectedin samples collected should, however, be retainedfor further consideration.

10.4.8 DEVELOPING A SET OFRADIONUCLIDE DATA ANDINFORMATION FOR USE IN ARISK ASSESSMENT

The process described in Section 5.8 forselection of chemical data for inclusion in thequantitative risk assessment generally applies forradionuclides as well. One exception is the lackof CLP qualifiers for radionuclides, as discussedpreviously. Radionuclides of concern shouldinclude those that are positively detected in atleast one sample in a given medium, at levelssignificantly above levels detected in blank samplesand significantly above local background levels.As discussed previously, the decision to includeradionuclides not detected in samples from anymedium but suspected at the site based onhistorical information should be made by aqualified health physicist.

10.4.9 GROUPING RADIONUCLIDES BYCLASS

Grouping radionuclides for consideration inthe quantitative risk assessment is generallyunnecessary and inappropriate. Radiation doseand resulting health risk is highly dependent onthe specific properties of each radionuclide. Insome cases, however, it may be acceptable togroup different radioisotopes of the same elementthat have similar radiological characteristics (e.g.,Pu-238/239/240, U-235/238) or belong to the samedecay series. Such groupings should be determinedvery selectively and seldom offer any significantadvantage.

10.4.10 FURTHER REDUCTION IN THENUMBER OF RADIONUCLIDES

For sites with a large number ofradionuclides detected in samples from one ormore media, the risk assessment should focus ona select group of radionuclides that dominate theradiation dose and health risk to the criticalreceptors. For example, when consideringtransport through ground water to distantreceptors, transit times may be very longconsequently, only radionuclides with longhalf-lives or radioactive progeny that are formedduring transport may be of concern for thatexposure pathway. For direct external exposures,high-energy gamma emitters are of principal

22

concern, whereas alpha-emitters may dominatedoses from the inhalation and ingestion pathways.The important radionuclides may differ for eachexposure pathway and must be determined ontheir relative concentrations, half-lives,environmental mobility, and dose conversionfactors (see Section 10.5 for discussion of doseconversion factors) for each exposure pathway ofinterest.

The total activity inventory and individualconcentrations of radionuclides at a Superfund sitewill change with time as some nuclides decayaway and others “grow in” as a result ofradioactive decay processes. Consequently, it maybe important to evaluate different time scales inthe risk assessment. For example, at a site whereRa-226 (half-life = 1600 years) is the onlycontaminant of concern in soil at some initialtime, the Pb-210 (half-life = 22.3 years) andPo-210 (half-life = 138 days) progeny will alsobecome dominant contributors to the activityonsite over a period of several hundred years.

10.4.11 SUMMARIZING AND PRESENTINGDATA

Presentation of results of the data collectionand evaluation process will be generally the samefor radionuclides and chemical contaminants. Thesample table formats presented in Exhibits 5-6 and5-7 are equally applicable to radionuclide data,except that direct radiation measurement datashould be added, if appropriate for theradionuclides and exposure pathways identified atthe site.

10.5 EXPOSURE AND DOSEASSESSMENT

This section describes a methodology forestimating the radiation dose equivalent tohumans from potential exposures to radionuclidesthrough all pertinent exposure pathways at aremedial site. These estimates of dose equivalentmay be used for comparison with radiationprotection standards and criteria. However, thismethodology has been developed for regulation ofoccupational radiation exposures for adults and isnot completely applicable for estimating healthrisk to the general population. Section 10.7.2,

therefore, describes a separate methodology forestimating health risk.

Chapter 6 describes the procedures forconducting an exposure assessment for chemicalcontaminants as part of the baseline riskassessment for Superfund sites. Though manyaspects of the discussion apply to radionuclides,the term “exposure” is used in a fundamentallydifferent way for radionuclides as compared tochemicals. For chemicals, exposure generallyrefers to the intake (e.g., inhalation, ingestion,dermal exposure) of the toxic chemical, expressedin units of mg/kg-day. These units are convenientbecause the toxicity values for chemicals aregenerally expressed in these terms. For example,the toxicity value used to assess carcinogeniceffects is the slope factor, expressed in units ofrisk of lifetime excess cancers per mg/kg-day. Asa result, the product of the intake estimate withthe slope factor yields the risk of cancer (withproper adjustments made for absorption, ifnecessary.

Intakes by inhalation, ingestion, andabsorption are also potentially important exposurepathways for radionuclides, although radionuclideintake is typically expressed in units of activity(i.e., Bq or Ci) rather than mass. Radionuclidesthat enter through these internal exposurepathways may become systemically incorporatedand emit alpha, beta, or gamma radiation withintissues or organs. Unlike chemical assessments,an exposure assessment for radioactivecontaminants can include an explicit estimation ofthe radiation dose equivalent. As discussedpreviously in Section 10.1, the dose equivalent isan expression that takes into consideration boththe amount of energy deposited in a unit mass ofa specific organ or tissue as a result of theradioactive decay of a specific radionuclide, aswell as the relative biological effectiveness of theradiations emitted by that nuclide. (Note that theterm dose has a different meaning forradionuclides [dose = energy imparted to a unitmass of tissue] than that used in Chapter 6 forchemicals [dose, or absorbed dose = masspenetrating into an organism].)

Unlike chemicals, radionuclides can havedeleterious effects on humans without being takeninto or brought in contact with the body. This isbecause high energy beta particles and photons

23

from radionuclides in contaminated air, water, orsoil can travel long distances with only minimumattenuation in these media before depositing their energy in human tissues. External radiationexposures can result from either exposure toradionuclides at the site area or to radionuclidesthat have been transported from the site to otherlocations in the environment. Gamma and x-raysare the most penetrating of the emitted radiations,and comprise the primary contribution to theradiation dose from external exposures. Alphaparticles are not sufficiently energetic to penetratethe outer layer of skin and do not contributesignificantly to the external dose. Externalexposure to beta particles primarily imparts a doseto the outer layer skin cells, although high-energybeta radiation can penetrate into the human body.

The quantification of the amount of energydeposited in living tissue due to internal andexternal exposures to radiation is termed radiationdosimetry. The amount of energy deposited inliving tissue is of concern because the potentialadverse effects of radiation are proportional toenergy deposition. The energy deposited in tissuesis proportional to the decay rate of a radionuclide,and not its mass. Therefore, radionuclidequantities and concentrations are expressed inunits of activity (e.g., Bq or Ci), rather than inunits of mass.

Despite the fundamental difference betweenthe way exposures are expressed for radionuclidesand chemicals, the approach to exposureassessment presented in Chapter 6 for chemicalcontaminants largely applies to radionuclidecontaminants. Specifically, the three steps of anexposure assessment for chemicals also apply toradionuclides: (1) characterization of the exposuresetting; (2) identification of the exposurepathways; and (3) quantification of exposure.However, some of the methods by which thesethree steps are carried out are different forradionuclides.

10.5.1 CHARACTERIZING THE EXPOSURESETTING

Initial characterization of the exposure settingfor radioactively contaminated sites is virtuallyidentical to that described in Chapter 6. Oneadditional consideration is that, at sites suspectedof having radionuclide contamination, a survey

should be conducted to determine externalradiation fields using any one of a number of fieldsurvey instruments (preferably, G-M tubes andNaI(T1) field detectors) (see Exhibit 10-2). Healthand safety plans should be implemented to reducethe possibility of radiation exposures that are inexcess of allowable limits.

10.5.2 IDENTIFYING EXPOSUREPATHWAYS

The identification of exposure pathways forradioactively contaminated sites is very similar tothat described in Chapter 6 for chemicallycontaminated sites, with the following additionalguidance.

In addition to the various ingestion,inhalation, and direct contact pathwaysdescribed in Chapter 6, external exposureto penetrating radiation should also beconsidered. Potential external exposure

24

pathways to be considered includeimmersion in contaminated air,immersion in contaminated water, andradiation exposure from ground surfacescontaminated with beta- and photon-emitting radionuclides.

A with nonradioactive chemicals,environmentally dispersed radionuclidesare subject to the same chemicalprocesses that may accelerate or retardtheir transfer rates and may increase ordecrease their bioaccumulationpotentials. These transformationprocesses must be taken intoconsideration during the exposureassessment.

Radionuclides undergo radioactive decaythat, in some respects, is similar to thechemical or biological degradation oforganic compounds. Both processesreduce the quantity of the hazardoussubstance in the environment andproduce other substances. (Note,however, that biological and chemicaltransformations can never alter, i.e.,either increase or decrease, theradioactivity of a radionuclide.)Radioactive decay products can alsocontribute significantly to the radiationexposure and must be considered in theassessment.

Chapter 6 presents a series of equations(Exhibits 6-11 through 6-19) forquantification of chemical exposures.These equations and suggested defaultvariable values may be used to estimateradionuclide intakes as a firstapproximation, if the equations aremodified by deleting the body weight andaveraging time from the denominator.However, depending upon thecharacteristics of the radionuclides ofconcern, consideration of radioactivedecay and ingrowth of radioactive decayproducts may be important additions, aswell as the external exposure pathways.

Chapter 6 also refers to a number ofcomputer models that are used to predictthe behavior and fate of chemicals in the

environment. While those models maybe suitable for evaluations of radioactivecontaminants in some cases, numerousmodels have been developed specificallyfor evaluating the transport ofradionuclides in the environment andpredicting the doses and risks to exposedindividuals, In general, modelsdeveloped specifically for radiologicalassessments should be used. Suchmodels include, for example, explicitconsideration of radioactive decay andingrowth of radioactive decay products.(Contact ORP for additional guidance onthe fate and transport modelsrecommended by EPA.)

10.5.3 QUANTIFYING EXPOSURE:GENERAL CONSIDERATIONS

One of the primary objectives of an exposureassessment is to make a reasonable estimate ofthe maximum exposure to individuals and criticalpopulation groups. The equation presented inExhibit 6-9 to calculate intake for chemicals maybe considered. to be applicable to exposureassessment for radionuclides, except that the bodyweight and averaging time terms in thedenominator should be omitted. However, asdiscussed previously, exposures to radionuclidesinclude both internal and external exposurepathways. In addition, radiation exposureassessments do not end with the calculation ofintake, but take the calculation an additional stepin order to estimate radiation dose equivalent.

The radiation dose equivalent to specifiedorgans and the effective dose equivalent due tointakes of radionuclides by inhalation or ingestionare estimated by multiplying the amount of eachradionuclide inhaled or ingested times appropriatedose conversion factors (DCFs), which representthe dose equivalent per unit intake. As notedpreviously, the effective dose equivalent is aweighted sum of the dose equivalents to allirradiated organs and tissues, and represents ameasure of the overall detriment. FederalGuidance Report No. 11 (EPA 1988) providesDCFs for each of over 700 radionuclides for bothinhalation and ingestion exposures. It isimportant to note, however, that these DCFs weredeveloped for regulation of occupational exposures

2S

to radiation and may not be appropriate for thegeneral population.

Radionuclide intake by inhalation andingestion is calculated in the same manner aschemical intake except that it is not divided bybody weight or averaging time. For radionuclides,a reference body weight is already incorporatedinto the DCFs, and the dose is an expression ofenergy deposited per gram of tissue.

If intake of a radionuclide is defined for aspecific time period (e.g., Bq/year), the doseequivalent will be expressed in correspondinglimits (e.g., Sv/year). Because systemicallyincorporated radionuclides can remain within thebody for long periods of time, internal dose isbest expressed in terms of the committed effectivedose equivalent, which is equal to the effectivedose equivalent over the 50-year period followingintake.

External exposures may be determined bymonitoring and sampling of the radionuclideconcentrations in environmental media, directmeasurement of radiation fields using portableinstrumentation, or by mathematical modeling.Portable survey instruments that have beenproperly calibrated can display dose rates (e.g.,Sv/hr), and dose equivalents can be estimated bymultiplying by the duration of exposure to theradiation field. Alternatively, measured orpredicted concentrations in environmental mediamay be multiplied by DCFs, which relateradionuclide concentrations on the ground, in air,or in water to external dose rates (e.g., Sv/hr perBq/m2 for ground contamination or Sv/hr perBq/m3 for air or water immersion).

The dose equivalents associated with externaland internal exposures are expressed in identicalunits (e.g., Sv), so that contributions from allpathways can be summed to estimate the totaleffective dose equivalent value and prioritize riskfrom different sources.

In general, radiation exposure assessmentsneed not consider acute toxicity effects. Acuteexposures are of less concern for radionuclidesthan for chemicals because the quantities ofradionuclides required to cause adverse effectsfrom acute exposure are extremely large and suchlevels are not normally encountered at Superfund

sites. Toxic effects from acute radiation exposuresare possible when humans are exposed to theradiation from large amounts of radioactivematerials released during a major nuclear plantaccident, such as Chernobyl, or duringabove-ground weapons detonations. Consequently,the exposure and risk assessment guidance forradionuclides presented in this chapter is limitedto situations causing chronic exposures to lowlevels of radioactive contaminants.

10.5.4 QUANTIFYING EXPOSUREDETERMINING EXPOSURE POINTCONCENTRATIONS

The preferred method for estimating theconcentration of chemical or radioactivecontaminants at those places where members ofthe public may come into contact with them is bydirect measurement. However, this will not bepossible in many circumstances and it may benecessary, therefore, to use environmental fate andtransport models to predict contaminantconcentrations. Such modeling would benecessary, for example: (1) when it is not possibleto obtain representative samples for allradionuclides of concern; (2) when thecontaminant has not yet reached the potentialexposure points; and (3) when the contaminantsare below the limits of detection but, if present,can still represent a significant risk to the public.

Numerous fate and transport models havebeen developed to estimate contaminantconcentrations in ground water, soil, air, surfacewater, sediments, and food chains. Modelsdeveloped for chemical contaminants, such asthose discussed in Chapter 6, may also be appliedto radionuclides with allowance for radioactivedecay and ingrowth of decay products. There arealso a number of models that have beendeveloped specifically for radionuclides. Thesemodels are similar to the models used for toxicchemicals but have features that make themconvenient to use for radionuclide pathwayanalysis, such as explicit consideration ofradioactive decay and daughter ingrowth.Available models for use in radiation riskassessments range in complexity from a series ofhand calculations to major computer codes. Forexample, NRC Regulatory Guide 1.109 presentsa methodology that may be used to manuallyestimate dose equivalents from a variety of

26

exposure pathways (NRC 1977). Examples ofcomputerized radiological assessment modelsinclude the AIRDOS-EPA code and theEPA-PRESTO family of codes, which are usedextensively by EPA to estimate exposures anddoses to populations following atmosphericreleases of radionuclides and releases from alow-level waste disposal facility, respectively.Guidance on selection and use of the variousmodels can be obtained from the EPA Office ofRadiation Programs.

Exhibit 6-10, Example of Table Format forSummarizing Exposure Concentrations, may beused for radionuclide contaminants, except thatradionuclide concentrations are expressed in termsof activity per unit mass or volume of theenvironmental medium (e.g., Bq/kg, Bq/L) ratherthan mass.

10.5.5 QUANTIFYING EXPOSURE:ESTIMATING INTAKE AND DOSEEQUIVALENT

Section 6.6 presents a description of themethods used to estimate intake rates ofcontaminants from the various exposure pathways.Exhibits 6-11 to 6-19 present the equations andinput assumptions recommended for use in intakecalculations. In concept, those equations andassumptions also apply generally to radionuclides,except that the body weight and averaging timeterms in the denominators should be omitted.However, as discussed previously, the product ofthese calculations for radionuclides is an estimateof the radionuclide intake, expressed in units ofactivity (e.g., Bq), as opposed to mg/kg-day. Inaddition, the endpoint of a radiation exposureassessment is radiation dose, which is calculatedusing DCFs as explained below. AS explainedpreviously, dose equivalents calculated in thefollowing manner should be used to compare withradiation protection standards and criteria, not toestimate risk.

Internal Exposure. Exhibits 6-11,6-12,6-14,6-17, 6-18, and 6-19 present simplified models for the ingestion of water, food, and soil as pathwaysfor the intake of environmental contaminants.The recommended assumptions for ingestion ratesand exposure durations are applicable toradionuclide exposures and may be used toestimate the intake rates of radionuclides by these

pathways. As noted previously, however, theseintake estimates for radionuclidcs should not bedivided by the body weight or averaging time.These intake rates must be multiplied byappropriate DCF values in order to obtaincommitted effective dose equivalent values. Themore rigorous and complex radionuclide pathwaymodels noted previously typically require muchmore extensive input data and may include defaultparameter values that differ somewhat from thevalues recommended in these exhibits.

Exhibit 6-16 presents the equation andassumptions used to estimate the contaminantintake from air. For radionuclides, the dose frominhalation of contaminated air is determined asthe product of the radionuclide concentration inair (Bq/m3), the breathing rate (m3 per day oryear), exposure duration (day or year), and theinhalation DCF (Sv per Bq inhaled). The resultof this calculation is the committed effective doseequivalent, in units of Sv.

Chapter 6 points out that dermal absorptionof airborne chemicals is not an important routeof uptake. This point is also true for mostradionuclides, except airborne tritiated watervapor, which is efficiently taken into the bodythrough dermal absorption. In order to accountfor this route of uptake, the inhalation DCF fortritium includes an adjustment factor tofor dermal absorption.

External Exposure. Immersioncontaining certain beta-emitting

account

in airand/or

photon-emitting radioactive contaminants can alsoresult in external exposures. Effective doseequivalents from external exposure are calculatedas the product of the airborne radionuclideconcentration (Bq/m3), the external DCF for airimmersion (Sv/hr per Bq/m3), and the duration ofexposure (hours).

Exhibits 6-13 and 6-15 illustrate the dermaluptake of contaminants resulting from immersionin water or contact with soil. This route ofuptake can be important for many organicchemicals; however, dermal uptake is generally notan important route of uptake for radionuclides,which have small dermal permeability constants.External radiation exposure due to submersion inwater contaminated with radionuclides is possibleand is similar to external exposure due to

27

immersion in air. However, because of theshielding effects of water and the generally shortdurations of such exposures, immersion in wateris typically of lesser significance. The product ofthe radionuclide concentration in water (Bq/m3),the relevant DCF (Sv/hr per Bq/m3), and theduration of exposure (hours) yields effective doseequivalent.

The third external exposure pathway ofpotential significance is irradiation fromradionuclides deposited on the ground surface.Effective dose equivalents resulting from thispathway may be estimated as the product of thesoil surface concentration ( B q / m2) ofphoton-emitting radionuclides of concern, theexternal DCF for ground surface exposure (Sv/hrper Bq/m2), and the duration of exposure (hours).

10.5.6 COMBINING INTAKES AND DOSES\ ACROSS PATHWAYS

The calculations described previously resultin estimates of committed effective doseequivalents (Sv) from individual radionuclides viaa large number of possible exposure pathways.Because a given population may be subject tomultiple exposure pathways, the results of theexposure assessment should be organized bygrouping all applicable exposure pathways for eachexposed population. Risks from various exposurepathways and contaminants then can be integratedduring the risk characterization step (see Section10.7).

10.5.7 EVALUATING UNCERTAINTY

The radiation exposure assessment shouldinclude a discussion of uncertainty, that, at aminimum, should include: (1) a tabular summaryof the values used to estimate exposures and dosesand the range of these values; and (2) a summaryof the major assumptions of the exposureassessment, including the uncertainty associatedwith each assumption and how it might affect theexposure and dose estimates. Sources ofuncertainty that must be addressed include: (1)how well the monitoring data represent actual siteconditions; (2) the exposure models, assumptions,and. input variables used to estimate exposurepoint concentrations; and (3) the values of thevariables used to estimate intakes and externalexposures. More comprehensive discussions of

uncertainty associated with radiological riskassessment are provided in the BackgroundInformation Document for the Draft EIS forProposed NESHAPS for Radionuclides (EPA1989a), Radiological Assessment (Till and Meyer1983), and NCRP Report No. 76 (NCRP 1984a).

10.5.8 SUMMARIZING AND PRESENTINGEXPOSURE ASSESSMENT RESULTS

Exhibit 6-22 presents a sample format forsummarizing the results of the exposureassessment. The format may also be used forradionuclide contaminants except that the entriesshould be specified as committed effective doseequivalents (Sv) and the annual estimated intakes(Bq) for each radionuclide of concern. Theintakes and dose estimates should be tabulatedfor each exposure pathway so that the mostimportant radionuclides and pathways contributingto the total health risk may be identified.

The information should be organized byexposure pathway, population exposed, and currentand future use assumptions. For radionuclides,however, it may not be necessary to summarizeshort-term and long-term exposures separately asspecified for chemical contaminants.

10.6 TOXICITY ASSESSMENT

Chapter 7 describes the two-step processemployed to assess the potential toxicity of a givenchemical contaminant. The first step, hazardidentification, is used to determine whetherexposure to a contaminant can increase theincidence of an adverse health effect. The secondstep, dose-response assessment, is used toquantitatively evaluate the toxicity information andcharacterize the relationship between the dose ofthe contaminant administered or received and theincidence of adverse health effects in the exposedpopulation.

There are certain fundamental differencesbetween radionuclides and chemicals thatsomewhat simplify toxicity assessment forradionuclides. As discussed in the previoussections, the adverse effects of exposure toradiation are due to the energy deposited insensitive tissue, which is referred to as the

28

radiation dose. In theory, any dose of radiationhas the potential to produce an adverse effect.Accordingly, exposure to my radioactivesubstances is, by definition, hazardous.

Dose-response assessment for radionuclidesis also more straightforward. The type of effectsand the likelihood of occurrence of any one of anumber of possible adverse effects from radiationexposure depends on the radiation dose. Therelationship between dose and effect is relativelywell characterized (at high doses) for most typesof radiations. As a result, the toxicity assessment,within the context that it is used in this manual,need not be explicitly addressed in detail forindividual radionuclides at each contaminated site.

The sections that follow provide a briefsummary of the human and experimental animalstudies that establish the hazard and dose-responserelationship for radiation exposure. More detaileddiscussions of radiation toxicity are provided inpublications of the National Academy of Sciencescommittee on Biological Effects of IonizingRadiation (BEIR), the United Nations ScientificCommittee on Effects of Atomic Radiation(UNSCEAR), NRC, NCRP, and ICRP listed inthe box on this page.

10.6.1 HAZARD IDENTIFICATION

The principal adverse biological effectsassociated with ionizing radiation exposures fromradioactive substances in the environment arecarcinogenicity, mutagenicity, and teratogenicity.Carcinogenicity is the ability to produce cancer.Mutagenicity is the property of being able toinduce genetic mutation, which may be in thenucleus of either somatic (body) or germ(reproductive) cells. Mutations in germ cells leadto genetic or inherited defects. Teratogenicityrefers to the ability of an agent to induce orincrease the incidence of congenital malformationsas a result of permanent structural or functionaldeviations produced during the growth anddevelopment of an embryo (more commonlyreferred to as birth defects). Radiation mayinduce other deleterious effects at acute dosesabove about 1 Sv, but doses of this magnitude arenot normally associated with radioactivecontamination in the environment.

As discussed in Section 10.1, ionizingradiation causes injury by breaking molecules intoelectrically charged fragments (i.e., free radicals),thereby producing chemical rearrangements thatmay lead to permanent cellular damage. Thedegree of biological damage caused by varioustypes of radiation varies according to how spatiallyclose together the ionizations occur. Someionizing radiations (e.g., alpha particles) producehigh density regions of ionization. For thisreason, they are called high-LET (linear energytransfer) particles. Other types of radiation (e.g.,x-rays, gamma rays, and beta particles) are calledlow-LET radiations because of the low densitypattern of ionization they produce. In equaldoses, the carcinogenicity and mutagenicity of

29

high-LET radiations may be an order ofmagnitude or more greater than those of low-LETradiations, depending on the endpoint beingevaluated. The variability in biologicaleffectiveness is accounted for by the quality factorused to calculate the dose equivalent (see Section10.1).

Carcinogenesis. An extensive body ofliterature exists on radiation carcinogenesis in manand animals. This literature has been reviewedmost recently by the United Nations ScientificCommittee on the Effects of Atomic Radiation(UNSCEAR) and the National Academy ofSciences Advisory Committee on the BiologicalEffects of Ionizing Radiations (NAS-BEIRCommittee) (UNSCEAR 1977, 1982, 1988; NAS1972, 1980, 1988). Estimates of the average riskof fatal cancer from low-LET radiation from thesestudies range from approximately 0.007 to 0.07fatal cancers per sievert.

An increase in cancer incidence or mortalitywith increasing radiation dose has beendemonstrated for many types of cancer in bothhuman populations and laboratory animals(UNSCEAR 1982, 1988; NAS 1980, 1988).Studies of humans exposed to internal or externalsources of ionizing radiation have shown that theincidence of cancer increases with increasedradiation exposure. This increased incidence,however, is usually associated with appreciablygreater doses and exposure frequencies than thoseencountered in the environment. Therefore, riskestimates from small doses obtained over longperiods of time are determined by extrapolatingthe effects observed at high, acute doses.Malignant tumors in various organs most oftenappear long after the radiation exposure, usually10 to 35 years later (NAS 1980, 1988; UNSCEAR1982, 1988). Radionuclide metabolism can resultin the selective deposition of certain radionuclidesin specific organs or tissues, which, in turn, canresult in larger radiation doses andhigher-than-normal cancer risk in these organs.

Ionizing radiation can be consideredpancarcinogenic, i.e., it acts as a completecarcinogen in that it serves as both initiator andpromoter, and it can induce cancers in nearly anytissue or organ. Radiation-induced cancers inhumans have been reported in the thyroid, femalebreast, lung, bone marrow (leukemia), stomach,

liver, large intestine, brain, salivary glands, bone,esophagus, small intestine, urinary bladder,pancreas, rectum, lymphatic tissues, skin, pharynx,uterus, ovary, mucosa of cranial sinuses, andkidney (UNSCEAR 1977, 1982, 1988; NAS 1972,1980, 1988). These data are taken primarily fromstudies of human populations exposed to highlevels of radiation, including atomic bombsurvivors, underground miners, radium dialpainters, patients injected with thorotrast orradium, and patients who received high x-ray dosesduring various treatment programs. Extrapolationof these data to much lower doses is the majorsource of uncertainty in determining low-levelradiation risks (see EPA 1989a). It is assumedthat no lower threshold exists for radiationcareinogenesis.

On average, approximately 50 percent of allof the cancers induced by radiation are lethal.The fraction of fatal cancers is different for eachtype of cancer, ranging from about 10 percent inthe case of thyroid cancer to 100 percent in thecase of liver cancer (NAS 1980, 1988). Femaleshave approximately 2 times as many total cancersas fatal cancers following radiation exposure, andmales have approximately 1.5 times as many (NAS1980).

Mutagenesis. Very few quantitative data areavailable on radiogenic mutations in humans,particularly from low-dose exposures. Somemutations are so mild they are not noticeable,while other mutagenic effects that do occur aresimilar to nonmutagenic effects and are thereforenot necessarily recorded as mutations. The bulkof data supporting the mutagenic character ofionizing radiation comes from extensive studies ofexperimental animals (UNSCEAR 1977, 1982,1988; NAS 1972, 1980, 1988). These studies havedemonstrated all forms of radiation mutagenesis,including lethal mutations, translocations,inversions, nondisjunction, and point mutations.Mutation rates calculated from these studies areextrapolated to humans and form the basis forestimating the genetic impact of ionizing radiationon humans (NAS 1980, 1988 UNSCEAR 19821988). The vast majority of the demonstratedmutations in human germ cells contribute to bothincreased mortality and illness (NAS 1980;UNSCEAR 1982). Moreover, the radiationprotection community is generally in agreementthat the probability of inducing genetic changes

30

increases linearly with dose and that no“threshold” dose is required to initiate heritabledamage to germ cells.

The incidence of serious genetic disease dueto mutations and chromosome aberrations induced by radiation is referred to as genetic detriment.Serious genetic disease includes inherited illhealth, handicaps, or disabilities. Genetic diseasemay be manifest at birth or may not becomeevident until some time in adulthood.Radiation-induced genetic detriment includesimpairment of life, shortened life span, andincreased hospitalization. The frequency ofradiation-induced genetic impairment is relativelysmall in comparison with the magnitude ofdetriment associated with spontaneously arisinggenetic diseases (UNSCEAR 1982, 1988).

Teratogenesis. Radiation is a well-known teratogenic agent. The developing fetus is much“more sensitive to radiation than the mother. Theage of the fetus at the time of exposure is themost important factor in determining the extentand type of damage from radiation. Themalformations produced in the embryo depend onwhich cells, tissues, or organs in the fetus aremost actively differentiating at the time ofradiation exposure. Embryos are relativelyresistant to radiation-induced teratogenic effectsduring the later stages of their development andare most sensitive from just after implantationuntil the end of organogenesis (about two weeksto eight weeks after conception) (UNSCEAR1986, Brent 1980). Effects on nervous system,skeletal system, eyes, genitalia, and skin have beennoted (Brent 1980). The brain appears to bemost sensitive during development of theneuroblast (these cells eventually become thenerve cells). The greatest risk of brain damagefor the human fetus occurs at 8 to 15 weeks,which is the time the nervous system isundergoing the most rapid differentiation andproliferation of cells (Otake 1984).

10.6.2 DOSE-RESPONSE RELATIONSHIPS

This section describes the relationship of therisk of fatal cancer, serious genetic effects, andother detrimental health effects to exposure to lowlevels of ionizing radiation. Most important fromthe standpoint of the total societal risk fromexposures to low-level ionizing radiation are the

risks of cancer and genetic mutations. Consistentwith our current understanding of their origins interms of DNA damage, these effects are believedto be stochastic that is, the probability (risk) ofthese effects increases with the dose of radiation,but the severity of the effects is independent ofdose. For neither induction of cancer nor geneticeffects, moreover, is there any convincing evidencefor a “threshold” (i.e., some dose level belowwhich the risk is zero). Hence, so far as isknown, any dose of ionizing radiation, no matterhow small, might give rise to a cancer or to agenetic effect in future generations. Conversely,there is no way to be certain that a given dose ofradiation, no matter how large, has caused anobserved cancer in an individual or will cause onein the future.

Exhibit 10-5 summarizes EPA’s currentestimates of the risk of adverse effects associatedwith human exposure to ionizing radiation (EPA1989a). Important points from this summary tableare provided below.

Very large doses (> 1 Sv) of radiationare required to induce acute andirreversible adverse effects. It is unlikelythat such exposures would occur in theenvironmental setting associated with apotential Superfund site.

The risks of serious noncarcinogeniceffects associated with chronic exposureto radiation include genetic andteratogenic effects. Radiation-inducedgenetic effects have not been observedin human populations, and extrapolationfrom animal data reveals risks per unitexposure that are smaller than, orcomparable to, the risk of cancer. Inaddition, the genetic risks are spreadover several generations. The risks perunit exposure of serious teratogeniceffects are greater than the risks ofcancer. However, there is a possibilityof a threshold, and the exposures mustoccur over a specific period of timeduring gestation to cause the effect.Teratogenic effects can be induced onlyduring the nine months of pregnancy.Genetic effects are induced during the30-year reproductive generation andcancer can be induced at any point

31

EXHIBIT 10-5

SUMMARY OF EPA’S RADIATION RISK FACTORSa

Risk Significant Exposure Period Risk Factor Range

Low LET (Gy-l)

Teratogenic:b

Severe mental retardation Weeks 8 to 15 of gestation

GeneticSevere hereditary defects, 30-year reproductive generationall generations

SomaticFatal cancers Lifetime

In uteroAll cancers Lifetime

0.25-0.55

0.006-0.11

0.012-0.120.029-0.100.019-0.19

High LET (Gy-1)

GeneticSevere hereditary defects, 30-year reproductive generation 0.016-0.29all generations

Somatic:Fatal cancers Lifetime 0.096-0.96All cancers Lifetime 0.15-1.5

Radon Decay Products (10-6 WLM-l).Fatal lung cancer Lifetime 140-720

a In addition to the stochastic risks indicated, acute toxicity may occur at a mean lethal dose of 3-5 Svwith a threshold in excess of 1 Sv.

b The range assumes a linear, non-threshold dose-response. However, it is plausible that a thresholdmay exist for this effect.

32

during the lifetime. If a radiation sourceis not controlled, therefore, thecumulative risk of cancer may be manytimes greater than the risk of genetic orteratogenic effects due to the potentiallylonger period of exposure.

Based on these observations, it appears thatthe risk of cancer is limiting and may be used asthe sole basis for assessing the radiation-relatedhuman health risks of a site contaminated withradionuclides.

For situations where the risk of cancer induction in a specific target organ is of primaryinterest, the committed dose equivalent to thatorgan may be multiplied by an organ-specific riskfactor. The relative radiosensitivity of variousorgans (i.e., the cancer induction rate per unitdose) differs markedly for different organs andvaries as a function of the age and sex of theexposed individual. Tabulations of such riskfactors as a function of age and sex are providedin the Background Information Document for theDraft Environmental Impact Statement for ProposedNESHAPS for Radionuclides (EPA 1989a) forcancer mortality and cancer incidence.

10.7 RISK CHARACTERIZATION

The final step in the risk assessment processis risk characterization. This is an integration stepin which the risks from individual radionuclidesand pathways are quantified and combined whereappropriate. Uncertainties also are examined anddiscussed in this step.

10.7.1 REVIEWING OUTPUTS FROM THETOXICITY AND EXPOSUREASSESSMENTS

The exposure assessment results should beexpressed as estimates of radionuclide intakes byinhalation and ingestion, exposure rates andduration for external exposure pathways, andcommitted effective dose equivalents to individualsfrom all relevant radionuclides and pathways. Therisk assessor should compile the supportingdocumentation to ensure that it is sufficient tosupport the analysis and to allow an independentduplication of the results. The review should alsoconfirm that the analysis is reasonably complete

in terms of the radionuclides and pathwaysaddressed.

In addition, the review should evaluate thedegree to which the assumptions inherent in theanalysis apply to the site and conditions beingaddressed. The mathematical models used tocalculate dose use a large number ofenvironmental transfer factors and dose conversionfactors that may not always be entirely applicableto the conditions being analyzed. For example,the standard dose conversion factors are based oncertain generic assumptions regarding thecharacteristics of the exposed individual and thechemical and physical properties of theradionuclides. Also, as is the case for chemicalcontaminants, the environmental transfer factorsused in the models may not apply to all settings.

Though the risk assessment models mayinclude a large number of radionuclides andpathways, the important radionuclides andpathways are usually few in number. As a result,it is often feasible to check the computer outputusing hand calculations. This type of review canbe performed by health physicists familiar withthe models and their limitations. Guidance onconducting such calculations is provided innumerous references, including Till and Meyer(1983) and NCRP Report No. 76 (NCRP 1984a).

10.7.2 QUANTIFYING RISKS

Given that the results of the exposureassessment are virtually complete, correct, andapplicable to the conditions being considered, thenext step in the process is to calculate andcombine risks. As discussed previously, the riskassessment for radionuclides is somewhatsimplified because only radiation carcinogenesisneeds to be considered.

Section 10.5 presents a methodology forestimating committed effective dose equivalentsthat may be compared with radiation protectionstandards and criteria. Although the product ofthese dose equivalents (Sv) and an appropriaterisk factor (risk per Sv) yields an estimate of riskthe health risk estimate derived in such a manneris not completely applicable for members of thegeneral public. A better estimate of risk may becomputed using age- and sex-specific coefficientsfor individual organs receiving significant radiation

33

doses. This information may be used along withorgan-specific dose conversion factors to deriveslope factors that represent the age-averagedlifetime excess cancer incidence per unit intake forthe radionuclides of concern. The Integrated RiskInformation System (IRIS) and the Health EffectsAssessment Summary Tables (HEAST) containslope factor values for radionuclides of concern atremedial sites for each of the four major exposurepathways (inhalation, ingestion, air immersion, andground-surface irradiation), along with supportingdocumentation for the derivation of these values(see Chapter 7 for more detail on IRIS).

The slope factors for the inhalation pathwayshould be multiplied by the estimated inhaledactivity (derived using the methods presented inSection 6.6.3 and Exhibit 6-16, without divisionof the body weight and averaging time) for eachradionuclide of concern to estimate risks from theinhalation pathway. Similarly, risks from the’ingestion pathway should be estimated bymultiplying the ingestion slope factors by theactivity ingested for each radionuclide of concern(derived using the methods presented in Exhibits6-11, 6-12, 6-14, 6-17, 6-18, and 6-19, withoutdivision by the body weight and averaging time).Estimates of the risk from the air immersionpathway should be computed by multiplying theappropriate slope factors by the airborneradionuclide concentration (Bq/m3) and theduration of exposure. Risk from the groundsurface pathway should be computed as theproduct of the slope factor, the soil concentration(Bq/m2), and the duration of exposure for eachradionuclide of concern.

The sum of the risks from all radionuclidesand pathways yields the lifetime risk from theoverall exposure. As discussed in Chapter 8,professional judgment must be used in combiningthe risks from various pathways, as it may not bephysically possible for one person to be exposedto the maximum radionuclide concentrations forall pathways.

10.7.3 COMBINING RADIONUCLIDE ANDCHEMICAL CANCER RISKS

Estimates of the lifetime risk of cancer toexposed individuals resulting from radiological andchemical risk assessments may be summed inorder to determine the overall potential human

health hazard associated with a site. Certainprecautions should be taken, however, beforesumming these risks. First, the risk assessorshould evaluate whether it is reasonable to assumethat the same individual can receive the maximumradiological and chemical dose. It is possible forthis to occur in some eases because many of theenvironmental transport processes and routes ofexposure are the same for radionuclides andchemicals.

In cases where different environmental fateand transport models have been used to predictchemical and radionuclide exposure, themathematical models may incorporate somewhatdifferent assumptions. These differences can resultin incompatibilities in the two estimates of risk.One important difference of this nature is how thecancer toxicity values (i.e., slope factors) weredeveloped. For both radionuclides and chemicals,cancer toxicity values are obtained by extrapolationfrom experimental and epidemiological data. Forradionuclides, however, human epidemiologiesdata form the basis of the extrapolation, while formany chemical carcinogens, laboratoryexperiments are the primary basis for theextrapolation. Another even more fundamentaldifference between the two is that slope factorsfor chemical carcinogens generally represent anupper bound or 95th percent confidence limitvalue, while radionuclide slope factors are bestestimate values.

In light of these limitations, the two sets ofrisk estimates should be tabulated separately inthe final baseline risk assessment.

10.7.4 ASSESSING AND PRESENTINGUNCERTAINTIES

Uncertainties in the risk assessment must beevaluated and discussed, including uncertainties inthe physical setting definition for the site, in themodels used, in the exposure parameters, and inthe toxicity assessment. Monte Carlo uncertaintyanalyses are frequently performed as part of theuncertainty and sensitivity analysis for radiologicalrisk assessments. A summary of the use ofuncertainty analyses in support of radiological riskassessments is provided in NCRP Report No. 76(NCRP 1984a), Radiological Assessment (Till andMeyer 1983), and in the Background Information

34

Document for the Draft EIS for Reposed NESHAPsfor Radionuclides (EPA 1989a).

10.7.5 SUMMARIZING AND PRESENTINGTHE BASELINE RISKCHARACTERIZATION RESULTS

The results of the baseline riskcharacterization should be summarized andpresented in art effective manner to assist indecision-making. The estimates of risk should besummarized in the context of the specific siteconditions. Information should include theidentity and concentrations of radionuclides, typesand magnitudes of health risks predicted,uncertainties in the exposure estimates and toxicityinformation, and characteristics of the site andpotentially exposed populations. A summary tableshould be provided in a format similar to thatshown in Exhibit 6-22, as well as graphicalpresentations of the predicted health risks (seeExhibit 8-7).

10.8 DOCUMENTATION,REVIEW, ANDMANAGEMENT TOOLSFOR THE RISK ASSESSOR,REVIEWER, ANDMANAGER

The discussion provided in Chapter 9 alsoapplies to radioactively contaminated sites. Thesuggested outline provided in Exhibit 9-1 may alsobe used for radioactively contaminated sites withonly minor modifications. For example, theportions that uniquely pertain to the CLPprogram and noncarcinogenic risks are not needed.In addition, because radionuclide hazard andtoxicity have been addressed adequately on ageneric basis, there is no need for an extensivediscussion of toxicity in the report.

35


American Public Health Association. 1987. QA Procedures for Health Labs Radiochemistry.

Beebe, G.W., Kate, H., and Land, GE. 1977. Mortalitv Experience of Atomic Bomb Survivors, 1950-1974, Life Span Studv Repott8. RERF TR 1-77. Radiation Effects Research Foundation. Japan.

Brent, R.L 1980. Radiation Teratogenesis. Teratology 21:281-298.

Cember, H. 1983. Introduction to Health Physics (2nd Ed.) Pergamon Press. New York NY.

Department of Energy (DOE). 1987. The Environmental Survey Manual. DOE/EH-0053.

Department of Energy (DOE). 1988. External Dose-Rate Conversion Factors for Calculation of Dose to the Public. DOE/EH-0070.

Depatment of Energy (DOE). 1989. Environmental Monitoring for Low-level Waste Disposal Sites. DOE/LLW-13Tg.

Eisenbud, M. 1987. Environmental Radioactivity (3rd Ed.) Academic Press. Orlando, FL

Environmental Protection Agency (EPA). 1972. Environmental Monitoring Surveillance Guide

Environmental Protection Agency (EPA). 1977a. Handbook of Analytical Quality Control in Radioanalytical Laboratories. Office ofResearch and Development. EPA/600/7-77/008.

Environmental Protection Agency (EPA). 1977b. Qualitv Control for Environmental Measurements Using Gamma-Ray Spectrometry.EPA/500/7-77/14.

Environmental Protection Agency (EPA). 1979a. Radiochemical Analytical Procedures for Analysis of Environmental Samples.EMSL-LV-0539-17.

Environmental Protection Agency (EPA). 1980. Upgrading Environmental Radiation Data. Office of Radiation Programs. EPA/520/l-80/012

Environmental Protection Agency (EPA). 1984a. Eastern Environmental Radiation Facilitv Radiochemistrv Procedures Manual. Officeof Radiation Programs. EPA/520/5-84/006.

Environmental Protection Agency (EPA). 1984b. Federal Guidance Report No. l0: The Radioactivity Concentration Guides. Officeof Radiation Programs. EPA/520/l-84/010.

Environmental Protection Agency (EPA). 1987a. Population Exposure to External National Radiation Background in the UnitedStates. Office of Radiation Programs. EPA ORP/SEPD-80-12.

Environmental Protection Agency (EPA). 1987b. Handbook of Analytical Quality Control in Radioanalytical Laboratories. Office ofResearch and Development. EPA/600/7-87/008,

Environmental Protection Agency (EPA). 1988. Federal Guidance Report No. 11: Limiting Values of Radionuclide Intake and AirConcentration and Dose Conversion Factors for Inhalation. Submersion. and Ingestion. Office of Radiation Programs. EPA/520/l-88/020.

Environmental Proctection Agency (EPA). 1989a. Background Information Document, Draft EIS for Proposed NESHAPS forRadionuclides, Volume 1, Risk Assessment Methodology. Offic of Radiation Programs. EPA/520/l-89/005.

Environmental Protection Agency (EPA). 1989b. Annual Report Fiscal Year 1988 Laboratory Intercomparison Studies forRadionuclides.

Environmental Protection Agency (EPA). 1989c. CERCLA Conpliance with Other Laws Manual. Part II. Office of Emergencyand Remedial Response. (OSWER Directive 9234.1-02).

International commission on Radiological Protection (ICRP). 1977. Recommendations of the ICRP. ICRP Publication 26.

International Commision on Radiological Protection (ICRP). 1979. Limits for Intake of Radionuctides by Workers. ICRP Publication30.

36

International Commission on Radiological Protection (ICRP). 1983. Principles for Limiting Exposure of the Public to Natural Sourcesof Radiation. ICRP Publication 39.

International Commission on Radiological Protection (ICRP). 1984. A Complilation of the Major Concepts and Quantities in Useby the ICRP. ICRP Publication 42.

International Commission on Radiological Protection (ICRP). 1985. Principles of Monitoring for the Radiation Protection of thePopulation. ICRP Publication 43.

International Commission on Radiological Protection (ICRP). 1987. Lung Cancer Risk from Indoor Exposures to Radon Daughters.ICRP Publication 50.

Kato, H. and Schull, W.J. 1982. Studies of the Mortality of A-Bomb Survivors. Report 7 Part 1, Cancer Mortality Among AtomicBomb Survivors, 1950-78. Radiation Research 90:395-432.

Kocher, D. 1981. Radioactive Decay Data Tables A Handbook of Decay Data for Application to Radiation Dosimetry and Radiological Assessments. DOE/TIC-11O26.

Miller. 1984. Models and Parameters for Environmental Radiological Assessments. DOE/TIC-11468.

National Academy of Sciences - National Research Council. 1972. The Effects on Populations of Exposuures to Low Levels of IonizingRadiation. (BEIR Report).

National Academy of Sciences - National Research Council. 1980. The Effects on Population of Exposures to Low Levels of IonizingRadiation. (BEIR Report).

National Academy of Sciences - National Research Council. 1988. Health Risks of Radon and Other Internally Deposited Alpha-Emitters. (BEIR Report).

National Council on Radiation Protection and Measurements (NCRP). 1989. screening Techniques for Determining Compliancewith Environmental Standards. NCRP Commentary No. 3.

National Council on Radiation Protection and Measurements (NCRP). 1976. Environmental Radiation Measurements. NCRP ReportNo. 50.

National Council on Radiation Protection and Measurements (NCRP). 1978. Instrumentation and Monitoring Methods for RadiationProtection. NCRP Report No. 57.

National Council on Radiation Protection and Measurements (NCRP). 1979. Tritium in the Environment. NCRP Report No. 62.

National Council on Radiation Protection and Measurements (NCRP). 1980. Influence of Dose and Its Distribution in Time onDose-rcsponse Relationships for Low-LET Radiations. NCRP Report No. 64.

National Council on Radiation Protection and Measurements (NCRP). 1984a. Radiological Assessment: Predicting theTransport, Bioaccumulation, and Uptake by Man of Radionuclides Released to the Environment. NCRP Report No. 76.

National Council on Radiation Protection and Measurements (NCRP). 1984b. Exposure from the Uranium Series with Emphasison Radon and its Daughters. NCRP Report No. 77.

National Council on Radiation Protection and Measurements (NCRP). 1985a. A Handbook of Radioactivity Measurement Procedures.NCRP Report No. 58.

National Council on Radiation Protection and Measurements (NCRP). 1985b. Induction of Thyroid Cancer by Ionizing Radiation.NCRP Report No. 80.

National Council on Radiation Protection and Measurements (NCRP). 1985c. Carbon-14 in the Environment. NCRP Report No.81.

National Council on Radiation Protection and Measurements (NCRP). 1987a. Ionizing Radiation Exposure of the Population of theUnited States. NCRP Report No. 93.

National Council on Radiation Protection and Measurements (NCRP). 1987b. Exposure of the Population of the United States andCanada from Natural Background Radiation. NCRP Report No. 94.

37

National Council on Radiation Protection and Measurements (NCRP). 1989. Screening Technicques for Determining Compliancewith Environmental Standards. NCRP Commentary No. 3.

Nuclear Regulatory Commission (NRC). 1977. Calculation of Annual Doses to Man from Routine Releases of Reactor Effluents forthe Purpose of Evaluating Compliance with 10 CFR 50. Appemdix I. Regulatory Guide 1.109.

Nuclear Regulatory Commission (NRC). 1979. Qualitv Assurance Monitoring Programs (Normal Operation) -- Effluent Streams andthe Environment. NRC Regulatory Guide 4.15.

Nuclear Regulatory Commission (NRC). 1989. Health Effects Modelsor Nuclear Power Plant Accident Consequence Analysis: Low-LET Radiation, Part II: Scientific Bases for Health Effects Models. NUREG/CR-4214, Rev. 1. Part II.

Otake, M. and Schull W. 1984. Mental Retardation in Children Exposed in Utero to the Atomic Bombs: A Reassessment. TechnicalReport RERF TR 1-83. Radiation Effects Research Foundation. Japan.

Schleien, B. and Terpilak, M., (Eds). 1984. The Health Physics and Radiolog ical Health Handbook. (7th Ed.) Nucleon LecternAssoc., Inc. Maryland.

Till, J.E. and Meyer, H.R. 1983. Radiological Assessment A Textbook on Environmental Dose Analysis. Prepared for Office ofNuclear Reactor Regulation. U.S. Nuclear Regulatory Commision. Washington, DC NUREG/CR-3332

Turner, J.E. 1986. Atoms, Radiation. and Radiation Protection. Pergamon Press. New York, NY.

United Nations Scientific Committee Report on the Effects of Atomic Radiation (UNSCEAR). 1958. Official Records: ThirteenthSession, Supplement No. 17(A/3838). United Nations. New York NY.

United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR). 1977. Sources and Effects of IonizingRadiation. United Nations. New York NY.

United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR). 1982 Ionizing Radiation Sources andEffects. United Nations. New York NY.

United Nations scientific Committee on the Effects of Atomic Radiation (UNSCEAR). 1986. Genetic and Somatic Effects of IonizingRadiation. United Nations. New York, NY.

United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR). 1988. Sources, Effects, and Risks of IonizingRadiation. United Nations. New York, NY.

Wakabayashi, T., et al. 1983. Studies of the Mortality of A-Bomb Survivors, Report 7, Part III, Incidence of Cancer in 1959-1978,Based on the Tumor Registry, Nagasaki. Radiat. Res. 93: 112-146.

APPENDIX A

ADJUSTMENTS FORABSORPTION

This appendix contains example calculationsfor absorption efficiency adjustments that might beneeded for Superfund site risk assessments.Absorption adjustments might be necessary in therisk characterization step to ensure that the siteexposure estimate and the toxicity value forcomparison are both expressed as absorbed dosesor both expressed as intakes.

Information concerning absorption effi-ciencies might be found in the sections describingabsorption toxicokinetics in HEAs, HEEDs,HEEPs, HADs, EPA drinking water qualitycriteria or ambient water quality criteriadocuments, or in ATSDR toxicological profiles.If there is no information on absorption efficiencyby the oral/inhalation routes, one can attempt tofind absorption efficiencies for chemically relatedsubstances. If no information is available,conservative default assumptions might be used.Contact ECAO for further guidance.

Adjustments may be necessary to match theexposure estimate with the toxicity value if one isbased on an absorbed dose and the other is basedon an intake (i.e., administered dose).Adjustments may also be necessary for differentvehicles of exposure (e,g., water, food, or soil).

For the dermal route of exposure, theprocedures outlined in Chapter 6 result in anestimate of the absorbed dose. Toxicity valuesthat are expressed as administered doses will needto be adjusted to absorbed doses for comparison.This adjustment is discussed in Section A.l.

For the other routes of exposure (i.e., oraland inhalation), the procedures outlined inChapter 6 result in an estimate of daily intake.If the toxicity value for comparison is expressed

EFFICIENCY

as an administered dose, no adjustment may benecessary (except, perhaps, for vehicle ofexposure). If the toxicity value is expressed as anabsorbed dose, however, adjustment of theexposure estimate (i.e., intake) to an absorbeddose is needed for comparison with the toxicityvalue. This adjustment is discussed in SectionA.2.

Adjustments also may be necessary fordifferent absorption efficiencies depending on themedium of exposure (e.g., contaminants ingestedwith food or soil might be less completelyabsorbed than contaminants ingested with water).This adjustment is discussed in Section A.3.

A.1 ADJUSTMENTS OF TOXICITYVALUE FROM ADMINISTEREDTO ABSORBED DOSE

Because there are few, if any, toxicityreference values for dermal exposure, oral valuesare frequently used to assess risks from dermal

Page A-2

exposure. Most RfDs and some slope factors areexpressed as the amount of substance administeredper unit time and unit body weight, whereasexposure estimates for the dermal route ofexposure are eventually expressed as absorbeddoses. Thus, for dermal exposure to contaminantsin water or in soil, it may be necessary to adjustan oral toxicity value from an administered to anabsorbed dose. In the boxes to the right and onthe next page are samples of adjustments for anoral RfD and an oral slope factor, respectively.If the oral toxicitv value is alreadv empressed as anabsorbed dose (e.g. trichloroethvlene), it is notnecessarv to adjust the toxicitv value.

In the absence of any information onabsorption for the substance or chemically relatedsubstances, one must assume an oral absorptionefficiency. Assuming 100 percent absorption in anoral administration study that serves as the basisfor an RfD or slope factor would be a non-conservative approach for estimating the dermal RfD or slope factor (i.e., depending on the typeof chemical, the true absorbed dose might havebeen much lower than 100 percent, and hence anabsorbed-dose RfD should similarly be much loweror the slope factor should be much higher). Forexample, some metals tend to be poorly absorbed

Page A-3

(less than 5 percent) by the gastrointestinal tract.A relatively conservative assumption for oralabsorption in the absence of appropriateinformation would be 5 percent.

A.2 ADJUSTMENT OF EXPOSUREESTIMATE TO AN ABSORBEDDOSE

If the toxicity value is expressed as anabsorbed rather than an administered dose, it maybe necessary to convert the exposure estimatefrom an intake into an absorbed dose forcomparison. An example of estimating anabsorbed dose from an intake using an absorptionefficiency factor is provided in the box in the topright corner. Do not adjust exposure estimatesfor absorption efficiency if the toxicity values arebased on administered doses.

A.3 ADJUSTMENT FOR MEDIUMOF EXPOSURE

If the medium of exposure in the siteexposure assessment differs from the medium of

exposure assumed by the toxicity value (e.g., RfDvalues usually are based on or have been adjustedto reflect exposure via drinking water, while thesite medium of concern may be soil), anabsorption adjustment may, on occasion, beappropriate. For example, a substance might bemore completely absorbed following exposure tocontaminated drinking water than followingexposure to contaminated food or soil (e.g., if thesubstance does not desorb from soil in thegastrointestinal tract). Similarly, a substancemight be more completely absorbed followinginhalation of vapors than following inhalation ofparticulate. The selection of adjustment methodwill depend upon the absorption efficiencyinherent in the RfD or slope factor used forcomparison. To adjust a food or soil ingestionexposure estimate to match an RfD or slopefactor based on the assumption of drinking wateringestion, an estimate of the relative absorptionof the substance from food or soil and from wateris needed. A sample maculation is provided inthe box on the next page.

In the absence of a strong argument formaking this adjustment or reliable informationon relative absorption efficiencies, assume that the

Page A-4

relative absorption efficiency between food or soiland water is 1.0.

If the RfD or slope factor is expressed as anabsorbed dose rather than an administered dose,it is only necessary to identify an absorptionefficiency associated with the medium of concernin the site exposure estimate. In the exampleabove, this situation would translate into a relativeabsorption of 0.3 (i.e., 30/100).

APPENDIX B

INDEX

AAbsorbed dose

calculation 6-34, 6-39, 7-8, 7-10, 7-12definition 6-2, 6-4, 6-32, 6-34, 7-10, 10-2following dermal contact with soil,

sediment or dust 6-39, 6-41 to 6-43, 7-16

following dermal contact with water 6-34,6-39, 7-16

radiation 10-1, 10-2, 10-6toxicity value 7-10, 7-16, 8-5, A-1, A-2

Absorption adjustment dermal exposures 8-5, A-1, A-2medium of exposure 8-5, A-3, A-4

Absorption efficiencydefault assumptions 6-34, 6-39, A-2 to A-4dermal 6-34, 6-39general 62,7-10,7-20, 8-5, 8-10

Acceptable daily intakes 7-1, 7-2, 7-6

Activity at time t 10-1

Activity patterns 6-2, 6-6, 6-7, 6-24, 7-3

Acute exposures See Exposure -- short-term

Acute toxicants 6-23, 6-28

ADIs. See Acceptable daily intakes

Administered dose 6-2, 6-4, 7-1, 7-2, 7-10, 8-2,8-5, A-1 to A-4

Agency for Toxic Substances and Disease Registry 1-8,2-1,2-3, 2-4, 2-8 to 2-11,61, 6-17, 7-14, 8-1, 8-15, 8-24

Air data collectionand soil 4-10background sampling 4-9

concentration variability 4-9emission sources 4-15flow 4-8meteorological conditions 4-15, 4-20monitoring 4-8, 4-9, 4-14radionuclides 10-11sample type 4-19sampling locations 4-19short-term 4-15spatial considerations 4-15temporal considerations 4-15, 4-20time and cost 4-21

Air exposuredispersion models 6-29indoor modeling 6-29outdoor modeling 6-29volatilization 6-29

Analytes 4-2, 5-2, 5-5, 5-7, 5-10, 5-27

Analytical methodsevaluation 5-5 to 5-7radionuclides 10-12, 10-13routine analytical services 4-22special analytical services 4-3, 4-22

Animal studies 7-1210-28, 10-29, 10-33

Applicable or relevant and appropriaterequirement 2-2, 2-7, 2-8, 8-1, 10-8 to 10-10

Applied dose 6-2, 6-4

ARAR. See Applicable or relevant andappropriate requirement

A(t). See Activity at time t

ATSDR, See Agency for Toxic Substances andDisease Registry

Averaging time 6-23

Page B-2

BBackground

anthropogenic 4-2, 4-5comparison to site related contamination 4-

9, 4-10, 4-18defining needs 4-5 to 4-10,6-29, 6-30information useful for data collection 4-1localized 4-5naturally occurring 4-2, 4-5, 8-25, 10-14sampling 4-5 to 4-10, 10-14ubiquitous 4-5

BCF. See Bioconcentration factor

Bench scale tests 4-3

Benthic oxygen conditions 4-7

Bioconcentration 4-11, 6-31, 6-32

Bioconcentration factor 6-1, 6-12, 6-31,6-32

Biota sampling 4-7, 4-10, 4-16

Blanksevaluation 5-17field 4-22, 4-23, 5-17, 10-20laboratory 4-22, 5-13, 5-17laboratory calibration 5-17laboratory reagent or method 5-17trip 4-22, 5-17

Body weight as an intake variable 6-22,6-23, 6-39, 7-8, 7-12, 10-26, 10-33

Bulk density 4-7, 4-12

CCancer risks

extrapolating to lower doses 7-11, 7-12linear low-dose equation 8-6multiple pathways 8-16multiple substances 8-12one-hit equation 8-11radiation 10-28 to 10-32summation of 8-12, 8-16

Carcinogenesis 7-10, 10-28 to 10-32

Carcinogen Risk Assessment VerificationEndeavor 7-1, 7-13

Carcinogens 5-8, 5-21, 6-23, 7-10, 8-6, 10-30, 10-33

CDI. See Chronic daily intake

CEAM. See Center for Exposure AssessmentModeling

Center for Exposure Assessment Modeling 6-1,6-25, 6-31

CERCLA See Comprehensive EnvironmentalResponse, Compensation, and Liability Act of1980

CERCLA Information System 2-4

CERCLIS. See CERCLA Information System

Checklist for manager involvement 9-14 to 9-17

Chemicals of potential concerndefinition 5-2listing 5-20preliminary assessment 5-8radionuclides 10-21reducing 5-20 to 5-24summary 5-24 to 5-27

Chronic daily intake 6-1, 6-2, 6-23, 7-1, 8-1, 8-6to 8-11

CLP. See Contract Laboratory Program

Combustible gas indicator 5-6

Common laboratory contaminants 5-2, 5-3, 5-13,5-16, 5-17

Comprehensive Environmental Response,Compensation, and Liability Act of 19801-1,1-3, 2-1 to 2-4

Concentration-toxicity screen 5-20, 5-23

Conceptual model 4-5, 4-10

Contact rate 6-2, 6-22

Contract Laboratory Programapplicability to radionuclides 10-16, 10-17,

10-20, 10-21

Page B-3

definition 4-2routine analytical services 4-22, 5-5, 5-7, 5-

15, 5-18,5-20special analytical services 4-3, 4-22, 5-5, 5-7

to 5-10, 5-18 to 5-20statements of work 5-5

Contract-required detection limit. SeeDetection limit

Contract-required quantitation limit. SeeQuantitation limit

CRAVE. See Carcinogen Risk AssessmentVerification Endeavor

CRDL. See Contract-required detection limit

Critical study. See Reference dose

Critical toxicity effect. See Reference dose

CRQL. See Contract-required quantitationlimit

Curie 10-2, 10-4, 10-6

DD. See Absorbed dose -- radiation

Datacodes 5-11 to 5-16positive 5-2qualifiers 5-11 to 5-16

Data quality objectives 3-4, 4-1 to 4-5, 4-19, 4-24, 10-14

DCF. See Dose conversion factor

Decay products 10-2, 10-7, 10-21, 10-24

Decision Summary 9-3

Declaration 9-3

Dermalabsorption efficiency 6-34, 6-39contact with soil, sediment, or dust 6-39, 6-

41 to 6-43, A-2contact with water 6-34, 6-37 to 6-39, A-2

exposure 4-10, 4-11, 4-14, 6-34, 6-37 to 6-39, 6-43, 8-5, A-2

external radiation exposure 10-22, 10-23,10-25, 10-26

toxicity values 7-16

Detection frequency 5-20, 5-22

Detection limitscontract-required 5-1, 5-2, 5-8definition 5-1, 5-2, 5-8evaluation 4-3 to 4-5, 5-7 to 5-11, 5-20, 6-

31instrument 4-1, 5-1, 5-7limitations to 4-15, 4-22, 5-8method 4-22, 5-1, 5-7radionuclides 10-17 to 10-20

Diffusivity 6-12

Dissolved oxygen 4-7

DL. See Detection limit

Documentation. See Preparingthe baseline risk assessment

Dose

and reviewing

absorbed vs administered 6-4, 7-10, 8-2, A-1 to A-3

absorption efficiency A-1 to A-3response curve 7-12response evaluation 7-1, 7-2, 7-11, 7-12

Dose conversion factor 10-1, 10-2, 10-24, 10-25,10-26

Dose equivalentcommitted 10-1, 10-2, 10-7, 10-24, 10-25,

10-26effective 10-1, 10-2, 10-7, 10-24, 10-25, 10-

26

DQO. See Data quality objectives

Dry weight 4-7

Dustexposure 6-39, 6-43fugitive dust generation 4-3, 4-5,4-15, 6-29transport indoors 6-29

EE. See Exposure level

ECAO. See Environmental Criteria andAssessment Office

Emission samplingrate 4-5, 4-7, 4-14strength 4-7

Endangernent Assessment Handbook 1-1, 2-9

Endangerment assessments 2-1, 2-8

Environmental Criteria and Assessment Office7-1, 7-15, 7-16, 7-19, 8-1, 8-5, A-1

Enironmental Evaluation Manual 1-1, 1-11,2-9,4-16

Environmental Photographic InterpretationCenter 4-4

EPIC. See Environmental PhotographicInterpretation Center

Epidemiologysite-specific studies 2-10, 8-22, 8-24toxicity assessment 7-3, 7-5

Essential nutrients 5-23

Estuary sampling 4-7, 4-13,4-14

Exposureaveraging time 6-23characterization of setting 6-2, 6-5, to 6-8definition 6-2, 8-2event 6-2expressed as absorbed doses 6-34, 6-39, A-1for dermal route 6-34, 6-39, 6-41 to 6-43frequency/duration 6-22general considerations 6-19 to 6-24level 8-1long-term 6-23parameter estimation 6-19 to 6-23pathway-specific exposures 6-32 to 6-47point 6-2, 6-11potentially exposed populations 6-6 to 6-8radionuclides vs chemicals 10-22route 6-2, 6-11, 6-17, 6-18, 8-2, A-1short-term 6-23, 8-11, 10-25, 10-28, 10-30

Exposure assessmentdefinition 1-6, 1-7, 6-1, 6-2, 8-2intake calculations 6-32 to 6-47objective 6-1output for dermal contact with

contaminated soil 6-39output for dermal exposure to

contaminated water 6-34preliminary 4-3,4-10 to 4-16radiation 10-22 to 10-27spatial considerations 6-24 to 6-26

Exposure concentrationsand the reasonable maximum exposure 6-19in air 6-28, 6-29in food 6-31, 6-32in ground water 6-26, 6-27in sediment 6-30in soil 6-27, 6-28in surface water 6-29, 6-30summarizing 6-32, 6-33, 6-50, 6-52

Exposure pathwayscomponents 6-8, 6-9definition 6-2, 8-2external radiation exposure 10-22, 10-23,

10-25, 10-26identification 6-8 to 6-19multiple 6-47summarizing 6-17, 6-20

FFate and transport assessment 6-11, 6-14 to 6-

16. See also Exposure assessment

Field blanks. See Blanks

Field investigation team 4-1, 4-16,4-20, 4-24,5-1, 5-2

Field sampling plan 4-1, 4-24-23, 4-24, 10-15

Field screen 4-11, 4-20,4-21, 5-5,5-6, 5-24

First-order analysis 8-20

FIT. See Field investigation team

Five-year review 2-3, 2-5

Food chain 2-3,4-7,4-10, 4-16,6-31, 6-32

Fraction organic content of soil 4-7

Page B-5

Frequency of detection. See Detectionfrequency

FS. See Remedial investigation/feasibility study

FSP. See Field sampling plan

GGround-water data collection

and air 4-13and soil 4-12filtered vs unfiltered samples 4-12, 6-27hydrogeologic properties 4-12sample type 4-19transport route 4-11well location and depth 4-12

Grouping chemicals by class 5-21, 10-21

HHADs. See Health Assessment Documents

HAs. See Health Advisories

Half-life 6-12, 10-2

Hazard identification 1-6, 7-1, 7-2, 10-28 to 10-30

Hazard indexchronic 8-13definition 8-1, 8-2multiple pathways 8-16, 8-17multiple substances 8-12, 8-13noncancer 8-12, 8-13segregation 8-14, 8-15short-term 8-13, 8-14subchronic 8-13, 8-14

Hazard quotient 8-2, 8-11

Hazard Ranking System 2-5, 2-6, 4-1, 4-4

HE See Dose equivalent

HE,50- See Dose equivalent

Head measurements 4-7

Health Advisories 2-10, 7-9, 7-10, 8-13

Health and Environmental Effects Documents7-1, 7-14, A-1

Health and Environmental Effects Profiles 7-1,7-14, A-1

Health Assessment Documents 7-1, 7-14, A-1

Health Effects Assessments 7-1, 7-14, A-1

Health Effects Assessment Summary Tables 7-1,7-14

Health physicist 10-3, 10-21

HEAs. See Health Effects Assessments

HEAST. See Health Effects AssessmentSummary Tables

HEEDs. See Health and Environmental EffectsDocuments

HEEPs. See Health and Environmental EffectsProfiles

Henry’s law constant 6-12

HI. See Hazard index

HNu organic vapor detector 5-6

Hot Spots 4-10 to 4-12, 4-17, 4-19, 5-27, 6-24, 6-28

HQ. See Hazard quotient

HRS. See Hazard Ranking System

HT. See Dose equivalent

Ht,50 See Dose equivalent

Hydraulic gradient 4-7

IIARC. See International Agency for Research

on Cancer

IDL. See Instrument detection limit

Page B-6

Ingestionof dairy products 4-16, 6-47, 6-48of fish and shellfish 4-3, 4-11, 4-14, 4-15, 4-

16,6-43,6-45 of ground water 6-34, 6-35of meat 4-15, 4-16, 6-47, 6-48of produce 4-16, 6-43, 6-46, 6-47of soil, sediment, or dust 6-39, 6-40of surface water 4-14, 6-34, 6-35while swimming 4-14, 6-34, 6-36

Instrument detection limit. See Detection limit

Inhalation 6-43, 6-44

Intake 6-2, 6-4,6-19, 6-21, 8-2, 10-26

Integrated Risk Information System 7-1, 7-27-6, 7-12 to 7-15, 8-1, 8-2, 8-7, 8-8, 10-33

International Agency for Research on Cancer 7-11

International System of Units 10-1

Ionizing radiation. See Radionuclides, radiation

IRIS. See Integrated Risk Information System

K

Kd 6-12

Koc 6-12

Kow 6-12, 6-31

Kriging 6-19

LLand use

and risk characterization 8-10, 8-20, 8-26current 6-6future 6-7

Lentic waters 4-14

LET. See Linear energy transfer

Level of effort 1-6 to 1-8, 3-3

Life history stage 4-7

Lifetime average daily intake 6-2, 6-23, 8-4

Linear energy transfer 10-1, 10-2, 10-28, 10-29,10-31

Linearized multistage model 7-12, 8-6

Lipid content 4-7, 10-14

LLD. See Lower limit of detection

LOAEL See Lowest-observed-adverse-effect-level

Lotic waters 4-13, 4-14

Lower limit of detection 10-1

Lowest-observed-adverse-effect-level 7-1, 7-2, 7-7, 8-1

MManagement tools 9-1, 9-14, 10-1, 10-34

Maximum contaminant levels 1-8, 5-8

MCLS. See Maximum contaminant levels

MDL. See Method detection limit

Media of concernair 4-14biota 4-15ground water 4-12sampling 4-2, 4-3, 4-10 to 4-16soil 4-11surface water/sediments 4-13

Metalsabsorption by gastrointestinal tract A-2, A-

3default assumptions for A-2

Method detection limit. See Detection limit

MeV. See Million electron volts

MF. See Modifying factor

Million electron volts 10-1, 10-5

Page B-7

Modeling 4-3 to 4-8, 5-8, 5-22, 5-27,6-25, 6-26,8-18 to 8-20

Modifying factor 7-7, 7-21, 8-4, 8-8, 10-1, 10-2,10-6

Monte Carlo simulation 8-19, 8-20

Multistage model. See Linearized multistagemodel

NN. See Dose equivalent

National Oceanographic and AtmosphericAdministration 6-1, 6-6

National Oil and Hazardous SubstancesPollution Contingency Plan 1-1, 2-2, 2-4, 2-5

National Priorities List 2-3, 2-5, 2-6, 10-1

National Response Center 2-4

National Technical Guidance Studies 6-1

NCP. See National Oil and HazardousSubstances Pollution Contingent Plan

ND. See Non-detect

NOAA. See National Oceanographic andAtmospheric Administration

NOAEL. See No-observed-adverse-effect-level

Noncancer hazard indices. See Hazard index

Noncancer hazard quotient. See Hazardquotient

NonCarcinogenic threshold toxicants 7-6

Non-detects 5-1, 5-2, 5-7, 5-10, 5-11,5-15, 5-16

No-observed-adverse-effect-level 7-1, 7-2, 7-7, 8-1

Normalized exposure rate 6-4, 8-2, A-2

NPL. See National Priorities List

NRC. See Nuclear Regulatory Commission

NTGS. See National Technical GuidanceStudies

Nuclear Regulatory Commission 8-1, 10-8

Nuclear transformation 10-2

0OAQPS. See Office of Air Quality Planning

and Standards

OERR. See Office of Emergency and RemedialResponse

Office of Air Quality Planning and Standards 6-1

Office of Emergency and Remedial Response 1-1

Office of Radiation Programs 10-3, 10-10, 10-14,10-24 to 10-26

Operable units 1-8, 1-9, 3-1, 3-2, 5-24

Oral absorption A-2, A-3

Oral cancer potency factor adjustment A-3

Oral reference dose adjustment A-2

Organic carbon content 4-7, 4-12, 5-5

Organic vapor analyzer 5-6

OVA. See Oxygen vapor analyzer

Oxygen-deficient atmosphere 5-6

PPA. See Preliminary assessment/site inspection

Partition coefficient 4-7, 6-31, 6-32

PA/SI. See Preliminary assessment/siteinspection

PC See Permeability constant

PE. See Performance evaluation

Page B-8

Performance evaluation 5-1, 5-5

Permeability constant 6-34, 10-26

Persistence 4-2, 5-21, 6-4, 6-23, 6-24

pH 4-7

PHE. See Public health evaluation

Porosity 4-7, 4-12

PQL. See Practical quantitation limit

Practical quantitation limit 5-1

Quality assurance/quality control 3-4, 4-1, 4-3, 5-1, 5-29

Quality factor 10-2, 10-6

Quantitation limitcompared to health-based concentrations 5-

2, 5-5, 5-7, 5-8, 5-11contract-required 5-1, 5-2, 5-8definitions 5-2, 5-5, 5-8evaluation 5-1 to 5-9, 10-20high 5-10radionuclides 10-17 to 10-20sample 5-8strategy 4-21unavailability 4-3, 5-10

Preliminary assessment/site inspection 2-4, 2-5, 2-6, 4-2, 4-4, 6-5

Preliminary remediation goals 1-3 to 1-5, 1-8, 8-1

Preparing and reviewing the baseline riskassessment

addressing the objectives 9-1, 9-2communicating the results 9-1, 9-2documentation tools 9-1 to 9-8other key reports 9-3review tools 9-3, 9-9 to 9-14scope 9-2, 9-3

PRGs. See Preliminary remediation goals

Primary balancing criteria 1-9

Proxy concentration 5-10

Public health evaluation 1-11

QQ. See Dose equivalent

QAPjP. See Quality assurance project plan

QA/QC. See Quality Assurance/Quality Control

QL. See Quantitation limit

Qualifiers. See Data

Quality assurance project plan 4-1, 4-2,4-23

RRA. See Remedial action

Radiation. See Radionuclides, radiation

Radiation advisory groupsInternational Commission on Radiation

Protection 10-3, 10-9, 10-28National Academy of Sciences 10-28, 10-29National Council on Radiation Protection

and Measurements 10-9, 10-28United Nations Scientific Committee on

the Effects of Atomic Radiation 10-28,10-29, 10-30

Radiation detection instrumentsgas proportional counters 10-12, 10-13Geiger-Mueller (G-M) counters 10-11, 10-

12ionization chambers 10-11 to 10-13scintillation detectors 10-11 to 10-13solid-state detectors 10-12, 10-13

Radiation unitsbecquerel 10-1, 10-2, 10-4, 10-6curie 10-1, 10-2, 10-4, 10-6picocurie 10-1rad 10-2, 10-6rem 10-2roentgen 10-2, 10-6sievert 10-1, 10-2, 10-6working level 10-7working level month 10-7

Page B-9

Radionuclides, radiationalpha particles 10-4, 10-5, 10-28beta particles 10-4, 10-5, 10-28decay products 10-2, 10-7, 10-21, 10-24definition 10-2external 10-2half-life 10-2internal 10-2ionizing 10-2linear energy transfer 10-2, 10-28, 10-29,

10-31lower limit of detection 10-17, 10-20neutrons 10-4photons 10-4, 10-5, 10-28positrons 10-4quality factors 10-2, 10-6, 10-29radioactive decay 10-2, 10-2radon decay products 10-7regulatory agencies 10-8, 10-9relative biological effectiveness 10-1, 10-6,

10-29risk characterization 10-32 to 10-34toxicity assessment 10-27 to 10-32

RAS. See Routine analytical services

RBE. See Relative biological effectiveness

RCRA. See Resource Conservation andRecovery Act

RD. See Remedial design

Reasonable maximum exposureand body weight 6-22, 6-23and contact rate 6-22and exposure concentration 6-19and exposure frequency and duration 6-22and risk characterization 8-1, 8-15, 8-16, 8-

26definition 6-1, 6-4, 6-5estimation of 6-19 to 6-23, 8-15, 8-16

Record of Decision 2-5, 9-3

Redox potential 4-7

Reference dosechronic 7-1, 7-2, 7-5, 8-1, 8-2, 8-8, 8-10, 8-

13, A-1, A-2critical toxic effect 7-7, 8-4, 8-10, 8-15critical study 7-7definition 7-1, 7-2, 8-2, A-2

developmental 7-1, 7-6, 7-9, 8-2inhalation 7-8oral 7-6, 7-7subchronic 7-1, 7-2, 7-6, 7-8, 7-9, 8-2, 8-9,

8-14verified 7-10

Regional Radiation Program Managers 10-3, 10-10

Relative biological effectiveness 10-1, 10-6, 10-29

Release sources 6-10

Remedial action 1-3, 1-8 to 1-10, 2-5, 2-7, 2-9,3-1, 3-2, 6-8, 10-8

Remedial action objectives 1-3, 1-8, 2-7

Remedial design 2-5, 2-6, 2-9

Remedial investigation/feasibility study 1-1 to 1-5, 1-8 to 1-10, 2-5 to 2-7, 3-1 to 3-3, 4-1 to4-5, 4-23, 8-1

Remedial project managerand background sampling 4-8and elimination of data 5-2, 5-17, 5-20, 5-

21and ground-water sampling 4-13and radiation 10-3and reasonable maximum exposure 6-5and scoping meeting 4-3definition 1-2management tools for 9-14 to 9-17

Remedy selection 1-9, 2-5

Resource Conservation and Recovery Act 2-7,10-8

Responsiveness Summary 9-3

Reviewing the risk assessment. See Preparingand reviewing the baseline risk assessment

RfD. See Reference dose

RfDdt See Reference dose

RfDs See Reference dose

Page B-10

RI. See Remedial investigation/feasibility study

RI/FS. See Remedial investigation/feasibilitystudy

Risk assessment reviewer 1-2, 9-1, 9-3, 9-9 to 9-14

Risk assessordefinition 1-2tools for documentation 9-1 to 9-8

Risk characterization 1-6, 1-7, 8-1

Risk information in the RI/FS process 1-3 to 1-10

Risk manager 1-2

RIME See Reasonable maximum exposure

ROD. See Record of Decision

Route-to-route extrapolation 7-16

Routine analytical services. See ContractLaboratory Program

RPM. See Remedial project manager

Ssalinity 4-7, 4-14, 6-5

Saltwater incursion extent 4-7

Sample Management Office 4-1, 4-2,5-1,5-5

Sample quantitation limitQuantitation limit

Samples. See Sampling

Sampling

5-1. See also

annual/seasonal cycle 4-20composite 4-11, 4-14, 4-19cost 4-10, 4-17, 4-18, 4-20, 4-21depth 4-7, 4-11, 4-12, 4-19devices 4-21grab 4-19purposive 4-9,4-10, 4-12,4-18,4-19radionuclides 10-10 to 10-16random 4-9, 4-12, 4-18 to 4-20

routes of contaminant transport 4-10 to 4-16

strategy 4-16systematic 4-18, 4-19

sampling and analysis plan 1-4, 4-1, 4-2, 4-3, 4-22 to 4-24

SAP. See Sampling and analysis plan

SARA. See Superfund Amendments andReauthorization Act of 1986

SAS. See Special analytical services

Scopingmeeting 4-3, 4-18, 4-22, 4-23, 9-15, 10-15of project 1-3 to 1-5, 1-8, 2-7, 3-2, 3-3

SDI. See SubChronic daily intake

SEAM. See Superfund Exposure AssessmentManual

Segregation of hazard indices 8-14, 8-15

Selection of remedy. See Remedy selection

Semi-volatile organic chemical 5-1

SI. See International System of Units,Preliminary assessment/site inspection

Site discovery or notification 2-4

Site inspection. See Preliminary assessment/siteinspection

Skin 5-29, 7-16, 10-4, 10-6, 10-22, 10-29. Seealso Dermal

Slope factor 5-9, 5-21, 7-3, 7-11 to 7-13, 7-16, 8-1, 8-2 to 8-7, 8-10 to 8-12, 10-2, 10-33, A-1to A-4

SMO. See Sample management office

Soil data collection 4-11and ground water 4-12depth of samples 4-12heterogeneity 4-11hot spots 4-11

Page B-11

Solubility 6-12

Sorption 6-27

SOW. See Statements of work

Special analytical services. See ContractLaboratory Program

Specific organ 4-7, 10-7, 10-22

SPHEM. See Superfund Public HealthEvaluation Manual

SQL. See Sample quantitation limit

Stability class 4-7

Statements of work. See Contract LaboratoryProgram

Statisticsand background 4-8 to 4-10, 5-18

certainty 4-8, 4-17, 4-18methods 4-8, 4-18power 4-9, 4-18sampling strategy 4-16 to 4-20variability 4-9, 4-18

Structure-activity studies 7-5

SubChronic daily intake 6-1, 6-2, 6-23, 7-1, 8-1

Superfund. See Comprehensive EnvironmentalResponse, Compensation, and Liability Act of1980

Superfund Amendments and ReauthorizationAct of 19861-11, 2-1 to 2-4

Superfund Exposure Assessment Manual 2-1, 2-8,6-1

Superfund Public Health Evaluation Manual 1-1,2-8

SVOC. See Semi-volatile organic chemical

TT. See Tissue

TAL. See Target analyte list

Target analyte list 4-1, 4-2, 5-5, 5-8,5-17

Target compound list 4-1, 4-2, 4-22, 5-1,5-5, 5-8, 5-17, 5-21, 10-20

TCL. See Target compound list

Tentatively identified compound 4-1, 5-1, 5-13,5-17, 5-18

Thermocline 4-7

TIC See Tentatively identified compound

Tidal cycle 4-7, 4-14

Tissue 10-1

TOC. See Total organic carbon

Toolsdocumentation 9-1 to 9-8management 9-13 to 9-17review 9-3, 9-9 to 9-14

Topography 4-7

Total organic carbon 5-1

Total organic halogens 5-1

TOX. See Total organic halogens

Toxicity assessment 1-6, 1-7, 7-1, 7-4, 10-27 to10-32

Toxicity valuesabsorbed vs administered dose 7-10, A-1definition 7-3generation of 7-16hierarchy of information 7-15oral 7-16, 10-33, A-2radiation 10-22, 10-32reducing number of chemicals 5-21, 5-23

Transfer coefficients 6-32

Transformation 5-20, 6-27, 7-5, 10-210-3, 10-5

Treatability 5-21

Trip blanks. See Blanks

Page B-12

U U.S. Geological survey 6-1, 6-6

UFs. See Uncertainty factorsUSGS. See U.S. Geological Survey

Uncertainty analysisexposure 6-17,6-34,6-47,6-49 to 6-51,8- V

18, 8-22 Vapor pressure 6-12factors 7-7 to 7-10, 8-4, 8-8, 8-9, 8-17,8-18,

8-20,8-22 VOC. See Volatile organic chemicalfirst-order analysis 8-20model applicability and assumptions 6-50, Volatile organic chemical 4-2, 5-1,5-17,6-31

8-18 to 8-22Monte Carlo simulation 8-20multiple substance exposure 8-22parameter value 8-19qualitative 8-20, 8-21quantitative 8-19, 8-20radiation 10-27, 10-33risk 8-17semiquantitative 8-20toxicity 7-19, 7-20, 8-22

Uncertainty factors. See Uncertainty analysis -- factors

Unit risk 7-13

WWater hardness 4-7

Weighting factor 10-1, 10-2, 10-7

Weight-of-evidence classification 5-20, 7-3, 7-9,7-11, 8-2, 8-4,8-7, 8-10

Whole body 4-7, 4-16, 6-31, 10-6, 10-7

Workplan 4-1, 4-4, 4-22 to 4-24, 9-15

WT See Weighting factor

* U . S . G . P . O . : 1 9 9 2 - 3 1 1 - 8 9 3 : 6 2 6 7 3

CHAPTER 1:

CHAPTER 2:

CHAPTER 3:

CHAPTER 4:

CHAPTER 5:

CHAPTER 6:

CHAPTER 7:

CHAPTER 8:

CHAPTER 9:

CHAPTER 10:

APPENDIX A:

APPENDIX B:

INTRODUCTION

STATUTES, REGULATIONS, GUIDANCE, AND STUDIES RELEVANTTO THE HUMAN HEALTH EVALUATION

GETTING STARTED: PLANNING FOR THE HUMAN HEALTHEVALUATION IN THE RI/FS

DATA COLLECTION

DATA EVALUATION

EXPOSURE ASSESSMENT

TOXICITY ASSESSMENT


DOCUMENTATION,REVIEW, ANDMANAGEMENT TOOLS FOR-THE RISKASSESSOR, REVIEWER, AND MANAGER

RADIATION RISK ASSESSMENT GUIDANCE

ADJUSTMENTS FOR ABSORPTION EFFICIENCY

INDEX

PART A BASELINERISK ASSESSMENT