United States
Environmental Protection
Agency
Office of Water
Office of Science and Technology
4304
EPA-822-B-00-004
October 2000
U>EPA Methodology for Deriving Ambient
Water Quality Criteria for the
Protection of Human Health (2000)
-------
EPA 822-B-00-004
October 2000
Methodology for Deriving
Ambient Water Quality Criteria
for the Protection of
Human Health (2000)
Final
Office of Science and Technology
Office of Water
U.S. Environmental Protection Agency
Washington, DC 20460
-------
NOTICE
The policies and procedures set forth in this document are intended solely to describe
EPA methods for developing or revising ambient water quality criteria to protect human health,
pursuant to Section 304(a) of the Clean Water Act, and to serve as guidance to States and
authorized Tribes for developing their own water quality criteria. This guidance does not
substitute for the Clean Water Act or EPA's regulations; nor is it a regulation itself. Thus, it
does not impose legally-binding requirements on EPA, States, Tribes or the regulated
community, and may not apply to a particular situation based upon the circumstances.
This document has been reviewed in accordance with U.S. Environmental Protection
Agency policy and approved for publication. Mention of trade names or commercial products
does not constitute endorsement or recommendation for use.
11
-------
FOREWORD
This document presents EPA's recommended Methodology for developing ambient water
quality criteria as required under Section 304(a) of the Clean Water Act (CWA). The
Methodology is guidance for scientific human health assessments used by EPA to develop,
publish, and from time to time revise, recommended criteria for water quality accurately
reflecting the latest scientific knowledge. The recommended criteria serve States and Tribes'
needs in their development of water quality standards under Section 303(c) of the CWA.
The term "water quality criteria" is used in two sections of the Clean Water Act, Section
304(a)(l) and Section 303(c)(2). The term has a different program impact in each section. In
Section 304, the term represents a scientific assessment of ecological and human health effects
that EPA recommends to States and authorized Tribes for establishing water quality standards
that ultimately provide a basis for controlling discharges or releases of pollutants. Ambient
water quality criteria associated with specific stream uses when adopted as State or Tribal water
quality standards under Section 303 define the maximum levels of a pollutant necessary to
protect designated uses in ambient waters. The water quality criteria adopted in the State or
Tribal water quality standards could have the same numerical limits as the criteria developed
under Section 304. However, in many situations States and authorized Tribes may want to
adjust water quality criteria developed under Section 304 to reflect local environmental
conditions and human exposure patterns before incorporation into water quality standards.
When adopting their water quality criteria, States and authorized Tribes have four options: (1)
adopt EPA's 304(a) recommendations; (2) adopt 304(a) criteria modified to reflect site-specific
conditions; (3) develop criteria based on other scientifically defensible methods; or (4) establish
narrative criteria where numeric criteria cannot be determined.
EPA will use this Methodology to develop new ambient water quality criteria and to
revise existing recommended water quality criteria. It also provides States and authorized Tribes
the necessary guidance to adjust water quality criteria developed under Section 304 to reflect
local conditions or to develop their own water quality criteria using scientifically defensible
methods consistent with this Methodology. EPA encourages States and authorized Tribes to use
this Methodology to develop or revise water quality criteria to appropriately reflect local
conditions. EPA believes that ambient water quality criteria inherently require several risk
management decisions that are, in many cases, better made at the State, Tribal, or regional level.
Additional guidance to assist States and authorized Tribes in the modification of criteria based
on the Methodology will accompany this document in the form of three companion Technical
Support Documents on Risk Assessment, Exposure Assessment, and Bioaccumulation
Assessment.
Geoffrey H. Grubbs
Director
Office of Science and Technology
in
-------
[This page left blank intentionally.]
IV
-------
ACKNOWLEDGMENTS
Project Leader
Denis Borum U.S. EPA Office of Science and Technology
Coauthors
Risk Assessment
Joyce M. Donohue, Ph.D.* U.S. EPA Office of Science and Technology
Julie T. Du, Ph.D.* U.S. EPA Office of Science and Technology
Charles O. Abernathy, Ph.D. U.S. EPA Office of Science and Technology
Exposure
Denis Borum * U.S. EPA Office of Science and Technology
Helen Jacobs, M.S. U.S. EPA Office of Science and Technology
Henry Kahn, D.Sc. U.S. EPA Office of Science and Technology
Bioaccumulation
Keith G. Sappington, M.S.* U.S. EPA Office of Science and Technology
Lawrence P. Burkhard, Ph.D. U.S. EPA Office of Research and Development
Philip M. Cook, Ph.D. U.S. EPA Office of Research and Development
Erik L. Winchester, M.S. U.S. EPA Office of Science and Technology
U.S. EPA Technical Reviewers
William Beckwith U. S. EPA Region 1
Jeff Bigler U.S. EPA Office of Science and Technology
Sally Brough U.S. EPA Region 10
Karen Clark U.S. EPA Office of General Counsel
Gregory Currey U.S. EPA Office of Wastewater Management
Vicki Dellarco U.S. EPA Office of Prevention, Pesticides, and Toxic Substances
Charles Delos U.S. EPA Office of Science and Technology
Arnold Den U.S. EPA Region 9
Catherine Eiden U.S. EPA Office of Prevention, Pesticides, and Toxic Substances
Michael Firestone U.S. EPA Office of Prevention, Pesticides, and Toxic Substances
Steven Galson U.S. EPA Office of Prevention, Pesticides, and Toxic Substances
Sue Gilbertson U.S. EPA Office of Science and Technology
Denise Hakowski U.S. EPA Region 3
Joel Hansel U.S. EPA Region 4
Wayne Jackson U.S. EPA Region 2
Annie Jarabek U.S. EPA Office of Research and Development
William Jordan U.S. EPA Office of Prevention, Pesticides, and Toxic Substances
-------
Margaret Kelly
Henry Lee
Sharon Lin
Roseanne Lorenzana
Gregory McCabe
Jennifer Mclain
Bruce Mintz
Dave Moon
William Morrow
Jacqueline Moya
Deirdre Murphy
Joseph Nabholz
Russell Nelson
Jennifer Orme-Zavaleta
Lynn Papa
Robert Pepin
David Pfeifer
Rita Schoeny
Charles Stephan
Linda Teuschler
David Tomey
Fritz Wagener
Jennifer Wigal
Jeanette Wiltse
Gary Wolinsky
Philip Woods
William Wuerthele
U.S. EPA Office of Children's Health Protection
U.S. EPA Office of Research and Development
U.S. EPA Office of Wetlands, Oceans, and Watersheds
U.S. EPA Region 10
U.S. EPA Region 7
U.S. EPA Office of Ground Water and Drinking Water
U.S. EPA Office of Research and Development
U.S. EPA Region 8
U.S. EPA Office of Science and Technology
U.S. EPA Office of Research and Development
U.S. EPA Office of Air Quality Planning and Standards
U.S. EPA Office of Prevention, Pesticides, and Toxic Substances
U.S. EPA Region 6
U.S. EPA Office of Research and Development
U.S. EPA Office of Research and Development
U.S. EPA Region 5
U.S. EPA Region 5
U.S. EPA Office of Science and Technology
U.S. EPA Office of Research and Development
U.S. EPA Office of Research and Development
U.S. EPA Region 1
U.S. EPA Region 4
U.S. EPA Office of Science and Technology
U.S. EPA Office of Scence and Technology
U.S. EPA Region 9
U.S. EPA Region 9
U.S. EPA Region 8
* Principal U.S. EPA Author and Contact
VI
-------
EXTERNAL
PEER REVIEW WORKGROUP
The following professionals were part of the External Peer Review Workgroup that provided
technical and scientific review regarding the content and technical approach in the July 1998
Draft Ambient Water Quality Criteria Derivation Methodology: Human Health. Their
comments were reviewed and incorporated where appropriate to develop this final document.
Kenneth T. Bogen, Ph.D. Lawrence Livermore National Laboratory
Paul E. Brubaker, Ph.D. P.E. Brubaker Associates
Peter L. DeFur, Ph.D. Virginia Commonwealth University
Karen Erstfeld, Ph.D. Rutgers University
Bob Fares, Ph.D. Environmental Standards, Inc.
Laura Green, Ph.D. Cambridge Environmental, Inc.
Robert Hales, Ph.D. Virginia Institute of Marine Science
Brendan Hickie, Ph.D. Trent University
Ernest Hodgson, Ph.D. North Carolina State University
Paul Locke, Ph.D. Johns Hopkins University
Lynn S. McCarty, Ph.D. LS McCarty Scientific Research and Consulting
Erik Rifkin, Ph.D. Rifkin and Associates, Inc.
Damian Shea, Ph.D. North Carolina State University
Nga Tran, Ph.D. Johns Hopkins University
Curtis Travis, Ph.D. Project Performance Corp.
Potential areas for conflict of interest were investigated via direct inquiry with the peer reviews
and review of their current affiliations. No conflicts of interest were identified.
vn
-------
[This page left blank intentionally.]
Vlll
-------
TABLE OF CONTENTS
Page
NOTICE ii
FOREWORD iii
ACKNOWLEDGMENTS v
EXTERNAL PEER REVIEW WORKGROUP vii
CONTENTS ix
TABLES AND FIGURES xiv
LIST OF ACRONYMS xv
1. INTRODUCTION 1-1
1.1 Water Quality Criteria and Standards 1-1
1.2 Purpose of This Document 1-1
1.3 History of the Ambient Water Quality Criteria (AWQC) Methodology 1-2
1.4 Relationship of Water Quality Standards to AWQC 1-4
1.5 Need for the AWQC Methodology Revisions 1-4
1.5.1 Group C Chemicals 1-6
1.5.2 Consideration of Non-Water Sources of Exposure 1-7
1.5.3 Cancer Risk Ranges 1-8
1.6 Overview of the AWQC Methodology Revisions 1-9
1.7 References 1-13
2. CLARIFICATIONS ON THE METHODOLOGY, RISK CHARACTERIZATION,
AND OTHER ISSUES FOR DEVELOPING CRITERIA 2-1
2.1 Identifying the Population Subgroup that the AWQC Should Protect 2-1
2.2 Science, Science Policy, and Risk Management 2-3
2.3 Setting Criteria to Protect Against Multiple Exposures From Multiple Chemicals
(Cumulative Risk) 2-4
2.4 Cancer Risk Range 2-6
2.5 Microbiological Ambient Water Quality Criteria 2-7
2.6 Risk Characterization Considerations 2-9
2.7 Discussion of Uncertainty 2-11
2.7.1 Observed Range of Toxicity Versus Range of Environmental Exposure .... 2-11
2.7.2 Continuum of Preferred Data/Use of Defaults 2-11
2.7.3 Significant Figures 2-11
2.8 Other Considerations 2-13
2.8.1 Minimum Data Considerations 2-13
2.8.2 Site-Specific Criterion Calculation 2-13
2.8.3 Organoleptic Criteria 2-14
2.8.4 Criteria for Chemical Classes 2-15
2.8.5 Criteria for Essential Elements 2-16
2.9 References 2-16
ix
-------
3. RISK ASSESSMENT 3-1
3.1 Cancer Effects 3-1
3.1.1 Background on EPA Cancer Risk Assessment Guidelines 3-1
3.1.2 EPA's Proposed Guidelines for Carcinogen Risk Assessment and the Subsequent
July, 1999 Draft Revised Cancer Guidelines 3-2
3.1.3 Methodology for Deriving AWQC by the 1999 Draft Revised
Cancer Guidelines 3-4
3.1.3.1 Weight-of-Evidence Narrative 3-5
3.1.3.2 Mode of Action-General Considerations and Framework for
Analysis 3-6
3.1.3.3 Dose Estimation 3-7
A. Determining the Human Equivalent Dose 3-7
B. Dose-Response Analysis 3-7
3.1.3.4 Characterizing Dose-Response Relationships in the Range of Observation
and at Low Environmentally Relevant Doses 3-8
A. Extrapolation to Low, Environmentally Relevant Doses 3-9
B. Biologically-Based Modeling Approaches 3-9
C. Default Linear Extrapolation Approach 3-10
D. Default Nonlinear Approach 3-11
E. Both Linear and Nonlinear Approaches 3-13
3.1.3.5 AWQC Calculation 3-13
A. Linear Approach 3-13
B. Nonlinear Approach 3-14
3.1.3.6 Risk Characterization 3-14
3.1.3.7 Use of Toxicity Equivalence Factors (TEF) and Relative Potency
Estimates 3-15
3.1.4 References for Cancer Section 3-16
3.2 NoncancerEffects 3-17
3.2.1 1980 AWQC National Guidelines for Noncancer Effects 3-17
3.2.2 Noncancer Risk Assessment Developments Since 1980 3-18
3.2.3 Issues and Recommendations Concerning the Derivation of AWQC for
Noncarcinogens 3-20
3.2.3.1 Using the Current NOAEL/UF-Based RfD Approach or Adopting More
Quantitative Approaches for Noncancer Risk Assessment 3-20
A. The Benchmark Dose 3-22
B. Categorical Regression 3-24
C. Summary 3-25
3.2.3.2 Presenting the RfD as a Single Point or as a Range for Deriving
AWQC 3-25
3.2.3.3 Guidelines to be Adopted for Derivation of Noncancer Health Effects
Values 3-27
3.2.3.4 Treatment of Uncertainty Factors/Severity of Effects During the RfD
Derivation and Verification Process 3-27
3.2.3.5 Use of Less-Than-90-Day Studies to Derive RfDs 3-27
-------
3.2.3.6 Use of Reproductive/Developmental, Immunotoxicity, and Neurotoxi city
Data as the Basis for Deriving RfDs 3-28
3.2.3.7 Applicability of Toxicokinetic Data in Risk Assessment 3-28
3.2.3.8 Consideration of Linearity (or Lack of a Threshold) for Noncarcinogenic
Chemicals 3-29
3.2.3.9 Minimum Data Guidance 3-29
3.2.4 References for Noncancer Effects 3-30
4. EXPOSURE 4-1
4.1. Exposure Policy Issues 4-1
4.1.1 Sources of Exposure Associated with Ambient Water 4-2
4.1.1.1 Appropriateness of Including the Drinking Water Pathway in
AWQC 4-2
4.1.1.2 Setting Separate AWQC for Drinking Water and Fish
Consumption 4-2
4.1.1.3 Incidental Ingestion from Ambient Surface Waters 4-3
4.2. Consideration of Non-Water Sources of Exposure When Setting AWQC 4-3
4.2.1 Policy Background 4-3
4.2.2 The Exposure Decision Tree Approach 4-5
4.2.2.1 Problem Formulation 4-9
4.2.2.2 Data Adequacy 4-10
4.2.2.3 Regulatory Actions 4-13
4.2.2.4 Apportionment Decisions 4-13
4.2.3 Additional Points of Clarification on the Exposure Decision Tree Approach for
Setting AWQC 4-15
4.2.4 Quantification of Exposure 4-16
4.2.5 Inclusion of Inhalation and Dermal Exposures 4-16
4.3 Exposure Factors Used in the AWQC Computation 4-17
4.3.1 Human Body Weight Values for Dose Calculations 4-18
4.3.1.1 Rate Protective of Human Health from Chronic Exposure 4-19
4.3.1.2 Rates Protective of Developmental Human Health Effects 4-20
4.3.2 Drinking Water Intake Rates 4-21
4.3.2.1 Rate Protective of Human Health from Chronic Exposure 4-23
4.3.2.2 Rates Protective of Developmental Human Health Effects 4-24
4.3.2.3 Rates Based on Combining Drinking Water Intake and Body
Weight 4-24
4.3.3 Fish Intake Rates 4-25
4.3.3.1 Rates Protective of Human Health from Chronic Exposure 4-25
4.3.3.2 Rates Protective of Developmental Human Health Effects 4-29
4.3.3.3 Rates Based on Combining Fish Intake and Body Weight 4-30
4.4 References for Exposure 4-30
5. BIOACCUMULATION 5-1
5.1 Introduction 5-1
xi
-------
5.1.1 Important Bioaccumulation and Bioconcentration Concepts 5-2
5.1.2 Goal of the National BAF 5-3
5.1.3 Changes to the 1980 Methodology 5-3
5.1.3.1 Overall Approach 5-4
5.1.3.2 Lipid Normalization 5-4
5.1.3.3 Bioavailability 5-5
5.1.3.4 Trophic Level Considerations 5-5
5.1.3.5 Site-Specific Adjustments 5-5
5.1.4 Organization of This Section 5-6
5.2 Definitions 5-6
5.3 Framework for Determining National Bioaccumulation Factors 5-10
5.3.1 Four Different Methods 5-10
5.3.2 Overview of BAF Derivation Framework 5-12
5.3.3 Defining the Chemical of Concern 5-14
5.3.4 Collecting and Reviewing Data 5-14
5.3.5 Classifying the Chemical of Concern 5-15
5.4 National Bioaccumulation Factors forNonionic Organic Chemicals 5-16
5.4.1 Overview 5-16
5.4.2 Selecting the BAF Derivation Procedure 5-18
5.4.2.1 Chemicals with Moderate to High Hydrophobicity 5-18
5.4.2.2 Chemicals with Low Hydrophobicity 5-19
5.4.2.3 Assessing Metabolism 5-20
5.4.3 Deriving National BAFs Using Procedure #1 5-22
5.4.3.1 Calculating Individual Baseline BAFfs 5-23
A. Baseline BAFf from Field-Measured BAFs 5-23
B. Baseline BAFf Derived from BSAFs 5-28
C. Baseline BAFf from a Laboratory-Measured BCF^
and FCM 5-32
D. Baseline BAFf from a Kow and FCM 5-38
5.4.3.2 Selecting Final Baseline BAFfs 5-39
5.4.3.3 Calculating National BAFs 5-41
5.4.4 Deriving National BAFs Using Procedure #2 5-44
5.4.4.1 Calculating Individual Baseline BAFfs 5-45
A. Baseline BAFf from Field-Measured BAFs 5-45
B. Baseline BAFf Derived from Field-Measured BSAFs 5-46
C. Baseline BAFf from a Laboratory-Measured BCF 5-46
5.4.4.2 Selecting Final Baseline BAFfs 5-46
5.4.4.3 Calculating the National BAFs 5-47
5.4.5 Deriving National BAFs Using Procedure #3 5-47
5.4.5.1 Calculating Individual Baseline BAFfs 5-47
A. Baseline BAFf from Field-Measured BAFs 5-48
B. Baseline BAFf from a Laboratory-Measured BCF 5-48
C. Baseline BAFf from aKow 5-49
5.4.5.2 Selecting Final Baseline BAFfs 5-49
5.4.5.3 Calculating the National BAFs 5-50
5.4.6 Deriving National BAFs Using Procedure #4 5-51
xii
-------
5.4.6.1 Calculating Individual Baseline BAFfs 5-52
A. Baseline BAFf from Field-measured BAFs 5-52
B. Baseline BAFf from a Laboratory-measured BCF 5-53
5.4.6.2 Selecting Final Baseline BAFfs 5-53
5.4.6.3 Calculating National BAFs 5-54
5.5 National Bioaccumulation Factors for Ionic Organic Chemicals 5-55
5.6 National Bioaccumulation Factors for Inorganic and Organometallic Chemicals . . . 5-57
5.6.1 Selecting the BAF Derivation Procedure 5-57
5.6.2 Bioavailability 5-58
5.6.3 Deriving BAFs Using Procedure #5 5-58
5.6.3.1 Determining Field-Measured BAFs 5-59
5.6.3.2 Determining Laboratory-Measured BCFs 5-60
5.6.3.3 Determining the National BAFs 5-60
5.6.4 Deriving BAFs Using Procedure #6 5-61
5.6.4.1 Determining Field-Measured BAFs 5-62
5.6.4.2 Determining Laboratory-Measured BCFs 5-62
5.6.4.3 Determining the National BAF 5-62
5.7 References 5-63
Xlll
-------
TABLES AND FIGURES
Page
Table 3-1. Uncertainty Factors and the Modifying Factor 3-19
Figure 4-1. Exposure Decision Tree for Defining Proposed RfD (or POD/UF)
Apportionment 4-8
Figure 5-1 Framework for Deriving a National BAF 5-13
Figure 5-2 BAF Derivation for Nonionic Organic Chemicals 5-17
Table 5-1 Food-Chain Multipliers for Trophic Levels 2, 3 and 4 5-36
xiv
-------
LIST OF ACRONYMS
ADI
ARAR
ASTM
AWQC
BAF
BAFf
BCF
BCFf
BCF
BMD
BMDL
BMF
BMR
BSAF
BW
c,
cw
CDC
CSFII
CWA
DDT
DDE
ODD
DI
DNA
DNOC
DOC
ED10
EPA
FCM
FEL
FI
FIFRA
GLI
HCBD
IARC
II
ILSI
Acceptable Daily Intake
Applicable or Relevant and Appropriate Requirements
American Society of Testing and Materials
Ambient Water Quality Criteria
Bioaccumulation Factor
Baseline Bioaccumulation Factor
Bioconcentration Factor
Baseline Bioconcentration Factor
Bioconcentration Factor Based on Total Concentrations in
Tissue and Water
Benchmark Dose
Lower-Bound Confidence Limit on the BMD
Biomagnification Factor
Benchmark Response
Biota-Sediment Accumulation Factors
Body Weight
Lipid-normalized Concentration
Organic Carbon-normalized Concentration
Concentration of the Chemical in the Specified Wet Tissue
Concentration of the Chemical in Water
U.S. Centers for Disease Control and Prevention
Continuing Survey of Food Intake by Individuals
Clean Water Act
1,1,1 -trichloro-2,2-bis(p-chlorophenyl)ethane
1,1 -dichloro-2,2-bis(p-chlorophenyl)ethylene
l,l-dichloro-2,2-bis(p-chlorophenyl)ethane
Drinking Water Intake
Deoxyribonucleic Acid
2,4-dinitro-o-cresol
Dissolved Organic Carbon
Dose Associated with a 10 Percent Extra Risk
Environmental Protection Agency
Fraction Freely Dissolved
Fraction Lipid
Food Chain Multiplier
Frank Effect Level
Fish Intake
Federal Insecticide, Fungicide, and Rodenticide Act
Great Lakes Water Quality Initiative
Hexachl orobutadi ene
International Agency for Research on Cancer
Incidental Ingestion
International Life Sciences Institute
xv
-------
IRIS
kg
Kow
L
LAS
LED10
LMS
LOAEL
M{
Mt
MCL
MCLG
MF
mg
ml
MOA
MOE
NCHS
NCI
NFCS
NHANES
NOAEL
NOEL
NPDES
PAH
PCB
POD
POC
RDA
RfC
RfD
RfDDT
RPF
RSC
RSD
SAB
SDWA
SF
STORE!
TEAM
TEF
TMDL
TSD
USDA
USEPA
Integration Risk Information System
kilogram
Octanol-Water Partition Coefficient
Liter
Linear Alkylbenzesulfonate
The Lower 95 Percent Confidence Limit on a Dose Associated with a 10
Percent Extra Risk
Linear Multistage Model
Lowest Observed Adverse Effect Level
Mass of Lipid in Specified Tissue
Mass of Specified Tissue (Wet Weight)
Maximum Contaminant Level
Maximum Contaminant Level Goal
Modifying Factor
Milligrams
Milliliters
Mode of Action
Margin of Exposure
National Center for Health Statistics
National Cancer Institute
Nationwide Food Consumption Survey
National Health and Nutrition Examination Survey
No Observed Adverse Effect Level
No Observed Effect Level
National Pollutant Discharge Elimination System
Polycyclic Aromatic Hydrocarbon
Polychlorinated Biphenyls
Point of Departure
Particulate Organic Carbon
Recommended Daily Allowance
Reference Concentration
Reference Dose
Reference Dose for Developmental Effects
Relative Potency Factor
Relative Source Contribution
Risk-Specific Dose
Science Advisory Board
Safe Drinking Water Act
Safety Factor
Storage Retrieval
Total Exposure Assessment Methodology
Toxicity Equivalency Factor
Total Maximum Daily Load
Technical Support Document
United States Department of Agriculture
United States Environmental Protection Agency
xvi
-------
UF Uncertainty Factor
WQBEL Water Quality-Based Effluent Limits
xvn
-------
[This page left blank intentionally.]
xvin
-------
1. INTRODUCTION
1.1 WATER QUALITY CRITERIA AND STANDARDS
Pursuant to Section 304(a)(l) of the Clean Water Act (CWA), the U.S. Environmental
Protection Agency (EPA) is required to publish, and from time to time thereafter revise, criteria
for water quality accurately reflecting the latest scientific knowledge on the kind and extent of
all identifiable effects on human health which may be expected from the presence of pollutants
in any body of water.
Historically, the ambient water quality criteria (AWQC or 304(a) criteria) provided two
essential types of information: (1) discussions of available scientific data on the effects of the
pollutants on public health and welfare, aquatic life, and recreation; and (2) quantitative
concentrations or qualitative assessments of the levels of pollutants in water which, if not
exceeded, will generally ensure adequate water quality for a specified water use. Water quality
criteria developed under Section 304(a) are based solely on data and scientific judgments on the
relationship between pollutant concentrations and environmental and human health effects. The
304(a) criteria do not reflect consideration of economic impacts or the technological feasibility
of meeting the criteria in ambient water. These 304(a) criteria may be used as guidance by
States and authorized Tribes to establish water quality standards, which ultimately provide a
basis for controlling discharges or releases of pollutants into ambient waters.
In 1980, AWQC were derived for 64 pollutants using guidelines developed by the
Agency for calculating the impact of waterborne pollutants on aquatic organisms and on human
health. Those guidelines consisted of systematic procedures for assessing valid and appropriate
data concerning a pollutant's acute and chronic adverse effects on aquatic organisms, nonhuman
mammals, and humans.
1.2 PURPOSE OF THIS DOCUMENT
The Methodology for Deriving Ambient Water Quality Criteria for the Protection of
Human Health (2000) (hereafter the "2000 Human Health Methodology") addresses the
development of AWQC to protect human health. The Agency intends to use the 2000 Human
Health Methodology both to develop new AWQC for additional pollutants and to revise existing
AWQC. Within the next several years, EPA intends to focus on deriving AWQC for chemicals
of high priority (including, but not limited to, mercury, arsenic, PCBs, and dioxin). Furthermore,
EPA anticipates that 304(a) criteria development in the future will be for bioaccumulative
chemicals and pollutants considered highest priority by the Agency. The 2000 Human Health
Methodology is also intended to provide States and authorized Tribes flexibility in establishing
water quality standards by providing scientifically valid options for developing their own water
quality criteria that consider local conditions. States and authorized Tribes are strongly
encouraged to use this Methodology to derive their own AWQC. However, the 2000 Human
Health Methodology also defines the default factors EPA intends to use in evaluating and
determining consistency of State water quality standards with the requirements of the CWA.
The Agency intends to use these default factors to calculate national water quality criteria under
1-1
-------
Section 304(a) of the Act. EPA will also use this Methodology as guidance when promulgating
water quality standards for a State or Tribe under Section 303(c) of the CWA.
This Methodology does not substitute for the CWA or EPA's regulations; nor is it a
regulation itself. Thus, the 2000 Human Health Methodology cannot impose legally-binding
requirements on EPA, States, Tribes or the regulated community, and may not apply to a
particular situation based upon the circumstances. EPA and State/Tribal decision-makers retain
the discretion to use different, scientifically defensible, methodologies to develop human health
criteria on a case-by-case basis that differ from this Methodology where appropriate. EPA may
change the Methodology in the future through intermittent refinements as advances in science or
changes in Agency policy occur.
The 2000 Human Health Methodology incorporates scientific advancements made over
the past two decades. The use of this Methodology is an important component of the Agency's
efforts to improve the quality of the Nation's waters. EPA believes the Methodology will
enhance the overall scientific basis of water quality criteria. Further, the Methodology should
help States and Tribes address their unique water quality issues and risk management decisions,
and afford them greater flexibility in developing their water quality programs.
There are three companion Technical Support Document (TSD) volumes for the 2000
Human Health Methodology: a Risk Assessment TSD; an Exposure Assessment TSD; and a
Bioaccumulation TSD. These documents are intended to further support States and Tribes in
developing AWQC to reflect local conditions. The Risk Assessment TSD (USEPA, 2000) is
being published concurrently with this Methodology. Publication of the Exposure Assessment
and Bioaccumulation TSDs are anticipated in 2001.
1.3 HISTORY OF THE AMBIENT WATER QUALITY CRITERIA (AWQC)
METHODOLOGY
In 1980, EPA published AWQC for 64 pollutants/pollutant classes identified in Section
307(a) of the CWA and provided a methodology for deriving the criteria (USEPA, 1980). These
1980 AWQC National Guidelines (or the "1980 Methodology") for developing AWQC for the
protection of human health addressed three types of endpoints: noncancer, cancer, and
organoleptic (taste and odor) effects. Criteria for protection against noncancer and cancer effects
were estimated by using risk assessment-based procedures, including extrapolation from animal
toxicity or human epidemiological studies. Basic human exposure assumptions were applied to
the criterion equation.
The risk assessment-based procedures used to derive the AWQC to protect human health
were specific to whether the endpoint was cancer or noncancer. When using cancer as the
critical risk assessment endpoint (which had been assumed not to have a threshold), the AWQC
were presented as a range of concentrations associated with specified incremental lifetime risk
1-2
-------
levels1. When using noncancer effects as the critical endpoint, the AWQC reflected an
assessment of a "no-effect" level, since noncancer effects were assumed to have a threshold.
The key features of each procedure are described briefly in the following paragraphs.
Cancer effects. If human or animal studies on a contaminant indicated that it induced a
statistically significant carcinogenic response, the 1980 AWQC National Guidelines treated the
contaminant as a carcinogen and derived a low-dose cancer potency factor from available animal
data using the linearized multistage model (LMS). The LMS, which uses a linear, nonthreshold
assumption for low-dose risk, was used by the Agency as a science policy choice in protecting
public health, and represented a plausible upper limit for low-dose risk. The cancer potency
factor, which expresses incremental, lifetime risk as a function of the rate of intake of the
contaminant, was then combined with exposure assumptions to express that risk in terms of an
ambient water concentration. In the 1980 AWQC National Guidelines, the Agency presented a
range of contaminant concentrations corresponding to incremental cancer risks of 10"7 to 10"5
(that is, a risk of one additional case of cancer in a population often million to one additional
cancer case in a population of one hundred thousand, respectively).
Noncancer effects. If the pollutant was not considered to have the potential for causing
cancer in humans (later defined as a known, probable, or possible human carcinogen by the 1986
Guidelines for Carcinogen Risk Assessment, USEPA, 1986d), the 1980 AWQC National
Guidelines treated the contaminant as a noncarcinogen; a criterion was derived using a threshold
concentration for noncancer adverse effects. The criteria derived from noncancer data were
based on the Acceptable Daily Intake (ADI) (now termed the reference dose [RfD]). ADI values
were generally derived using a no-observed-adverse-effect level (NOAEL) from animal studies,
although human data were used whenever available. The ADI was calculated by dividing the
NOAEL by an uncertainty factor to account for uncertainties inherent in extrapolating limited
toxicological data to humans. In accordance with the National Research Council
recommendations of 1977 (NRC, 1977), safety factors (SFs) (later redefined as uncertainty
factors) of 10, 100, or 1,000 were used, depending on the quality of the data.
Organoleptic effects. Organoleptic characteristics were also used in developing criteria
for some contaminants to control undesirable taste and/or odor imparted by them to ambient
water. In some cases, a water quality criterion based on Organoleptic effects would be more
stringent than a criterion based on toxicologic endpoints. The 1980 AWQC National Guidelines
emphasized that criteria derived for Organoleptic endpoints are not based on toxicological
information, have no direct relationship to adverse human health effects and, therefore, do not
necessarily represent approximations of acceptable risk levels for humans.
'Throughout this document, the term "risk level" regarding a cancer assessment using linear approach refers to an
upper-bound estimate of excess lifetime cancer risk.
1-3
-------
1.4 RELATIONSHIP OF WATER QUALITY STANDARDS TO AWQC
Under Section 303(c) of the CWA, States have the primary responsibility for establishing
water quality standards, defined under the Act as designated beneficial uses of a water segment
and the water quality criteria necessary to support those uses. Additionally, Native American
Tribes authorized to administer the water quality standards program under 40 CFR 131.8
establish water quality standards for waters within their jurisdictions. This statutory framework
allows States and authorized Tribes to work with local communities to adopt appropriate
designated uses and to adopt criteria to protect those designated uses. Section 303(c) provides
for EPA review of water quality standards and for promulgation of a superseding Federal rule in
cases where State or Tribal standards are not consistent with the applicable requirements of the
CWA and the implementing Federal regulations, or where the Agency determines Federal
standards are necessary to meet the requirements of the Act. Section 303(c)(2)(B) specifically
requires States and authorized Tribes to adopt water quality criteria for toxics for which EPA has
published criteria under Section 304(a) and for which the discharge or presence could reasonably
be expected to interfere with the designated use adopted by the State or Tribe. In adopting such
criteria, States and authorized Tribes must establish numerical values based on one of the
following: (1) 304(a) criteria; (2) 304(a) criteria modified to reflect site-specific conditions; or,
(3) other scientifically defensible methods. In addition, States and authorized Tribes can
establish narrative criteria where numeric criteria cannot be determined.
It must be recognized that the Act uses the term "criteria" in two different ways. In
Section 303(c), the term is part of the definition of a water quality standard. Specifically, a water
quality standard is composed of designated uses and the criteria necessary to protect those uses.
Thus, States and authorized Tribes are required to adopt regulations which contain legally
enforceable criteria. However, in Section 304(a) the term criteria is used to describe the
scientific information that EPA develops to be used as guidance by States, authorized Tribes and
EPA when establishing water quality standards pursuant to 303(c). Thus, two distinct purposes
are served by the 304(a) criteria. The first is as guidance to the States and authorized Tribes in
the development and adoption of water quality criteria which will protect designated uses, and
the second is as the basis for promulgation of a superseding Federal rule when such action is
necessary.
1.5 NEED FOR THE AWQC METHODOLOGY REVISIONS
Since 1980, EPA risk assessment practices have evolved significantly in all of the major
Methodology areas: that is, cancer and noncancer risk assessments, exposure assessments, and
bioaccumulation. When the 1980 Methodology was developed, EPA had not yet developed
formal cancer or noncancer risk assessment guidelines. Since then, EPA has published several
risk assessment guidelines. In cancer risk assessment, there have been advances in the use of
mode of action (MO A) information to support both the identification of potential human
carcinogens and the selection of procedures to characterize risk at low, environmentally relevant
exposure levels. EPA published Proposed Guidelines for Carcinogen Risk Assessment (USEPA,
1996a, hereafter the "1996 proposed cancer guidelines"). These guidelines presented revised
procedures to quantify cancer risk at low doses, replacing the current default use of the LMS
model. Following review by the Agency's Science Advisory Board (SAB), EPA published the
1-4
-------
revised Guidelines for Carcinogen Risk Assessment-Review Draft in July 1999 (USEPA, 1999a,
hereafter the "1999 draft revised cancer guidelines"). In noncancer risk assessment, the Agency
is moving toward the use of the benchmark dose (BMD) and other dose-response approaches in
place of the traditional NOAEL approach to estimate an RfD or Reference Concentration (RfC).
Guidelines for Mutagenicity Risk Assessment were published in 1986 (USEPA, 1986b). In 1991,
the Agency published Guidelines for Developmental Toxicity Risk Assessment (USEPA, 1991),
and it issued Guidelines for Reproductive Toxicity Risk Assessment in 1996 (USEPA, 1996b). In
1998, EPA published final Guidelines for Neurotoxicity Risk Assessment (USEPA, 1998), and in
1999 it issued the draft Guidance for Conducting Health Risk Assessment of Chemical Mixtures
(USEPA, 1999b).
In 1986, the Agency made available to the public the Integrated Risk Information System
(IRIS). IRIS is a database that contains risk information on the cancer and noncancer effects of
chemicals. The IRIS assessments are peer reviewed and represent EPA consensus positions
across the Agency's program and regional offices.
New studies have addressed water consumption and fish tissue consumption. These
studies provide a more current and comprehensive description of national, regional, and special-
population consumption patterns that EPA has reflected in the 2000 Human Health
Methodology. In addition, more formalized procedures are now available to account for human
exposure from multiple sources when setting health goals such as AWQC that address only one
exposure source. In 1986, the Agency published the Total Exposure Assessment Methodology
(TEAM) Study: Summary and Analysis, Volume I, Final Report (USEPA, 1986c), which presents
a process for conducting comprehensive evaluation of human exposures. In 1992, EPA
published the revised Guidelines for Exposure Assessment (USEPA, 1992), which describe
general concepts of exposure assessment, including definitions and associated units, and provide
guidance on planning and conducting an exposure assessment. The Exposure Factors Handbook
was updated in 1997 (USEPA, 1997a). Also in 1997, EPA developed Guiding Principles for
Monte Carlo Analysis (USEPA, 1997b) and published its Policy for Use of Probabilistic
Analysis in Risk Assessment (see http://www.epa.gov/ncea/mcpolicy.htm). The Monte Carlo
guidance can be applied to exposure assessments and risk assessments. The Agency has recently
developed the Relative Source Contribution (RSC) Policy for assessing total human exposure to
a contaminant and apportioning the RfD among the media of concern, published for the first time
in this Methodology.
The Agency has moved toward the use of a bioaccumulation factor (B AF) to reflect the
uptake of a contaminant from all sources (e.g., ingestion, sediment) by fish and shellfish, rather
than just from the water column as reflected by the use of a bioconcentration factor (BCF) in the
1980 Methodology. The Agency has also developed detailed procedures and guidelines for
estimating BAF values.
Another reason for the 2000 Human Health Methodology is the need to bridge the gap
between the differences in the risk assessment and risk management approaches used by EPA's
Office of Water for the derivation of AWQC under the authority of the CWA and Maximum
Contaminant Level Goals (MCLGs) under the Safe Drinking Water Act (SOWA). Three notable
differences are the treatment of chemicals designated as Group C, possible human carcinogens
1-5
-------
under the 1996 proposed cancer guidelines, the consideration of non-water sources of exposure
when setting an AWQC or MCLG for a noncarcinogen, and cancer risk ranges. Those three
differences are described in the three subsections below, respectively.
1.5.1 Group C Chemicals
Chemicals were typically classified as Group C-i.e., possible human carcinogens-under
the existing (1986) EPA cancer classification scheme for any of the following reasons:
1) Carcinogenicity has been documented in only one test species and/or only one
cancer bioassay and the results do not meet the requirements of "sufficient
evidence."
2) Tumor response is of marginal statistical significance due to inadequate design or
reporting.
3) Benign, but not malignant, tumors occur with an agent showing no response in a
variety of short-term tests for mutagenicity.
4) There are responses of marginal statistical significance in a tissue known to have
a high or variable background rate.
The 1986 Guidelines for Carcinogen Risk Assessment (hereafter the "1986 cancer
guidelines") specifically recognized the need for flexibility with respect to quantifying the risk of
Group C, possible human carcinogens. The 1986 cancer guidelines noted that agents judged to
be in Group C, possible human carcinogens, may generally be regarded as suitable for
quantitative risk assessment, but that case-by-case judgments may be made in this regard.
The EPA Office of Water has historically treated Group C chemicals differently under
the CWA and the SDWA. It is important to note that the 1980 AWQC National Guidelines for
setting AWQC under the CWA predated EPA's carcinogen classification system, which was
proposed in 1984 (USEPA, 1984) and finalized in 1986 (USEPA, 1986a). The 1980 AWQC
National Guidelines did not explicitly differentiate among agents with respect to the weight of
evidence for characterizing them as likely to be carcinogenic to humans. For all pollutants
judged as having adequate data for quantifying carcinogenic risk-including those now classified
as Group C-AWQC were derived based on data on cancer incidence. In the 1980 AWQC
National Guidelines, EPA emphasized that the AWQC for carcinogens should state that the
recommended concentration for maximum protection of human health is zero. At the same time,
the criteria published for specific carcinogens presented water concentrations for these pollutants
corresponding to individual lifetime excess cancer risk levels in the range of 10"7 to 10"5.
In the development of national primary drinking water regulations under the SDWA,
EPA is required to promulgate a health-based MCLG for each contaminant. The Agency policy
has been to set the MCLG at zero for chemicals with strong evidence of carcinogenicity
associated with exposure from water. For chemicals with limited evidence of carcinogenicity,
including many Group C agents, the MCLG was usually obtained using an RfD based on the
1-6
-------
pollutant's noncancer effects with the application of an additional uncertainty factor of 1 to 10 to
account for carcinogenic potential of the chemical. If valid noncancer data for a Group C agent
were not available to establish an RfD but adequate data are available to quantify the cancer risk,
then the MCLG was based upon a nominal lifetime excess cancer risk in the range of 10"6 tolO"5
(ranging from one case in a population of one million to one case in a population of one hundred
thousand). Even in those cases where the RfD approach has been used for the derivation of the
MCLG for a Group C agent, the drinking water concentrations associated with excess cancer
risks in the range of 10"6 to 10"5 were also provided for comparison.
It should also be noted that EPA's pesticides program has applied both of the previously
described methods for addressing Group C chemicals in actions taken under the Federal
Insecticide, Fungicide, and Rodenticide Act (FIFRA) and finds both methods applicable on a
case-by-case basis. Unlike the drinking water program, however, the pesticides program does
not add an extra uncertainty factor to account for potential carcinogenicity when using the RfD
approach.
In the 1999 draft revised cancer guidelines, there are no more alphanumeric categories.
Instead, there will be longer narratives for hazard characterization that will use consistent
descriptive terms when assessing cancer risk.
1.5.2 Consideration of Non-water Sources of Exposure
The 1980 AWQC National Guidelines recommended that contributions from non-water
sources, namely air and non-fish dietary intake, be subtracted from the Acceptable Daily Intake
(ADI), thus reducing the amount of the ADI "available" for water-related sources of intake. In
practice, however, when calculating human health criteria, these other exposures were generally
not considered because reliable data on these exposure pathways were not available.
Consequently, the AWQC were usually derived such that drinking water and fish ingestion
accounted for the entire ADI (now called RfD).
In the drinking water program, a similar "subtraction" method was used in the derivation
of MCLGs proposed and promulgated in drinking water regulations through the mid-1980s.
More recently, the drinking water program has used a "percentage" method in the derivation of
MCLGs for noncarcinogens. In this approach, the percentage of total exposure typically
accounted for by drinking water, referred to as the relative source contribution (RSC), is applied
to the RfD to determine the maximum amount of the RfD "apportioned" to drinking water
reflected by the MCLG value. In using this percentage procedure, the drinking water program
also applies a ceiling level of 80 percent of the RfD and a floor level of 20 percent of the RfD.
That is, the MCLG cannot account for more than 80 percent of the RfD, nor less than 20 percent
of the RfD.
The drinking water program usually takes a conservative approach to public health by
applying an RSC factor of 20 percent to the RfD when adequate exposure data do not exist,
assuming that the major portion (80 percent) of the total exposure comes from other sources,
such as diet.
1-7
-------
In the 2000 Human Health Methodology, guidance for the routine consideration of non-
water sources of exposure [both ingestion exposures (e.g., food) and exposures other than the
oral route (e.g., inhalation)] is presented. The approach is called the Exposure Decision Tree.
Relative source contribution estimates will be made by EPA using this approach, which allows
for use of either the subtraction or percentage methods, depending on chemical-specific
circumstances, within the 20 to 80 percent range described above.
1.5.3 Cancer Risk Ranges
In addition to the different risk assessment approaches discussed above for deriving
AWQC and MCLGs for Group C agents, there have been different risk management approaches
by the drinking water and surface water programs on lifetime excess risk values when setting
health-based criteria for carcinogens. The surface water program has derived AWQC for
carcinogens that generally corresponded to lifetime excess cancer risk levels of 10"7 to 10"5. The
drinking water program has set MCLGs for Group C agents based on a slightly less stringent risk
range of 10"6 to 10"5, while MCLGs for chemicals with strong evidence of carcinogenicity (that
is, classified as Group A, known, or B probable, human carcinogen) are set at zero. The drinking
water program is now following the principles of the 1999 draft revised cancer guidelines to
determine the type of low-dose extrapolation based on mode of action.
It is also important to note that under the drinking water program, for those substances
having an MCLG of zero, enforceable Maximum Contaminant Levels (MCLs) have generally
been promulgated to correspond with cancer risk levels ranging from 10"6 to 10"4. Unlike AWQC
and MCLGs which are strictly health-based criteria, MCLs are developed with consideration
given to the costs and technological feasibility of reducing contaminant levels in water to meet
those standards.
With the 2000 Human Health Methodology, EPA will publish its national 304(a) water
quality criteria at a 10"6 risk level, which EPA considers appropriate for the general population.
EPA is increasing the degree of consistency between the drinking water and ambient water
programs, given the somewhat different requirements of the CWA and SDWA.
-------
1.6 OVERVIEW OF THE AWQC METHODOLOGY REVISIONS
The following equations for deriving AWQC include toxicological and exposure
assessment parameters which are derived from scientific analysis, science policy, and risk
management decisions. For example, values for parameters such as a field-measured BAF or a
point of departure from an animal study [in the form of a lowest-observed-adverse-effect level
(LOAEL)/no-observed -adverse-effect level (NOAEL)/lower 95 percent confidence limit on a
dose associated with a 10 percent extra risk (LED10)] are empirically measured using scientific
methods. By contrast, the decision to use animal effects as surrogates for human effects involves
judgment on the part of the EPA (and similarly, by other agencies) as to the best practice to
follow when human data are lacking. Such a decision is, therefore, a matter of science policy.
The choice of default fish consumption rates for protection of a certain percentage (i.e., the 90th
percentile) of the general population is clearly a risk management decision. In many cases, the
Agency has selected parameter values using its best judgment regarding the overall protection
afforded by the resulting AWQC when all parameters are combined. For a longer discussion of
the differences between science, science policy, and risk management, please refer to Section 2
of this document. Section 2 also provides further details with regard to risk characterization for
this Methodology, with emphasis placed on explaining the uncertainties in the overall risk
assessment.
The generalized equations for deriving AWQC based on noncancer effects are:
Noncancer Effects2
BW
AWQC = RfD • RSC
DI + £ (FIj • BAFj)
i=2
(Equation 1-1)
Cancer Effects: Nonlinear Low-Dose Extrapolation
BW
AWQC =
UF
RSC
DI + E (FIj
i=2
(Equation 1-2)
2 Although appearing in this equation as a factor to be multiplied, the RSC can also be an amount subtracted. Refer to
the explanation key below the equations.
1-9
-------
Cancer Effects: Linear Low-Dose Extrapolation
BW
AWQC = RSD
DI + £ (FIj • BAF;)
i=2
(Equation 1-3)
where:
AWQC
RfD
POD
UF
RSD
RSC
BW
DI
FL
BAF;
Ambient Water Quality Criterion (mg/L)
Reference dose for noncancer effects (mg/kg-day)
Point of departure for carcinogens based on a nonlinear low-dose
extrapolation (mg/kg-day), usually a LOAEL, NOAEL, or LED10
Uncertainty Factor for carcinogens based on a nonlinear low-dose
extrapolation (unitless)
Risk-specific dose for carcinogens based on a linear low-dose
extrapolation (mg/kg-day) (dose associated with a target risk, such
as 1Q-6)
Relative source contribution factor to account for non-water
sources of exposure. (Not used for linear carcinogens.) May be
either a percentage (multiplied) or amount subtracted, depending
on whether multiple criteria are relevant to the chemical.
Human body weight (default = 70 kg for adults)
Drinking water intake (default = 2 L/day for adults)
Fish intake at trophic level (TL) I (I = 2, 3, and 4) (defaults for
total intake = 0.0175 kg/day for general adult population and sport
anglers, and 0.1424 kg/day for subsistence fishers). Trophic level
breakouts for the general adult population and sport anglers are:
TL2 = 0.0038 kg/day; TL3 = 0.0080 kg/day; and TL4 = 0.0057
kg/day.
Bioaccumulation factor at trophic level I (1=2, 3 and 4), lipid
normalized (L/kg)
For highly bioaccumulative chemicals where ingestion from water might be considered
negligible, EPA is currently evaluating the feasibility of developing and implementing AWQCs
that are expressed in terms of concentrations in tissues of aquatic organisms. Such tissue residue
criteria might be used as an alternative to AWQCs which are expressed as concentrations in
water, particularly in situations where AWQCs are at or below the practical limits for
quantifying a chemical in water. Even though tissue residue criteria would not require the use of
a BAF in their derivation, implementing such criteria would still require a mechanism for
relating chemical loads and concentrations in water and sediment to concentrations in tissues of
appropriate fish and shellfish (e.g., a BAF or bioaccumulation model). At this time, no revisions
are planned to the Methodology to provide specific guidance on developing fish tissue-based
water quality criteria. However, guidance may be provided in the future either as a separate
document or integrated in a specific 304(a) water quality criteria document for a chemical that
warrants such an approach.
1-10
-------
AWQC for the protection of human health are designed to minimize the risk of adverse
effects occurring to humans from chronic (lifetime) exposure to substances through the ingestion
of drinking water and consumption offish obtained from surface waters. The Agency is not
recommending the development of additional water quality criteria similar to the "drinking water
health advisories" that focus on acute or short-term effects; these are not seen as routinely having
a meaningful role in the water quality criteria and standards program. However, as discussed
below, there may be some instances where the consideration of acute or short-term toxicity and
exposure in the derivation of AWQC is warranted.
Although the AWQC are based on chronic health effects data (both cancer and noncancer
effects), the criteria are intended to also be protective against adverse effects that may reasonably
be expected to occur as a result of elevated acute or short-term exposures. That is, through the
use of conservative assumptions with respect to both toxicity and exposure parameters, the
resulting AWQC should provide adequate protection not only for the general population over a
lifetime of exposure, but also for special subpopulations who, because of high water- or fish-
intake rates, or because of biological sensitivities, have an increased risk of receiving a dose that
would elicit adverse effects. The Agency recognizes that there may be some cases where the
AWQC based on chronic toxicity may not provide adequate protection for a subpopulation at
special risk from shorter-term exposures. The Agency encourages States, Tribes, and others
employing the 2000 Human Health Methodology to give consideration to such circumstances in
deriving criteria to ensure that adequate protection is afforded to all identifiable subpopulations.
(See Section 4.3, Factors Used in the AWQC Computation, for additional discussion of these
subpopulations.)
The EPA is in the process of revising its cancer guidelines, including its descriptions of
human carcinogenic potential. Once final guidelines are published, they will be the basis for
assessment under this methodology. In the meanwhile, the 1986 guidelines are used and
extended with principles discussed in EPA's 1999 Guidelines for Carcinogen Risk Assessment -
Review Draft (hereafter "1999 draft revised cancer guidelines"). These principles arise from
new science about cancer discovered in the last 15 years and from EPA policy of recent years
supporting full characterization of hazard and risk both for the general population and
potentially sensitive groups such as children. These principles are incorporated in recent and
ongoing assessments such as the reassessment of dioxin, consistent with the 1986 guidelines.
Until final guidelines are published, information is presented to describe risk under both the old
guidelines and draft revisions. Dose-response assessment under the 1986 guidelines employs a
linearized multistage model to extrapolate tumor dose-response observed in animal or human
studies down to zero dose, zero extra risk. The dose-response assessment under EPA's 1999
draft revised cancer guidelines is a two-step process. In the first step, the response data are
modeled in the range of empirical observation. Modeling in the observed range is done with
biologically based or appropriate curve-fitting modeling. In the second step, extrapolation below
the range of observation is accomplished by biologically based modeling if there are sufficient
data or by a default procedure (linear, nonlinear, or both). A point of departure (POD) for
extrapolation is estimated from modeling observed data. The lower 95 percent confidence limit
on a dose associated with 10 percent extra risk (LED10) is the standard POD for low-dose
extrapolation. The linear default procedure is a straight line extrapolation to the origin (i.e., zero
dose, zero extra risk) from the POD, which is the LED10 identified in the observable response
1-11
-------
range. The result of this procedure is generally comparable (within 2-fold) to that of using a
linearized multistage model under existing, 1986 guidelines. The linear low-dose extrapolation
applies to agents that are best characterized by the assumption of linearity (e.g., direct DNA
reactive mutagens) for their MOA. A linear approach would also be applied when inadequate or
no information is available to explain the carcinogenic MOA; this is a science policy choice in
the interest of public health. If it is determined that the MOA understanding fully supports a
nonlinear extrapolation, the AWQC is derived using the nonlinear default which is based on a
margin of exposure (MOE) analysis using the LED10 as the POD and applying uncertainty
factors (UFs) to arrive at an acceptable MOE. There may be situations where it is appropriate to
apply both the linear and nonlinear default procedures (e.g., for an agent that is both DNA
reactive and active as a promoter at higher doses).
For substances that are carcinogenic, particularly those for which the MOA suggests
nonlinearity at low doses, the Agency recommends that an integrated approach be taken in
looking at cancer and noncancer effects. If one effect does not predominate, AWQC values
should be determined for both carcinogenic and noncarcinogenic endpoints. The lower of the
resulting values should be used for the AWQC.
When deriving AWQC for noncarcinogens and carcinogens based on a nonlinear low-
dose extrapolation, a factor is included to account for other non-water exposure sources [both
ingestion exposures (e.g., food) and exposures other than the oral route (e.g., inhalation)] so that
the entire RfD, or POD/UF, is not apportioned to drinking water and fish consumption alone.
Guidance is provided in the 2000 Human Health Methodology for determining the factor (i.e.,
the RSC) to be used for a particular chemical. The Agency is recommending the use of an
Exposure Decision Tree procedure to support the determination of the appropriate RSC value for
a given water contaminant. In the absence of data, the Agency intends to use 20 percent of the
RfD (or POD/UF) as the default RSC in calculating 304(a) criteria or promulgating State or
Tribal water quality standards under Section 303(c).
With AWQC derived for carcinogens based on a linear low-dose extrapolation, the
Agency will publish recommended criteria values at a 10"6 risk level. States and authorized
Tribes can always choose a more stringent risk level, such as 10"7. EPA also believes that
criteria based on a 10"5 risk level are acceptable for the general population as long as States and
authorized Tribes ensure that the risk to more highly exposed subgroups (sportfishers or
subsistence fishers) does not exceed the 10"4 level. Clarification on this risk management
decision is provided in Section 2 of this document.
The default fish consumption value for the general adult population in the 2000 Human
Health Methodology is 17.5 grams/day, which represents an estimate of the 90th percentile
consumption rate for the U.S. adult population based on the U.S. Department of Agriculture's
(USDA's) Continuing Survey of Food Intake by Individuals (CSFII) 1994-96 data (USDA,
1998). EPA will use this default intake rate with future national 304(a) criteria derivations or
revisions. This default value is chosen to be protective of the majority of the general population.
However, States and authorized Tribes are urged to use a fish intake level derived from local
data on fish consumption in place of this default value when deriving AWQC, ensuring that the
fish intake level chosen is protective of highly exposed individuals in the population. EPA has
1-12
-------
provided default values for States and authorized Tribes that do not have adequate information
on local or regional consumption patterns, based on numerous studies that EPA has reviewed on
sport anglers and subsistence fishers. EPA's defaults for these population groups are estimates
of their average consumption. EPA recommends a default of 17.5 grams/day for sport anglers as
an approximation of their average consumption and 142.4 grams/day for subsistence fishers,
which falls within the range of averages for this group. Consumption rates for women of
childbearing age and children younger than 14 are also provided to maximize protection in those
cases where these subpopulations may be at greatest risk.
In the 2000 Human Health Methodology, criteria are derived using a B AF rather than a
BCF. To derive the BAF, States and authorized Tribes may use EPA's Methodology or any
method consistent with this Methodology. EPA's highest preference in developing BAFs are
BAFs based on field-measured data from local/regional fish.
1.7 REFERENCES
NRC (National Research Council). 1977. Drinking Water and Health. Safe Drinking Water
Committee. National Academy of Sciences, National Academy Press. Washington, DC.
USDA. 1998. U.S. Department of Agriculture. 1994-1996 Continuing Survey of Food Intakes
by Individuals and 1994 1996 Diet and Health Knowledge Survey. Agricultural
Research Service, USDA. NTIS CD-ROM, accession number PB98-500457.
USEPA (U.S. Environmental Protection Agency). 1980. Guidelines and methodology used in
the preparation of health effect assessment chapters of the consent decree water criteria
documents. Federal Register 45: 79347, Appendix 3.
USEPA (U.S. Environmental Protection Agency). 1984. Proposed guidelines for carcinogen
risk assessment. Federal Register 49:46294.
USEPA (U.S. Environmental Protection Agency). 1986a. Guidelines for carcinogen risk
assessment. Federal Register 51:33992-34003.
USEPA (U.S. Environmental Protection Agency). 1986b. Guidelines for mutagenicity risk
assessment. Federal Register 51:34006-34012.
USEPA (U.S. Environmental Protection Agency). 1986c. Total Exposure Assessment Model
(TEAM) Study: Summary and Analysis, Volume I. Final Report. EPA/600/6-87/002a.
USEPA (U.S. Environmental Protection Agency). 1986d. Guidelines for exposure assessment.
Federal Register 51:34042-34054.
USEPA (U.S. Environmental Protection Agency). 1991. Guidelines for developmental toxicity
risk assessment. Federal Register 56:63789- 63826.
1-13
-------
USEPA (U.S. Environmental Protection Agency). 1992. Guidelines for exposure assessment.
Federal Register 57:22888-22938.
USEPA (U.S. Environmental Protection Agency). 1996a. Proposed guidelines for carcinogen
risk assessment. Federal Register 61:17960-18011.
USEPA (U.S. Environmental Protection Agency). 1996b. Guidelines for reproductive toxicity
risk assessment. Federal Register 61: 6274-56322.
USEPA (U.S. Environmental Protection Agency). 1997a. Exposure Factors Handbook. Office
of Research and Development. Washington, DC. EPA/600/P-95/002Fa..
USEPA (U.S. Environmental Protection Agency). 1997b. Guiding Principles for Monte Carlo
Analysis. Risk Assessment Forum. Washington, DC. EPA/630/R-97/001.
USEPA (U.S. Environmental Protection Agency). 1998. Guidelines for neurotoxicity risk
assessment. Federal Register 63: 26926.
USEPA (U.S. Environmental Protection Agency). 1999a. 1999 Guidelines for Carcinogen Risk
Assessment. Review Draft. Office of Research and Development. Washington, DC.
NCEA-F-0644.
USEPA (U.S. Environmental Protection Agency). 1999b. Guidance for Conducting Health Risk
Assessment of Chemical Mixtures. Final Draft. Risk Assessment Forum Technical
Panel. Washington, DC. EPA/NCEA-C-0148. September. Website:
http ://www. epa.gov/ncea/raf/rafpub. htm
USEPA (U.S. Environmental Protection Agency). 2000. Methodology for Deriving Ambient
Water Quality Criteria for the Protection of Human Health (2000). Technical Support
Document Volume 1: Risk Assessment. Office of Science and Technology, Office of
Water. Washington, DC. EPA-822-B-00-005. August.
1-14
-------
2. CLARIFICATIONS ON THE METHODOLOGY, RISK CHARACTERIZATION,
AND OTHER ISSUES FOR DEVELOPING CRITERIA
2.1 IDENTIFYING THE POPULATION SUBGROUP THAT THE AWQC SHOULD
PROTECT
Water quality criteria are derived to establish ambient concentrations of pollutants which,
if not exceeded, will protect the general population from adverse health impacts from those
pollutants due to consumption of aquatic organisms and water, including incidental water
consumption related to recreational activities. For each pollutant, chronic criteria are derived to
reflect long-term consumption of food and water. An important decision to make when setting
AWQC is the choice of the particular population to protect. For instance, criteria could be set to
protect those individuals who have average or "typical" exposures, or the criteria could be set so
that they offer greater protection to those individuals who are more highly exposed. EPA has
selected default parameter values that are representative of several defined populations: adults in
the general population; sport (recreational) fishers; subsistence fishers; women of childbearing
age (defined as ages 15-44); and children (up to the age of 14). In deciding on default parameter
values, EPA is aware that multiple parameters are used in combination when calculating AWQC
(e.g., intake rates and body weight). EPA describes the estimated population percentiles that are
represented by each of the default exposure parameter values in Section 4.
EPA's national 304(a) criteria are usually derived to protect the majority of the general
population from chronic adverse health effects. EPA has used a combination of median values,
mean values, and percentile estimates for the parameter value defaults to calculate its national
304(a) criteria. EPA believes that its assumptions afford an overall level of protection targeted at
the high end of the general population (i.e., the target population or the criteria-basis population).
EPA also believes that this is reasonably conservative and appropriate to meet the goals of the
CWA and the 304(a) criteria program. EPA considers that its target protection goal is satisfied if
the population as a whole will be adequately protected by the human health criteria when the
criteria are met in ambient water. However, associating the derived criteria with a specific
population percentile is far more difficult, and such a quantitative descriptor typically requires
detailed distributional exposure and dose information. EPA's Guidelines For Exposure
Assessment (USEPA, 1992) describes the extreme difficulty in making accurate estimates of
exposures and indicates that uncertainties at the more extreme ends of the distribution increase
greatly. On quantifying population exposures/risks, the guidelines specifically state:
In practice, it is difficult even to establish an accurate mean health effect risk for
a population. This is due to many complications, including uncertainties in using
animal data for human dose-response relationships, nonlinearities in the dose-
response curve, projecting incidence data from one group to another dissimilar
group, etc. Although it has been common practice to estimate the number of
cases of disease, especially cancer, for populations exposed to chemicals, it
should be understood that these estimates are not meant to be accurate estimates
of real (or actuarial) cases of disease. The estimate's value lies in framing
2-1
-------
hypothetical risk in an understandable way rather than in any literal
interpretation of the term "cases. "
Although it is not possible to subject the estimates to such a rigorous analysis (say, for
example, to determine what criterion value provides protection of exactly the 90th percentile of
the population), EPA believes that the combination of parameter value assumptions achieves its
target goal, without being inordinately conservative. The standard assumptions made for the
national 304(a) criteria are as follows. The assumed body weight value used is an arithmetic
mean, as are the RSC intake estimates of other exposures (e.g., non-fish dietary), when data are
available. The BAF component data (e.g., for lipid values, for paniculate and dissolved organic
carbon) are based on median (i.e., 50th percentile) values. The drinking water intake values are
approximately 90th percentile estimates and fish intake values are 90th percentile estimates. EPA
believes the use of these values will result in 304(a) criteria that are protective of a majority of
the population; this is EPA's goal.
However, EPA also strongly believes that States and authorized Tribes should have the
flexibility to develop criteria, on a site-specific basis, that provide additional protection
appropriate for highly exposed populations. EPA is aware that exposure patterns in general, and
fish consumption in particular, vary substantially. EPA understands that highly exposed
populations may be widely distributed geographically throughout a given State or Tribal area.
EPA recommends that priority be given to identifying and adequately protecting the most highly
exposed population. Thus, if the State or Tribe determines that a highly exposed population is
at greater risk and would not be adequately protected by criteria based on the general population,
and by the national 304(a) criteria in particular, EPA recommends that the State or Tribe adopt
more stringent criteria using alternative exposure assumptions.
EPA has provided recommended default intake rates for various population groups for
State and Tribal consideration. EPA does not intend for these alternative default values to be
prescriptive. EPA strongly emphasizes its preference that States and Tribes use local or regional
data over EPA's defaults, if they so choose, as being more representative of their population
groups of concern.
In the course of updating the 2000 Human Health Methodology, EPA received some
questions regarding the population groups for which the criteria would be developed. EPA does
not intend to derive multiple 304(a) criteria for all subpopulation groups for every chemical. As
stated above, criteria that address chronic adverse health effects are most applicable to the CWA
Section 304(a) criteria program and the chemicals evaluated for this program. If EPA
determined that pregnant women/fetuses or young children were the target population (or criteria
basis population) of a chemical's RfD or POD/UF, then the 304(a) criteria would be developed
using exposure parameters for that subgroup. This would only be relevant for acute or
subchronic toxicity situations. This does not conflict with the fact that chronic health effects
potentially reflect a person's exposure during both childhood and adult years.
For RfD-based and POD/UF-based chemicals, EPA's policy is that, in general, the RfD
(or POD/UF) should not be exceeded and the exposure assumptions used should reflect the
population of concern. It is recommended that when a State or authorized Tribe sets a
2-2
-------
waterbody-specific AWQC, they consider the populations most exposed via water and fish.
EPA's policy on cancer risk management goals is discussed in Section 2.4.
Health Risks to Children
In recognition that children have a special vulnerability to many toxic substances, EPA's
Administrator directed the Agency in 1995 to explicitly and consistently take into account
environmental health risks to infants and children in all risk assessments, risk characterizations,
and public health standards set for the United States. In April 1997, President Clinton signed
Executive Order 13045 on the protection of children from environmental health risks, which
assigned a high priority to addressing risks to children. In May 1997, EPA established the Office
of Children's Health Protection to ensure the implementation of the President's Executive Order.
EPA has increased efforts to ensure its guidance and regulations take into account risks to
children. Circumstances where risks to children should be considered in the context of the 2000
Human Health Methodology are discussed in the Section 3.2, Noncancer Effects (in terms of
developmental and reproductive toxicity) and in Section 4, Exposure (for appropriate exposure
intake parameters).
Details on risk characterization and the guiding principles stated above are included in
EPA's March 21, 1995 policy statement and the discussion of risk characterization (USEPA,
1995) and the 1999 Guidelines for Carcinogen Risk Assessment. Review Draft (USEPA, 1999a)
and the Reproductive and Toxicity Risk Assessment Guidelines of 1996 (USEPA, 1996b).
2.2 SCIENCE, SCIENCE POLICY, AND RISK MANAGEMENT
An important part of risk characterization, as described later in Section 2.7, is to make
risk assessments transparent. This means that conclusions drawn from the science are identified
separately from policy judgments and risk management decisions, and that the use of default
values or methods, as well as the use of assumptions in risk assessments, are clearly articulated.
In this Methodology, EPA has attempted to separate scientific analysis from science policy and
risk management decisions for clarity. This should allow States and Tribes (who are also
prospective users of this Methodology) to understand the elements of the Methodology
accurately and clearly, and to easily separate out the scientific decisions from the science policy
and risk management decisions. This is important so that when questions are asked regarding
the scientific merit, validity, or apparent stringency or leniency of AWQC, the implementer of
the criteria can clearly explain what judgments were made to develop the criterion in question
and to what degree these judgments were based on science, science policy, or risk management.
To some extent this process will also be displayed in future AWQC documents.
When EPA speaks of science or scientific analysis, it is referring to the extraction of data
from toxicological or exposure studies and surveys with a minimum of judgment being used to
make inferences from the available evidence. For example, if EPA is describing a POD from an
animal study (e.g., a LOAEL), this is usually determined as a lowest dose that produces an
observable adverse effect. This would constitute a scientific determination. Judgments applying
science policy, however, may enter this determination. For example, several scientists may
differ in their opinion of what is adverse, and this in turn can influence the selection of a LOAEL
2O
-3
-------
in a given study. The use of an animal study to predict effects in a human in the absence of
human data is an inherent science policy decision. The selection of specific UFs when
developing an RfD is another example of science policy. In any risk assessment, a number of
decision points occur where risk to humans can only be inferred from the available evidence.
Both scientific judgments and policy choices may be involved in selecting from among several
possible inferences when conducting a risk assessment.
Risk management is the process of selecting the most appropriate guidance or regulatory
actions by integrating the results of risk assessment with engineering data and with social,
economic, and political concerns to reach a decision. In this Methodology, the choice of a
default fish consumption rate which is protective of 90 percent of the general population is a risk
management decision. The choice of an acceptable cancer risk by a State or Tribe is a risk
management decision.
Many of the components in the 2000 Human Health Methodology are an amalgam of
science, science policy, and/or risk management. For example, most of the default values chosen
by EPA are based on examination of scientific data and application of either science policy or
risk management. This includes the default assumption of 2 liters a day of drinking water; the
assumption of 70 kilograms for an adult body weight; the use of default percent lipid and
particulate organic carbon/dissolved organic carbon (POC/DOC) for developing national BAFs;
the default fish consumption rates for the general population and sport and subsistence anglers;
and the choice of a default cancer risk level. Some decisions are more grounded in science and
science policy (such as the choice of default BAFs) and others are more obviously risk
management decisions (such as the determination of default fish consumption rates and cancer
risk levels). Throughout the 2000 Human Health Methodology, EPA has identified the kind of
decision necessary to develop defaults and what the basis for the decision was. More details on
the concepts of science analysis, science policy, risk management, and how they are introduced
into risk assessments are included in Risk Assessment in the Federal Government: Managing the
Process (NRC, 1983).
2.3 SETTING CRITERIA TO PROTECT AGAINST MULTIPLE EXPOSURES
FROM MULTIPLE CHEMICALS (CUMULATIVE RISK)
EPA is very much aware of the complex issues and implications of cumulative risk and
has endeavored to begin developing an overall approach at the Agency-wide level. Assuming
that multiple exposures to multiple chemicals are additive is scientifically sound if they exhibit
the same toxic endpoints and modes of action. There are numerous publications relevant to
cumulative risk that can assist States and Tribes in understanding the complex issues associated
with cumulative risk. These include the following:
» Durkin, P.R., R.C. Hertzberg, W. Stiteler, and M. Mumtaz. 1995. The identification and
testing of interaction patterns. Toxicol. Letters 79:251-264.
*• Hertzberg, R.C., G. Rice, and L.K. Teuschler. 1999. Methods for health risk assessment
of combustion mixtures. In: Hazardous Waste Incineration: Evaluating the Human
2-4
-------
Health and Environmental Risks. S. Roberts, C. Teaf and J. Bean, (eds). CRC Press
LLC, Boca Raton, FL. Pp. 105-148.
> Rice, G., J. Swartout, E. Brady-Roberts, D. Reisman, K. Mahaffey, and B. Lyon. 1999.
Characterization of risks posed by combustor emissions. Drug and Chem. Tox. 22:221-
240.
*• USEP A. 1999. Guidance for Conducting Health Risk Assessment of Chemical Mixtures.
Final Draft. Risk Assessment Forum Technical Panel. Washington, DC. NCEA-C-
0148. September. Web site: http://www.epa.gov/ncea/raf/rafpub.htm
*• USEP A. 1998. Methodology for Assessing Health Risks Associated with Multiple
Pathways of Exposure to Combustor Emissions. (Update to EPA/600/6-90/003
Methodology for Assessing Health Risks Associated with Indirect Exposure to Combustor
Emissions). National Center for Environmental Assessment. Washington, DC. EPA-
600-R-98-137. Website http://www.epa.gov/ncea/combust.htm
*• USEPA. 1996. PCBs: Cancer Dose-Response Assessment and Application to
Environmental Mixtures. National Center for Environmental Assessment. Washington,
DC. EPA/600/P-96/001F.
*• USEPA. 1993. Review Draft Addendum to the Methodology for Assessing Health Risks
Associated with Indirect Exposure to Combustor Emissions. Office of Health and
Environmental Assessment, Office of Research and Development. Washington, DC.
EPA/600/AP-93/003. November.
*• USEPA. 1993. Provisional Guidance for Quantitative Risk Assessment of'Polycyclic
Aromatic Hydrocarbons. Office of Research and Development. Washington, DC.
EPA/600/R-93/089. July.
*• USEPA. 1990. Technical Support Document on Health Risk Assessment of Chemical
Mixtures. Office of Research and Development. Washington, DC. EPA/600/8/90/064.
August.
*• USEPA. 1989a. Risk Assessment Guidance for Superfund. Vol. 1. Human Health
Evaluation Manual (Part A). Office of Emergency and Remedial Response.
Washington, DC. EPA/540/1-89/002.
*• USEPA. 1989b. Interim Procedures for Estimating Risks Associated with Exposures to
Mixtures of Chlorinated Dibenzo-p-Dioxins and -Dibenzofurans (CDDs and CDFs) and
1989 Update. Risk Assessment Forum. Washington, DC. EPA/625/3-89/016. March.
The Agency's program offices are also engaged in on-going discussions of the great
complexities, methodological challenges, data adequacy needs and other information gaps, as
well as the science policy and risk management decisions that will need to be made, as they
pursue developing a sound strategy and, eventually, specific guidance for addressing cumulative
2-5
-------
risks. As a matter of internal policy, EPA is committed to refining the Methodology as advances
in relevant aspects of the science improve, as part of the water quality criteria program.
2.4 CANCER RISK RANGE
For deriving 304(a) criteria or promulgating water quality criteria for States and Tribes
under Section 303(c) based on the 2000 Human Health Methodology, EPA intends to use the 10"
6 risk level, which the Agency believes reflects an appropriate risk for the general population.
EPA's program office guidance and regulatory actions have evolved in recent years to target a
10~6 risk level as an appropriate risk for the general population. EPA has recently reviewed the
policies and regulatory language of other Agency mandates (e.g., the Clean Air Act
Amendments of 1990, the Food Quality Protection Act) and believes the target of a 10~6 risk
level is consistent with Agency-wide practice.
EPA believes that both 10~6 and 10~5 may be acceptable for the general population and
that highly exposed populations should not exceed a 10~4 risk level. States or Tribes that have
adopted standards based on criteria at the 10~5 risk level can continue to do so, if the highly
exposed groups would at least be protected at the 10~4 risk level. However, EPA is not
automatically assuming that 10~5 will protect "the highest consumers" at the 10~4 risk level. Nor
is EPA advocating that States and Tribes automatically set criteria based on assumptions for
highly exposed population groups at the 10~4 risk level. The Agency is simply endeavoring to
add that a specific determination should be made to ensure that highly exposed groups do not
exceed a 10~4 risk level. EPA understands that fish consumption rates vary considerably,
especially among subsistence populations, and it is such great variation among these population
groups that may make either 10~6 or 10~5 protective of those groups at a 10~4 risk level.
Therefore, depending on the consumption patterns in a given State or Tribal jurisdiction, a 10~6
or 10~5 risk level could be appropriate. In cases where fish consumption among highly exposed
population groups is of a magnitude that a 10~4 risk level would be exceeded, a more protective
risk level should be chosen. Such determinations should be made by the State or Tribal
authorities and are subject to EPA's review and approval or disapproval under Section 303(c) of
the CWA.
Adoption of a 10~6 or 10~5 risk level, both of which States and authorized Tribes have
chosen in adopting water quality standards to date, represents a generally acceptable risk
management decision, and EPA intends to continue providing this flexibility to States and
Tribes. EPA believes that such State or Tribal decisions are consistent with Section 303(c) if the
State or authorized Tribe has identified the most highly exposed subpopulation, has
demonstrated that the chosen risk level is adequately protective of the most highly exposed
subpopulation, and has completed all necessary public participation. States and authorized
Tribes also have flexibility in how they demonstrate this protectiveness and obtain such
information. A State or authorized Tribe may use existing information as well as collect new
information in making this determination. In addition, if a State or authorized Tribe does not
believe that the 10~6 risk level adequately protects the exposed subpopulations, water quality
criteria based on a more stringent risk level may be adopted. This discretion includes combining
the 10~6 risk level with fish consumption rates for highly exposed population groups.
2-6
-------
It is important to understand that criteria for carcinogens are based on chosen risk levels
that inherently reflect, in part, the exposure parameters used to derive those values. Therefore,
changing the exposure parameters also changes the risk. Specifically, the incremental cancer
risk levels are relative, meaning that any given criterion associated with a particular cancer risk
level is also associated with specific exposure parameter assumptions (e.g., intake rates, body
weights). When these exposure parameter values change, so does the relative risk. For a
criterion derived on the basis of a cancer risk level of 1CT6, individuals consuming up to 10 times
the assumed fish intake rate would not exceed a 1CT5 risk level. Similarly, individuals
consuming up to 100 times the assumed rate would not exceed a 10~4 risk level. Thus, for a
criterion based on EPA's default fish intake rate (17.5 gm/day) and a risk level of 10~6, those
consuming a pound per day (i.e., 454 grams/day) would potentially experience between a 10~5
and a 10~4 risk level (closer to a 10"5 risk level). (Note: Fish consumers of up to 1,750 gm/day
would not exceed the 10~4 risk level.) If a criterion were based on high-end intake rates and the
relative risk of 10~6, then an average fish consumer would be protected at a cancer risk level of
approximately 10~8. The point is that the risks for different population groups are not the same.
2.5 MICROBIOLOGICAL AMBIENT WATER QUALITY CRITERIA
Guidance for deriving microbiological AWQC is not a part of this Methodology. In
1986, EPA published Ambient Water Quality Criteria for Bacteria - 1986 (USEPA, 1986a),
which updated and revised bacteriological criteria previously published in 1976 in Quality
Criteria for Water (USEPA, 1976). The inclusion of guidance for deriving microbiological
AWQC was considered in the 1992 national workshop that initiated the effort to revise the 1980
Methodology and was recommended by the SAB in 1993. Since that time, however, efforts
separate from these Methodology revisions have addressed microbiological AWQC concerns.
The purpose of this section is to briefly describe EPA's current recommendations and activities.
EPA's Ambient Water Quality Criteria for Bacteria - 1986 recommends the use of
Escherichia coli and enterococci rather than fecal coliforms (USEPA, 1986a). EPA's criteria
recommendations are:
• Fresh water: E. coli not to exceed 126/100 ml or enterococci not to exceed 33/100 ml;
and
• Marine water: enterococci not to exceed 35/100 ml.
These criteria should be calculated as the geometric mean based on five equally spaced samples
taken over a 30-day period.
In addition, EPA recommends that States adopt a single sample maximum, based on the
expected frequency of use. No sample taken should exceed this value. EPA specifies
appropriate single sample maximum values in the 1986 criteria document.
Current Activities and Plans for Future Work
2-7
-------
EPA has identified development of microbial water quality criteria as part of its strategy
to control waterborne microbial disease, by controlling pathogens in waterbodies and by
protecting designated uses, such as recreation and public water supplies. The program fosters an
integrated approach to protect both ground-water and surface water sources. EPA plans to
conduct additional monitoring for Cryptosporidium parvum and E. coli, and determine action
plans in accordance with the results of this monitoring.
EPA recommends no change at this time in the stringency of its bacterial criteria for
recreational waters; existing criteria and methodologies from 1986 will still apply. The
recommended methods for E. coli and enterococci have been improved. As outlined in the
Action Plan for Beaches and Recreational Waters (Beach Action Plan, see below), the Agency
plans to conduct national studies on improving indicators together with epidemiology studies for
new criteria development (USEPA, 1999b). The Agency is also planning to establish improved
temporal and spatial monitoring protocols.
In the Beach Action Plan, EPA identifies a multi-year strategy for monitoring
recreational water quality and communicating public health risks associated with potentially
pathogen-contaminated recreational rivers, lakes, and ocean beaches. It articulates the Agency's
rationale and goals in addressing specific problems and integrates all associated program, policy,
and research needs and directions. The Beach Action Plan also provides information on timing,
products and lead organization for each activity. These include activities and products in the
areas of program development, risk communication, water quality indicator research, modeling
and monitoring research, and exposure and health effects research.
Recently, EPA approved new 24-hour E. coli and enterococcus tests for recreational
waters that may be used as an alternative to the 48-hour test (USEPA, 1997). EPA anticipates
proposing these methods for inclusion in the 40 CRF 136 in the Fall of 2000. EPA has also
published a video with accompanying manual on the original and newer methods for enterococci
and E. coli (USEPA, 2000).
As part of the Beach Action Plan, EPA made the following recommendations for further
Agency study:
• Future criteria development should consider the risk of diseases other than
gastroenteritis. EPA intends to consider and evaluate such water-related exposure routes
as inhalation and dermal absorption when addressing microbial health effects. The
nature and significance of other than the classical waterborne pathogens are to some
degree tied to the particular type of waste sources.
• A new set of indicator organisms may need to be developed for tropical water if it is
proven that the current fecal indicators can maintain viable cell populations in the soil
and water for significant periods of time in uniform tropical conditions. Some potential
alternative indicators to be fully explored are coliphage, other bacteriophage, and
Clostridium perfringens.
2-8
-------
• Because animal sources of pathogens of concern for human infection such as Giardia
lamblia, Cryptosporidiumparvum, and Escherichia coli 0157:H7 may be waterborne or
washed into water and thus become a potential source for infection, they should not be
ignored in risk assessment. A likely approach would be phylogenetic differentiation; that
is, indicators that are specific to, or can discriminate among, animal sources.
EPA intends to develop additional data on secondary infection routes and infection rates
from prospective epidemiology studies and outbreaks from various types of exposure
(e.g., shellfish consumption, drinking water, recreational exposure).
• EPA needs to improve sampling strategies for recreational water monitoring including
consideration of rainfall and pollution events to trigger sampling.
2.6 RISK CHARACTERIZATION CONSIDERATIONS
On March 21, 1995, EPA's Administrator issued the EPA Risk Characterization Policy
and Guidance (USEPA, 1995). This policy and guidance is intended to ensure that
characterization information from each stage of a risk assessment is used in forming conclusions
about risk and that this information is communicated from risk assessors to risk managers, and
from EPA to the public. The policy also provides the basis for greater clarity, transparency,
reasonableness, and consistency in risk assessments across EPA programs. The fundamental
principles which form the basis for a risk characterization are as follows:
• Risk assessments should be transparent, in that the conclusions drawn from the science
are identified separately from policy judgments, and the use of default values or methods
and the use of assumptions in the risk assessment are clearly articulated.
Risk characterizations should include a summary of the key issues and conclusions of
each of the other components of the risk assessments, as well as describe the likelihood
of harm. The summary should include a description of the overall strengths and
limitations (including uncertainties) of the assessment and conclusions.
• Risk characterizations should be consistent in general format, but recognize the unique
characteristics of each specific situation.
Risk characterizations should include, at least in a qualitative sense, a discussion of how
a specific risk and its context compares with similar risks. This may be accomplished by
comparisons with other pollutants or situations on which the Agency has decided to act,
or other situations with which the public may be familiar. The discussion should
highlight the limitations of such comparisons.
• Risk characterization is a key component of risk communication, which is an interactive
process involving exchange of information and expert opinion among individuals,
groups, and institutions.
Additional guiding principles include:
2-9
-------
• The risk characterization integrates the information from the hazard identification, dose-
response, and exposure assessments, using a combination of qualitative information,
quantitative information, and information regarding uncertainties.
The risk characterization includes a discussion of uncertainty and variability in the risk
assessment.
• Well-balanced risk characterizations present conclusions and information regarding the
strengths and limitations of the assessment for other risk assessors, EPA decision-makers,
and the public.
In developing the methodology presented here, EPA has closely followed the risk
characterization guiding principles listed above. As States and Tribes adopt criteria using the
2000 Human Health Methodology, they are strongly encouraged to follow EPA's risk
characterization guidance. There are a number of areas within the Methodology and criteria
development process where risk characterization principles apply:
• Integration of cancer and noncancer assessments with exposure assessments, including
bioaccumulation potential determinations, in essence, weighing the strengths and
weaknesses of the risk assessment as a whole when developing a criterion.
Selecting a fish consumption rate, either locally derived or the national default value,
within the context of a target population (e.g., sensitive subpopulations) as compared to
the general population.
• Presenting cancer and/or noncancer risk assessment options.
Describing the uncertainty and variability in the hazard identification, the dose-response,
and the exposure assessment.
2.7 DISCUSSION OF UNCERTAINTY
2.7.1 Observed Range of Toxicity Versus Range of Environmental Exposure
When characterizing a risk assessment, an important distinction to make is between the
observed range of adverse effects (from an epidemiology or animal study) and the
environmentally observed range of exposure (or anticipated human exposure) to the
contaminant. In many cases, EPA intends to apply default factors to account for uncertainties or
incomplete knowledge in developing RfDs or cancer risk assessments using nonlinear low-dose
extrapolation to provide a margin of protection. In reality, the actual effect level and the
environmental exposure levels may be separated by several orders of magnitude. The difference
between the dose causing some observed response and the anticipated human exposure should be
described by risk assessors and managers, especially when comparing criteria to environmental
levels of a contaminant.
2.7.2 Continuum of Preferred Data/Use of Defaults
2-10
-------
In both toxicological and exposure assessments, EPA has defined a continuum of
preferred data for toxicological assessments ranging from a highest preference for chronic
human data (e.g., studies that examine a long-term exposure of humans to a chemical, usually
from occupational and/or residential exposure) and actual field data for many of the exposure
parameter values (e.g., locally derived fish consumption rates, waterbody-specific
bioaccumulation rates), to default values which are at the lower end of the preference continuum.
EPA has supplied default values for all of the risk assessment parameters in the 2000 Human
Health Methodology; however, it is important to note that when default values are used, the
uncertainty in the final risk assessment may be higher, and the final resulting criterion may not
be as applicable to local conditions, than is a risk assessment derived from human/field data.
Using defaults assumes generalized conditions and may not capture the actual variability in the
population (e.g., sensitive subpopulations/high-end consumers). If defaults are chosen as the
basis for criteria, these inherent uncertainties should be communicated to the risk manager and
the public. While this continuum is an expression of preference on the part of EPA, it does not
imply in any way that any of the choices are unacceptable or scientifically indefensible.
2.7.3 Significant Figures
The number of significant figures in a numeric value is the number of certain digits plus
one estimated digit. Digits should not be confused with decimal places. For example, 15.1,
0.0151, and 0.0150 all have 3 significant figures. Decimal places may have been used to
maintain the correct number of significant figures, but in themselves they do not indicate
significant figures (Blinker, 1984). Since the number of significant figures must include only
one estimated digit, the sources of input parameters (e.g., fish consumption and water
consumption rates) should be checked to determine the number of significant figures associated
with data they provide. However, the original measured values may not be available to
determine the number of significant figures in the input parameters. In these situations, EPA
recommends utilizing the data as presented.
When developing criteria, EPA recommends rounding the number of significant figures
at the end of the criterion calculation to the same number of significant figures in the least
precise parameter. This is a generally accepted practice which can be found described in greater
detail in APHA (1992) and Brinker (1984). The general rule is that for multiplication or
division, the resulting value should not possess any more significant figures than is associated
with the factor in the calculation with the least precision. When numbers are added or
subtracted, the number that has the fewest decimal places, not necessarily the fewest significant
figures, puts the limit on the number of places that justifiably may be carried in the sum or
difference. Rounding off a number is the process of dropping one or more digits so that the
value contains only those digits that are significant or necessary in subsequent computations
(Brinker, 1984). The following rounding procedures are recommended: (1) if the digit 6, 7, 8, or
9 is dropped, increase the preceding digit by one unit; (2) if the digit 0, 1, 2, 3, or 4 is dropped,
do not alter the preceding digit; and (3) if the digit 5 is dropped, round off the preceding digit to
the nearest even number (e.g., 2.25 becomes 2.2 and 2.35 becomes 2.4) (APHA, 1992; Brinker,
1984).
2-11
-------
EPA recommends that calculations of water quality criteria be performed without
rounding of intermediate step values. The resulting criterion may be rounded to a manageable
number of decimal places. However, in no case should the number of digits presented exceed
the number of significant figures implied in the data and calculations performed on them. The
term "intermediate step values" refers to values of the parameters in Equations 1-1 through 1-3.
The final step is considered the resulting AWQC. Although AWQC are, in turn, used for
purposes of establishing water quality-based effluent limits (WQBELs) in National Pollutant
Discharge Elimination System (NPDES) permits, calculating total maximum daily loads
(TMDLs), and applicable or relevant and appropriate requirements (ARARs) for Superfund, they
are considered the final step of this Methodology and, for the purpose of this discussion, where
the rounding should occur.
The determination of appropriate significant figures inevitably involves some judgment
given that some of the equation parameters are adopted default exposure values. Specifically,
the default drinking water intake rate of 2 L/day is a value adopted to represent a majority of the
population over the course of a lifetime. Although supported by drinking water consumption
survey data, this value was adopted as a policy decision and, as such, does not have to be
considered in determining the parameter with the least precision. That is, the resulting AWQC
need not always be reduced to one significant digit. Similarly, the 70-kg adult body weight has
been adopted Agency-wide and represents a default policy decision.
The following example with a simplified AWQC equation illustrates the rule described
above. The example is for hexachlorobutadiene (HCBD), which EPA used to demonstrate the
1998 draft Methodology revisions (USEPA, 1998b). The parameters that were calculated (i.e.,
not policy adopted values) include values with significant figures of two (the POD and RSC),
three (the UF), and four (the FI and B AF). Based on the 2000 Human Health Methodology, the
final criterion should be rounded to two significant figures. The bold numbers in parentheses
indicate the number of significant figures and those with asterisks also indicate Agency adopted
policy values.
AWQC .. • RSC • (Elation 2-,)
Example [Refer to draft HCBD document for details on the POD/UF, RSC and BAF data (EPA
822-R-98-004). Also note that the fish intake rate in this example is the revised value.]:
AWQC = [ °-°54(2> - 1.2 x lO-4(2)] x(
( 300(3) ') (
70(2*)
2(1*) + (0.01750(4) x 3,180(4)),
2-12
-------
AWQC = 7.3 x 1Q-5 mg/L (0.073 |ig/L, rounded from 7.285 x 10'2 |ig/L)
* represents Agency adopted policy value
A number of the values used in the equation may result in intermediate step values that
have more than four figures past the decimal place and may be carried throughout the
calculation. However, carrying more than four figures past the decimal place (equivalent to the
most precise parameter) is unnecessary as it has no effect on the resulting criterion value.
2.8 OTHER CONSIDERATIONS
2.8.1 Minimum Data Considerations
For many of the preceding technical areas, considerations have been presented for data
quality in developing toxicological and exposure assessments. For greater detail and discussion
of minimum data recommendations, the reader is referred to the specific sections in the
Methodology on cancer and noncancer risk assessments (and especially to the referenced EPA
risk assessment guidelines documents), exposure assessment, and bioaccumulation assessment,
in addition to the TSD volumes for each.
2.8.2 Site-Specific Criterion Calculation
The 2000 Human Health Methodology allows for site-specific modifications by States
and Tribes to reflect local environmental conditions and human exposure patterns. "Local" may
refer to any appropriate geographic area where common aquatic environmental or exposure
patterns exist. Thus "local" may signify Statewide, regional, a river reach, or an entire river.
Such site-specific criteria may be developed as long as the site-specific data, either
toxicological or exposure-related, is justifiable. For example, when using a site-specific fish
consumption rate, a State should use a value that represents at least the central tendency of the
population surveyed (either sport or subsistence, or both). If a site-specific fish consumption rate
for sport anglers or subsistence anglers is lower than an EPA default value, it may be used in
calculating AWQC. However, to justify such a level (either higher or lower than EPA defaults),
the State should assemble appropriate survey data to arrive at a defensible site-specific fish
consumption rate.
Such data must also be submitted to EPA for its review when approving or disapproving
State or Tribal water quality standards under Section 303(c). The same conditions apply to site-
specific calculations of BAF, percent fish lipid, or the RSC. In the case of deviations from
toxicological values (i.e., IRIS values: verified noncancer and cancer assessments), EPA strongly
recommends that the data upon which the deviation is based be presented to and approved by the
Agency before a criterion is developed.
Additional guidance on site-specific modifications to the 2000 Human Health
Methodology is provided in each of the three TSD volumes.
2.8.3 Organoleptic Criteria
2-13
-------
Organoleptic criteria define concentrations of chemicals or materials which impart
undesirable taste and/or odor to water. Organoleptic effects, while significant from an aesthetic
standpoint, are not a significant health concern. In developing and utilizing such criteria, two
factors must be appreciated: (1) the limitations of most Organoleptic data; and (2) the human
health significance of Organoleptic properties. In the past, EPA has developed Organoleptic
criteria if Organoleptic data were available for a specific contaminant. The 1980 AWQC
National Guidelines made a clear distinction that Organoleptic criteria and toxicity-based criteria
are derived from completely different endpoints, and that Organoleptic criteria have no
demonstrated relationship to potential adverse human health effects because there is no
toxicological basis. EPA acknowledges that if Organoleptic effects (i.e., objectionable taste and
odor) cause people to reject the water and its designated uses, then the public is effectively
deprived of the natural resource. It is also possible that intense Organoleptic characteristics could
result in depressed fluid intake which, in turn, might lead to an indirect human health effect via
decreased fluid consumption. Although EPA has developed Organoleptic criteria in the past and
may potentially do so in the future, this will not be a significant part of the water quality criteria
program. EPA encourages the development of Organoleptic criteria when States and Tribes
believe they are needed. However, EPA cautions States and Tribes that the quality of
Organoleptic data is often significantly less than that of toxicologic data used in establishing
health-based criteria. Therefore, a comprehensive evaluation of available Organoleptic data
should be made, and the selection of the most appropriate database for the criterion should be
based on sound scientific judgment.
In 1980, EPA provided recommended criteria summary language when both types of data
are available. The following format was used and is repeated here:
For comparison purposes, two approaches were used to derive criterion levels for
. Based on available toxicity data, for the protection of public health the
derived level is . Using available Organoleptic data, for controlling
undesirable taste and odor quality of ambient water the estimated level is .
It should be recognized that Organoleptic data as a basis for establishing a water
quality criteria have no demonstrated relationship to potential adverse human
health effects.
Similarly, the 1980 Methodology recommended that in those instances where a level to
limit toxicity cannot be derived, the following statement should be provided:
Sufficient data are not available for to derive a level which would protect
against the potential toxicity of this compound.
2.8.4 Criteria for Chemical Classes
The 2000 Human Health Methodology also allows for the development of a criterion for
classes of chemicals, as long as a justification is provided through the analysis of mechanistic
data, toxicokinetic data, structure-activity relationship data, and limited acute and chronic
toxicity data. When potency differences between members of a class is great (such as in the case
2-14
-------
of chlorinated dioxins and furans), toxicity equivalency factors (TEFs) may be more
appropriately developed than one class criterion.
A chemical class is defined as any group of chemical compounds which are similar in
chemical structure and biological activity, and which frequently occur together in the
environment usually because they are generated by the same commercial process. In criterion
development, isomers should be regarded as part of a chemical class rather than as a single
compound. A class criterion, therefore, is an estimate of risk/safety which applies to more than
one member of a class. It involves the use of available data on one or more chemicals of a class
to derive criteria for other compounds of the same class in the event that there are insufficient
data available to derive compound-specific criteria. The health-based criterion may apply to the
water concentration of each member of the class, or may apply to the sum of the water
concentrations of the compounds within the class. Because relatively minor structural changes
within the class of compounds can have pronounced effects on their biological activities, reliance
on class criteria should be minimized depending on the data available.
The following guidance should also be followed when considering the development of a
class criterion.
• A detailed review of the chemical and physical properties of the chemicals within the
group should be made. A close relationship within the class with respect to chemical
activity would suggest a similar potential to reach common biological sites within tissues.
Likewise, similar lipid solubilities would suggest the possibility of comparable
absorption and distribution.
• Qualitative and quantitative toxicological data for chemicals within the group should be
examined. Adequate toxicological data on a number of compounds within a group
provides a more reasonable basis for extrapolation to other chemicals of the same class
than minimal data on one chemical or a few chemicals within the group.
Similarities in the nature of the toxicological response to chemicals in the class provides
additional support for the prediction that the response to other members of the class may
be similar. In contrast, where the biological response has been shown to differ markedly
on a qualitative and quantitative basis for chemicals within a class, the extrapolation of a
criterion to other members is not appropriate.
Additional support for the validity of extrapolation of a criterion to other members of a
class could be provided by evidence of similar metabolic and toxicokinetic data for some
members of the class.
Additional guidance is described in the Technical Support Document on Health Risk
Assessment of Chemical Mixtures (USEPA, 1990).
2-15
-------
2.9.5 Criteria for Essential Elements
2-16
-------
Developing criteria for essential elements, particularly metals, must be a balancing act
between toxicity and the requirement for good health. The AWQC must consider essentiality
and cannot be established at levels that would result in deficiency of the element in the human
population. The difference between the recommended daily allowance (RDA) and the daily
doses causing a specified risk level for carcinogens or the RfDs for noncarcinogens defines the
spread of daily doses within which the criterion may be derived. Because errors are inherent in
defining both essential and adverse-effect levels, the criterion is derived from a dose level near
the center of such dose ranges.
The process for developing criteria for essential elements should be similar to that used
for any other chemical with minor modifications. The RfD represents concern for one end of the
exposure spectrum (toxicity), whereas the RDA represents the other end (minimum essentiality).
While the RDA and RfD values might occasionally appear to be similar in magnitude to one
another, it does not imply incompatibility of the two methodological approaches, nor does it
imply inaccuracy or error in either calculation.
2.9 REFERENCES
APHA. American Public Health Association. 1992. Standard Methods: For the Examination
of Water and Wastewater. 18th Edition. Prepared and published jointly by: American
Public Health Association, American Water Works Association, and Water Environment
Federation. Washington, DC.
Brinker, R.C. 1984. Elementary Surveying. 7th Edition. Cliff Robichaud and Robert Greiner,
Eds. Harper and Row Publishers, Inc. New York, NY.
NRC (National Research Council). 1983. Risk Assessment in the Federal Government:
Managing the Process. National Academy Press. Washington, DC.
USEPA (U.S. Environmental Protection Agency). 1976. Quality Criteria for Water. Office of
Water and Hazardous Materials. Washington, DC. July.
USEPA (U.S. Environmental Protection Agency). 1986a. Ambient Water Quality Criteria for
Bacteria - 1986. Office of Water Regulations and Standards. Washington, DC.
EPA/440/5-84/002. January.
USEPA (U.S. Environmental Protection Agency). 1986b. Test Methods for Escherichia coll
and Enterococci in Water by the Membrane Filter Procedure. Office of Research and
Development. Cincinnati, OH. EPA/600/4-85/076.
USEPA (U.S. Environmental Protection Agency). 1990. Technical Support Document on
Health Risk Assessment of Chemical Mixtures. Office of Research and Development.
Washington, DC. EPA/600/8-90/064. August.
USEPA (U.S. Environmental Protection Agency). 1992. Guidelines for exposure assessment.
Federal Register 57:22888-22938.
2-17
-------
USEPA (U.S. Environmental Protection Agency). 1995. Policy for Risk Characterization.
Memorandum of Carol M. Browner, Administrator. March 21, 1995. Washington, DC.
USEPA (U.S. Environmental Protection Agency). 1996a. Draft revisions to guidelines for
carcinogen risk assessment. Federal Register 61:17960.
USEPA (U.S. Environmental Protection Agency). 1996b. Guidelines for reproductive toxicity
risk assessment. Federal Register 61:6274-56322.
USEPA (U.S. Environmental Protection Agency). 1997. Method 1600: Membrane Filter Test
Method for Enter ococci in Water. Office of Water. Washington, DC. EPA/821/R-
97/004. May.
USEPA (U.S. Environmental Protection Agency). 1998a. Draft Water Quality Criteria
Methodology: Human Health. Office of Water. Washington, DC. EPA-822-Z-98-001.
(FederalRegister 63:43756)
USEPA (U.S. Environmental Protection Agency). 1998b. Ambient Water Quality Criteria for
the Protection of Human Health. Hexachlorobutadiene (HCBD). Draft. Office of
Water. Washington, DC. EPA 882-R-98-004. July.
USEPA (U.S. Environmental Protection Agency). 1999a. 1999 Guidelines for Carcinogen Risk
Assessment. Review Draft. Office of Research and Development. Washington, DC.
NCEA-F-0644.
USEPA (U.S. Environmental Protection Agency). 1999b. Action Plan for Beaches and
Recreational Waters. Reducing Exposures to Waterborne Pathogens. Office of
Research and Development and Office of Water. Washington, DC. EPA-600-R-98-079.
March.
USEPA (U.S. Environmental Protection Agency). 2000. Improved Enumeration Methods for
the Recreational Water Quality Indicators: Enterococci andEscherichia coli. Office of
Water, Office of Science and Technology. Washington, DC. EPA-821-R-97-004.
March.
2-18
-------
3. RISK ASSESSMENT
This section describes the methods used to estimate ambient water quality criteria
(AWQC) for the protection of human health for carcinogenic chemicals (Section 3.1) and for
noncarcinogenic chemicals (Section 3.2).
3.1 CANCER EFFECTS
3.1.1 Background on EPA Cancer Risk Assessment Guidelines
The current EPA Guidelines for Carcinogen Risk Assessment were published in 1986
(USEPA, 1986a, hereafter the "1986 cancer guidelines"). The 1986 cancer guidelines categorize
chemicals into alpha-numerical Groups: A, known human carcinogen (sufficient evidence from
epidemiological studies or other human studies); B, probable human carcinogen (sufficient
evidence in animals and limited or inadequate evidence in humans); C, possible human
carcinogen (limited evidence of carcinogenicity in animals in the absence of human data); D, not
classifiable (inadequate or no animal evidence of carcinogenicity); and E, evidence of
noncarcinogenicity for humans (no evidence of carcinogenicity in at least two adequate animal
tests in different species or in both adequate epidemiological and animal studies). Within Group
B there are two subgroups, Groups Bl and B2. Group Bl is reserved for agents for which there
is limited evidence of carcinogenicity from epidemiological studies. Group B2 is generally for
agents for which there is sufficient evidence from animal studies and for which there is
inadequate evidence or no data from epidemiological studies (USEPA, 1986). The system was
similar to that used by the International Agency for Research on Cancer (IARC).
The 1986 cancer guidelines include guidance on what constitutes sufficient, limited, or
inadequate evidence. In epidemiological studies, sufficient evidence indicates a causal
relationship between the agent and human cancer; limited evidence indicates that a causal
relationship is credible, but that alternative explanations, such as chance, bias, or confounding,
could not adequately be excluded; inadequate evidence indicates either lack of pertinent data, or
a causal interpretation is not credible. In general, although a single study may be indicative of a
cause-effect relationship, confidence in inferring a causal association is increased when several
independent studies are concordant in showing the association. In animal studies, sufficient
evidence includes an increased incidence of malignant tumors or combined malignant and
benign tumors:
In multiple species or strains;
• In multiple experiments (e.g., with different routes of administration or using different
dose levels);
• To an unusual degree in a single experiment with regard to high incidence, unusual site
or type of tumor, or early age at onset;
Additional data on dose-response, short-term tests, or structural activity relationships.
3-1
-------
In the 1986 cancer guidelines, hazard identification and the weight-of-evidence process
focus on tumor findings. The weight-of-evidence approach for making judgments about cancer
hazard analyzes human and animal tumor data separately, then combines them to make the
overall conclusion about potential human carcinogenicity. The next step of the hazard analysis
is an evaluation of supporting evidence (e.g., mutagenicity, cell transformation) to determine
whether the overall weight-of-evidence conclusion should be modified.
For cancer risk quantification, the 1986 cancer guidelines recommend the use of
linearized multistage model (LMS) as the only default approach. The 1986 cancer guidelines
also mention that a low-dose extrapolation model other than the LMS might be considered more
appropriate based on biological grounds. However, no guidance is given in choosing other
approaches. The 1986 cancer guidelines recommended the use of body weight raised to the 2/3
power (BW2/3) as a dose scaling factor between species.
3.1.2 EPA's Proposed Guidelines for Carcinogen Risk Assessment and the
Subsequent July, 1999 Draft Revised Cancer Guidelines
In 1996, EPA published Proposed Guidelines for Carcinogen Risk Assessment (USEPA,
1996a, hereafter the "1996 proposed cancer guidelines"). After the publication of the 1996
proposed cancer guidelines and a February, 1997 and January, 1999 Science Advisory Board
(SAB) review, a revision was made in July, 1999 Guidelines for Carcinogen Risk Assessment -
Review Draft (hereafter the "1999 draft revised cancer guidelines"; USEPA, 1999a), and an SAB
meeting was convened to review this revised document. When final guidelines are published,
they will replace the 1986 cancer guidelines. These revisions are designed to ensure that the
Agency's cancer risk assessment methods reflect the most current scientific information and
advances in risk assessment methodology.
In the meanwhile, the 1986 guidelines are used and extended with principles discussed in
the 1999 draft revised cancer guidelines. These principles arise from scientific discoveries
concerning cancer made in the last 15 years and from EPA policy of recent years supporting full
characterization of hazard and risk both for the general population and potentially sensitive
groups such as children. These principles are incorporated in recent and ongoing assessments
such as the reassessment of dioxin, consistent with the 1986 guidelines. Until final guidelines
are published, information is presented to describe risk under both the 1986 guidelines and 1999
draft revisions.
The 1999 draft revised cancer guidelines call for the full use of all relevant information to
convey the circumstances or conditions under which a particular hazard is expressed (e.g., route,
duration, pattern, or magnitude of exposure). They emphasize understanding the mode of action
(MOA) whereby the agent induces tumors. The MOA underlies the hazard assessment and
provides the rationale for dose-response assessments.
The key principles in the 1999 draft revised cancer guidelines include:
5-2
-------
a) Hazard assessment is based on the analysis of all biological information rather
than just tumor findings.
b) An agent's MOA in causing tumors is emphasized to reduce the uncertainty in
describing the likelihood of harm and in determining the dose-response
approach(es).
c) The 1999 draft revised cancer guidelines emphasize the conditions under which
the hazard may be expressed (e.g., route, pattern, duration and magnitude of
exposure). Further, the guidelines call for a hazard characterization to integrate
the data analysis of all relevant studies into a weight-of-evidence conclusion of
hazard and to develop a working conclusion regarding the agent's mode of action
in leading to tumor development.
d) A weight-of-evidence narrative with accompanying descriptors (listed in Section
3.1.3.1 below) would replace the current alphanumeric classification system. The
narrative summarizes the key evidence for carcinogenicity, describes the agent's
MOA, characterizes the conditions of hazard expression, including route of
exposure, describes any disproportionate effects on subgroups of the human
population (e.g., children), and recommends appropriate dose-response
approach(es). Significant strengths, weaknesses, and uncertainties of contributing
evidence are also highlighted.
e) Biologically based extrapolation models are the preferred approach for
quantifying risk. These models integrate data and conclusions about events in the
carcinogenic process throughout the dose-response range from high to low doses.
It is anticipated, however, that the necessary data for the parameters used in such
models will not be available for most chemicals. The 1999 draft revised cancer
guidelines allow for alternative quantitative methods, including several default
approaches.
f) Dose-response assessment is a two-step process. In the first step, response data
are modeled in the observable range of data and a determination is made of the
point of departure (POD) from the observed range to extrapolate to low doses.
The second step is extrapolation from the POD to estimate dose-response at lower
doses. In addition to modeling tumor data, the 1999 draft revised cancer
guidelines call for the use and modeling of other kinds of responses if they are
considered to be more informed measures of carcinogenic risk. Nominally, these
responses reflect key events in the carcinogenic process integral to the MOA of
the agent.
g) Three default approaches are provided-linear, nonlinear, or both when adequate
data are unavailable to generate a biologically based model. As the first step for
all approaches, curve fitting in the observed range is used to determine a POD. A
standard POD is the effective dose corresponding to the lower 95 percent limit on
-------
a dose associated with 10 percent extra risk (LED10).3 Linear: The linear default
is a straight line extrapolation from the response at LED10 to the origin (zero dose,
zero extra risk). Nonlinear: The nonlinear default begins with the identified POD
and provides a margin of exposure (MOE) analysis rather than estimating the
probability of effects at low doses. The MOE analysis is used to determine the
appropriate margin between the POD and the exposure level of interest, in this
Methodology, the AWQC. The key objective of the MOE analysis is to describe
for the risk manager how rapidly responses may decline with dose. Other factors
are also considered in the MOE analysis (i.e., nature of the response, slope of the
dose-response curve, human sensitivity compared with experimental animals,
nature and extent of human variability in sensitivity and human exposure).
Linear and nonlinear: Section 3.1.3.4E describes the situations when both linear
and nonlinear defaults are used.
h) The approach used to calculate an oral human equivalent dose when assessments
are based on animal bioassays has been refined and includes a change in the
default assumption for interspecies dose scaling. The 1999 draft revised cancer
guidelines use body weight raised to the 3/4 power.
EPA health risk assessment practices for both cancer and noncancer endpoints are
beginning to come together with recent proposals to emphasize MOA understanding in risk
assessment and to model response data in the observable range to derive PODs for data sets and
benchmark doses (BMDs) for individual studies. The modeling of observed response data to
identify PODs in a standard way will help to harmonize cancer and noncancer dose-response
approaches and permit comparisons of cancer and noncancer risk estimates.
3.1.3 Methodology for Deriving AWQC4 by the 1999 Draft Revised Cancer Guidelines
Following the publication of the Draft Water Quality Criteria Methodology: Human
Health (USEPA, 1998a) and the accompanying TSD (USEPA, 1998b), EPA received comments
from the public. EPA also held an external peer review of the draft Methodology. Both the peer
reviewers and the public recommended that EPA incorporate the new approaches into the
AWQC Methodology.
Until new guidelines are published, the 1986 cancer guidelines will be used along with
principles of the 1999 draft revised cancer guidelines. The 1986 guidelines are the basis for IRIS
risk numbers which were used to derive the current AWQC. Each new assessment applying the
principles of the 1999 draft revised cancer guidelines will be subject to peer review before being
used as the basis of AWQC.
3 Use of the LED10 as the point of departure is recommended with this Methodology, as it is with the 1999 draft revised
cancer guidelines.
4 Additional information regarding the revised method for assessing carcinogens may be found in the Methodology for
Deriving Ambient Water Quality Criteria for the Protection of Human Health (2000). Technical Support Document, Volume 1:
Risk Assessment (USEPA, 2000).
3-4
-------
The remainder of Section 3 illustrates the methodology for deriving numerical AWQC
for carcinogens applying the 1999 draft revised cancer guidelines (USEPA, 1999a). This
discussion of the revised methodology for carcinogens focuses primarily on the quantitative
aspects of deriving numerical AWQC values. It is important to note that the cancer risk
assessment process outlined in the 1999 draft revised cancer guidelines is not limited to the
quantitative aspects. A numerical AWQC value derived for a carcinogen is to be based on
appropriate hazard characterization and accompanied by risk characterization information.
This section contains a discussion of the weight-of-evidence narrative, that describes all
information relevant to a cancer risk evaluation, followed by a discussion of the quantitative
aspects of deriving numerical AWQC values for carcinogens. It is assumed that data from an
appropriately conducted animal bioassay or human epidemiological study provide the underlying
basis for deriving the AWQC value. The discussion focuses on the following: (1) the weight-of-
evidence narrative; (2) general considerations and framework for analysis of the MO A; (3) dose
estimation; (4) characterizing dose-response relationships in the range of observation and at low,
environmentally relevant doses; (5) calculating the AWQC value; (6) risk characterization; and
(7) use of Toxicity Equivalent Factors (TEF) and Relative Potency Estimates. The first three
topics encompass the quantitative aspects of deriving AWQC for carcinogens.
3.1.3.1 Weight-of-Evidence Narrative5
The 1999 draft revised cancer guidelines include a weight-of-evidence narrative that is
based on an overall judgment of biological and chemical/physical considerations. Hazard
assessment information accompanying an AWQC value for a carcinogen in the form of a weight-
of-evidence narrative is described in the footnote. Of particular importance is that the weight-of-
evidence narrative explicitly provides adequate support based on human studies, animal
bioassays, and other key evidence for the conclusion whether the substance is or is likely to be
carcinogenic to humans from exposures through drinking water and/or fish ingestion. The
Agency emphasizes the importance of providing an explicit discussion of the MOA for the
substance in the weight-of-evidence narrative if data are available, including a discussion that
relates the MOA to the quantitative procedures used in the derivation of the AWQC.
3.1.3.2 Mode of Action - General Considerations and Framework for Analysis
5 The weight-of-evidence narrative is intended for the risk manager, and thus explains in nontechnical language the key
data and conclusions, as well as the conditions for hazard expression. Conclusions about potential human carcinogenicity are
presented by route of exposure. Contained within this narrative are simple likelihood descriptors that essentially distinguish
whether there is enough evidence to make a projection about human hazard (i.e., Carcinogenic to humans; Likely to be
carcinogenic to humans; Suggestive evidence of carcinogenicity but not sufficient to assess human carcinogenic potential; Data
are inadequate for an assessment of human carcinogenic potential; and Not likely to be carcinogenic to humans). Because one
encounters a variety of data sets on agents, these descriptors are not meant to stand alone; rather, the context of the weight-of-
evidence narrative is intended to provide a transparent explanation of the biological evidence and how the conclusions were
derived. Moreover, these descriptors should not be viewed as classification categories (like the alphameric system), which often
obscure key scientific differences among chemicals. The new weight-of-evidence narrative also presents conclusions about how
the agent induces tumors and the relevance of the mode of action to humans, and recommends a dose-response approach based
on the MOA understanding (USEPA, 1996a, 1999a).
3-5
-------
An MOA is composed of key events and processes starting with the interaction of an
agent with a cell, through operational and anatomical changes, resulting in cancer formation.
"Mode" of action is contrasted with "mechanism" of action, which implies a more detailed,
molecular description of events than is meant by MOA.
Mode of action analysis is based on physical, chemical, and biological information that
helps to explain key events6 in an agent's influence on development of tumors. Inputs to MOA
analysis include tumor data in humans, animals, and among structural analogues as well as the
other key data.
There are many examples of possible modes of carcinogenic action, such as
mutagenicity, mitogenesis, inhibition of cell death, cytotoxicity with reparative cell proliferation,
and immune suppression. All pertinent studies are reviewed in analyzing an MOA, and an
overall weighing of evidence is performed, laying out the strengths, weaknesses, and
uncertainties of the case as well as potential alternative positions and rationales. Identifying data
gaps and research needs is also part of the assessment.
Mode of action conclusions are used to address the question of human relevance of
animal tumor responses, to address differences in anticipated response among humans such as
between children and adults or men and women, and as the basis of decisions about the
anticipated shape of the dose-response relationship.
In reaching conclusions, the question of "general acceptance" of an MOA will be tested
as part of the independent peer review that EPA obtains for its assessment and conclusions.
Framework for Evaluating a Postulated Carcinogenic Mode(s) of Action
The framework is intended to be an analytic tool for judging whether available data
support a mode of carcinogenic action postulated for an agent and includes nine elements:
1. Summary description of postulated MOA
2. Identification of key events
3. Strength, consistency, specificity of association
4. Dose-response relationship
5. Temporal relationship
6. Biological plausibility and coherence
7. Other modes of action
8. Conclusion
9. Human relevance, including subpopulations
3.1.3.3 Dose Estimation
6A "key event" is an empirically observable, precursor step that is itself a necessary element of the mode of action, or
is a marker for such an element.
5-6
-------
A. Determining the Human Equivalent Dose by the Oral Route
An important objective in the dose-response assessment is to use a measure of internal or
delivered dose at the target site where possible. This is particularly important in those cases
where the carcinogenic response information is being extrapolated to humans from animal
studies. Generally, by the oral exposure route, the measure of a dose provided in the underlying
human studies or animal bioassays is the applied dose, typically given in terms of unit mass per
unit body weight per unit time, (e.g., mg/kg-day). When animal bioassay data are used, it is
necessary to make adjustments to the applied dose values to account for differences in
toxicokinetics between animals and humans that affect the relationship between applied dose and
delivered dose at the target organ.
In the estimation of a human equivalent dose, the 1999 draft revised cancer guidelines
recommend that when adequate data are available, the doses used in animal studies can be
adjusted to equivalent human doses using toxicokinetic information on the particular agent.
However, in most cases, there are insufficient data available to compare dose between species.
In these cases, the estimate of a human equivalent dose is based on science policy default
assumptions. To derive an equivalent human oral dose from animal data, the default procedure
in the 1999 draft revised cancer guidelines is to scale daily applied oral doses experienced for a
lifetime in proportion to body weight raised to the 3/4 power (BW3/4). The adjustment factor is
used because metabolic rates, as well as most rates of physiological processes that determine the
disposition of dose, scale this way. Thus, the rationale for this factor rests on the empirical
observation that rates of physiological processes consistently tend to maintain proportionality
with body weight raised to 3/4 power (USEPA, 1992a, 1999a).
The use of BW3/4 is a departure from the scaling factor of BW2/3 that was based on
surface area adjustment and was included in the 1980 AWQC National Guidelines as well as the
1986 cancer guidelines.
B. Dose-Response Analysis
If data on the agent are sufficient to support the parameters of a biologically based model
and the purpose of the assessment is such as to justify investing resources supporting its use, this
is the preferred approach for both the observed tumor and related response data and for
extrapolation below the range of observed data in either animal or human studies.
3.1.3.4 Characterizing Dose-Response Relationships in the Range of Observation and at
Low Environmentally Relevant Doses
The first quantitative component in the derivation of AWQC for carcinogens is the dose-
response assessment in the range of observation. For most agents, in the absence of adequate
data to generate a biologically based model, dose-response relationships in the observed range
can be addressed through curve-fitting procedures for response data. It should be noted that the
1999 draft revised cancer guidelines call for modeling of not only tumor data in the observable
range, but also other responses thought to be important events preceding tumor development
(e.g., DNA adducts, cellular proliferation, receptor binding, hormonal changes). The modeling of
3-7
-------
these data is intended to better inform the dose-response assessment by providing insights into
the relationships of exposure (or dose) below the observable range for tumor response. These
non-tumor response data can only play a role in the dose-response assessment if the agent's
carcinogenic mode of action is reasonably understood, as well as the role of that precursor event.
The 1999 draft revised cancer guidelines recommend calculating the lower 95 percent
confidence limit on a dose associated with an estimated 10 percent increased tumor or relevant
non-tumor response (LED10) for quantitative modeling of dose-response relationships in the
observed range. The estimate of the LED10 is used as the POD for low-dose extrapolations
discussed below. This standard point of departure (LED10) is adopted as a matter of science
policy to remain as consistent and comparable from case to case as possible. It is also a
convenient comparison point for noncancer endpoints. The rationale supporting use of the
LED10 is that a 10 percent response is at or just below the limit of sensitivity for discerning a
statistically significant tumor response in most long-term rodent studies and is within the
observed range for other toxicity studies. Use of lower limit takes experimental variability and
sample size into account. The ED10 (central estimate) is also presented as a reference for
comparison uses, especially for use in relative hazard/potency ranking among agents for priority
setting.
For some data sets, a choice of the POD other than the LED10 may be appropriate. The
objective is to determine the lowest reliable part of the dose-response curve for the beginning of
the second step of the dose-response assessment—determine the extrapolation range. Therefore,
if the observed response is below the LED10, then a lower point may be a better choice (e.g.,
LED5). Human studies more often support a lower POD than animal studies because of greater
sample size.
The POD may be a NOAEL when a margin of exposure analysis is the nonlinear dose-
response approach. The kinds of data available and the circumstances of the assessment both
contribute to deciding to use a NOAEL or LOAEL which is not as rigorous or as ideal as curve
fitting, but can be appropriate. If several data sets for key events and tumor response are
available for an agent, and they are a mixture of continuous and incidence data, the most
practicable way to assess them together is often through a NOAEL/LOAEL approach.
When an LED value estimated from animal data is used as the POD, it is adjusted to the
human equivalent dose using an interspecies dose adjustment or a toxicokinetic analysis as
described in Section 3.1.3.3.
Analysis of human studies in the observed range is designed on a case-by-case basis
depending on the type of study and how dose and response are measured in the study.
A. Extrapolation to Low, Environmentally Relevant Doses
In most cases, the derivation of an AWQC will require an evaluation of carcinogenic risk
at environmental exposure levels substantially lower than those used in the underlying study.
Various approaches are used to extrapolate risk outside the range of observed experimental data.
In the 1999 draft revised cancer guidelines, the choice of extrapolation method is largely
-------
dependent on the mode of action. It should be noted that the term "mode of action" (MOA) is
deliberately chosen in the 1999 draft revised cancer guidelines in lieu of the term "mechanism"
to indicate using knowledge that is sufficient to draw a reasonable working conclusion without
having to know the processes in detail as the term mechanism might imply. The 1999 draft
revised cancer guidelines favor the choice of a biologically based model, if the parameters of
such models can be calculated from data sources independent of tumor data. It is anticipated that
the necessary data for such parameters will not be available for most chemicals. Thus, the 1999
draft revised cancer guidelines allow for several default extrapolation approaches (low-dose
linear, nonlinear, or both).
B. Biologically Based Modeling Approaches
If a biologically based approach has been used to characterize the dose-response
relationships in the observed range, and the confidence in the model is high, it may be used to
extrapolate the dose-response relationship to environmentally relevant doses. For the purposes
of deriving AWQC, the environmentally relevant dose would be the risk-specific dose (RSD)
associated with incremental lifetime cancer risks in the 10"6 to 10"4 range for carcinogens for
which a linear extrapolation approach is applied.7 The use of the RSD and the POD/UF to
compute the AWQC is presented in Section 3.1.3.5, below. Although biologically-based
approaches are appropriate both for characterizing observed dose-response relationships and
extrapolating to environmentally relevant doses, it is not expected that adequate data will be
available to support the use of such approaches for most substances. In the absence of such data,
the default linear approach, the nonlinear (MOE) approach, or both linear and nonlinear
approaches will be used.
7 For discussion of the cancer risk range, see Section 2.4.
3-9
-------
C. Default Linear Extrapolation Approach
The default linear approach replaces the LMS approach that has served as the default for
EPA cancer risk assessments. Any of the following conclusions leads to selection of a linear
dose-response assessment approach:
• There is an absence of sufficient tumor MOA information.
• The chemical has direct DNA mutagenic reactivity or other indications of DNA
effects that are consistent with linearity.
• Human exposure or body burden is high and near doses associated with key
events in the carcinogenic process (e.g., 2,3,7,8-tetrachlorodibenzo-p-dioxin).
• Mode of action analysis does not support direct DNA effects, but the dose-
response relationship is expected to be linear (e.g., certain receptor-mediated
effects).
The procedures for implementing the default linear approach begin with the estimation of
a POD as described above. The point of departure, LED10, reflects the interspecies conversion to
the human equivalent dose and the other adjustments for less-than-lifetime experimental
duration. In most cases, the extrapolation for estimating response rates at low, environmentally
relevant exposures is accomplished by drawing a straight line between the POD and the origin
(i.e., zero dose, zero extra risk). This is mathematically represented as:
y = mx + b (Equation 3-1)
b = 0
where:
y = Response or incidence
m = Slope of the line (cancer potency factor) = Ay/AX
x = Dose
b = Slope intercept
The slope of the line, "m" (the estimated cancer potency factor at low doses), is
computed as:
0.10
m=
LED10 (Equation 3-2)
The RSD is then calculated for a specific incremental targeted lifetime cancer risk (in the range
oflO-6tolO-4)as:
3-10
-------
RSD = Target Incremental Cancer Risk (Equation 3-3)
m
where:
RSD = Risk-specific dose (mg/kg-day)
Target Incremental
Cancer Risk8 = Value in the range of 10'6tolO'4
m = Cancer potency factor (mg/kg-day)"1
The use of the RSD to compute the AWQC is described in Section 3.1.3.5 below.
D. Default Nonlinear Approach
As discussed in the 1999 draft revised cancer guidelines, any of the following
conclusions leads to a selection of a nonlinear (MOE) approach to dose-response assessment:
• A tumor MOA supporting nonlinearity applies (e.g., some cytotoxic and hormonal agents
such as disrupters of hormonal homeostasis), and the chemical does not demonstrate
mutagenic effects consistent with linearity.
An MOA supporting nonlinearity has been demonstrated, and the chemical has some
indication of mutagenic activity, but it is judged not to play a significant role in tumor
causation.
Thus, a default assumption of nonlinearity is appropriate when there is no evidence for
linearity and sufficient evidence to support an assumption of nonlinearity. The MOA may lead
to a dose-response relationship that is nonlinear, with response falling much more quickly than
linearly with dose, or being most influenced by individual differences in sensitivity.
Alternatively, the MOA may theoretically have a threshold (e.g., the carcinogenicity may be a
secondary effect of toxicity or of an induced physiological change that is itself a threshold
phenomenon).
The nonlinear approach may be used, for instance, in the case of a bladder tumor inducer,
where the chemical is not mutagenic and causes only stone formation in male rat bladders at high
doses. This dynamic leads to tumor formation only at the high doses. Stone and subsequent
tumor formation are not expected to occur at doses lower than those that induce the
physiological changes that lead to stone formation. (More detail on this chemical is provided in
the cancer section of the Risk Assessment TSD; USEPA, 2000). EPA does not generally try to
distinguish between modes of action that might imply a "true threshold" from others with a
8In 1980, the target lifetime cancer risk range was set at 10-7 to 10-5. However, both the expert panel for the AWQC
workshop (USEPA, 1993) and the peer review workshop experts (USEPA,1999c) recommended that EPA change the risk range
to 10-6 to 10-4, to be consistent with SDWA program decisions. See Section 2.4 for more details.
3-11
-------
nonlinear dose-response relationship, because there is usually not sufficient information to
distinguish between those possibilities empirically.
The nonlinear MOE approach in the 1986 proposed cancer guidelines compares an
observed response rate such as the LED10, NOAEL, or LOAEL with actual or nominal
environmental exposures of interest by computing the ratio between the two. In the context of
deriving AWQC, the environmentally relevant exposures are nominal targets rather than actual
exposures.
If the evidence for an agent indicates nonlinearity (e.g., when carcinogenicity is
secondary to another toxicity for which there is a threshold), the MOE analysis for the toxicity is
similar to what is done for a noncancer endpoint, and an RfD or RfC for that toxicity may also be
estimated and considered in the cancer assessment. However, a threshold of carcinogenic
response is not necessarily assumed. It should be noted that for cancer assessment, the MOE
analysis begins from a POD that is adjusted for toxicokinetic differences between species to give
a human equivalent dose.
To support the use of the MOE approach, risk assessment information provides
evaluation of the current understanding of the phenomena that may be occurring as dose
(exposure) decreases substantially below the observed data. This gives information about the
risk reduction that is expected to accompany a lowering of exposure. The various factors that
influence the selection of the UF in an MOE approach are also discussed below.
There are two main steps in the MOE approach. The first step is the selection of a POD.
The POD may be the LED10 for tumor incidence or a precursor, or in some cases, it may also be
appropriate to use a NOAEL or LOAEL value. When animal data are used, the POD is a human
equivalent dose or concentration arrived at by interspecies dose adjustment (as discussed in
Section 3.1.3.3) or toxicokinetic analysis.
The second step in using MOE analysis to establish AWQC is the selection of an
appropriate margin or UF to apply to the POD. This is supported by analyses in the MOE
discussion in the risk assessment. The following issues should be considered when establishing
the overall UF for the derivation of AWQC using the MOE approach (others may be found
appropriate in specific cases):
The nature of the response used for the dose-response assessment, for instance, whether it
is a precursor effect or a tumor response. The latter may support a greater MOE.
The slope of the observed dose-response relationship at the POD and its uncertainties and
implications for risk reduction associated with exposure reduction. (A steeper slope
implies a greater reduction in risk as exposure decreases. This may support a smaller
MOE).
• Human sensitivity compared with that of experimental animals.
Nature and extent of human variability and sensitivity.
3-12
-------
• Human exposure. The MOE evaluation also takes into account the magnitude,
frequency, and duration of exposure. If the population exposed in a particular scenario is
wholly or largely composed of a subpopulation of special concern (e.g., children) for
whom evidence indicates a special sensitivity to the agent's MO A, an adequate MOE
would be larger than for general population exposure.
E. Both Linear and Nonlinear Approaches
Any of the following conclusions leads to selection of both a linear and nonlinear
approach to dose-response assessment. Relative support for each dose-response method and
advice on the use of that information needs to be documented for the AWQC. In some cases,
evidence for one MOA is stronger than for the other, allowing emphasis to be placed on that
dose-response approach. In other cases, both modes of action are equally possible, and both
dose-response approaches should be emphasized.
• Modes of action for a single tumor type support both linear and nonlinear dose response
in different parts of the dose-response curve (e.g., 4,4' methylene chloride).
• A tumor mode of action supports different approaches at high and low doses; e.g., at high
dose, nonlinearity, but, at low dose, linearity (e.g., formaldehyde).
• The agent is not DNA-reactive and all plausible modes of action are consistent with
nonlinearity, but not fully established.
• Modes of action for different tumor types support differing approaches, e.g., nonlinear
for one tumor type and linear for another due to lack of MOA information (e.g.,
tri chl oroethy 1 ene).
3.1.3.5 AWOC Calculation
A. Linear Approach
The following equation is used for the calculation of the AWQC for carcinogens where
an RSD is obtained from the linear approach:
AWQC = RSD BW
4
DI + E (FIj • BAF;)
i=2
AWQC = Ambient water quality criterion (mg/L)
RSD = Risk-specific dose (mg/kg-day)
BW = Human body weight (kg)
DI = Drinking water intake (L/day)
3-13
(Equation 3-4)
-------
FI; = Fish intake at trophic level I (I = 2, 3, and 4) (kg/day)
BAF; = Bioaccumulation factor for trophic level I (I = 2, 3, and 4), lipid
normalized (L/kg)
B. Nonlinear Approach
In those cases where the nonlinear, MOE approach is used, a similar equation is used to
calculate the AWQC 9
AWQC = • RSC
UF
BW
DI + E (FI; • BAF;)
i=2
(Equation 3-5)
where variables are defined as for Equation 3-4 and:
POD = Point of departure (mg/kg-day)
UF = Uncertainty factor (unitless)
RSC = Relative source contribution (percentage or subtraction)
Differences between the AWQC values obtained using the linear and nonlinear
approaches should be noted. First, the AWQC value obtained using the default linear approach
corresponds to a specific estimated incremental lifetime cancer risk level in the range of 10"4 to
10"6. In contrast, the AWQC obtained using the nonlinear approach does not describe a specific
cancer risk. The AWQC calculations shown above are appropriate for waterbodies that are used
as sources of drinking water.
The actual AWQC chosen for the protection of human health is based on a review of all
relevant information, including cancer and noncancer data. The AWQC may, or may not, utilize
the value obtained from the cancer analysis in the final AWQC value. The endpoint selected for
the AWQC will be based on consideration of the weight of evidence and a complete analysis of
all toxicity endpoints.
3.1.3.6 Risk Characterization
Risk assessment is an integrative process that is documented in a risk characterization
summary. Risk characterization is the final step of the risk assessment process in which all
preceding analyses (i.e., hazard, dose-response, and exposure assessments) are tied together to
convey the overall conclusions about potential human risk. This component of the risk
assessment process characterizes the data in nontechnical terms, explaining the extent and
weight of evidence, major points of interpretation and rationale, and strengths and weaknesses of
9 Although appearing in this equation as a factor to be multiplied, the RSC can also be an amount subtracted.
3-14
-------
the evidence, and discussing alternative approaches, conclusions, uncertainties, and variability
that deserve serious consideration.
Risk characterization information accompanies the numerical AWQC value and
addresses the major strengths and weaknesses of the assessment arising from the availability of
data and the current limits of understanding the process of cancer causation. Key issues relating
to the confidence in the hazard assessment and the dose-response analysis (including the low-
dose extrapolation procedure used) are discussed. Whenever more than one interpretation of the
weight of evidence for carcinogenicity or the dose-response characterization can be supported,
and when choosing among them is difficult, the alternative views are provided along with the
rationale for the interpretation chosen in the derivation of the AWQC value. Where possible,
quantitative uncertainty analyses of the data are provided; at a minimum, a qualitative discussion
of the important uncertainties is presented.
3.1.3.7 Use of Toxicity Equivalence Factors and Relative Potency Estimates
The 1999 draft revised cancer guidelines state:
A toxicity equivalence factor (TEF) procedure is one used to derive quantitative
dose-response estimates for agents that are members of a category or class of
agents. TEFs are based on shared characteristics that can be used to order the
class members by carcinogenic potency when cancer bioassay data are
inadequate for this purpose. The ordering is by reference to the characteristics
and potency of a well-studied member or members of the class. Other class
members are indexed to the reference agent(s) by one or more shared
characteristics to generate their TEFs.
In addition, the 1999 draft revised cancer guidelines state that TEFs are generated and used for
the limited purpose of assessment of agents or mixtures of agents in environmental media when
better data are not available. When better data become available for an agent, the TEF should be
replaced or revised. To date, adequate data to support use of TEFs have been found only for
dibenzofurans (dioxins) and coplanar poly chlorinated biphenyls (PCBs) (USEPA, 1989, 1999b).
The uncertainties associated with TEFs must be described when this approach is used.
This is a default approach to be used when tumor data are not available for individual
components in a mixture. Relative potency factors (RPFs) can be similarly derived and used for
agents with carcinogenicity or other supporting data. The RPF is conceptually similar to TEFs,
but does not have the same level of data to support it and thus has a less rigorous definition
compared with the TEF. TEFs and RPFs are used only when there is no better alternative.
When they are used, assumptions and uncertainties associated with them are discussed. As of
today, there are only three classes of compounds for which relative potency approaches have
been examined by EPA: dibenzofurans (dioxins), polychlorinated biphenyls (PCBs), and
poly cyclic aromatic hydrocarbons (PAHs). There are limitations to the use of TEF and RFP
approaches, and caution should be exercised when using them. More guidance can be found in
the draft document for conducting health risk assessment of chemical mixtures, published by the
EPA Risk Assessment Forum (USEPA,1999b).
3-15
-------
3.1.4 References for Cancer Section
Barnes, D.G., G.P Daston, J.S. Evans, A.M. Jarabek, RJ. Kavlock, C.A. Kimmel, C. Park, and
H.L. Spitzer. 1995. Benchmark dose workshop: Criteria for use of a benchmark dose to
estimate a reference dose. Regul. Toxicol. Pharmacol. 21:296-306.
USEPA (U.S. Environmental Protection Agency). 1980. Water quality criteria documents.
Federal Register 45: 79318-79379.
USEPA (U.S. Environmental Protection Agency). 1986. Guidelines for carcinogen risk
assessment. Federal Register 51:33992-34003.
USEPA (U.S. Environmental Protection Agency). 1989. Interim Procedures for Estimating
Risks Associated with Exposures to Mixtures of Chlorinated Dibenzo-p-dioxins and -
Dibenzofurans (CDDs and CDFs) and 1989 Update. Risk Assessment Forum.
Washington, DC. EPA/625/3-89/016.
USEPA (U.S. Environmental Protection Agency). 1992a. Draft report: a cross-species scaling
factor for carcinogen risk assessment based on equivalence of mg/kg3/4/day. Federal
Register 57': 24152-24173.
USEPA (U.S. Environmental Protection Agency). 1993. Revision of Methodology for Deriving
National Ambient Water Quality Criteria for the Protection of Human Health: Report of
Workshop and EPA 's Preliminary Recommendations for Revision. Submitted to EPA
Science Advisory Board Drinking Water Committee, January 8, 1993. Office of Science
and Technology, Office of Water. Water Docket W-97-20.
USEPA (U.S. Environmental Protection Agency). 1996. Proposed Guidelines for Carcinogen
Risk Assessment. Office of Research and Development. Washington, DC. EPA/600/P-
92/003C. (FederalRegister 61:17960)
USEPA (U.S. Environmental Protection Agency). 1998a. Draft Water Quality Criteria
Methodology: Human Health. Federal Register Notice. Office of Water. Washington,
DC. EPA-822-Z-98-001.
USEPA (U.S. Environmental Protection Agency). 1998b. Ambient Water Quality Criteria
Derivation Methodology - Human Health. Technical Support Document. Final Draft.
Office of Water. Washington, DC. EPA-822-B-98-005.
USEPA (U.S. Environmental Protection Agency). 1999a. Guidelines for Carcinogen Risk
Assessment. Review Draft. Risk Assessment Forum. Washington, DC. EPA/NCEA-F-
0644. July.
USEPA (U.S. Environmental Protection Agency). 1999b. Guidance for Conducting Health Risk
Assessment of Chemical Mixtures. External Peer Review Draft. Risk Assessment Forum.
Washington, DC. EPA/NCEA-C-0148. April.
3-16
-------
USEPA (U.S. Environmental Protection Agency). 1999c. Revisions to the Methodology for
Deriving Ambient Water Quality Criteria for the Protection of Human Health. Peer
Review Workshop Summary Report. Office of Water. Washington, DC. EPA-822-R-99-
015. September.
USEPA (U.S. Environmental Protection Agency). 2000. Methodology for Deriving Ambient
Water Quality Criteria for the Protection of Human Health (2000). Technical Support
Document Volume 1: Risk Assessment. Office of Science and Technology, Office of
Water. Washington, DC. EPA-822-B-00-005. August.
3.2 NONCANCER EFFECTS
3.2.1 1980 AWQC National Guidelines for Noncancer Effects
In the 1980 AWQC National Guidelines, the Agency evaluated noncancer human health
effects from exposure to chemical contaminants using Acceptable Daily Intake (ADI) levels.
ADIs were calculated by dividing NOAELs by safely factors (SFs) to obtain estimates of doses
of chemicals that would not be expected to cause adverse effects over a lifetime of exposure. In
accordance with the National Research Council report of 1977 (NRC, 1977), EPA used SFs of
10, 100, or 1,000, depending on the quality and quantity of the overall database. In general, a
factor of 10 was suggested when good-quality data identifying a NOAEL from human studies
were available. A factor of 100 was suggested if no human data were available, but the database
contained valid chronic animal data. For chemicals with no human data and scant animal data, a
factor of 1,000 was recommended. Intermediate SFs could also be used for databases that fell
between these categories.
AWQC were calculated using the ADI levels together with standard exposure
assumptions about the rates of human ingestion of water and fish, and also accounting for intake
from other sources (see Equation 1-1 in the Introduction). Surface water concentrations at or
below the calculated criteria concentrations would be expected to result in human exposure
levels at or below the ADI. Inherent in these calculations is the assumption that, generally,
adverse effects from noncarcinogens exhibit a threshold.
3.2.2 Noncancer Risk Assessment Developments Since 1980
Since 1980, the risk assessment of noncarcinogenic chemicals has changed. To remove
the value judgments implied by the words "acceptable" and "safety," the ADI and SF terms have
been replaced with the terms RfD and UF/modifying factor (MF), respectively.
For the risk assessment of general systemic toxicity, the Agency currently uses the
guidelines contained in the IRIS background document entitled Reference Dose (RfD):
Description and Use in Health Risk Assessments (hereafter the "IRIS background document".
That document defines an RfD as "an estimate (with uncertainty spanning approximately an
order of magnitude) of a daily exposure to the human population (including sensitive subgroups)
that is likely to be without appreciable risk of deleterious effects over a lifetime" (USEPA,
3-17
-------
1993a). The most common approach for deriving the RfD does not involve dose-response
modeling. Instead, an RfD for a given chemical is usually derived by first identifying the
NOAEL for the most sensitive known toxicity endpoint, that is, the toxic effect that occurs at the
lowest dose. This effect is called the critical effect. Factors such as the study protocol, the
species of experimental animal, the nature of the toxicity endpoint assessed and its relevance to
human effects, the route of exposure, and exposure duration are critically evaluated in order to
select the most appropriate NOAEL from among all available studies in the chemical's database.
If no appropriate NOAEL can be identified from any study, then the LOAEL for the critical
effect endpoint is used and an uncertainly factor for LOAEL-to-NOAEL extrapolation is applied.
Using this approach, the RfD is equal to the NOAEL (or LOAEL) divided by the product of UFs
and, occasionally, anMF:
vfn ( /i iA \ NOAEL (or LOAEL)
RfD (mg/kg/day) = ^ >- (Equation 3-6)
Ur • Mr
The definitions and guidance for use of the UFs and the MFs are provided in the IRIS
background document and are repeated in Table 3-1.
The IRIS background document on the RfD (USEPA, 1993a) provides guidance for
critically assessing noncarcinogenic effects of chemicals and for deriving the RfD. Another
reference on this topic is Dourson (1994). Furthermore, the Agency has also published separate
guidelines for assessing specific toxic endpoints, such as developmental toxicity (USEPA,
199la), reproductive toxicity (USEPA, 1996a), and neurotoxicity risk assessment (USEPA,
1995). These endpoint-specific guidelines will be used for their respective areas in the hazard
assessment step and will complement the overall toxicological assessment. It should be noted,
however, that an RfD, derived using the most sensitive known endpoint, is considered protective
against all noncarcinogenic effects.
3-18
-------
TABLE 3-1. UNCERTAINTY FACTORS AND THE MODIFYING FACTOR
Uncertainty Factor
UFH
UFA
UF,
Definition
Use a 1, 3, or 10-fold factor when extrapolating from valid data in studies
using long-term exposure to average healthy humans. This factor is intended
to account for the variation in sensitivity (intraspecies variation) among the
members of the human population.
Use an additional factor of 1, 3, or 10 when extrapolating from valid results of
long-term studies on experimental animals when results of studies of human
exposure are not available or are inadequate. This factor is intended to account
for the uncertainty involved in extrapolating from animal data to humans
(interspecies variation).
Use an additional factor of 1, 3, or 10 when extrapolating from less-than-
chronic results on experimental animals when there are no useful long-term
human data. This factor is intended to account for the uncertainty involved in
extrapolating from less-than-chronic NOAELs to chronic NOAELs.
Use an additional factor of 1, 3, or 10 when deriving an RfD from a LOAEL,
instead of a NOAEL. This factor is intended to account for the uncertainty
involved in extrapolating from LOAELs to NOAELs.
Use an additional 3- or 10-fold factor when deriving an RfD from an
"incomplete" database. This factor is meant to account for the inability of any
single type of study to consider all toxic endpoints. The intermediate factor of
3 (approximately !/2 Iog10 unit, i.e., the square root of 10) is often used when
there is a single data gap exclusive of chronic data. It is often designated as
UFD.
Modifying Factor
Use professional judgment to determine the MF, which is an additional uncertainty factor that is
greater than zero and less than or equal to 10. The magnitude of the MF depends upon the
professional assessment of scientific uncertainties of the study and database not explicitly treated
above (e.g., the number of species tested). The default value for the MF is 1.
Note: With each UF or MF assignment, it is recognized that professional scientific judgment must
be used. The total product of the uncertainty factors and modifying factor should not exceed 3,000.
UFT
UFr
3-19
-------
Similar to the procedure used in the 1980 AWQC National Guidelines, the revised
method of deriving AWQC for noncarcinogens uses the RfD together with various assumptions
concerning intake of the contaminant from both water and non-water sources of exposure. The
objective of an AWQC for noncarcinogens is to ensure that human exposure to a substance
related to its presence in surface water, combined with exposure from other sources, does not
exceed the RfD. The algorithm for deriving AWQC for noncarcinogens using the RfD is
presented as Equation 1-1 in the Introduction.
3.2.3 Issues and Recommendations Concerning the Derivation of AWQC for
Noncarcinogens
During a review of the 1980 AWQC National Guidelines (USEPA, 1993b), the Agency
identified several issues that must be resolved in order to develop a final revised methodology
for deriving AWQC based on noncancer effects. These issues, as discussed below, mainly
concern the derivation of the RfD as the basis for such an AWQC. Foremost among these issues
is whether the Agency should revise the present method or adopt entirely new procedures that
use quantitative dose-response modeling for the derivation of the RfD. Other issues include the
following:
Presenting the RfD as a single point value or as a range to reflect the inherent imprecision
of the RfD;
• Selecting specific guidance documents for derivation of noncancer health effect levels;
• Considering severity of effect in the development of the RfD;
Using less-than-90-day studies as the basis for RfDs;
Integrating reproductive/developmental, immunotoxicity, and neurotoxicity data into the
RfD calculation;
• Applying toxicokinetic data in risk assessments; and
Considering the possibility that some noncarcinogenic effects do not exhibit a threshold.
3.2.3.1 Using the Current NOAEL/UF-Based RfD Approach or Adopting More
Quantitative Approaches for Noncancer Risk Assessment
The current NOAEL/UF-based RfD methodology, or its predecessor ADI/SF
methodology, have been used since 1980. This approach assumes that there is a threshold
exposure below which adverse noncancer health effects are not expected to occur. Exposures
above this threshold are believed to pose some risk to exposed individuals; however, the current
approach does not address the nature and magnitude of the risk above the threshold level (i.e.,
the shape of the dose-response curve above the threshold). The NO AEL/UF-based RfD
approach is intended primarily to ensure that the RfD value derived from the available data falls
below the population effects threshold. However, the NO AEL/UF-based RfD procedure has
3-20
-------
limitations. In particular, this method requires that one of the actual experimental doses used by
the researchers in the critical study be selected as the NOAEL or LOAEL value. The
determination that a dose is a NOAEL or LOAEL will depend on the biological endpoints used
and the statistical significance of the data. Statistical significance will depend on the number and
spacing of dose groups and the numbers of animals used in each dose group. Studies using a
small number of animals can limit the ability to distinguish statistically significant differences
among measurable responses seen in dose groups and control groups. Furthermore, the
determination of the NOAEL or LOAEL also depends on the dose spacing of the study. Doses
are often widely spaced, typically differing by factors of three to ten. A study can identify a
NOAEL and a LOAEL from among the doses studied, but the "true" effects threshold cannot be
determined from those results. The study size and dose spacing limitations also limit the ability
to characterize the nature of the expected response to exposures between the observed NOAEL
and LOAEL values.
The limitations of the NOAEL/UF approach have prompted development of alternative
approaches that incorporate more quantitative dose-response information. The traditional
NOAEL approach for noncancer risk assessment has often been a source of controversy and has
been criticized in several ways. For example, experiments involving fewer animals tend to
produce higher NOAELs and, as a consequence, may produce higher RfDs. Larger sample sizes,
on the other hand, should provide greater experimental sensitivity and lower NOAELs. The
focus of the NOAEL approach is only on the dose that is the NOAEL, and the NOAEL must be
one of the experimental doses. It also ignores the shape of the dose-response curve. Thus, the
slope of the dose-response plays little role in determining acceptable exposures for human
beings. Therefore, in addition to the NOAEL/UF-based RfD approach described above, EPA
will accept other approaches that incorporate more quantitative dose-response information in
appropriate situations for the evaluation of noncancer effects and the derivation of RfDs.
However, the Agency wishes to emphasize that it still believes the NOAEL/UF RfD
methodology is valid and can continue to be used to develop RfDs.
Two alternative approaches that may have relevance in assisting in the derivation of the
RfD for a chemical are the BMD and the categorical regression approaches. These alternative
approaches may overcome some of the inherent limitations in the NOAEL/UF approach. For
example, the BMD analyses for developmental effects show that NOAELs from studies correlate
well with a 5 percent response level (Allen et al., 1994). The BMD and the categorical
regression approaches usually have greater data requirements than the RfD approach. Thus, it is
unlikely that any one approach will apply to every circumstance; in some cases, different
approaches may be needed to accommodate the varying databases for the range of chemicals for
which water quality criteria must be developed. Acceptable approaches will satisfy the
following criteria: (1) meet the appropriate risk assessment goal; (2) adequately describe the
toxicity database and its quality; (3) characterize the endpoints properly; (4) provide a measure
of the quality of the "fit" of the model when a model is used for dose-response analysis; and (5)
describe the key assumptions and uncertainties.
3-21
-------
A. The Benchmark Dose
3-22
-------
The BMD is defined as the dose estimated to produce a predetermined level of change in
response (the Benchmark Response level, or BMR) relative to control. The BMDL is defined as
the statistical lower confidence limit on the BMD. In the derivation of an RfD, the BMDL is
used as the dose to which uncertainty factors are applied instead of the NOAEL. The BMD
approach first models a dose-response curve for the critical effect(s) using available
experimental data. Several mathematical algorithms can be used to model the dose-response
curve, such as polynomial or Weibull functions. To define a BMD from the modeled curve for
quantal data, the assessor first selects the BMR. The choice of the BMR is critical. For quantal
endpoints, a particular level of response is chosen (e.g., 1 percent, 5 percent, or 10 percent). For
continuous endpoints, the BMR is the degree of change from controls and is based on what is
considered a biologically significant change. The BMD is derived from the BMR dose by
applying the desired confidence limit calculation. The RfD is obtained by dividing the BMD by
one or more uncertainty factors, similar to the NOAEL approach. Because the BMD is used like
the NOAEL to obtain the RfD, the BMR should be selected at or near the low end of the range of
increased risks that can be detected in a study of typical size. Generally, this falls in the range
between the ED01 and the ED10.
The Agency will accept use of a BMD approach to derive RfDs for those agents for
which there is an adequate database. There are a number of technical decisions associated with
the application of the BMD technique. These include the following:
The definition of an adverse response;
• Selection of response data to model;
• The form of the data used (continuous versus quantal);
The choice of the measures of increased risk (extra risk versus additional risk);
The choice of mathematical model (including use of nonstandard models for unusual data
sets);
• The selection of the BMR;
• Methods for calculating the confidence interval;
Selection of the appropriate BMD as the basis for the RfD (when multiple endpoints are
modeled from a single study, when multiple models are applied to a single response, and
when multiple BMDs are calculated from different studies); and
• The use of uncertainty factors with the BMD approach.
These topics are discussed in detail in Crump et al. (1995) and in the Risk Assessment
TSD Volume (USEPA, 2000). The use of the BMD approach has been discussed in general
terms by several authors (Gaylor, 1983; Crump, 1984; Dourson et al., 1985; Kimmel and Gaylor,
1988; Brown and Erdreich, 1989; Kimmel, 1990). The International Life Sciences Institute
3-23
-------
(ILSI) also held a major workshop on the BMD in September 1993; the workshop proceedings
are summarized in ILSI (1993) and in Barnes et al. (1995). For further information on these
technical issues, the reader is referred to the publications referenced above.
The BMD approach addresses several of the quantitative or statistical criticisms of the
NOAEL approach. These are discussed at greater length in Crump et al. (1995) and are
summarized here. First, the BMD approach uses all the dose-response information in the
selected study rather than just a single data point, such as the NOAEL or LOAEL. By using
response data from all of the dose groups to model a dose-response curve, the BMD approach
allows for consideration of the steepness of the slope of the curve when estimating the ED10.
The use of the full data set also makes the BMD approach less sensitive to small changes in data
than the NOAEL approach, which relies on the statistical comparison of individual dose groups.
The BMD approach also allows consistency in the consideration of the level of effect (e.g., a 10
percent response rate) across endpoints.
The BMD approach accounts more appropriately for the size of each dose group than the
NOAEL approach. Laboratory tests with fewer animals per dose group tend to yield higher
NOAELs, and thus higher RfDs, because statistically significant differences in response rates are
harder to detect. Therefore, in the NOAEL approach, dose groups with fewer animals lead to a
higher (less conservative) RfD. In contrast, with the BMD approach, smaller dose groups will
tend to have the effect of extending the confidence interval around the ED10; therefore, the lower
confidence limit on the ED10 (the BMD) will be lower. With the BMD approach, greater
uncertainty (smaller test groups) leads to a lower (more conservative) RfD.
There are some issues to be resolved before the BMD approach is used routinely. These
were identified in a 1996 Peer Consultation Workshop (USEPA, 1996b). Methods for routine
use of the BMD are currently under development by EPA. Several RfCs and RfDs based on the
BMD approach are included in EPA's IRIS database. These include reference values for
methylmercury based on delayed postnatal development in humans; carbon disulfide based on
neurotoxicity; 1,1,1,2-tetrafluoroethane based on testicular effects in rats; and antimony trioxide
based on chronic pulmonary interstitial inflammation in female rats.
Various mathematical approaches have been proposed for modeling developmental
toxicity data (e.g., Crump, 1984; Kimmel and Gaylor, 1988; Rai and Van Ryzin, 1985; Faustman
et al., 1989), which could be used to calculate a BMD. Similar methods can be used to model
other types of toxicity data, such as neurotoxicity data (Gaylor and Slikker, 1990, 1992; Glowa
and MacPhail, 1995). The choice of the mathematical model may not be critical, as long as
estimation is within the observed dose range. Since the model fits a mathematical equation to
the observed data, the assumptions in a particular model regarding the existence or absence of a
threshold for the effect may not be pertinent (USEPA, 1997). Thus, any model that suitably fits
the empirical data is likely to provide a reasonable estimate of a BMD. However, research has
shown that flexible models that are nonsymmetric (e.g., the Weibull) are superior to symmetric
models (e.g., the probit) in estimating the BMD because the data points at the higher doses have
less influence on the shape of the curve than at low doses. In addition, models should
incorporate fundamental biological factors where such factors are known (e.g., intralitter
correlation for developmental toxicity data) in order to account for as much variability in the
3-24
-------
data as possible. The Agency is currently using the BMD approach in risk assessments where
the data support its use. Draft guidelines for application of the BMD approach also are being
developed by the Agency.
Use of BMD methods involves fitting mathematical models to dose-response data
obtained primarily from toxicology studies. When considering available models to use for a
BMD analysis, it is important to select the model that fits the data the best and is the most
biologically appropriate. EPA has developed software following several years of research and
development, expert peer review, public comment, subsequent revision, and quality assurance
testing. The software (BMDS, Version 1.2) can be downloaded from
http://www.epa.gov/ncea/bmds.htm. BMDS facilitates these operations by providing simple
data-management tools, a comprehensive help manual, an online help system, and an easy-to-use
interface to run multiple models on the same dose-response data.
As part of this software package, EPA has included sixteen (16) different models that are
appropriate for the analysis of dichotomous (quantal) data (Gamma, Logistic, Log-Logistic,
Multistage, Probit, Log-Probit, Quantal-Linear, Quantal-Quadratic, Weibull), continuous data
(Linear, Polynomial, Power, Hill), and nested developmental toxicology data (NLogistic, NCTR,
Rai & Van Ryzin). Results from all models include a reiteration of the model formula and
model run options chosen by the user, goodness-of-fit information, the BMD, and the estimate of
the lower-bound confidence limit on the benchmark dose (BMDL). Model results are presented
in textual and graphical output files which can be printed or saved and incorporated into other
documents.
B. Categorical Regression
Categorical regression is an emerging technique that may have relevance for the
derivation of RfDs or for estimating risk above the RfD (Dourson et al., 1997; Guth et al., 1997).
The categorical regression approach, like the BMD approach, can be used to estimate a dose that
corresponds to a given probability of adverse effects. This dose would then be divided by UFs to
establish an RfD. However, unlike the BMD approach, the Categorical regression approach can
incorporate information on different health endpoints in a single dose-response analysis. For
those health effects for which studies exist, responses to the substance in question are grouped
into severity categories; for example (1) no effect, (2) no adverse effect, (3) mild-to-moderate
adverse effect, and (4) frank effect. These categories correspond to the dose categories currently
used in setting the RfD, namely, the no-observed-effect level (NOEL), NOAEL, LOAEL, and
frank-effect level (PEL), respectively. Logistic transformation or other applicable mathematical
operations are used to model the probability of experiencing effects in a certain category as a
function of dose (Harrell, 1986; Hertzberg, 1989). The "acceptability" of the fit of the model to
the data can be judged using several statistical measures, including the x2 statistic, correlation
coefficients, and the statistical significance of its model parameter estimates.
The resulting mathematical equation can be used to find a dose (or the lower confidence
bound on the dose) at which the probability of experiencing adverse effects does not exceed a
selected level, e.g., 10 percent. This dose (like the NOAEL or BMD) would then be divided by
3-25
-------
relevant UFs to calculate an RfD. For more detail on how to employ the categorical regression
approach, see the discussion in the Risk Assessment TSD (USEPA, 2000).
As with the BMD approach, the categorical regression approach has the advantage of
using more of the available dose-response data to account for response variability as well as
accounting for uncertainty due to sample size through the use of confidence intervals.
Additional advantages of categorical regression include the combining of data sets prior to
modeling, thus allowing the calculation of the slope of a dose-response curve for multiple
adverse effects rather than only one effect at a time. Another advantage is the ability to estimate
risks for different levels of severity from exposures above the RfD.
On the other hand, as with BMD, opinions differ over the amount and adequacy of data
necessary to implement the method. The categorical regression approach also requires
judgments regarding combining data sets, judging goodness-of-fit, and assigning severity to a
particular effect. Furthermore, this approach is still in the developmental stage. It is not
recommended for routine use, but may be used when data are available and justify the extensive
analyses required.
C. Summary
Whether a NOAEL/UF-based methodology, a BMD, a categorical regression model, or
other approach is used to develop the RfD, the dose-response-evaluation step of a risk
assessment process should include additional discussion about the nature of the toxicity data and
its applicability to human exposure and toxicity. The discussion should present the range of
doses that are effective in producing toxicity for a given agent; the route, timing, and duration of
exposure; species specificity of effects; and any toxicokinetic or other considerations relevant to
extrapolation from the toxicity data to human-health-based AWQC. This information should
always accompany the characterization of the adequacy of the data.
3.2.3.2 Presenting the RfD as a Single Point or as a Range for Deriving AWOC
Although the RfD has traditionally been presented and used as a single point, its
definition contains the phrase "... an estimate (with uncertainty spanning perhaps an order of
magnitude) . . ." (USEPA, 1993a). Underlying this concept is the reasoning that the selection of
the critical effect and the total uncertainty factor used in the derivation of the RfD is based on the
"best" scientific judgment, and that competent scientists examining the same database could
derive RfDs which varied within an order of magnitude.
In one instance, IRIS presented the RfD as a point value within an accompanying range.
EPA derived a single number as the RfD for arsenic (0.3 |_ig/kg-day), but added that "strong
scientific arguments can be made for various values within a factor of 2 or 3 of the currently
recommended RfD value, i.e., 0.1 to 0.8 |ag/kg/day" (USEPA, 1993c). EPA noted that
regulatory managers should be aware of the flexibility afforded them through this action.
There are situations in which the risk manager can select an alternative value to use in
place of the RfD in the AWQC calculations. The domain from which this alternative value can
3-26
-------
be selected is restricted to a defined range around the point estimate. As explained further
below, the Agency is recommending that sometimes the use of a value other than the calculated
RfD point estimate is appropriate in characterizing risk. The selection of an alternative value
within an appropriate range must be determined for each individual situation, since several
factors affect the selection of the alternative value. Observing similar effects in several animal
species, including humans, can increase confidence in the selection of the critical effect and
thereby narrow the range of uncertainty. There are other factors that can affect the precision.
These include the slope of the dose-response curve, seriousness of the observed effect, dose
spacing, and possibly the route for the experimental doses. Dose spacing and the number of
animals in the study groups used in the experiment can also affect the confidence in the RfD.
To derive the AWQC, the calculated point estimate of the RfD is the default. Based on
consideration of the available data, the use of another number within the range defined by the
product of the UF(s) (and MF, if used) could be justified in some specific situations. This means
that there are risk considerations which indicate that some value in the range other than the point
estimate may be more appropriate, based on human health or environmental fate considerations.
For example, the bioavailability of the contaminant in fish tissues is one factor to consider. If
bioavailability from fish tissues is much lower than that from water and the RfD was derived
from a study in which the contaminant exposure was from drinking water, the alternative to the
calculated RfD could be selected from the high end of the range and justified using the
quantitative difference in bioavailability.
Most inorganic contaminants, particularly divalent cations, have bioavailability values of
20 percent or less from a food matrix, but are much more available (about 80 percent or higher)
from drinking water. Accordingly, the external dose necessary to produce a toxic internal dose
would likely be higher for a study where the exposure occurred through the diet rather than the
drinking water. As a result, the RfD from a dietary study would likely be higher than that for the
drinking water study if equivalent external doses had been used. Conversely, in cases where the
NOAEL that was the basis for the RfD came from a dietary study, the alternative value could be
slightly lower than the calculated RfD.
Because the uncertainty around the dose-response relationship increases as extrapolation
below the observed data increases, the use of an alternative point within the range may be more
appropriate in characterizing the risk than the use of the calculated RfD, especially in situations
when the uncertainty is high. Therefore, as a matter of policy, the 2000 Human Health
Methodology permits the selection of a single point within a range about the calculated RfD to
be used as the basis of the AWQC if an adequate justification of the alternative point is provided.
More complete discussion of this option, including limitations on the span of the range, is
provided in the Risk Assessment TSD (USEPA, 2000).
3.2.3.3 Guidelines to be Adopted for Derivation of Noncancer Health Effects Values
The Agency currently is using the IRIS background document as the general basis for the
risk assessment of noncarcinogenic effects of chemicals (USEPA, 1993a). EPA recommends
continued use of this document for this purpose. However, it should be noted that the process
for evaluating chemicals for inclusion in IRIS is undergoing revision (USEPA, 1996c). The
3-27
-------
revised assessments for many chemicals are now available on IRIS and can be consulted as
examples of the RfD development process and required supporting documentation.
3.2.3.4 Treatment of Uncertainty Factors/Severity of Effects During the RfD Derivation
and Verification Process
During the RfD derivation and toxicology review process, EPA considers the uncertainty
in extrapolating between animal species and within individuals of a species, as well as specific
uncertainties associated with the completeness of the database. The Agency's RfD Work Group
has always considered the severity of the observed effects induced by the chemical under review
when choosing the value of the UF with a LOAEL. For example, during the derivation and
verification of the RfD for zinc (USEPA, 1992), an uncertainty factor less than the standard
factor of 10 (UF of 3) was assigned to the relatively mild decrease in erythrocyte superoxide
dismutase activity in human subjects. EPA recommends that the severity of the critical effect be
assessed when deriving an RfD and that risk managers be made aware of the severity of the
effect and the weight placed on this attribute of the effect when the RfD was derived.
3.2.3.5 Use of Less-Than-90-Dav Studies to Derive RfDs
Generally, less-than-90-day experimental studies are not used to derive an RfD. This is
based on the rationale that studies lasting for less than 90 days may be too short to detect various
toxic effects. However, EPA, has in certain circumstances, derived an RfD based on a less-than-
90-day study. For example, the RfD for nonradioactive effects of uranium is based on a 30-day
rabbit study (USEPA, 1989). The short-term exposure period was used, because it was adequate
for determining doses that cause chronic toxicity. In other cases, it may be appropriate to use a
less-than-90-day study because the critical effect is expressed in less than 90 days. For example,
the RfD for nitrate was derived and verified using studies that were less than 3-months in
duration (USEPA, 1991b). For nitrate, the critical effect of methemoglobinemia in infants
occurs in less than 90 days. When it can be demonstrated from other data in the toxicological
database that the critical adverse effect is expressed within the study period and that a longer
exposure duration would not exacerbate the observed effect or cause the appearance of some
other adverse effect, the Agency may choose to use less-than-90-day studies as the basis of the
RfD. Such values would have to be used with care because of the uncertainty in determining if
other effects might be expressed if exposure was of greater duration than 90 days.
3.2.3.6 Use of Reproductive/Developmental, Immunotoxicity, and Neurotoxicity Data as the
Basis for Deriving RfDs
All relevant toxicity data have some bearing on the RfD derivation and verification and
are considered by EPA. The "critical" effect is the adverse effect most relevant to humans or, in
the absence of an effect known to be relevant to humans, the adverse effect that occurs at the
lowest dose in animal studies. If the critical effect is neurotoxicity, EPA will use that endpoint
as the basis for the derivation and verification of an RfD, as it did for the RfD for acrylamide.
Moreover, the Agency is continually revising its procedures for noncancer risk assessment. For
example, EPA has released guidelines for deriving developmental RfDs (RfDDT, USEPA,
199la), for using reproductive toxicity (USEPA, 1996a), and neurotoxicity (USEPA, 1995) data
3-28
-------
in risk assessments. The Agency is currently working on guidelines for using immunotoxicity
data to derive RfDs. In addition, the Agency is proceeding with the process of generating
acceptable emergency health levels for hazardous substances in acute exposure situations based
on established guidelines (NRC, 1993).
3.2.3.7 Applicability of Toxicokinetic Data in Risk Assessment
All pertinent toxicity data should be used in the risk assessment process, including
toxicokinetic and mechanistic data. The Agency has used toxicokinetic data in deriving the RfD
for cadmium and other compounds and currently is using toxicokinetic data to better characterize
human inhalation exposures from animal inhalation experiments during derivation/verification of
RfCs. In analogy to the RfD, the RfC is considered to be an estimate of a concentration in the air
that is not anticipated to cause adverse noncancer effects over a lifetime of inhalation exposure
(USEPA, 1994; Jarabek, 1995a). For RfCs, different dosimetry adjustments are made to account
for the differences between laboratory animals and humans in gas uptake and disposition or in
particle clearance and retention. This procedure results in calculation of a "human equivalent
concentration." Based on the use of these procedures, an interspecies UF of 3 (i.e.,
approximately 1005), instead of the standard factor of 10, is used in the RfC derivation (Jarabek,
1995b).
Toxicokinetics and toxicodynamics of a chemical each contribute to a chemical's
observed toxicity, and specifically, to observed differences among species in sensitivity.
Toxicokinetics describes the disposition (i.e., deposition, absorption, distribution, metabolism,
and elimination of chemicals in the body) and can be approximated using toxicokinetic models.
Toxicodynamics describes the toxic interaction of the agent with the target cell. In the absence
of specific data on their relative contributions to the toxic effects observed in species, each is
considered to account for approximately one-half of the difference in observed effects for
humans compared with laboratory animals. The implication of this assumption is that an
interspecies uncertainty factor of 3 rather than 10 could be used for deriving an RfD when valid
toxicokinetic data and models can be applied to obtain an oral "human equivalent applied dose"
(Jarabek, 1995b). If specific data exist on the relative contribution of either element to observed
effects, that proportion will be used. The role exposure duration may play, and whether or not
the chemical or its damage may accumulate over time in a particular scenario, also requires
careful consideration (Jarabek, 1995c).
3.2.3.8 Consideration of Linearity (or Lack of a Threshold) for Noncarcinogenic Chemicals
It is quite possible that there are chemicals with noncarcinogenic endpoints that have no
threshold for effects. For example, in the case of lead, it has not been possible to identify a
threshold for effects on neurological development. Other examples could include genotoxic
teratogens and germline mutagens. Genotoxic teratogens act by causing mutational events
during organogenesis, histogenesis, or other stages of development. Germline mutagens interact
with germ cells to produce mutations which may be transmitted to the zygote and expressed
during one or more stages of development. However, there are few chemicals which currently
have sufficient mechanistic information about these possible modes of action. It should be
recognized that although an MOA consistent with linearity is possible (especially for agents
3-29
-------
known to be mutagenic), this has yet to be reasonably demonstrated for most toxic endpoints
other than cancer.
EPA has recognized the potential for nonthreshold noncarcinogenic endpoints and
discussed this issue in the Guidelines for Developmental Toxicity Risk Assessment (USEPA,
199 la) and in the 1986 Guidelines for Mutagenicity Risk Assessment (USEPA, 1986). An
awareness of the potential for such teratogenic/mutagenic effects should be established in order
to deal with such data. However, without adequate data to support a genetic or mutational basis
for developmental or reproductive effects, the default becomes a UF or MOA approach, which
are procedures utilized for noncarcinogens assumed to have a threshold. Therefore, genotoxic
teratogens and germline mutagens should be considered an exception while the traditional
uncertainty factor approach is the general rule for calculating criteria or values for chemicals
demonstrating developmental/reproductive effects. For the exceptional cases, since there is no
well-established mechanism for calculating criteria protective of human health from the effects
of these agents, criteria will be established on a case-by-case basis. Other types of nonthreshold
noncarcinogens must also be handled on a case-by-case basis.
3.2.3.9 Minimum Data Guidance
For details on minimum data guidance for RfD development, see the Risk Assessment
TSD (USEPA, 2000).
3-30
-------
3.2.4 References for Noncancer Effects
Allen, B.C., R.T. Kavlock, C.A. Kimmel, and E.M. Faustman. 1994. Dose-response assessment
for developmental toxicity. Fund. Appl. Toxicol. 23:496-509.
Barnes, D.G., G.P Daston, J.S. Evans, A.M. Jarabek, RJ. Kavlock, C.A. Kimmel, C. Park, and
H.L. Spitzer. 1995. Benchmark dose workshop: criteria for use of a benchmark dose to
estimate a reference dose. Reg. Toxicol. Pharmacol. 21:296-306.
Brown, K.G. and L.S. Erdreich. 1989. Statistical uncertainty in the no-observed-adverse-effect
level. Fund. Appl. Toxicol. 13:235-244.
Crump, K.S., B. Allen, and E. Faustman. 1995. The Use of the Benchmark Dose Approach in
Health Risk Assessment. Prepared for U.S. Environmental Protection Agency's Risk
Assessment Forum. EPA/630/R-94/007.
Crump, K.S. 1984. A new method for determining acceptable daily intakes. Fund. Appl.
Toxicol. 4:854-871.
Dourson, M.L. 1994. Methodology for establishing oral reference doses (RfDs). In: Risk
Assessment of Essential Elements. W. Mertz, C.O. Abernathy, and S.S. Olin (eds.) ILSI
Press. Washington, DC. Pp. 51-61.
Dourson, M.L., R.C. Hertzberg, R. Hartung and K. Blackburn. 1985. Novel approaches for the
estimation of acceptable daily intake. Toxicol. Ind. Health 1:23-41.
Dourson, M.L., L.K. Teuschler, P.R. Durkin, and W.M. Stiteler. 1997. Categorical regression
of toxicity data, a case study using aldicarb. Regul. Toxicol. Pharmacol. 25:121-129.
Faustman, E.M., D.G. Wellington, W.P. Smith and C.A. Kimmel. 1989. Characterization of a
developmental toxicity dose-response model. Environ. Health Perspect. 79:229-241.
Gaylor, D.W. 1983. The use of safety factors for controlling risk. J. Toxicol. Environ. Health
11:329-336.
Gaylor, D.W. and W. Slikker. 1990. Risk assessment for neurotoxic effects. Neurotoxicology
11:211-218.
Gaylor, D.W. and W. Slikker. 1992. Risk assessment for neurotoxicants. In: Neurotoxicology.
H. Tilson and C. Mitchel (eds). Raven Press. New York, NY. Pp. 331-343.
Glowa, J.R. and R.C. MacPhail. 1995. Quantitative approaches to risk assessment in
neurotoxi col ogy. In: Neurotoxicology: Approaches and Methods. Academic Press. New
York, NY. Pp. 777-787.
3-31
-------
Guth, D.J., RJ. Carroll, D.G. Simpson, and H. Zhou. 1997. Categorical regression analysis of
acute exposure to tetrachloroethylene. Risk Anal. 17(3):321-332.
Harrell, F. 1986. The legist procedure. SUGI Supplemental Library Users Guide., Ver. 5th ed.
SAS Institute. Cary, NC.
Hertzberg, R.C. 1989. Fitting a model to categorical response data with application to species
extrapolation of toxicity. Health Physics 57: 405-409.
ILSI (International Life Sciences Institute). 1993. Report of the Benchmark Dose Workshop.
ISLI Risk Science Institute. Washington, DC.
Jarabek, A.M. 1995a. The application of dosimetry models to identify key processes and
parameters for default dose-response assessment approaches. Toxicol. Lett. 79:171-184.
Jarabek, A.M. 1995b. Interspecies extrapolation based on mechanistic determinants of chemical
disposition. Human Eco. Risk Asses. l(5):41-622.
Jarabek, A.M. 1995c. Consideration of temporal toxicity challenges current default
assumptions. Inhalation Toxicol. 7:927-946.
Kimmel, C.A. 1990. Quantitative approaches to human risk assessment for noncancer health
effects. Neurotoxicology 11: 189-198.
Kimmel, C.A. and D.W. Gaylor. 1988. Issues in qualitative and quantitative risk analysis for
developmental toxicity. Risk Anal. 8: 15-20.
NRC (National Research Council). 1977. Decision Making in the Environmental Protection
Agency. Vol. 2. National Academy of Sciences. Washington, DC. Pp. 32-33 and 241-
242.
NRC (National Research Council). 1993. Guidelines for Developing Emergency Exposure
Levels for Hazardous Substances. Subcommittee on Guidelines for Developing
Community Emergency Exposure Levels (CEELs) for Hazardous Substances.
Committee on Toxicology, NRC. National Academy Press. Washington, DC.
Rai, K. and J. Van Ryzin. 1985. A dose-response model for teratological experiments involving
quantal responses. Biometrics 41: 1-10.
USEPA (U.S. Environmental Protection Agency). 1986. Guidelines for mutagenicity
assessment. Federal Register 51:34006-34012. September 24.
USEPA (U.S. Environmental Protection Agency). 1989. Reference dose (RfD) for oral exposure
for uranium (soluble salts). Integrated Risk Information System (IRIS). Online.
(Verification date 10/1/89). Office of Health and Environmental Assessment,
Environmental Criteria and Assessment Office. Cincinnati, OH.
3-32
-------
USEPA (U.S. Environmental Protection Agency). 199la. Final guidelines for developmental
toxicity risk assessment. Federal Register 56:63798-63826. December 5.
USEPA (U.S. Environmental Protection Agency). 1991b. Reference dose (RfD) for oral
exposure for nitrate. Integrated Risk Information System (IRIS). Online. (Verification
date 10/01/91). Office of Health and Environmental Assessment, Environmental Criteria
and Assessment Office. Cincinnati, OH.
USEPA (U.S. Environmental Protection Agency). 1992. Reference dose (RfD) for oral
exposure for inorganic zinc. Integrated Risk Information System (IRIS). Online.
(Verification date 10/1/92). Office of Health and Environmental Assessment,
Environmental Criteria and Assessment Office. Cincinnati, OH.
USEPA (U.S. Environmental Protection Agency). 1993a. Reference dose (RfD): Description
and use in health risk assessments. Integrated Risk Information System (IRIS). Online.
Intra-Agency Reference Dose (RfD) Work Group, Office of Health and Environmental
Assessment, Environmental Criteria and Assessment Office. Cincinnati, OH. March 15.
USEPA (U.S. Environmental Protection Agency). 1993b. Revision of Methodology for
Deriving National Ambient Water Quality Criteria for the Protection of Human Health:
Report of Workshop and EPA 's Preliminary Recommendations for Revision. Submitted
to the EPA Science Advisory Board by the Human Health Risk Assessment Branch,
Health and Ecological Criteria Division, Office of Science and Technology, Office of
Water. Washington, DC. January 8.
USEPA (U.S. Environmental Protection Agency). 1993c. Reference dose (RfD) for oral
exposure for inorganic arsenic. Integrated Risk Information System (IRIS). Online.
(Verification date 02/01/93). Office of Health and Environmental Assessment,
Environmental Criteria and Assessment Office. Cincinnati, OH.
USEPA (U.S. Environmental Protection Agency). 1994. Methods for Derivation of Inhalation
Reference Concentrations and Application of Inhalation Dosimetry. Office of Health and
Environmental Assessment, Environmental Criteria and Assessment Office. Research
Triangle Park, NC. EPA/600/8-90/066F.
USEPA (U.S. Environmental Protection Agency). 1995. Proposed guidelines for neurotoxicity
risk assessment. Federal Register 60:52032-52056. October 4.
USEPA (U.S. Environmental Protection Agency). 1996a. Reproductive toxicity risk assessment
guidelines. Federal Register 61:56274-56322. October 31.
USEPA (U.S. Environmental Protection Agency). 1996b. Report on the Benchmark Dose Peer
Consultation Workshop. Risk Assessment Forum. Washington, DC. EPA/630/R-
96/011.
3-33
-------
USEPA (U.S. Environmental Protection Agency). 1996c. Integrated Risk Information System
(IRIS); announcement of pilot program; request for information. Federal Register. 61:
14570. April 2.
USEPA (U.S. Environmental Protection Agency). 1997. Mercury Study: Report to Congress.
Volume 5: Health Effects of Mercury and Mercury Compounds. Office of Air Quality
Planning and Standards, and Office of Research and Development. Research Triangle
Park, NC. EPA-452-R-97-007.
USEPA (U.S. Environmental Protection Agency). 2000. Methodology for Deriving Ambient
Water Quality Criteria for the Protection of Human Health (2000). Technical Support
Document Volume 1: Risk Assessment. Office of Science and Technology, Office of
Water. Washington, DC. EPA-822-B-00-005. August.
3-34
-------
[This page left blank intentionally.]
3-35
-------
4. EXPOSURE
The derivation of AWQC for the protection of human health requires information about
both the toxicological endpoints of concern for water pollutants and the pathways of human
exposure to those pollutants. The two primary pathways of human exposure to pollutants
present in a particular ambient waterbody that have been considered in deriving AWQC are
direct ingestion of drinking water obtained from that waterbody and the consumption of
fish/shellfish obtained from that waterbody. The water pathway also includes other exposures
from household uses (e.g., showering). The derivation of an AWQC involves the calculation of
the maximum water concentration for a pollutant (i.e., the water quality criteria level) that
ensures drinking water and/or fish ingestion exposures will not result in human intake of that
pollutant in amounts that exceed a specified level based upon the toxicological endpoint of
concern.
The equation for noncancer effects is presented again here, in simplified form, to
emphasize the exposure-related parameters (in bold). [Note: the RSC parameter also applies to
nonlinear low-dose extrapolation for cancer effects and the other exposure parameters apply to
all three of the equations (see Section 1.6).]
(BW)
AWQC- RJD. ™C' (Equa«ion4-l)
where:
AWQC = Ambient Water Quality Criterion (mg/L)
RfD = Reference dose for noncancer effects (mg/kg-day)
RSC = Relative source contribution factor to account for non-water
sources of exposure
BW = Human body weight (kg)
DI = Drinking water intake (L/day)
FI = Fish intake (kg/day)
BAF = Bioaccumulation factor (L/kg)
The following subsections discuss exposure issues relevant to the 2000 Human Health
Methodology: exposure policy issues; consideration of non-water sources of exposure (the
Relative Source Contribution approach); and the factors used in AWQC computation. In
relevant sections, science policy and risk management decisions made by EPA are discussed.
4.1 EXPOSURE POLICY ISSUES
This section discusses broad policy issues related to exposure concerning the major
objectives that the Agency believes should be met in setting AWQC.
4-1
-------
An Exposure Assessment TSD provides greater detail on numerous topics discussed in
this guidance: suggested sources of contaminant concentration and exposure intake information;
suggestions of survey methods for obtaining and analyzing exposure data necessary for deriving
AWQC; summaries of studies on fish consumption among sport fishers and subsistence fishers;
more detailed presentation of parameter values (e.g., fish consumption rates, body weights); and
additional guidance on the application of the RSC approach.
4.1.1 Sources of Exposure Associated With Ambient Water
4.1.1.1 Appropriateness of Including the Drinking Water Pathway in AWOC
EPA intends to continue including the drinking water exposure pathway in the derivation
of its national default human health criteria (AWQC), as has been done since the 1980 AWQC
National Guidelines were first published.
EPA recommends inclusion of the drinking water exposure pathway where drinking
water is a designated use for the following reasons: (1) Drinking water is a designated use for
surface waters under the CWA and, therefore, criteria are needed to assure that this designated
use can be protected and maintained. (2) Although rare, there are some public water supplies
that provide drinking water from surface water sources without treatment. (3) Even among the
majority of water supplies that do treat surface waters, existing treatments may not necessarily
be effective for reducing levels of particular contaminants. (4) In consideration of the Agency's
goals of pollution prevention, ambient waters should not be contaminated to a level where the
burden of achieving health objectives is shifted away from those responsible for pollutant
discharges and placed on downstream users to bear the costs of upgraded or supplemental water
treatment.
This policy decision has been supported by the States, most of the public stakeholders,
and by external peer reviewers. As with the other exposure parameters, States and authorized
Tribes have the flexibility to use alternative intake rates if they believe that drinking water
consumption is substantively different than EPA's recommended default assumptions of 2 L/day
for adults and 1 L/day for children. EPA recommends that States and authorized Tribes use an
intake rate that would be protective of a majority of consumers and will consider whether an
alternative assumption is adequately protective of a State's or Tribe's population based on the
information or rationale provided at the time EPA reviews State and Tribal water quality
standards submissions.
4.1.1.2 Setting Separate AWOC for Drinking Water and Fish Consumption
In conjunction with the issue of the appropriateness of including the drinking water
pathway explicitly in the derivation of AWQC for the protection of human health, EPA intends
to continue its practice of setting a single AWQC for both drinking water and fish/shellfish
consumption, and a separate AWQC based on ingestion offish/shellfish alone. This latter
criterion applies in those cases where the designated uses of a waterbody include supporting
fishable uses under Section 101(a) of the CWA and, thus, fish or shellfish for human
consumption, but not as a drinking water supply source (e.g., non-potable estuarine waters).
4-2
-------
EPA does not believe that national water quality criteria for protection of drinking water
uses only are particularly useful for two reasons. First, State and Tribal standards for human
health are set to protect Section 101(a) uses (e.g., "fishable, swimmable uses") under the CWA.
Second, most waters have multiple designated uses. Additionally, the water quality standards
program protects aquatic life. The 2000 Human Health Methodology revisions do not change
EPA's policy to apply aquatic life criteria to protect aquatic species where they are more
sensitive (i.e., when human health criteria would not be protective enough) or where human
health via fish or water ingestion is not an issue.
4.1.1.3 Incidental Ingestion from Ambient Surface Waters
The 2000 Human Health Methodology does not routinely include criteria to address
incidental ingestion of water from recreational uses. EPA has considered whether there are cases
where water quality criteria for the protection of human health based only on fish ingestion (or
only criteria for the protection of aquatic life) may not adequately protect recreational users from
health effects resulting from incidental water ingestion.
EPA reviewed information that provided estimates of incidental water ingestion rates
averaged over time. EPA generally believes that the averaged amount is negligible and will not
have any impact on the chemical criteria values representative of both drinking water and fish
ingestion. A lack of impact on the criteria values would likely also be true for chemical criteria
based on fish consumption only, unless the chemical exhibits no bioaccumulation potential.
However, EPA also believes that incidental/accidental water ingestion could be important for the
development of microbial contaminant water quality criteria, and for either chemical or
microbial criteria for States where recreational uses such as swimming and boating are
substantially higher than the national average. EPA also notes that some States have indicated
they already have established incidental ingestion rates for use in developing criteria. Therefore,
although EPA will not use this intake parameter when deriving its national 304(a) chemical
criteria, limited guidance is provided in the Exposure Assessment TSD volume in order to assist
States and authorized Tribes that face situations where this intake parameter could be of
significance.
4.2 CONSIDERATION OF NON-WATER SOURCES OF EXPOSURE WHEN
SETTING AWQC
4.2.1 Policy Background
The 2000 Human Health Methodology uses different approaches for addressing non-
water exposure pathways in setting AWQC for the protection of human health depending upon
the toxicological endpoint of concern. With those substances for which the appropriate toxic
endpoint is carcinogenicity based on a linear low-dose extrapolation, only the two water sources
(i.e., drinking water and fish ingestion) are considered in the derivation of the AWQC. Non-
water sources are not considered explicitly. In the case of carcinogens based on linear low-dose
extrapolation, the AWQC is being determined with respect to the incremental lifetime risk posed
by a substance's presence in water, and is not being set with regard to an individual's total risk
from all sources of exposure. Thus, the AWQC represents the water concentration that would be
4-3
-------
expected to increase an individual's lifetime risk of carcinogenicity from exposure to the
particular pollutant by no more than one chance in one million, regardless of the additional
lifetime cancer risk due to exposure, if any, to that particular substance from other sources.
Furthermore, health-based criteria values for one medium based on linear low-dose
extrapolation typically vary from values for other media in terms of the concentration value, and
often the associated risk level. Therefore, the RSC concept could not even theoretically apply
unless all risk assessments for a particular carcinogen based on linear low-dose extrapolation
resulted in the same concentration value and same risk level; that is, an apportionment would
need to be based on a single risk value and level.
In the case of substances for which the AWQC is set on the basis of a carcinogen based
on a nonlinear low-dose extrapolation or for a noncancer endpoint where a threshold is assumed
to exist, non-water exposures are considered when deriving the AWQC using the RSC approach.
The rationale for this approach is that for pollutants exhibiting threshold effects, the objective of
the AWQC is to ensure that an individual's total exposure does not exceed that threshold level.
There has been some discussion of whether it is, in fact, necessary in most cases to
explicitly account for other sources of exposure when computing the AWQC for pollutants
exhibiting threshold effects. It has been argued that because of the conservative assumptions
generally incorporated in the calculation of RfDs (or POD/UF values) used as the basis for the
AWQC derivation, total exposures slightly exceeding the RfD are unlikely to produce adverse
effects.
EPA emphasizes that the purpose of the RSC is to ensure that the level of a chemical
allowed by a criterion or multiple criteria, when combined with other identified sources of
exposure common to the population of concern, will not result in exposures that exceed the RfD
or the POD/UF. The policy of considering multiple sources of exposure when deriving health-
based criteria has become common in EPA's program office risk characterizations and criteria
and standard-setting actions. Numerous EPA workgroups have evaluated the appropriateness of
factoring in such exposures, and the Agency concludes that it is important for adequately
protecting human health. Consequently, EPA risk management policy has evolved significantly
over the last six years. Various EPA program initiatives and policy documents regarding
aggregate exposure and cumulative risk have been developed, including the consideration of
inhalation and dermal exposures. Additionally, accounting for other exposures has been
included in recent mandates (e.g., the Food Quality Protection Act) and, thus, is becoming a
requirement for the Agency. The Exposure Decision Tree approach has been shared with other
EPA offices, and efforts to coordinate policies on aggregate exposure, where appropriate, have
begun. EPA intends to continue developing policy guidance on the RSC issue and guidance to
address the concern that human health may not be adequately protected if criteria allow for
higher levels of exposure that, combined, may exceed the RfD or POD/UF. EPA also intends to
refine the 2000 Human Health Methodology in the future to incorporate additional guidance on
inhalation and dermal exposures. As stated previously, EPA is required to derive national water
quality criteria under Section 304(a) of the CWA and does not intend to derive site-specific
criteria. However, States and authorized Tribes have the flexibility to make alternative exposure
and RSC estimates based on local data, and EPA strongly encourages this.
4-4
-------
Uncertainty factors used in the derivation of the RfD (or POD/UF) to account for intra-
and interspecies variability and the incompleteness of the toxicity data set(s)/animal studies are
specifically relevant to the chemical's internal toxicological action, irrespective of the sources of
exposure that humans may be experiencing. The Agency's policy is to consider and account for
other sources of exposure in order to set protective health criteria. EPA believes that multiple
route exposures may be particularly important when uncertainty factors associated with the RfD
are small. Although EPA is well aware that RfDs are not all equivalent in their derivation, EPA
does not believe that uncertainty in the toxicological data should result in less stringent criteria
by ignoring exposure sources. However, the RSC policy approach does allow less stringent
assumptions when multiple sources of exposure are not anticipated.
The AWQC are designed to be protective criteria, generally applicable to the waters of
the United States. While EPA cannot quantitatively predict the actual human health risk
associated with combined exposures above the RfD or POD/UF, a combination of health criteria
for multiple media exceeding the RfD or POD/UF may not be sufficiently protective. Therefore,
EPA's policy is to routinely account for all sources and routes of non-occupational exposure
when setting AWQC for noncarcinogens and for carcinogens based on nonlinear low-dose
extrapolations. EPA believes that maintaining total exposure below the RfD (or POD/UF) is a
reasonable health goal and that there are circumstances where health-based criteria for a
chemical should not exceed the RfD (or POD/UF), either alone (if only one criterion is relevant,
along with other intake sources considered as background exposures) or in combination. EPA
believes its RSC policy ensures this goal.
Also, given the inability to reasonably predict future changes in exposure patterns, the
uncertainties in the exposure estimates due to typical data inadequacy, possible unknown sources
of exposure, and the potential for some populations to experience greater exposures than
indicated by the available data, EPA believes that utilizing the entire RfD (or POD/UF) does not
ensure adequate protection.
4.2.2 The Exposure Decision Tree Approach
As indicated in Section 1, EPA has, in the past, used a "subtraction" method to account
for multiple sources of exposure to pollutants. In the subtraction method, other sources of
exposure (i.e., those other than the drinking water and fish exposures) are subtracted from the
RfD (or POD/UF). However, EPA also previously used a "percentage" method for the same
purpose. In this approach, the percentage of total exposure typically accounted for by the
exposure source for which the criterion is being determined, referred to as the relative source
contribution (RSC), is applied to the RfD to determine the maximum amount of the RfD
"apportioned" to that source. With both procedures, a "ceiling" level of 80 percent of the RfD
and a "floor level" of 20 percent of the RfD are applied.
The subtraction method is considered acceptable when only one criterion is relevant for a
particular chemical. The percentage method is recommended in the context of the above goals
when multiple media criteria are at issue. The percentage method does not simply depend on the
amount of a contaminant in the prospective criterion source only. It is intended to reflect health
considerations, the relative portions of other sources, and the likelihood for ever-changing levels
4-5
-------
in each of those multiple sources (due to ever-changing sources of emissions and discharges).
Rather than simply defaulting in every instance, the Agency attempts to compare multiple source
exposures with one another to estimate their relative contribution to the total-given that
understanding the degree to which their concentrations vary, or making any distributional
analysis, is often not possible. The criteria levels, when multiple criteria are at issue, are based
on the actual levels, with an assumption that there may be enough relative variability such that
an apportionment (relating that percentage to the RfD) is a reasonable way of accounting for the
uncertainly regarding that variability.
The specific RSC approach recommended by EPA, which we will use for the derivation
of AWQC for noncarcinogens and carcinogens assessed using nonlinear low-dose extrapolation,
is called the Exposure Decision Tree and is described below. To account for exposures from
other media when setting an AWQC (i.e., non-drinking water/non-fish ingestion exposures, and
inhalation or dermal exposures), the Exposure Decision Tree for determining proposed RfD or
POD/UF apportionments represents a method of comprehensively assessing a chemical for water
quality criteria development. This method considers the adequacy of available exposure data,
levels of exposure, relevant sources/media of exposure, and regulatory agendas (i.e., whether
there are multiple health-based criteria or regulatory standards for the same chemical). The
Decision Tree addresses most of the disadvantages associated with the exclusive use of either the
percentage or subtraction approaches, because they are not arbitrarily chosen prior to
determining the following: specific population(s) of concern, whether these populations are
relevant to multiple-source exposures for the chemical in question (i.e., whether the population is
actually or potentially experiencing exposure from multiple sources), and whether levels of
exposure, regulatory agendas, or other circumstances make apportionment of the RfD or
POD/UF desirable. Both subtraction and percentage methods are potentially utilized under
different circumstances with the Exposure Decision Tree approach, and the Decision Tree is
recommended with the idea that there is enough flexibility to use other procedures if information
on the contaminant in question suggests it is not appropriate to follow the Decision Tree. EPA
recognizes that there may be other valid approaches in addition to the Exposure Decision Tree.
The Exposure Decision Tree approach allows flexibility in the RfD (or POD/UF)
apportionment among sources of exposure. When adequate data are available, they are used to
make protective exposure estimates for the population(s) of concern. When other sources or
routes of exposure are anticipated but data are not adequate, there is an even greater need to
make sure that public health protection is achieved. For these circumstances, a series of
qualitative alternatives is used (with the less adequate data or default assumptions) that allow for
the inadequacies of the data while protecting human health. Specifically, the Decision Tree
makes use of chemical information when actual monitoring data are inadequate. It considers
information on the chemical/physical properties, uses of the chemical, and environmental fate
and transformation, as well as the likelihood of occurrence in various media. Review of such
information, when available, and determination of a reasonable exposure characterization for the
chemical will result in a water quality criterion that more accurately reflects exposures than
automatically using a default value. Although the 20 percent default will still generally be used
when information is not adequate, the need for using it should be reduced. There may also be
some situations where EPA would consider the use of an 80 percent default (see Section 4.2.3).
4-6
-------
The Decision Tree also allows for use of either the subtraction or percentage method to
account for other exposures, depending on whether one or more health-based criterion is relevant
for the chemical in question. The subtraction method is considered acceptable when only one
criterion is relevant for a particular chemical. In these cases, other sources of exposure can be
considered "background" and can be subtracted from the RfD (or POD/UF).
EPA cautions States and Tribes when using the subtraction method in these
circumstances. The subtraction method results in a criterion allowing the maximum possible
chemical concentration in water after subtracting other sources. As such, it removes any cushion
between pre-criteria levels (i.e., actual "current" levels) and the RfD, thereby setting criteria at
the highest levels short of exceeding the RfD. It is somewhat counter to the goals of the CWA
for maintaining and restoring the nation's waters. It is also directly counter to Agency policies,
explicitly stated in numerous programs, regarding pollution prevention. EPA has advocated that
it is good health policy to set criteria such that exposures are kept low when current levels are
already low. The subtraction method generally results in criteria levels of a contaminant in a
particular medium at significantly higher levels than the percentage method and, in this respect,
is contradictory to such goals. In fact, many chemicals have pre-criteria levels in environmental
media substantially lower (compared to the RfD) than the resulting criteria allow.
When more than one criterion is relevant to a particular chemical, apportioning the RfD
(or POD/UF) via the percentage method is considered appropriate to ensure that the combination
of criteria and, thus, the potential for resulting exposures do not exceed the RfD (or POD/UF).
The Exposure Decision Tree (with numbered boxes) is shown in Figure 4-1. The explanation in
the text on the following pages must be read in tandem with the Decision Tree figure; the text in
each box of the figure only nominally identifies the process and conditions for determining the
outcome for that step of the Decision Tree. The underlying objective is to maintain total
exposure below the RfD (or POD/UF) while generally avoiding an extremely low limit in a
single medium that represents just a nominal fraction of the total exposure. To meet this
objective, all proposed numeric limits lie between 80 percent and 20 percent of the RfD (or
POD/UF). Again, EPA will use the Exposure Decision Tree approach when deriving its AWQC
but also recognizes that departures from the approach may be appropriate in certain cases. EPA
understands that there may be situations where the Decision Tree procedure is not practicable or
4-7
-------
2.
4.
Figure 4-1
Exposure Decision Tree for Defining Proposed RfD (or POD/UF) Apportionment
Identify population(s) of
concern.
I
Identify relevant exposure
sources/pathways. *
Problem
Formulation
Are adequate data available
to describe central
tendencies and high-ends
for relevant exposure
source s/pathway s?
9.
Yes
I
No
Are exposures from
multiple sources (due to a
sum of sources or an
individual source)
potentially at levels near
(i.e., over 80%), at or in
excess of the RfD (or
POD/UF)?
10.
Yes
V "\J^»
Describe exposures,
uncertainties, toxicity-
related information,
control issues, and
other information for
management decision.
Perform calculations
associated with Boxes
12 or 13 as applicable.
Are there sufficient data, physical/chemical
property information, fate and transport
information, and/or generalized information
available to characterize the likelihood of
exposure to relevant sources?
11.
Is there more than one regulatory action
(i.e., criteria, standard, guidance) relevant
for the chemical in question?
5B.
6.
12.
I
No
Yes
Use subtraction of appropriate
intake levels from sources other
than source of concern, including
80% ceiling/20% floor.
Yes
Are there significant known or
potential uses/sources other
than the source of concern?
13.
* Sources and
pathways include both
ingestion and routes
other than oral for
water-related
exposures, and
nonwater sources of
exposure, including
ingestion exposures
(e.g., food), inhalation,
and/or dermal.
8A.
1
Yes
Is there some information
available on each source
to make a characteri-
zation of exposure?
Apportion the RfD (or
POD/ UF) including
80% ceiling/20% floor
using the percentage
approach (with ceiling
and floor).
No
Yes
8C.
Use 20% of the RfD
(or POD/UF).
Perform apportionment as described in
Box 12 or 13, with a 50% ceiling/
20% floor.
4-8
-------
may be simply irrelevant after considering the properties, uses, and sources of the chemical in
question. EPA endorses such flexibility by States and authorized Tribes when developing
alternative water quality criteria in order to choose other procedures that are more appropriate
for setting health-based criteria and, perhaps, apportioning the RfD or POD/UF, as long as
reasons are given as to why it is not appropriate to follow the Exposure Decision Tree approach
and as long as the steps taken to evaluate the potential sources and levels of exposure are clearly
described. Often, however, the common situation of multiple exposure sources for a chemical is
likely to merit a Decision Tree evaluation for the purpose of developing human health water
quality criteria for a given chemical.
It is clear that this will be an interactive process; input by exposure assessors will be
provided to, and received from, risk managers throughout the process, given that there may be
significant implications regarding control issues (i.e., cost/feasibility), environmental justice
issues, etc. In cases where the Decision Tree is not chosen, communication and concurrence
about the decision rationale and the alternative water quality criteria are of great importance.
Descriptions of the boxes within the Decision Tree are separated by the following
process headings to facilitate an understanding of the major considerations involved. The
decision to perform, or not to perform, an apportionment could actually be made at several points
during the Decision Tree process. Working through the process is most helpful for identifying
possible exposure sources and the potential for exposure, determining the relevancy of the
Decision Tree to developing an AWQC for a particular chemical and, possibly, determining the
appropriateness of using an alternative approach to account for overall exposure. "Relevancy"
here means determining whether more than one criterion, standard, or other guidance is being
planned or is in existence for the chemical in question. Additional guidance for States and
Tribes that wish to use the Exposure Decision Tree is provided in the Exposure Assessment
TSD.
4.2.2.1 Problem Formulation
Initial Decision Tree discussion centers around the first two boxes: identification of
population(s) of concern (Box 1) and identification of relevant exposure sources and pathways
(Box 2). The term "problem formulation" refers to evaluating the population(s) and sources of
exposure in a manner that allows determination of the potential for the population of concern to
experience exposures from multiple sources for the chemical in question. Also, the data for the
chemical in question must be representative of each source/medium of exposure and be relevant
to the identified population(s). Evaluation includes determining whether the levels, multiple
criteria or regulatory standards, or other circumstances make apportionment of the RfD or
POD/UF reasonable. The initial problem formulation also determines the exposure parameters
chosen, the intake assumptions chosen for each route, and any environmental justice or other
social issues that aid in determining the population of concern. The term "data," as used here
and discussed throughout this section, refers to ambient sampling data (whether from Federal,
regional, State, or area-specific studies) and not internal human exposure measurements.
4-9
-------
4.2.2.2 Data Adequacy
In Box 3, it is necessary that adequate data exist for the relevant sources/pathways of
exposure if one is to avoid using default procedures. The adequacy of data is a professional
judgment for each individual chemical of concern, but EPA recommends that the minimum
acceptable data for Box 3 are exposure distributions that can be used to determine, with an
acceptable 95 percent confidence interval, the central tendency and high-end exposure levels for
each source. In fact, distributional data may exist for some or most of the sources of exposure.
There are numerous factors to consider in order to determine whether a dataset is
adequate. These include: (1) sample size (i.e., the number of data points); (2) whether the data
set is a random sample representative of the target population (if not, estimates drawn from it
may be biased no matter how large the sample); (3) the magnitude of the error that can be
tolerated in the estimate (estimator precision); (4) the sample size needed to achieve a given
precision for a given parameter (e.g., a larger sample is needed to precisely estimate an upper
percentile than a mean or median value); (5) an acceptable analytical method detection limit; and
(6) the functional form and variability of the underlying distribution, which determines the
estimator precision (e.g., whether the distribution is normal or lognormal and whether the
standard deviation is 1 or 10). Lack of information may prevent assessment of each of these
factors; monitoring study reports often fail to include background information or sufficient
summary statistics (and rarely the raw data) to completely characterize data adequacy. Thus, a
case-by-case determination of data adequacy may be necessary.
That being stated, there are some guidelines, as presented below, that lead to a rough
rule-of-thumb on what constitutes an "adequate" sample size for exposure assessment. Again,
first and foremost, the representativeness of the data for the population evaluated and the
analytical quality of the data must be acceptable. If so, the primary objective then becomes
estimating an upper percentile (e.g., say the 90th) and a central tendency value of some exposure
distribution based on a random sample from the distribution. Assuming that the distribution of
exposures is unknown, a nonparametric estimate of the 90th percentile is required. The required
estimate, based on a random sample of n observations from a target population, is obtained by
ranking the data from smallest to largest and selecting the observation whose rank is 1 greater
than the largest integer in the product of 0.9 times n. For example, in a data set of 25 points, the
nonparametric estimate of the 90th percentile is the 23rd largest observation.
In addition to this point estimate, it is useful to have an upper confidence bound on the
90th percentile. To find the rank of the order statistic that gives an upper 95 percent confidence
limit on the 90th percentile, the smallest value of r that satisfies the following formula is
determined:
(Equation 4-2)
4-10
-------
where:
r = the rank order of the observation
n = the number of observations
I = integer from 0 to r - 1
For relatively small data sets, the above formula will lead to selecting the largest
observation as the upper confidence limit on the 90th percentile. However, the problem with
using the maximum is that, in many environmental datasets, the largest observation is an outlier
and would provide an unrealistic upper bound on the 90th percentile. It would, therefore, be
preferable if the sample size n were large enough so that the formula yielded the second largest
observation as the confidence limit (see for example Gibbons, 1971).
This motivates establishing the following criterion for setting an "adequate" sample size:
pick the smallest n such that the nonparametric upper 95 percent confidence limit on the 90th
percentile is the second largest value. Application of the above formula with r set to n-l yields n
= 45 for this minimum sample size.
For the upper 95 percent confidence limit to be a useful indicator of a high-end exposure,
it must not be overly conservative (too large relative to the 90th percentile). It is, therefore, of
interest to estimate the expected magnitude of the ratio of the upper 95 percent confidence limit
to the 90th percentile. This quantity generally cannot be computed, since it is a function of the
unknown distribution. However, to get a rough idea of its value, consider the particular case of a
normal distribution. If the coefficient of variation (i.e., the standard deviation divided by the
mean) is between 0.5 and 2.0, the expected value of the ratio in samples of 45 will be
approximately 1.17 to 1.31; i.e., the upper 95 percent confidence limit will be only about 17 to
31 percent greater than the 90th percentile on the average.
It should be noted that the nonparametric estimate of the 95 percent upper confidence
limit based on the second largest value can be obtained even if the data set has only two detects
(it is assumed that the two detects are greater than the detection limit associated with all non-
detects). This is an argument for using nonparametric rather than parametric estimation, since
use of parametric methods would require more detected values. On the other hand, if non-
detects were not a problem and the underlying distribution were known, a parametric estimate of
the 90th percentile would generally be more precise.
As stated above, adequacy also depends on whether the samples are relevant to and
representative of the population at risk. Data may, therefore, be adequate for some decisions and
inadequate for others; this determination requires some professional judgment.
If the answer to Box 3 is no, based on the above determination of adequacy, then the
decision tree moves to Box 4. As suggested by the separate boxes, the available data that will be
reviewed as part of Box 4 do not meet the requirements necessary for Box 3. In Box 4, any
limited data that are available (in addition to information about the chemical/physical properties,
uses, and environmental fate and transformation, as well as any other information that would
characterize the likelihood of exposure from various media for the chemical) are evaluated to
4-11
-------
make a qualitative determination of the relation of one exposure source to another. Although
this information should always be reviewed at the outset, it is recommended that this information
also be used to estimate the health-based water quality criteria. The estimate should be rather
conservative (as indicated in the Decision Tree), given that it is either not based on actual
monitoring data or is based on data that has been considered to be inadequate for a more accurate
quantitative estimate. Therefore, greater uncertainties exist and accounting for variability is not
really possible. Whether the available data are adequate and sufficiently representative will
likely vary from chemical to chemical and may depend on the population of concern. If there are
some data and/or other information to make a characterization of exposure, a determination can
be made as to whether there are significant known or potential uses for the chemical/sources of
exposure other than the source of concern (i.e., in this case, the drinking water and fish intakes
relevant to developing an AWQC) that would allow one to anticipate/quantify those exposures
(Box 6). If there are not, then it is recommended that 50 percent of the RfD or POD/UF can be
safely apportioned to the source of concern (Box 7). While this leaves half of the RfD or
POD/UF unapportioned, it is recommended as the maximum apportionment due to the lack of
data needed to more accurately quantify actual or potential exposures. If the answer to the
question in Box 6 is yes (there is multiple source information available for the exposures of
concern), and some information is available on each source of exposure (Box 8A), apply the
procedure in either Box 12 or Box 13 (depending on whether one or more criterion is relevant to
the chemical), using a 50 percent ceiling (Box 8C)-again due to the lack of adequate data. If the
answer to the question in Box 8A is no (there is no available information to characterize
exposure), then the 20 percent default of the RfD or POD/UF is used (Box 8B).
If the answer to the question in Box 4 is no; that is, there are not sufficient
data/information to characterize exposure, EPA intends to generally use the "default" assumption
of 20 percent of the RfD or POD/UF (Box 5A) when deriving or revising the AWQC. It may be
better to gather more data or information and re-review when this information becomes available
(Box 5B). EPA has done this on occasion when resources permit the acquisition of additional
data to enable better estimates of exposure instead of the default. If this is not possible, then the
assumption of 20 percent of the RfD or POD/UF (Box 5 A) should be used. Box 5 A is likely to
be used infrequently with the Exposure Decision Tree approach, given that the information
described in Box 4 should be available in most cases. However, EPA intends to use 20 percent
of the RfD (or POD/UF), which has also been used in past water program regulations, as the
default value.
4.2.2.3 Regulatory Actions
If there are adequate data available to describe the central tendencies and high ends from
each exposure source/pathway, then the levels of exposure relative to the RfD or POD/UF are
compared (Box 9). If the levels of exposure for the chemical in question are not near (currently
defined as greater than 80 percent), at, or in excess of the RfD or POD/UF, then a subsequent
determination is made (Box 11) as to whether there is more than one health-based criterion or
regulatory action relevant for the given chemical (i.e., more than one medium-specific criterion,
4-12
-------
standard or other guidance being planned, performed or in existence for the chemical). The
subtraction method is considered acceptable when only one criterion (standard, etc.) is relevant
for a particular chemical. In these cases, other sources of exposure can be considered
"background" and can be subtracted from the RfD (or POD/UF). When more than one criterion
is relevant to a particular chemical, apportioning the RfD (or POD/UF) via the percentage
method is considered appropriate to ensure that the combination of health criteria, and thus the
potential for resulting exposures, do not exceed the RfD (or POD/UF).
As indicated in Section 2, for EPA's national 304(a) criteria, the RSC intake estimates of
non-water exposures (e.g., non-fish dietary exposures) will be based on arithmetic mean values
when data are available. The assumed body weight used in calculating the national criteria will
also be based on average values. The drinking water and fish intake values are 90th percentile
estimates. EPA believes that these assumptions will be protective of a majority of the population
and recommends them for State and Tribal use. However, States and authorized Tribes have the
flexibility to choose alternative intake rate and exposure estimate assumptions to protect specific
population groups that they have chosen.
4.2.2.4 Apportionment Decisions
If the answer to the question in Box 11 is no (there is not more than one relevant
medium-specific criterion/regulatory action), then the recommended method for setting a health-
based water quality criterion is to utilize a subtraction calculation (Box 12). Specifically,
appropriate intake values for each exposure source other than the source of concern are
subtracted out. EPA will rely on average values commonly used in the Agency for food
ingestion and inhalation rates, combined with mean contaminant concentration values, for
calculating RSC estimates to subtract. Alternatively, contaminant concentrations could be
selected based on the variability associated with those concentrations for each source. This
implies that a case-by-case determination of the variability and the resulting intake chosen would
be made, as each chemical evaluated can be expected to have different variations in
concentration associated with each source of intake. However, EPA anticipates that the
available data for most contaminants will not allow this for determination (based on past
experience). Guidance addressing this possibility is addressed in the Exposure Assessment TSD.
EPA does not recommend that high-end intakes be subtracted for every exposure source, since
the combination may not be representative of any actually exposed population or individual.
The subtraction method would also include an 80 percent ceiling and a 20 percent floor.
If the answer to the question in Box 11 is yes (there is more than one medium-specific
criterion/regulation relevant), then the recommended method for setting health-based water
quality criteria is to apportion the RfD or POD/UF among those sources for which health-based
criteria are being set (Box 13). This is done via a percentage approach (with a ceiling and floor).
This simply refers to the percentage of overall exposure contributed by an individual exposure
source. For example, if for a particular chemical, drinking water were to represent half of total
exposure and diet were to represent the other half, then the drinking water contribution (or RSC)
would be 50 percent. The health-based criteria would, in turn, be set at 50 percent of the RfD or
POD/UF. This method also utilizes an appropriate combination of intake values for each
4-13
-------
exposure source based on values commonly used in the Agency for food ingestion and inhalation
rates, combined with mean contaminant concentration values.
Finally, if the levels of exposure for the chemical in question are near (currently defined
as greater than 80 percent), at, or in excess of the RfD or POD/UF (i.e., the answer in Box 9 is
yes), then the estimates of exposures and related uncertainties, recommended apportionment
(either box 12 or 13), toxicity-related information, control issues, and other information are to be
presented to managers for a decision (Box 10). The high levels referred to in Box 9 may be due
to one source contributing that high level (while other sources contribute relatively little) or due
to more than one source contributing levels that, in combination, approach or exceed the RfD or
POD/UF. Management input may be necessary due to the control issues (i.e., cost and feasibility
concerns), especially when multiple criteria are at issue. In practice, risk managers are routinely
a part of decisions regarding regulatory actions and will be involved with any recommended
outcome of the Exposure Decision Tree or, for that matter, any alternative to the Exposure
Decision Tree. However, because exposures approach or exceed the RfD or POD/UF and
because the feasibility of controlling different sources of exposure are complicated issues, risk
managers will especially need to be directly involved in final decisions in these circumstances.
It is emphasized here that the procedures in these circumstances are not different than the
procedures when exposures are not at or above the RfD (or POD/UF). Therefore, in these cases,
estimates should be performed as with Boxes 11, 12, and 13. The recommendation should be
made based on health-based considerations only, just as when the chemical in question was not a
Box 10 situation. If the chemical is relevant to one health criterion or regulatory action only, the
other sources of exposure could be subtracted from the RfD or POD/UF to determine if there is
any leftover amount for setting the criterion. If the chemical is a multiple media criteria issue,
then an apportionment should be made, even though it is possible that all sources would need to
be reduced. Regardless of the outcome of Box 9, all apportionments made (via the methods of
Boxes 12 or 13) should include a presentation of the uncertainty in the estimate and in the RfD
or POD/UF for a more complete characterization.
The process for a Box 10 situation (versus a situation that is not) differs in that the
presentations for Boxes 12 and 13 are based on apportionments (following the review of
available information and a determination of appropriate exposure parameters) that must address
additional control issues and may result in more selective reductions. With Box 10, one or
several criteria possibilities ("scenarios") could be presented for comparison along with
implications of the effects of various control options. It is appropriate to present information in
this manner to risk managers given the complexity of these additional control issues.
4.2.3 Additional Points of Clarification on the Exposure Decision Tree Approach for
Setting AWQC
As with Box 9, if a determination is made in Box 8A (i.e., information is available to
characterize exposure) that exposures are near, at, or above the RfD (or POD/UF) based on the
available information, the apportionments made need to be presented to risk managers for
decision. If information is lacking on some of the multiple exposure sources, then EPA would
use a default of 20 percent of the RfD or POD/UF (Box 8B).
4-14
-------
Results of both Boxes 12 and 13 rely on the 80 percent ceiling and 20 percent floor. The
80 percent ceiling was implemented to ensure that the health-based goal will be low enough to
provide adequate protection for individuals whose total exposure to a contaminant is, due to any
of the exposure sources, higher than currently indicated by the available data. This also
increases the margin of safety to account for possible unknown sources of exposure. The 20
percent floor has been traditionally rationalized to prevent a situation where small fractional
exposures are being controlled. That is, below that point, it is more appropriate to reduce other
sources of exposure, rather than promulgating standards for de minimus reductions in overall
exposure.
If it can be demonstrated that other sources and routes of exposure are not anticipated for
the pollutant in question (based on information about its known/anticipated uses and
chemical/physical properties), then EPA would use the 80 percent ceiling. EPA qualifies this
policy with the understanding that as its policy on cumulative risk assessment continues to
develop, the 80 percent RSC may prove to be underprotective.
In the cases of pollutants for which substantial data sets describing exposures across all
anticipated pathways of exposure exist, and probabilistic analyses have been conducted based on
those data, consideration will be given to the results of those assessments as part of the Exposure
Decision Tree approach for setting AWQC.
For many chemicals, the rate of absorption from ingestion can differ substantially from
absorption by inhalation. There is also available information for some chemicals that
demonstrates appreciable differences in gastrointestinal absorption depending on whether the
chemical is ingested from water, soil, or food. For some contaminants, the absorption of the
contaminant from food can differ appreciably for plant compared with animal food products.
Regardless of the apportionment approach used, EPA recommends using existing data on
differences in bioavailability between water, air, soils, and different foods when estimating total
exposure for use in apportioning the RfD or POD/UF. The Agency has developed such exposure
estimates for cadmium (USEPA, 1994). In the absence of data, EPA will assume equal rates of
absorption from different routes and sources of exposure.
4.2.4 Quantification of Exposure
When selecting contaminant concentration values in environmental media and exposure
intake values for the RSC analysis, it is important to realize that each value selected (including
those recommended as default assumptions in the AWQC equation) may be associated with a
distribution of values for that parameter. Determining how various subgroups fall within the
distributions of overall exposure and how the combination of exposure variables defines what
population is being protected is a complicated and, perhaps, unmanageable task, depending on
the amount of information available on each exposure factor included. Many times, the default
assumptions used in EPA risk assessments are derived from the evaluation of numerous studies
and are considered to generally represent a particular population group or a national average.
Therefore, describing with certainty the exact percentile of a particular population that is
protected with a resulting criteria is often not possible.
4-15
-------
By and large, the AWQC are derived to protect the majority of the general population
from chronic adverse health effects. However, as stated above in Section 4.1.1.1, States and
authorized Tribes are encouraged to consider protecting population groups that they determine
are at greater risk and, thus, would be better protected using alternative exposure assumptions.
The ultimate choice of the contaminant concentrations used in the RSC estimate and the
exposure intake rates requires the use of professional judgment. This is discussed in greater
detail in the Exposure Assessment TSD.
4.2.5 Inclusion of Inhalation and Dermal Exposures
EPA intends to develop policy guidelines to apply to this Methodology for explicitly
incorporating inhalation and dermal exposures. When estimating overall exposure to pollutants
for AWQC development, EPA believes that the sources of inhalation and dermal exposures
considered should include, on a case-by-case basis, both non-oral exposures from water and
other inhalation and dermal sources (e.g., ambient or indoor air, soil). When the policy
guidelines are completed, this Methodology will be refined to include that guidance.
A number of drinking water contaminants are volatile and thus diffuse from water into
the air where they may be inhaled. In addition, drinking water is used for bathing and, thus,
there is at least the possibility that some contaminants in water may be dermally absorbed.
Volatilization may increase exposure via inhalation and decrease exposure via ingestion and
dermal absorption. The net effect of volatilization and dermal absorption upon total exposure to
volatile drinking water contaminants is unclear in some cases and varies from chemical to
chemical. Dermal exposures are also important to consider for certain population groups, such
as children and other groups with high soil contact.
With regard to additional non-water related exposures, it is clear that the type and
magnitude of toxicity produced via inhalation, ingestion, and dermal contact may differ; that is,
the route of exposure can affect absorption of a chemical and can otherwise modify its toxicity.
For example, an inhaled chemical such as hydrogen fluoride may produce localized effects on
the lung that are not observed (or only observed at much higher doses) when the chemical is
administered orally. Also, the active form of a chemical (and principal toxicity) can be the
parent compound and/or one or more metabolites. With this Methodology, EPA recommends
that differences in absorption and toxicity by different routes of exposure be determined and
accounted for in dose estimates and applied to the exposure assessment. EPA acknowledges that
the issue of whether the doses received from inhalation and ingestion exposures are cumulative
(i.e., toward the same threshold of toxicity) is complicated. Such a determination involves
evaluating the chemical's physical characteristics, speciation, and reactivity. A chemical may
also exhibit different metabolism by inhalation versus oral exposure and may not typically be
metabolized by all tissues. In addition, a metabolite may be much more or much less toxic than
the parent compound. Certainly with a systemic effect, if the chemical absorbed via different
routes enters the bloodstream, then there is some likelihood that it will contact the same target
organ. Attention also needs to be given to the fact that both the RfD and RfC are derived based
on the administered level. Toxicologists generally believe that the effective concentration of the
active form of a chemical(s) at the site(s) of action determines the toxicity. If specific
differences between routes of exposure are not known, it may be reasonable to assume that the
4-16
-------
internal concentration at the site from any route contributes as much to the same effect as any
other route. A default of assuming equal absorption has often been used. However, for many of
the chemicals that the Agency has reviewed, there is a substantial amount of information already
known to determine differences in rates of absorption. For example, absorption is, in part, a
function of blood solubility (i.e., Henry's Constant) and better estimations than the default can
be made.
The RSC analyses that accompany the 2000 Human Health Methodology accommodate
inclusion of inhalation exposures. Even if different target organs are involved between different
routes of exposure, a conservative policy may be appropriate to keep all exposures below a
certain level. A possible alternative is to set allowable levels (via an equation) such that the total
of ingestion exposures over the ingestion RfD added to the total of inhalation exposures over the
inhalation RfC is not greater than 1 (Note: the RfD is typically presented in mg/kg-day and the
RfC is in mg/m3). Again, EPA intends to develop guidance for this Methodology to explicitly
incorporate inhalation and dermal exposures, and will refine the Methodology when that
guidance is completed.
4.3 EXPOSURE FACTORS USED IN THE AWQC COMPUTATION
This section presents values for the specific exposure factors that EPA will use in the
derivation of AWQC. These include human body weight, drinking water consumption rates, and
fish ingestion rates.
When choosing exposure factor values to include in the derivation of a criterion for a
given pollutant, EPA recommends considering values that are relevant to population(s) that is
(are) most susceptible to that pollutant. In addition, highly exposed populations should be
considered when setting criteria. In general, exposure factor values specific to adults and
relevant to lifetime exposures are the most appropriate values to consider when determining
criteria to protect against effects from long-term exposure which, by and large, the human health
criteria are derived to protect. However, infants and children may have higher rates of water and
food consumption per unit body weight compared with adults and also may be more susceptible
to some pollutants than adults (USEPA, 1997a). There may be instances where acute or
subchronic developmental toxicity makes children the population group of concern. In addition,
exposure of pregnant women to certain toxic chemicals may cause developmental effects in the
fetus (USEPA, 1997b). Exposures resulting in developmental effects may be of concern for
some contaminants and should be considered along with information applicable to long-term
health effects when setting AWQC. (See Section 3.2 for further discussion of this issue.) Short-
term exposure may include multiple intermittent or continuous exposures occurring over a week
or so. Exposure factor values relevant for considering chronic toxicity, as well as exposure
factor values relevant for short-term exposure developmental concerns, that could result in
adverse health effects are discussed in the sections below. In appropriate situations, EPA may
consider developing criteria for developmental health effects based on exposure factor values
specific to children or to women of childbearing age. EPA encourages States and Tribes to do
the same when health risks are associated with short-term exposures.
4-17
-------
EPA believes that the recommended exposure factor default intakes for adults in chronic
exposure situations are adequately protective of the population over a lifetime. In providing
additional exposure intake values for highly exposed subpopulations (e.g., sport anglers,
subsistence fishers), EPA is providing flexibility for States and authorized Tribes to establish
criteria specifically targeted to provide additional protection using adjusted values for exposure
parameters for body weight, drinking water intake, and fish consumption. The exposure factor
values provided for women of childbearing age and children would only be used in the
circumstances indicated above.
Each of the following sections recommends exposure parameter values for use in
developing AWQC. These are based on both science policy decisions that consider the best
available data, as well as risk management judgments regarding the overall protection afforded
by the choice in the derivation of AWQC. These will be used by EPA to derive new, or revise
existing, 304(a) national criteria.
4.3.1 Human Body Weight Values for Dose Calculations
The source of data for default human body weights used in deriving the AWQC is the
third National Health and Nutrition Examination Survey (NHANES III). NHANES III
represents a very large interview and examination endeavor of the National Center for Health
Statistics (NCHS) and included participation from the Centers for Disease Control (CDC). The
NHANES III was conducted on a nationwide probability sample of over 30,000 persons from the
civilian, non-institutionalized population of the United States. The survey began in October
1988 and was completed in October 1994 (WESTAT, 2000; McDowell, 2000). Body weight
data were taken from the NHANES III Examination Data File. Sampling weights were applied
to all persons examined in the Mobile Examination Centers (MECs) or at home, as was
recommended by the NHANES data analysts (WESTAT, 2000).
The NHANES III survey has numerous strengths and very few weaknesses. Its primary
strengths are the national representativeness, large sample size, and precise estimates due to this
large sample size. Another strength is its high response rate; the examination rate was 73
percent overall, 89 percent for children under 1 year old, and approximately 85 percent for
children 1 to 5 years old (McDowell, 2000). Interview response rates were even higher, but the
body weight data come from the NHANES examinations; that is, all body weights were carefully
measured by survey staff, rather than the use of self-reported body weights. The only significant
potential weakness of the NHANES data is the fact that the data are now between 6 and 12 years
old. Given that there were upward trends in body weight from NHANES II to NHANES III, and
that NCHS has indicated the prevalence of overweight people increased in all age groups, the
data could underestimate current body weights if that trend has continued (WESTAT, 2000).
The NHANES III collected standard body measurements of sample subjects, including
height and weight, that were made at various times of the day and in different seasons of the
year. This technique was used because one's weight may vary between winter and summer and
may fluctuate with recency of food and water intake and other daily activities (McDowell, 2000).
4-18
-------
As with the other exposure assumptions, States and authorized Tribes are encouraged to
use alternative body weight assumptions for population groups other than the general population
and to use local or regional data over default values as more representative of their target
population group(s).
4.3.1.1 Rate Protective of Human Health from Chronic Exposure
EPA recommends maintaining the default body weight of 70 kg for calculating AWQC
as a representative average value for both male and female adults. As previously indicated,
exposure factor values specific to adults are recommended to protect against effects from long-
term exposure. The value of 70 kg is based on the following information. In the analysis of the
NHANES III database, median and mean values for female adults 18-74 years old are 65.8 and
69.5 kg, respectively (WESTAT, 2000). For males in the same age range, the median and mean
values are 79.9 and 82.1 kg, respectively. The mean body weight value for men and women ages
18 to 74 years old from this survey is 75.6 kg (WESTAT, 2000). This mean value is higher than
the mean value for adults ages 20-64 years old of 70.5 kg from a study by the National Cancer
Institute (NCI) which primarily measured drinking water intake (Ershow and Cantor, 1989). The
NCI study is described in the subsection on Drinking Water Intake Rates that follows (Section
4.3.2). The value from the NHANES III database is also higher than the value given in the
revised EPA Exposure Factors Handbook (USEPA, 1997b), which recommends 71.8 kg for
adults, based on the older NHANES II data. The Handbook also acknowledges the commonly
used 70 kg value and encourages risk assessors to use values which most accurately reflect the
exposed population. However, the point is also made that the 70 kg value is used in the
derivation of cancer slope factors and unit risks that appear in IRIS. Consistency is advocated
between the dose-response relationship and exposure factors assumed. Therefore, if a value
higher than 70 kg is used, the assessor needs to adjust the dose-response relationship as
described in the Appendix to Chapter 1, Volume 1 of the Handbook (USEPA, 1997b).
4.3.1.2 Rates Protective of Developmental Human Health Effects
As noted above, pregnant women may represent a more appropriate population for which
to assess risks from exposure to chemicals in ambient waters in some cases, because of the
potential for developmental effects in fetuses. In these cases, body weights representative of
women of childbearing age may be appropriate to adequately protect offspring from such health
effects. To determine a mean body weight value appropriate to this population, separate body
weight values for women in individual age groups within the range of 15 to 44 years old were
analyzed from the NHANES III data (WESTAT, 2000). The resulting median and mean body
weight values are 63.2 and 67.3 kg, respectively. Ershow and Cantor (1989) present body
weight values specifically for pregnant women included in the survey; median and mean weights
are 64.4 and 65.8 kilograms, respectively. Ershow and Cantor (1989), however, do not indicate
the ages of these pregnant women. Based on this information for women of childbearing age and
pregnant women, EPA recommends use of a body weight value of 67 kg in cases where pregnant
women are the specific population of concern and the chemical of concern exhibits reproductive
and/or developmental effects (i.e., the critical effect upon which the RfD or POD/UF is based).
Using the 67 kg assumption would result in lower (more protective) criteria than criteria based
on 70 kg.
4-19
-------
As discussed earlier, because infants and children generally have a higher rate of water
and food consumption per unit body weight compared with adults, a higher intake rate per unit
body weight may be needed when comparing estimated exposure doses with critical doses when
RfDs are based on health effects in children. To calculate intake rates relevant to such effects,
the body weight of children should be used. As with the default body weight for pregnant
women, EPA is not recommending the development of additional AWQC (i.e., similar to
drinking water health advisories) that focus on acute or short-term effects, since these are not
seen routinely as having a meaningful role in the water quality criteria program. However, there
may be circumstances where the consideration of exposures for these groups is warranted.
Although the AWQC generally are based on chronic health effects data, they are intended to also
be protective with respect to adverse effects that may reasonably be expected to occur as a result
of elevated shorter-term exposures. EPA acknowledges this as a potential course of action and
is, therefore, recommending these default values which EPA would consider in an appropriate
circumstance and for States and authorized Tribes to utilize in such situations.
EPA is recommending an assumption of 30 kg as a default child's body weight to
calculate AWQC to provide additional protection for children when the chemical of concern
indicates health effects in children are of predominant concern (i.e., test results show children are
more susceptible due to less developed immune systems, neurological systems, and/or lower
body weights). The value is based on the mean body weight value of 29.9 kg for children ages 1
to!4 years old, which combines body weight values for individual age groups within this larger
group. The mean value is based on body weight information from NHANES III for individual-
year age groups between one and 14 years old (WESTAT, 2000). A mean body weight of 28 kg
is obtained using body weight values from Ershow and Cantor (1989) for five age groups within
this range of 0-14 years and applying a weighting method for different ages by population
percentages from the U.S. Bureau of the Census. The 30 kg assumption is also consistent with
the age range for children used with the estimated fish intake rates. Unfortunately, fish intake
rates for finer age group divisions are not possible due to the limited sampling base from the fish
intake survey; there is limited confidence in calculated values (e.g., the mean) for such fine age
groups. Given this limitation, the broad age category of body weight for children is suitable for
use with the default fish intake assumption.
Given the hierarchy of preferences regarding the use offish intake information (see
Section 4.3.3), States may have more comprehensive data and prefer to target a more narrow,
younger age group. If States choose to specifically evaluate toddlers, EPA recommends using 13
kg as a default body weight assumption for children ages 1 to 3 years old. The median and mean
values of body weight for children 1 to 3 years old are 13.2 and 13.1 kg, respectively, based on
an analysis of the NHANES III database (WESTAT, 2000). The NHANES III median and mean
values for females between 1 and 3 years old are 13.0 and 12.9 kg, respectively, and are 13.4 and
13.4 kg for males, respectively. Median and mean body weight values from the earlier Ershow
and Cantor (1989) study for children ages 1 to 3 years old were 13.6 and 14.1 kg, respectively.
Finally, if infants are specifically evaluated, EPA recommends a default body weight of 7 kg
based on the NHANES III analysis. Median and mean body weights for both male and female
infants (combined) 2 months old were 6.3 and 6.3 kg, respectively, and for infants 3 months old
were 7.0 and 6.9 kg, respectively. With the broader age category of males and females 2 to 6
months old, median and mean body weights were 7.4 and 7.4 kg, respectively. The NHANES
4-20
-------
analysis did not include infants under 2 months of age. Although EPA is not recommending
body weight values for newborns, the NCHS National Vital Statistics Report indicates that, for
1997, the median birth weight ranged from 3 to 3.5 kg, according to WESTAT (2000).
Body weight values for individual ages within the larger range of 0-14 years are listed in
the Exposure Assessment TSD for those States and authorized Tribes who wish to use body
weight values for these individual groups. States and Tribes may wish to consider certain
general developmental ages (e.g., infants, pre-adolescents, etc.), or certain specific
developmental landmarks (e.g., neurological development in the first four years), depending on
the chemical of concern. EPA encourages States and authorized Tribes to choose a body weight
intake from the tables presented in the TSD, if they believe a particular age subgroup is more
appropriate.
4.3.2 Drinking Water Intake Rates
The basis for the drinking water intake rates (also for the fish intake rates presented in
Section 4.3.3) is the 1994-96 Continuing Survey of Food Intake by Individuals fCSFII)
conducted by the U.S. Department of Agriculture (USD A, 1998). The CSFII survey collects
dietary intake information from nationally representative samples of non-institutionalized
persons residing in United States households. Households in these national surveys are sampled
from the 50 states and the District of Columbia. Each survey collects daily consumption records
for approximately 10,000 food codes across nine food groups. These food groups are (1) milk
and milk products; (2) meat, poultry, and fish; (3) eggs; (4) dry beans, peas, legumes, nuts, and
seeds; (5) grain products; (6) fruit; (7) vegetables; (8) fats, oils, and salad dressings; and (9)
sweets, sugars, and beverages. The survey also asks each respondent how many fluid ounces of
plain drinking water he or she drank during each of the survey days. In addition, the CSFII
collects household information, including the source of plain drinking water, water used to
prepare beverages, and water used to prepare foods. Data provide "up-to-date information on
food intakes by Americans for use in policy formation, regulation, program planning and
evaluation, education, and research." The survey is "the cornerstone of the National Nutritional
Monitoring and Related Research Program, a set of related federal activities intended to provide
regular information on the nutritional status of the United States population" (USD A, 1998).
The 1994-96 CSFII was conducted according to a stratified, multi-area probability
sample organized using estimates of the 1990 United States population. Stratification accounted
for geographic location, degree of urbanization, and socioeconomics. Each year of the survey
consisted of one sample with oversampling for low-income households.
Survey participants provided two non-consecutive, 24-hour days of dietary data. Both
days' dietary recall information was collected by an in-home interviewer. Interviewers provided
participants with an instructional booklet and standard measuring cups and spoons to assist them
in adequately describing the type and amount of food ingested. If the respondent referred to a
cup or bowl in their own home, a 2-cup measuring cup was provided to aid in the calculation of
the amount consumed. The sample person could fill their own bowl or cup with water to
represent the amount eaten or drunk, and the interviewer could then measure the amount
consumed by pouring it into the 2-cup measure. The Day 2 interview occurred three to 10 days
4-21
-------
after the Day 1 interview, but not on the same day of the week. The interviews allowed
participants "three passes" through the daily intake record to maximize recall (USDA, 1998).
Proxy interviews were conducted for children aged six and younger and sampled individuals
unable to report due to mental or physical limitations. The average questionnaire administration
time for Day 1 intake was 30 minutes, while Day 2 averaged 27 minutes.
Two days of dietary recall data were provided by 15,303 individuals across the three
survey years. This constitutes an overall two-day response rate of 75.9 percent. Survey weights
were corrected by the USDA for nonresponse.
All three 1994-96 CSFII surveys are multistage, stratified-cluster samples. Sample
weights, which project the data from a sampled individual to the population, are based on the
probability of an individual being sampled at each stage of the sampling design. The sample
weights associated with each individual reporting two days of consumption data were adjusted to
correct for nonresponse bias.
The 1994-96 CSFII surveys have advantages and limitations for estimating per capita
water (or fish) consumption. The primary advantage of the CSFII surveys is that they were
designed and conducted by the USDA to support unbiased estimation of food consumption
across the population in the United States and the District of Columbia. Second, the survey is
designed to record daily intakes of foods and nutrients and support estimation of food
consumption.
One limitation of the 1994-96 CSFII surveys is that individual food consumption data
were collected for only two days-a brief period which does not necessarily depict "usual intake."
Usual dietary intake is defined as "the long-run average of daily intakes by an individual."
Upper percentile estimates may differ for short-term and longer-term data because short-term
food consumption data tend to be inherently more variable. It is important to note, however, that
variability due to duration of the survey does not result in bias of estimates of overall mean
consumption levels. Also, the multistage survey design does not support interval estimates for
many of the subpopulations of interest because of sparse representation in the sample.
Subpopulations with sparse representation include Native Americans on reservations and certain
ethnic groups. While these individuals are participants in the survey, they are not present in
sufficient numbers to support consumption estimates.
Despite these limitations, the CSFII is considered one of the best sources of current
information on consumption of water and fish-containing foods. The objective of estimating per
capita water and fish consumption by the United States population is compatible with the
statistical design and scope of the CSFII survey.
4.3.2.1 Rate Protective of Human Health from Chronic Exposure
EPA recommends maintaining the default drinking water intake rate of 2 L/day to protect
most consumers from contaminants in drinking water. EPA believes that the 2 L/day assumption
is representative of a majority of the population over the course of a lifetime. EPA also notes
that there is comparatively little variability in water intake within the population compared with
4-22
-------
fish intake (i.e., drinking water intake varies, by and large, by about a three-fold range, whereas
fish intake can vary by 100-fold). EPA believes that the 2 L/day assumption continues to
represent an appropriate risk management decision. The results of the 1994-96 CSFII analysis
indicate that the arithmetic mean, 75th, and 90th percentile values for adults 20 years and older are
1.1, 1.5, and 2.2 L/day, respectively (USEPA, 2000a). The 2 L/day value represents the 86th
percentile for adults. These values can also be compared to data from an older National Cancer
Institute (NCI) study, which estimated intakes of tapwater in the United States based on the
USDA's 1977-78 Nationwide Food Consumption Survey (NFCS). The arithmetic mean, 75th,
and 90th percentile values for adults 20 - 64 years old were 1.4, 1.7, and 2.3 L/day, respectively
(Ershow and Cantor, 1989). The 2 L/day value represents the 88th percentile for adults from the
NCI study.
The 2 L/day assumption was used with the original 1980 AWQC National Guidelines and
has also been used in EPA's drinking water program. EPA believes that the newer studies
continue to support the use of 2 L/day as a reasonable and protective consumption rate that
represents the intake of most water consumers in the general population. However, individuals
who work or exercise in hot climates could have water consumption rates significantly above 2
L/day, and EPA believes that States and Tribes should consider regional or occupational
variations in water consumption.
4.3.2.2 Rates Protective of Developmental Human Health Effects
Based on the 1994-96 CSFII study data, EPA also recommends 2 L/day for women of
childbearing age. The analysis for women of childbearing age (ages 15-44) indicate mean, 75th,
and 90th percentile values of 0.9, 1.3, and 2.0 L/day, respectively. These rates compare well with
those based on an analysis of tapwater intake by pregnant and lactating women by Ershow et al.
(1991), based on the older USDA data, for women ages 15-49. Arithmetic mean, 75th and 90th
percentile values were 1.2, 1.5, and 2.2 L/day, respectively, for pregnant women. For lactating
women, the arithmetic mean, 75th and 90th percentile values were 1.3, 1.7, and 1.9 L/day,
respectively.
As noted above, because infants and children have a higher daily water intake per unit
body weight compared with adults, a water consumption rate measured for children is
recommended for use when RfDs are based on health effects in children. Use of this water
consumption rate should result in adequate protection for infants and children when setting
criteria based on health effects for this target population. EPA recommends a drinking water
intake of 1 L/day to, again, represent a majority of the population of children that consume
drinking water. The results of the 1994-96 CSFII analysis indicate that for children from 1 to 10
years of age, the arithmetic mean, 75th, and 90th percentile values are 0.4, 0.6, and 0.9 L/day,
respectively (USEPA, 2000a). The 1 L/day value represents the 93rd percentile for this group.
The arithmetic mean, 75th, and 90th percentile values for smaller children, ages 1 to 3 years, are
0.3, 0.5, and 0.7 L/day, respectively. The 1 L/day value represents the 97th percentile of the
group ages 1 to 3 years old. For the category of infants under 1 year of age, the arithmetic mean,
75th, and 90th percentile values are 0.3, 0.7, and 0.9 L/day, respectively. These data can similarly
be compared to those of the older National Cancer Institute (NCI) study. The arithmetic mean,
75th, and 90th percentile values for children 1 to 10 years old were 0.74, 0.96, and 1.3 L/day,
4-23
-------
respectively. The mean, 75th, and 90th percentile values for children 1 to 3 years old in the NCI
study were 0.6, 0.8, and 1.2 L/day, respectively. Finally, the mean, 75th, and 90th percentile
values for infants less than 6 months old were 0.3, 0.3, and 0.6 L/day, respectively (Ershow and
Cantor, 1989).
4.3.2.3 Rates Based on Combining Drinking Water Intake and Body Weight
As an alternative to considering body weight and drinking water intake rates separately,
EPA is providing rates based on intake per unit body weight data (in units of ml/kg) in the
Exposure Assessment TSD, with additional discussion on their use. These rates are based on
self-reported body weights from the CSFII survey respondents for the 1994-96 data. While EPA
intends to derive or revise national default criteria on the separate intake values and body
weights, in part due to the strong input received from its State stakeholders, the ml/kg-BW/day
values are provided in the TSD for States or authorized Tribes that prefer their use. It should be
noted that in their 1993 review, EPA's Science Advisory Board (SAB) felt that using drinking
water intake rate assumptions on a per unit body weight basis would be more accurate, but did
not believe this change would appreciably affect the criteria values (USEPA, 1993).
4.3.3 Fish Intake Rates
The basis for the fish intake rates is the 1994-96 CSFII conducted by the USD A, and
described above in Section 4.3.2.
4.3.3.1 Rates Protective of Human Health from Chronic Exposure
EPA recommends a default fish intake rate of 17.5 grams/day to adequately protect the
general population offish consumers, based on the 1994 to 1996 data from the USDA's CSFII
Survey. EPA will use this value when deriving or revising its national 304(a) criteria. This
value represents the 90th percentile of the 1994-96 CSFII data. This value also represents the
uncooked weight estimated from the CSFII data, and represents intake of freshwater and
estuarine finfish and shellfish only. For deriving AWQC, EPA has also considered the States'
and Tribes' needs to provide adequate protection from adverse health effects to highly exposed
populations such as recreational and subsistence fishers, in addition to the general population.
Based on available studies that characterize consumers offish, recreational fishers and
subsistence fishers are two distinct groups whose intake rates may be greater than the general
population. It is, therefore, EPA's decision to discuss intakes for these two groups, in addition to
the general population.
EPA recommends default fish intake rates for recreational and subsistence fishers of 17.5
grams/day and 142.4 grams/day, respectively. These rates are also based on uncooked weights
for fresh/estuarine finfish and shellfish only. However, because the level offish intake in highly
exposed populations varies by geographical location, EPA suggests a four preference hierarchy
for States and authorized Tribes to follow when deriving consumption rates that encourages use
of the best local, State, or regional data available. A thorough discussion of the development of
this policy method and relevant data sources is contained in the Exposure Assessment TSD. The
hierarchy is also presented here because EPA strongly emphasizes that States and authorized
4-24
-------
Tribes should consider developing criteria to protect highly exposed population groups and use
local or regional data over the default values as more representative of their target population
group(s). The four preference hierarchy is: (1) use of local data; (2) use of data reflecting similar
geography/population groups; (3) use of data from national surveys; and (4) use of EPA's default
intake rates.
The recommended four preference hierarchy is intended for use in evaluating fish intake
from fresh and estuarine species only. Therefore, to protect humans who additionally consume
marine species offish, the marine portion should be considered an other source of exposure
when calculating an RSC for dietary intake. Refer to the Exposure Assessment TSD for further
discussion. States and Tribes need to ensure that when evaluating overall exposure to a
contaminant, marine fish intake is not double-counted with the other dietary intake estimate
used. Coastal States and authorized Tribes that believe accounting for total fish consumption
(i.e., fresh/estuarine and marine species) is more appropriate for protecting the population of
concern may do so, provided that the marine intake component is not double-counted with the
RSC estimate. Tables offish consumption intakes based on the CSFII in the TSD provide rates
for fresh/estuarine species, marine species, and total (combined) values to facilitate this option
for States and Tribes. Throughout this section, the terms "fish intake" or "fish consumption" are
used. These terms refer to the consumption of finfish and shellfish, and the CSFII survey
includes both. States and Tribes should ensure that when selecting local or regionally-specific
studies, both finfish and shellfish are included when the population exposed are consumers of
both types.
EPA's first preference is that States and authorized Tribes use the results from fish intake
surveys of local watersheds within the State or Tribal jurisdiction to establish fish intake rates
that are representative of the defined populations being addressed for the particular waterbody.
Again, EPA recommends that data indicative of fresh/estuarine species only be used which is, by
and large, most appropriate for developing AWQC. EPA also recommends the use of uncooked
weight intake values, which is discussed in greater detail with the fourth preference. States and
authorized Tribes may use either high-end values (such as the 90th or 95th percentile values) or
average values for an identified population that they plan to protect (e.g., subsistence fishers,
sport fishers, or the general population). EPA generally recommends that arithmetic mean
values should be the lowest value considered by States or Tribes when choosing intake rates for
use in criteria derivation. When considering geometric mean (median) values from fish
consumption studies, States and authorized Tribes need to ensure that the distribution is based on
survey respondents who reported consuming fish because surveys based on both consumers and
nonconsumers can often result in median values of zero. If a State or Tribe chooses values
(whether the central tendency or high-end values) from studies that particularly target high-end
consumers, these values should be compared to high-end fish intake rates for the general
population to make sure that the high-end consumers within the general population would be
protected by the chosen intake rates. EPA believes this is a reasonable procedure and is also
consistent with the recent Great Lakes Water Quality Initiative (known as the "GLI") (USEPA,
1995). States and authorized Tribes may wish to conduct their own surveys offish intake, and
EPA guidance is available on methods to conduct such studies in Guidance for Conducting Fish
and Wildlife Consumption Surveys (USEPA, 1998). Results from broader geographic regions in
which the State or Tribe is located can also be used, but may not be as applicable as results from
4-25
-------
local watersheds. Since such studies would ultimately form the basis of a State or Tribe's
AWQC, EPA would review any surveys offish intake for consistency with the principles of
EPA's guidance as part of the Agency's review of water quality standards under Section 303(c).
If surveys conducted in the geographic area of the State or Tribe are not available, EPA's
second preference is that States and authorized Tribes consider results from existing fish intake
surveys that reflect similar geography and population groups (e.g., from a neighboring State or
Tribe or a similar watershed type), and follow the method described above regarding target
values to derive a fish intake rate. Again, EPA recommends the use of uncooked weight intake
values and the use of fresh/estuarine species data only. Results of existing local and regional
surveys are discussed in greater detail in the TSD.
If applicable consumption rates are not available from local, State, or regional surveys,
EPA's third preference is that States and authorized Tribes select intake rate assumptions for
different population groups from national food consumption surveys. EPA has analyzed one
such national survey, the 1994-96 CSFII. As described in Section 4.3.2, this survey, conducted
annually by the USD A, collects food consumption information from a probability sample of the
population of all 50 states. Respondents to the survey provide two days of dietary recall data. A
detailed description of the combined 1994-96 CSFII survey, the statistical methodology, and the
results and uncertainties of the EPA analyses are provided in a separate EPA report (USEPA,
2000b). The Exposure Assessment TSD for this Methodology presents selected results from this
report including point and interval estimates of combined finfish and shellfish consumption for
the mean, 50th (median), 90th, 95th, and 99th percentiles. The estimated fish consumption rates are
by fish habitat (i.e., freshwater/estuarine, marine and all habitats) for the following population
groups: (1) all individuals; (2) individuals age 18 and over; (3) women ages 15-44; and (4)
children age 14 and under. Three kinds of estimated fish consumption rates are provided: (1) per
capita rates (i.e., rates based on consumers and nonconsumers offish from the survey period-
refer to the TSD for further discussion); (2) consumers-only rates (i.e., rates based on
respondents who reported consuming finfish or shellfish during the two-day reporting period);
and (3) per capita consumption by body weight (i.e., per capita rates reported as milligrams of
fish per kilogram of body weight per day).
EPA's fourth preference is that States and authorized Tribes use as fish intake
assumptions the following default rates, based on the 1994-96 CSFII data, that EPA believes are
representative offish intake for different population groups: 17.5 grams/day for the general adult
population and sport fishers, and 142.4 grams/day for subsistence fishers. These are risk
management decisions that EPA has made after evaluating numerous fish intake surveys. These
values represent the uncooked weight intake of freshwater/estuarine finfish and shellfish. As
with the other preferences, EPA requests that States and authorized Tribes routinely consider
whether there is a substantial population of sport fishers or subsistence fishers when developing
site-specific estimates, rather than automatically basing them on the typical individual. Because
the combined 1994-96 CSFII survey is national in scope, EPA will use the results from this
survey to estimate fish intake for deriving national criteria. EPA has recognized the data gaps
and uncertainties associated with the analysis of the 1994-96 CSFII survey in the process of
making its default recommendations. The estimated mean of freshwater and estuarine fish
ingestion for adults is 7.50 grams/day, and the median is 0 grams/day. The estimated 90th
4-26
-------
percentile is 17.53 grams/day; the estimated 95th percentile is 49.59 grams/day; and the estimated
99th percentile is 142.41 grams/day. The median value of 0 grams/day may reflect the portion of
individuals in the population who never eat fish as well as the limited reporting period (2 days)
over which intake was measured. By applying as a default 17.5 grams/day for the general adult
population, EPA intends to select an intake rate that is protective of a majority of the population
(again, the 90th percentile of consumers and nonconsumers according to the 1994-96 CSFII
survey data). Trophic level breakouts are: TL2 = 3.8 grams/day; TL3 = 8.0 grams/day; and TL4
= 5.7 grams/day. EPA further considers 17.5 grams/day to be indicative of the average
consumption among sport fishers based on averages in the studies reviewed, which are presented
in the Exposure Assessment TSD. Similarly, EPA believes that the assumption of 142.4
grams/day is within the range of average consumption estimates for subsistence fishers based on
the studies reviewed. Experts at the 1992 National Workshop that initiated the effort to revise
this Methodology acknowledged that the national survey high-end values are representative of
average rates for highly exposed groups such as subsistence fishermen, specific ethnic groups, or
other highly exposed people. EPA is aware that some local and regional studies indicate greater
consumption among Native American, Pacific Asian American, and other subsistence
consumers, and recommends the use of those studies in appropriate cases, as indicated by the
first and second preferences. Again, States and authorized Tribes have the flexibility to choose
intake rates higher than an average value for these population groups. If a State or authorized
Tribe has not identified a separate well-defined population of high-end consumers and believes
that the national data from the 1994-96 CSFII are representative, they may choose these
recommended rates.
As indicated above, the default intake values are based on the uncooked weights of the
fish analyzed. There has been some question regarding whether to use cooked or uncooked
weights offish intake for deriving the AWQC. Studies show that, typically, with a filet or steak
offish, the weight loss in cooking is about 20 percent; that is, the uncooked weight is
approximately 20 percent higher (Jacobs et al., 1998). This obviously means that using
uncooked weights results in a slightly higher intake rate and slightly more stringent AWQC. In
researching consumption surveys for this proposal, EPA has found that some surveys have
reported rates for cooked fish, others have reported uncooked rates, and many more are unclear
as to whether cooked or uncooked rates are used. The basis of the CSFII survey was prepared or
as consumed intakes; that is, the survey respondents estimated the weight offish that they
consumed. This was also true with the GLI (which was specifically based on studies describing
consumption rates of cooked fish) and, by and large, cooked fish is what people consume.
However, EPA's Guidance For Assessing Chemical Contaminant Data For Use In Fish
Advisories recommends analysis and advisories based on uncooked fish (USEPA, 1997a). EPA
considered the potential confusion over the fact that the uncooked weights are used in the fish
advisory program. Further, the measures of a contaminant in fish tissue samples that are
applicable to compliance monitoring and the permitting program are related to the uncooked
weights. The choice of intakes is also complicated by factors such as the effect of the cooking
process, the different parts of a fish where a chemical may accumulate, and the method of
preparation.
After considering all of the above (in addition to public input received), EPA will derive
its national default criteria based on the uncooked weight fish intakes. The Exposure
4-27
-------
Assessment TSD provides additional guidance on site-specific modifications. Specifically, an
alternate approach is described for calculating AWQC with the as consumed weight-which is
more directly associated with human exposure and risk-and then adjusting the value by the
approximate 20 percent loss to an uncooked equivalent (thereby representing the same relative
risk as the as consumed value). This approach results in a different AWQC value (than using the
uncooked weights) and represents a more direct translation of the as consumed risk to the
uncooked equivalent. However, EPA understands that it is more scientifically rigorous and may
be too intensive of a process for States and Tribes to rely on. The option is presented in the TSD
to offer States and authorized Tribes greater flexibility with their water quality standards
program.
The default fish intake values also reflect specific designations of species classified in
accordance with information regarding the life history of the species or based on landings
information form the National Marine Fisheries Service. Most significantly, salmon has been
reclassified from a freshwater/estuarine species to a marine species. As marine harvested salmon
represents approximately 99 percent of salmon consumption in the 1994-96 CSFII Survey,
removal reduces the overall fresh/estuarine fish consumption rate by 13 percent. Although they
represent a very small percentage of freshwater/estuarine intake, land-locked and farm-raised
salmon consumed by 1994-96 CSFII respondents are still included. The rationale for the default
intake species designations is explained in the Exposure Assessment TSD. Once again, EPA
emphasizes the flexibility for States and authorized Tribes to use alternative assumptions based
on local or regional data to better represent their population groups of concern.
4.3.3.2 Rates Protective of Developmental Human Health Effects
Exposures resulting in health effects in children or developmental effects in fetuses may
be of primary concern. As discussed at the beginning of this section on exposure factors used, in
a situation where acute or sub-chronic toxicity and exposure are the basis of an RfD (or
POD/UF), EPA will consider basing its national default criteria on children or women of
childbearing age, depending on the target population at greatest risk. EPA recommends that
States and authorized Tribes use exposure factors for children or women of childbearing age in
these situations. As stated previously, EPA is not recommending the development of additional
AWQC but is acknowledging that basing a criterion on these population groups is a potential
course of action and is, therefore, recommending the following default intake rates for such
situations.
EPA's preferences for States and authorized Tribes in selecting values for intake rates
relevant for children is the same as that discussed above for establishing values for average daily
consumption rates for chronic effects; i.e., in decreasing order of preference, results from fish
intake surveys of local watersheds, results from existing fish intake surveys that reflect similar
geography and population groups, the distribution of intake rates from nationally based surveys
(e.g., the CSFII), or lastly, the EPA default rates. When an RfD is based on health effects in
children, EPA recommends a default intake rate of 156.3 grams/day for assessing those
contaminants that exhibit adverse effects. This represents the 90th percentile consumption rate
for actual consumers of freshwater/estuarine finfish and shellfish for children ages 14 and under
using the combined 1994 to 1996 results from the CSFII survey. The value was calculated based
4-28
-------
on data for only those children who ate fish during the 2-day survey period, and the intake was
averaged over the number of days during which fish was actually consumed. EPA believes that
by selecting the data for consumers only, the 90th percentile is a reasonable intake rate to
approximate consumption of fresh/estuarine finfish and shellfish within a short period of time for
use in assessments where adverse effects in children are of primary concern. As discussed
previously, EPA will use a default body weight of 30 kg to address potential acute or subchronic
effects from fish consumption by children. EPA is also providing these default intake values for
States and authorized Tribes that choose to provide additional protection when developing
criteria that they believe should be based on health effects in children. This is consistent with
the rationale in the recent GLI (USEPA, 1995) and is an approach that EPA believes is
reasonable. Distributional information on intake values relevant for assessing exposure when
health effects to children are of concern is presented in the Exposure Assessment TSD.
There are also cases in which pregnant women may be the population of most concern,
due to the possibility of developmental effects that may result from exposures of the mother to
toxicants. In these cases, fish intake rates specific to females of childbearing age are most
appropriate when assessing exposures to developmental toxicants. When an RfD is based on
developmental toxicity, EPA proposes a default intake rate of 165.5 grams/day for assessing
exposures for women of childbearing age from contaminants that cause developmental effects.
This is equivalent to the 90th percentile consumption rate for actual consumers of freshwater/
estuarine finfish and shellfish for women ages 15 to 44 using the combined 1994 to!996 results
from the CSFII survey. As with the rate for children, this value represents only those women
who ate fish during the 2-day survey period. As discussed previously, EPA will use a default
body weight of 67 kg for women of childbearing age.
4.3.3.3 Rates Based on Combining Fish Intake and Body Weight
As with the drinking water intake values, EPA is providing values for fish intake based
on a per unit body weight basis (in units of mg/kg) in the Exposure Assessment TSD. These
rates use the self-reported body weights of the 1994-96 CSFII survey. Again, while EPA intends
to derive or revise national default criteria on the separate intake values and body weights, the
mg/kg-BW/day values are provided in the TSD for States or authorized Tribes that prefer their
use.
4.4 REFERENCES FOR EXPOSURE
Ershow A.G., Brown L.M. and Cantor K.P. 1991. Intake of tapwater and total water by
pregnant and lactating women. Am. J. Public Health. 81:328-334.
Ershow A.G. and K.P. Cantor. 1989. Total Water and Tap Water Intake in the United States:
Population-based Estimates of Quantities and Sources. National Cancer Institute.
Bethesda, MD. Order #263-MD-810264.
Gibbons, J.D. 1971. Nonparametric Statistical Inference. Chapter 2: Order Statistics.
McGraw-Hill, Inc. New York, NY.
4-29
-------
Jacobs, H.L., H.D. Kahn, K.A. Stralka, and D.B. Phan. 1998. Estimates of per capita fish
consumption in the U.S. based on the continuing survey of food intake by individuals
(CSFII). Risk Analysis: An International Journal 18(3).
McDowell, M. 2000. Personal communication between Denis R. Borum, U.S. Environmental
Protection Agency, and Margaret McDowell, Health Statistician, National Health and
Nutrition Examination Survey, National Center for Health Statistics. March 24, 2000.
USDA. 1998. U.S. Department of Agriculture. 1994-1996 Continuing Survey of Food Intakes
by Individuals and 1994 1996 Diet and Health Knowledge Survey. Agricultural
Research Service, USDA. NTIS CD-ROM, accession number PB98-500457.
[Available from the National Technical Information Service, 5285 Port Royal Road,
Springfield, VA 22161. Phone: (703)487-4650.]
USEPA. 1993. Review of the Methodology for Developing Ambient Water Quality Criteria for
the Protection of Human Health. Prepared by the Drinking Water Committee of the
Science Advisory Board. EPA-SAB-DWC
USEPA. 1994. Reference dose (RfD) for oral exposure for cadmium. Integrated Risk
Information System (IRIS). Online. (Verification date 02/01/94.) Office of Health and
Environmental Assessment, Environmental Criteria and Assessment Office. Cincinnati,
OH.
USEPA. 1995. Great Lakes Water Quality Initiative Technical Support Document for the
Procedure to Determine Bioaccumulation Factors. Office of Water. Washington, DC.
EPA/820/B-95/005.
USEPA. 1997'a. Guidance for Assessing Chemical Contaminant Data for Use in Fish
Advisories. Volume II: Risk Assessment and Fish Consumption Limits. Second Edition.
Office of Water. Washington DC. EPA/823/B-97/009.
USEPA. 1997b. Exposure Factors Handbook. National Center for Environmental Assessment,
Office of Research and Development. Washington, DC. EPA/600/P-95/002Fa. August.
USEPA. 1998. Guidance for Conducting Fish and Wildlife Consumption Surveys. Office of
Science and Technology, Office of Water. Washington, DC. EPA-823-B-98-007.
November.
USEPA. 2000a. Estimated Per Capita Water Inge stion in the United States: Based on Data
Collected by the United States Department of Agriculture's 1994-96 Continuing Survey
of Food Intakes by Individuals. Office of Science and Technology, Office of Water.
Washington, DC. EPA-822-00-008. April.
USEPA. 2000b. Estimated Per Capita Fish Consumption in the United States: Based on Data
Collected by the United States Department of Agriculture's 1994-1996 Continuing
4-30
-------
Survey of Food Intake by Individuals. Office of Science and Technology, Office of
Water, Washington, DC. March.
WESTAT. 2000. Memorandum on Body Weight Estimates Based on NHANES III data,
Including Data Tables and Graphs. Analysis conducted and prepared by WESTAT,
under EPA Contract No. 68-C-99-242. March 3, 2000.
4-31
-------
5. BIOACCUMULATION
5.1 INTRODUCTION
Aquatic organisms can accumulate certain chemicals in their bodies when exposed to
these chemicals through water, their diet, and other sources. This process is called
bioaccumulation. The magnitude of bioaccumulation by aquatic organisms varies widely
depending on the chemical but can be extremely high for some highly persistent and
hydrophobic chemicals. For such highly bioaccumulative chemicals, concentrations in aquatic
organisms may pose unacceptable human health risks from fish and shellfish consumption even
when concentrations in water are too low to cause unacceptable health risks from drinking water
consumption alone. These chemicals may also biomagnify in aquatic food webs, a process
whereby chemical concentrations increase in aquatic organisms of each successive trophic level
due to increasing dietary exposures (e.g., increasing concentrations from algae, to zooplankton,
to forage fish, to predatory fish).
In order to prevent harmful exposures to waterborne chemicals through the consumption
of contaminated fish and shellfish, national 304(a) water quality criteria for the protection of
human health must address the process of chemical bioaccumulation in aquatic organisms. For
deriving national 304(a) criteria to protect human health, EPA accounts for potential
bioaccumulation of chemicals in fish and shellfish through the use of national bioaccumulation
factors (BAFs). A national BAF is a ratio (in L/kg) that relates the concentration of a chemical
in water to its expected concentration in commonly consumed aquatic organisms in a specified
trophic level. An illustration of how national B AFs are used in the derivation of 304(a) criteria
for carcinogens using linear low-dose extrapolation is shown in the following equation:
AWQC = RSD • BW
(Equation 5-1)
^rij • Df^r-j
i=2
where:
RSD = Risk specific dose (mg/kg-day)
BW = Human body weight (kg)
DI = Drinking water intake (L/day)
FI; = Fish intake at trophic level I, where 1=2, 3, and 4;
BAF; = National bioaccumulation factor at trophic level I,
where 1=2, 3, and 4
The purpose of this chapter is to present EPA's recommended methodology for deriving
national bioaccumulation factors for setting national 304(a) water quality criteria to protect
human health. A detailed scientific basis of the recommended national BAF methodology is
provided in the Bioaccumulation TSD. While the methodology detailed in this chapter is
5-1
-------
intended to be used by EPA for deriving national BAFs, EPA encourages States and authorized
Tribes to derive BAFs that are specific to certain regions or waterbodies, where appropriate.
Guidance to States and authorized Tribes for deriving site-specific BAFs is provided in the
Biaccumulation TSD.
5.1.1 Important Bioaccumulation and Bioconcentration Concepts
Several attributes of the bioaccumulation process are important to understand when
deriving national BAFs for use in setting national 304(a) criteria. First, the term
"bioaccumulation" refers to the uptake and retention of a chemical by an aquatic organism from
all surrounding media (e.g., water, food, sediment). The term "bioconcentration" refers to the
uptake and retention of a chemical by an aquatic organism from water only. For some chemicals
(particularly those that are highly persistent and hydrophobic), the magnitude of bioaccumulation
by aquatic organisms can be substantially greater than the magnitude of bioconcentration. Thus,
an assessment of bioconcentration alone would underestimate the extent of accumulation in
aquatic biota for these chemicals. Accordingly, EPA's guidelines presented in this chapter
emphasize the measurement of chemical bioaccumulation by aquatic organisms, whereas EPA's
1980 Methodology emphasized the measurement of bioconcentration.
Another noteworthy aspect of bioaccumulation process is the issue of steady-state
conditions. Specifically, both bioaccumulation and bioconcentration can be viewed simply as
the result of competing rates of chemical uptake and depuration (chemical loss) by an aquatic
organism. The rates of chemical uptake and depuration can be affected by various factors
including the properties of the chemical, the physiology of the organism in question, water
quality and other environmental conditions, ecological characteristics of the waterbody (e.g.,
food web structure), and the concentration and loadings history of the chemical. When the rates
of chemical uptake and depuration are equal, tissue concentrations remain constant over time and
the distribution of the chemical between the organism and its source(s) is said to be at steady-
state. For constant chemical exposures and other conditions, the steady-state concentration in
the organism represents the highest accumulation potential of the chemical in that organism
under those conditions. The time required for a chemical to achieve steady state has been shown
to vary according to the properties of the chemical and other factors. For example, some highly
hydrophobic chemicals can require long periods of time to reach steady state between
environmental compartments (e.g., many months), while highly hydrophilic chemicals usually
reach steady-state relatively quickly (e.g., hours to days).
Since national 304(a) criteria for the protection of human health are typically designed to
protect humans from harmful lifetime or long-term exposures to waterborne contaminants, the
assessment of bioaccumulation that equals or approximates steady-state accumulation is one of
the principles underlying the derivation of national BAFs. For some chemicals that require
relatively long periods of time to reach steady-state in tissues of aquatic organisms, changes in
water column concentrations may occur on a much more rapid time scale compared to the
corresponding changes in tissue concentrations. Thus, if the system departs substantially from
steady-state conditions and water concentrations are not averaged over a sufficient time period,
the ratio of the tissue concentration to a water concentration may have little resemblance to the
steady-state ratio and have little predictive value of long-term bioaccumulation potential.
5-2
-------
Therefore, BAF measurements should be based on water column concentrations which are
averaged over a sufficient period of time (e.g., a duration comparable to the time required for the
chemical to reach steady-state). In addition, BAF measurements should be based on adequate
spatial averaging of both tissue and water column concentrations for use in deriving 304(a)
criteria for the protection of human health.
For this reason, a BAF is defined in this Methodology as representing the ratio (in L/kg-
tissue) of a concentration of a chemical in tissue to its concentration in the surrounding water in
situations where the organism and its food are exposed and the ratio does not change
substantially over time (i.e., the ratio which reflects bioaccumulation at or near steady-state). A
bioconcentration factor (BCF) is the ratio (in L/kg-tissue) of the concentration of a substance in
tissue of an aquatic organism to its concentration in the ambient water, in situations where the
organism is exposed through the water only and the ratio does not change substantially over
time.
5.1.2 Goal of the National BAF
The goal of EPA's national BAF is to represent the long-term, average bioaccumulation
potential of a chemical in edible tissues of aquatic organisms that are commonly consumed by
humans throughout the United States. National BAFs are not intended to reflect fluctuations in
bioaccumulation over short time periods (e.g., a few days) because 304(a) human health criteria
are generally designed to protect humans from long-term exposures to waterborne chemicals.
National BAFs are also intended to account for some major chemical, biological, and ecological
attributes that can affect bioaccumulation in bodies of water across the United States. For
example, separate procedures are provided for deriving national BAFs depending on the type of
chemical (i.e., nonionic organic, ionic organic, inorganic and organometallic). In addition,
EPA's national BAFs are derived separately for each trophic level to account for potential
biomagnification of some chemicals in aquatic food webs and broad physiological differences
between trophic levels that may influence bioaccumulation. Because lipid content of aquatic
organisms and the amount of organic carbon in the water column have been shown to affect
bioaccumulation of nonionic organic chemicals, EPA's national BAFs are adjusted to reflect the
lipid content of commonly consumed fish and shellfish and the freely dissolved fraction of the
chemical in ambient water for these chemicals.
5.1.3 Changes to the 1980 Methodology
Numerous scientific advances have occurred in the area of bioaccumulation since the
publication of the 1980 Methodology for deriving AWQC for the protection of human health
(USEPA, 1980). These advances have significantly increased our ability to assess and predict
the bioaccumulation of chemicals in aquatic biota. As a result, EPA has revised the
bioaccumulation portion of the 1980 Methodology to reflect the current state of the science and
to improve accuracy in assessing bioaccumulation for setting 304(a) criteria for the protection of
human health. The changes contained in the bioaccumulation portion of the 2000 Human Health
Methodology are mostly designed to:
5O
-------
• Improve the ability to incorporate chemical exposure from sediments and aquatic food
webs in assessing bioaccumulation potential,
• Expand the ability to account for site-specific factors which affect bioaccumulation, and
• Incorporate new data and assessment tools into the bioaccumulation assessment process.
A summary of the key changes that have been incorporated into the bioaccumulation
portion of the 2000 Human Health Methodology and appropriate comparisons to the!980
Methodology are provided below.
5.1.3.1 Overall Approach
The 1980 Methodology for deriving 304(a) criteria for the protection of human health
emphasized the assessment of bioconcentration (uptake from water only) through the use of the
BCF. Based on the 1980 Methodology, measured BCFs were usually determined from
laboratory data unless field data demonstrated consistently higher or lower accumulation
compared with laboratory data. In these cases, "field BCFs" (currently termed field-measured
BAFs) were recommended for use. For lipophilic chemicals where lab or field-measured data
were unavailable, EPA recommended predicting BCFs from the octanol-water partition
coefficient and the following equation from Veith et al. (1979): "log BCF = (0.85 log Kow) -
0.70".
The 2000 Human Health Methodology revisions contained in this chapter emphasize the
measurement of bioaccumulation (uptake from water, sediment, and diet) through the use of the
BAF. Consistent with the 1980 Methodology, measured data are preferred over predictive
approaches for determining the BAF (i.e., field-measured BAFs are generally preferred over
predicted BAFs). However, the 2000 Human Health Methodology contains additional methods
for deriving a national BAF that were not available in 1980. The preference for using the BAF
methods also differs depending on the type and properties of the chemical. For example, the
BAF derivation procedure differs for each of three broadly defined chemical categories: (1)
nonionic organic, (2) ionic organic, and (3) inorganic and organometallic chemicals.
Furthermore, within the category of nonionic organic chemicals, different procedures are used to
derive the BAF depending on a chemicals' hydrophobicity and extent of chemical metabolism
that would be expected to occur in aquatic biota.
5.1.3.2 Lipid Normalization
In the 1980 Methodology, BCFs for lipophilic chemicals were normalized by the lipid
fraction in the tissue offish and shellfish used to determine the BCF. Lipid normalization
enabled BCFs to be averaged across tissues and organisms. Once the average lipid-normalized
BCF was determined, it was adjusted by the consumption-weighted lipid content of commonly
consumed aquatic organisms in the United States to obtain an overall consumption-weighted
BCF. A similar procedure has been retained in the 2000 Human Health Methodology, whereby
BAFs for nonionic organic chemicals are lipid normalized and adjusted by the consumption-
weighted lipid content of commonly consumed organisms to obtain a BAF for criteria
5-4
-------
calculations. However, the 2000 Human Health Methodology uses more up-to-date lipid data
and consumption data for deriving the consumption-weighted BAFs.
5.1.3.3 Bioavailability
Bioconcentration factors derived according to the 1980 Methodology were based on the
total concentration of the chemical in water, for both lipophilic and nonlipophilic chemicals. In
the 2000 Human Health Methodology, BAFs for nonionic organic chemicals are derived using
the most bioavailable fraction (i.e., the freely dissolved fraction) to account for the influence of
particulate and dissolved organic carbon on a chemical's bioavailability. Such BAFs are then
adjusted to reflect the expected bioavailability at the sites of interest (i.e., by adjusting for
organic carbon concentrations at the sites of interest). Procedures for accounting for the effect of
organic carbon on bioaccumulation were published previously by EPA under the Great Lakes
Water Quality Initiative (GLWQI or GLI) rulemaking (USEPA, 1995a,b). Bioavailability is also
considered in developing BAFs for the other chemical classes defined in the 2000 Human Health
Methodology (e.g., ionic organics, inorganics/organometallics) but is done so on a chemical-by-
chemical basis.
5.1.3.4 Trophic Level Considerations
In the 1980 Methodology, BCFs were determined and used for criteria derivation without
explicit regard to the trophic level of the aquatic organism (e.g., benthic filter feeder, forage fish,
predatory fish). Over the past two decades, much information has been assembled which
demonstrates that an organism's trophic position in the aquatic food web can have an important
effect on the magnitude of bioaccumulation of certain chemicals. In order to account for the
variation in bioaccumulation that is due to trophic position of the organism, the 2000 Human
Health Methodology recommends that BAFs be determined and applied on a trophic level-
specific basis.
5.1.3.5 Site-Specific Adjustments
The 1980 Methodology contained little guidance for making adjustments to the national
BCFs to reflect site- or region-specific conditions. The 2000 Human Health Methodology has
greatly expanded the guidance to States and authorized Tribes for making adjustments to
national BAFs to reflect local conditions. This guidance is contained in the Bioaccumulation
TSD. In the Bioaccumulation TSD, guidance and data are provided for adjusting national BAFs
to reflect the lipid content in locally consumed aquatic biota and the organic carbon content in
the waterbodies of concern. This guidance also allows the use of appropriate bioaccumulation
models for deriving site-specific BAFs. EPA also plans to publish detailed guidance on
designing and conducting field bioaccumulation studies for measuring BAFs and biota-sediment
accumulation factors (BSAFs). In general, EPA encourages States and authorized Tribes to
make site-specific modifications to EPA's national BAFs provided such adjustments are
scientifically defensible and adequately protect the designated use of the waterbody.
While the aforementioned revisions are new to EPA's Methodology for deriving national
304(a) criteria for the protection of human health, many of these refinements have been
5-5
-------
incorporated in prior Agency guidance and regulations. For example, the use of food chain
multipliers to account for the biomagnification of nonionic organic chemicals in aquatic food
webs when measured data are unavailable was introduced by EPA in three documents: Technical
Support Document for Water Quality-Based Toxics Control (USEPA, 1991), a draft document
entitled Assessment and Control of Bioconcentratable Contaminants in Surface Waters (USEPA,
1993), and in the Great Lakes Water Quality Initiative (GLI) (USEPA, 1995b). Similarly,
procedures for predicting BAFs using BSAFsand incorporating the effect of organic carbon on
bioavailability were used to derive water quality criteria under the GLI.
5.1.4 Organization of This Section
The methodology for deriving national BAFs for use in deriving National 304(a) Human
Health AWQC is provided in the following sections. Important terms used throughout this
chapter are defined in Section 5.2. Section 5.3 provides an overview of the BAF derivation
guidelines. Detailed procedures for deriving national BAFs are provided in Section 5.4 for
nonionic organic chemicals, in Section 5.5 for ionic organic chemicals, and in Section 5.6 for
inorganics and organometallic chemicals. Literature cited is provided in Section 5.7.
5.2 DEFINITIONS
The following terms and definitions are used throughout this chapter.
Bioaccumulation. The net accumulation of a substance by an organism as a result of uptake
from all environmental sources.
Bioconcentration. The net accumulation of a substance by an aquatic organism as a result of
uptake directly from the ambient water, through gill membranes or other external body surfaces.
Bioaccumulation Factor (BAF). The ratio (in L/kg-tissue) of the concentration of a substance
in tissue to its concentration in the ambient water, in situations where both the organism and its
food are exposed and the ratio does not change substantially over time. The BAF is calculated
as:
Ct
BAF = — (Equation 5-2)
where:
Ct = Concentration of the chemical in the specified wet tissue
CL = Concentration of chemical in water
5-6
-------
Bioconcentration Factor (BCF). The ratio (in L/kg-tissue) of the concentration of a substance
in tissue of an aquatic organism to its concentration in the ambient water, in situations where the
organism is exposed through the water only and the ratio does not change substantially over
time. The BCF is calculated as:
Ct
BCF = — (Equation 5-3)
where:
Ct = Concentration of the chemical in the specified wet tissue
Cw = Concentration of chemical in water
Baseline BAF (BAF[d). For nonionic organic chemicals (and certain ionic organic chemicals
where similar lipid and organic carbon partitioning behavior applies), a BAF (in L/kg-lipid) that
is based on the concentration of freely dissolved chemical in the ambient water and the lipid
normalized concentration in tissue.
Baseline BCF (BCF[d). For nonionic organic chemicals (and certain ionic organic chemicals
where similar lipid and organic carbon partitioning behavior applies), a BCF (in L/kg-lipid) that
is based on the concentration of freely dissolved chemical in the ambient water and the lipid
normalized concentration in tissue.
Biomagnification. The increase in tissue concentration of a chemical in organisms at successive
trophic levels through a series of predator-prey associations, primarily through the mechanism of
dietary accumulation.
Biomagnification Factor (BMF). The ratio (unitless) of the tissue concentration of a chemical
in a predator at a particular trophic level to the tissue concentration in its prey at the next lower
trophic level for a given waterbody and chemical exposure. For nonionic organic chemicals (and
certain ionic organic chemicals where similar lipid and organic carbon partitioning behavior
applies), a BMF can be calculated using lipid-normalized concentrations in the tissue of
organisms at two successive trophic levels as:
BMF(TL, „) = rt'n (Equation 5-4)
U« (TL, n-1)
where:
(TL, n) = Lipid-normalized concentration in appropriate tissue of predator organism at
a given trophic level (TL "n")
5-7
-------
Q (TL,n-i) = Lipid-normalized concentration in appropriate tissue of prey
organism at the next lower trophic level from the predator (TL "n-1")
For inorganic, organometallic, and certain ionic organic chemicals where lipid and organic
carbon partitioning does not apply, a BMP can be calculated using chemical concentrations in
the tissue of organisms at two successive trophic levels as:
BMFCTL, n) = ' n (Equation 5-5)
Ut (TL, n-1)
where:
Ct (TL, n) = Concentration in appropriate tissue of predator organism at trophic
level "n" (may be either wet weight or dry weight concentration so
long as both the predator and prey concentrations are expressed in the
same manner)
Ct (TL, n-i) = Concentration in appropriate tissue of prey organism at the next lower
trophic level from the predator (may be either wet weight or dry
weight concentration so long as both the predator and prey
concentrations are expressed in the same manner)
Biota-Sediment Accumulation Factor (BSAF). For nonionic organic chemicals (and certain
ionic organic chemicals where similar lipid and organic carbon partitioning behavior applies),
the ratio of the lipid-normalized concentration of a substance in tissue of an aquatic organism to
its organic carbon-normalized concentration in surface sediment (expressed as kg of sediment
organic carbon per kg of lipid), in situations where the ratio does not change substantially over
time, both the organism and its food are exposed, and the surface sediment is representative of
average surface sediment in the vicinity of the organism. The BSAF is defined as:
C.
BSAF = —— (Equation 5-6)
soc
where:
Cc = The lipid-normalized concentration of the chemical in tissues of the biota
(|ig/g lipid)
Csoc = The organic carbon-normalized concentration of the chemical in the
surface sediment (|ig/g sediment organic carbon)
Depuration. The loss of a substance from an organism as a result of any active or passive
process.
5-S
-------
Food Chain Multiplier (FCM). For nonionic organic chemicals (and certain ionic organic
chemicals where similar lipid and organic carbon partitioning behavior applies), the ratio of a
baseline BAFf for an organism of a particular trophic level to the baseline BCFf (usually
determined for organisms in trophic level one). For inorganic, organometallic, and certain ionic
organic chemicals where lipid and organic carbon partitioning does not apply, a FCM is based on
total (wet or dry weight) concentrations of the chemical in tissue.
Freely Dissolved Concentration. For nonionic organic chemicals, the concentration of the
chemical that is dissolved in ambient water, excluding the portion sorbed onto paniculate or
dissolved organic carbon. The freely dissolved concentration is considered to represent the most
bioavailable form of an organic chemical in water and, thus, is the form that best predicts
bioaccumulation. The freely dissolved concentration can be determined as:
(ffd) (Equation 5-7)
where:
C^d = Freely dissolved concentration of the organic chemical in ambient water
C^ = Total concentration of the organic chemical in ambient water
ffd = Fraction of the total chemical in ambient water that is freely dissolved
Hydrophilic. A term that refers to the extent to which a chemical is attracted to partitioning into
the water phase. Hydrophilic organic chemicals have a greater tendency to partition into polar
phases (e.g., water) compared to chemicals of hydrophobic chemicals.
Hydrophobic. A term that refers to the extent to which a chemical avoids partitioning into the
water phase. Highly hydrophobic organic chemicals have a greater tendency to partition into
nonpolar phases (e.g., lipid, organic carbon) compared with chemicals of lower hydrophobicity.
Lipid-normalized Concentration (Ct). The total concentration of a contaminant in a tissue or
whole organism divided by the lipid fraction in that tissue or whole organism. The lipid-
normalized concentration can be calculated as:
Ct
(Equation 5-8)
where:
Ct = Concentration of the chemical in the wet tissue (either whole organism or
specified tissue)
5-9
-------
fc = Fraction lipid content in the organism or specified tissue
Octanol-water Partition Coefficient (Kow). The ratio of the concentration of a substance in the
n-octanol phase to its concentration in the aqueous phase in an equilibrated two-phase octanol-
water system. For log Kow, the log of the octanol-water partition coefficient is a base 10
logarithm.
Organic Carbon-normalized Concentration (Csoc). For sediments, the total concentration of a
contaminant in sediment divided by the fraction of organic carbon in sediment. The organic
carbon-normalized concentration can be calculated as:
Cs
Csoc = ~ (Equation 5-9)
where:
Cs = Concentration of chemical in sediment
foc = Fraction organic carbon in sediment
Uptake. Acquisition by an organism of a substance from the environment as a result of any
active or passive process.
5.3 FRAMEWORK FOR DETERMINING NATIONAL BIO ACCUMULATION
FACTORS
5.3.1 Four Different Methods
Bioaccumulation factors used to derive national BAFs can be measured or predicted
using some or all of the following four methods, depending on the type of chemical and its
properties. These methods are:
(1) a measured BAF obtained from a field study (i.e., a field-measured BAF);
(2) a BAF predicted from a field-measured BSAF;
(3) a BAF predicted from a laboratory-measured BCF (with or without adjustment by an
FCM); and
(4) a BAF predicted from a chemical's octanol-water partition coefficient (K ow ), with or
without adjustment using an FCM.
5-10
-------
A brief summary of each of the four methods is provided below. Additional details on
the use of these four methods is provided in Section 5.4 (for nonionic organics), Section 5.5 (for
ionic organics) and Section 5.6 (for inorganics and organometallics).
1. Field-Measured BAF. Use of a field-measured BAF, which is the most direct measure
of bioaccumulation, is the only method that can be used to derive a national BAF for all
types of chemicals (i.e., nonionic organic, ionic organic, and inorganic and
organometallic chemicals). A field-measured BAF is determined from a field study using
measured chemical concentrations in the aquatic organism and its surrounding water.
Because field studies are conducted in natural aquatic ecosystems, a field-measured BAF
reflects an organism's exposure to a chemical through all relevant exposure pathways
(i.e., water, sediment, and diet). A field-measured BAF also reflects any metabolism of a
chemical that might occur in the aquatic organism or its food web. Therefore, field-
measured BAFs are appropriate for all chemicals, regardless of the extent of chemical
metabolism in biota.
2. Field-measured BSAF. For nonionic organic chemicals (and certain ionic organic
chemicals where similar lipid and organic carbon partitioning behavior applies), a BAF
can also be predicted from BSAFs. A BSAF is similar to a field-measured BAF in that
the concentration of a chemical in biota is measured in the field and reflects an
organism's exposure to all relevant exposure routes. A BSAF also reflects any chemical
metabolism that might occur in the aquatic organism or its food web. However, unlike a
field-measured BAF which references the biota concentration to the water concentration,
a BSAF references the biota concentration to the sediment concentration. Use of the
BSAF procedure is restricted to organic chemicals which are classified as being
moderately to highly hydrophobic.
3. Lab-measured BCF. A laboratory-measured BCF can also be used to estimate a BAF
for organic and inorganic chemicals. However, unlike a field-measured BAF or a BAF
predicted from a field-measured BSAF, a laboratory-measured BCF only reflects the
accumulation of chemical through the water exposure route. Laboratory-measured BCFs
may therefore under estimate BAFs for chemicals where accumulation from sediment or
dietary sources is important. In these cases, laboratory-measured BCFs can be multiplied
by a FCM to reflect accumulation from non-aqueous (i.e., food chain) pathways of
exposure. Since a laboratory-measured BCF is determined using the measured
concentration of a chemical in an aquatic organism and its surrounding water, a
laboratory-measured BCF reflects any metabolism of the chemical that occurs in the
organism, but not in the food web.
4. K,,w. A chemical's octanol-water partition coefficient, or Kow, can also be used to predict
a BAF for nonionic organic chemicals. This procedure is appropriate only for nonionic
organic chemicals (and certain ionic organic chemicals where similar lipid and organic
carbon partitioning behavior applies). The Kow has been extensively correlated with the
BCF for nonionic organic chemicals that are poorly metabolized by aquatic organisms.
Therefore, where substantial metabolism is known to occur in biota, the Kmi, is not used
5-11
-------
to predict the BAF. For nonionic organic chemicals where chemical exposure through
the food web is important, use of the Kow alone will under predict the BAF. In such
cases, the Kow is adjusted with a FCM similar to the BCF procedure above.
5.3.2 Overview of BAF Derivation Framework
Although up to four methods can be used to derive a BAF as described in the previous
section, it is evident that these methods do not apply equally to all types of chemicals. In
addition, experience demonstrates that the required data will usually not be available to derive a
BAF value using all of the applicable methods. As a result, EPA has developed the following
guidelines to direct users in selecting the most appropriate method(s) for deriving a national
BAF.
Figure 5-1 shows the overall framework of EPA's national BAF methodology. This
framework illustrates the major steps and decisions that will ultimately lead to calculating a
national BAF using one of six hierarchical procedures shown at the bottom of Figure 5-1. Each
procedure contains a hierarchy of the BAF derivation methods discussed above, the composition
of which depends on the chemical type and certain chemical properties (e.g., its degree of
hydrophobicity and expected degree of metabolism and biomagnification). The number assigned
to each BAF method within a procedure indicates its general order of preference for deriving a
national BAF value. The goal of the framework and accompanying guidelines is to enable full
use of available data and methods for deriving a national BAF value while appropriately
restricting the use of certain methods to reflect their inherent limitations.
The first step in the framework is to define the chemical of concern. As described in
Section 5.3.3, the chemical used to derive the national BAF should be consistent with the
chemical used to derive the critical health assessment value. The second step is to collect and
review all relevant data on bioconcentration and bioaccumulation of the chemical of concern
(see Section 5.3.4). Once pertinent data are reviewed, the third step is to classify the chemical of
concern into one of three broadly defined chemical categories: (1) nonionic organic chemicals,
(2) ionic organic chemicals, and (3) and inorganic and organometallic chemicals. Guidance for
classifying chemicals into these three categories is provided in Section 5.3.5.
After a chemical has been classified into one of the three categories, other information is
used to select one of six hierarchical procedures to derive the national BAF. The specific
procedures for deriving a BAF for each chemical group are discussed in Section 5.4 for nonionic
organics, Section 5.5 for ionic organics, and Section 5.6 for inorganics and organometallics.
5-12
-------
Figure 5-1. Framework for Deriving a National BAF
( DEFINE CHEMICAL }
\^ OF CONCERN J
( COLLECTS.REVIEW
\ DATA
1
r
Nonionic Organic
i
CLASSIFY CHEMICAL
OF CONCERN
Moderate-High
(Log Kow >_A)
^
f
PROCEDURE #1
1. Field BAF
2. BSAF
3. Lab BCF"FCM
4. Kow"FCM
^
r
Inorganic &
Organometalic
4 BIOMAGNIFICA TION7J
PROCEDURE #5
1. Field BAF or
Lab BCF
PROCEDURE #6
1. Field BAF
2. Lab BCF"FCM
5-13
-------
Detailed guidance concerning the first three steps of the derivation process (i.e, defining the
chemical of concern, collecting and reviewing data, and classifying the chemical of concern) is
provided in the following three sections.
5.3.3 Defining the Chemical of Concern
Defining the chemical of concern is the first step in deriving a national BAF. This step
involves precisely defining the form(s) of the chemical upon which the national BAF value will
be derived. Although this step is usually straightforward for single chemicals, complications can
arise when the chemical of concern occurs as a mixture. The following guidelines should be
followed for defining the chemical of concern.
1. Information for defining the chemical of concern should be obtained from the health and
exposure assessment portions of the criteria derivation effort. The chemical(s) used to
derive the national BAF should be consistent with the chemical(s) used to derive the
reference dose (RfD), point of departure/uncertainty factor (POD/UF), or cancer potency
factor.
2. In most cases, the RfD, POD/UF, or cancer potency factor will be based on a single
chemical. In some cases, the RfD, POD/UF, or cancer potency factor will be based on a
mixture of compounds, typically within the same chemical class (e.g., toxaphene,
chlordane). In these situations, the national BAF should be derived in a manner that is
consistent with the mixture used to express the health assessment.
a. If sufficient data are available to reliably assess the bioaccumulation of each
relevant compound contained in the mixture, then the national BAF(s) should be
derived using the BAFs for the individual compounds of the mixture and
appropriately weighted to reflect the mixture composition used to establish the
RfD, POD/UF, or cancer potency factor. An example of this approach is shown
in the derivation of BAFs for PCBs in the GLI Rulemaking (USEPA, 1997).
b. If sufficient data are not available to reliably assess the bioaccumulation of
individual compounds of the mixture, then the national BAF(s) should be derived
using BAFs for the same or appropriately similar chemical mixture as that used to
establish the RfD, POD/UF, or cancer potency value.
5.3.4 Collecting and Reviewing Data
The second step in deriving a national BAF is to collect and review all relevant
bioaccumulation data for the chemical of concern. The following guidance should be followed
for collecting and reviewing bioaccumulation data for deriving national BAFs.
1. All data on the occurrence and accumulation of the chemical of concern in aquatic
animals and plants should be collected and reviewed for adequacy.
5-14
-------
2. A comprehensive literature search strategy should be used for gathering
bioaccumulation-related data. An example of a comprehensive literature search strategy
is provided in the Bioaccumulation TSD.
3. All data that are used should contain sufficient supporting information to indicate that
acceptable measurement procedures were used and that the results are probably reliable.
In some cases it may be appropriate to obtain additional written information from the
investigator.
4. Questionable data, whether published or unpublished, should not be used. Guidance for
assessing the acceptability of bioaccumulati on and bioconcentration studies is found in
Sections 5.4, 5.5, and 5.6.
5.3.5 Classifying the Chemical of Concern
The next step in deriving a national BAF consists of classifying the chemical of concern
into one of three categories: nonionic organic, ionic organic, and inorganic and organometallic
(Figure 5-1). This step helps to determine which of the four methods described in Section 5.3.1
are appropriate for deriving BAFs. The following guidance applies for classifying the chemical
of concern.
1. Nonionic Organic Chemicals. For the purposes of the 2000 Human Health
Methodology, nonionic organic chemicals are those organic compounds that do not
ionize substantially in natural bodies of water. These chemicals are also referred to as
neutral or nonpolar organics in the scientific literature. Due to their neutrality, nonionic
organic chemicals tend to associate with other neutral (or near neutral) compartments in
aquatic ecosystems (e.g., lipid, organic carbon). Examples of nonionic organic chemicals
which have been widely studied in terms of their bioaccumulati on include
polychlorinated biphenyls (PCBs), polychlorinated dibenzo-p-dioxins and furans, many
chlorinated pesticides, and polynuclear aromatic hydrocarbons (PAHs). Procedures for
deriving a national BAF for nonionic organic chemicals are provided in Section 5.4.
2. Ionic Organic Chemicals. For the purposes of the 2000 Human Health Methodology,
ionic organic chemicals are considered to include those chemicals that contain functional
groups with exchangeable protons such as hydroxyl, carboxylic, and sulfonic groups and
functional groups that readily accept protons such as amino and aromatic heterocyclic
nitrogen (pyridine) groups. Ionic organic chemicals undergo ionization in water, the
extent of which depends on pH and the pKa of the chemical. Because the ionized species
of these chemicals behave differently from the neutral species, separate guidance is
provided for deriving BAFs for ionic organic chemicals. Procedures for deriving
national BAFs for ionic organic chemicals are provided in Section 5.5.
3. Inorganic and Organometallic Chemicals. The inorganic and organometallic category
is considered to include inorganic minerals, other inorganic compounds and elements,
metals (e.g., copper, cadmium, chromium, zinc), metalloids (selenium, arsenic) and
5-15
-------
organometallic compounds (e.g., methylmercury, tributyltin, tetraalkyllead). Procedures
for deriving BAFs for inorganic and organometallic chemicals are provided in Section
5.6.
5.4 NATIONAL BIO ACCUMULATION FACTORS FOR NONIONIC ORGANIC
CHEMICALS
5.4.1 Overview
This section contains the methodology for deriving national BAFs for nonionic organic
chemicals as defined in Section 5.3.5. The four general steps of this methodology are:
1. Selecting the BAF derivation procedure,
2. Calculating individual baseline BAFf s,
3. Selecting the final baseline BAFf s, and
4. Calculating the national BAFs from the final baseline BAFf s.
A schematic of this four-step process is shown in Figure 5-2.
Step 1 of the methodology (selecting the BAF derivation procedure) determines which of
the four BAF procedures summarized in Figure 5-1 will be appropriate for deriving the national
BAF. Step 2 involves calculating individual, species-specific BAFf s using all of the methods
available within the selected BAF derivation procedure. Calculating the individual baseline
BAFf s involves using data from the field site or laboratory where the original data were
collected to account for site-specific factors which affect the bioavailability of the chemical to
aquatic organisms (e.g., lipid content of study organisms and freely dissolved concentration in
study water). Step 3 of the methodology consists of selecting the final baseline BAFf s from the
individual baseline BAFf s by taking into account the uncertainty in the individual BAFs and the
data preference hierarchy selected in Step 1. The final step is to calculate a BAF (or BAFs) that
will be used in the derivation of 304(a) criteria (i.e., referred to as the national BAF). This step
involves adjusting the final baseline BAFf (s) to reflect certain factors that affect bioavailablity
of the chemical to aquatic organisms in waters to which the national 304(a) criteria will apply
(e.g., the freely dissolved fraction expected in U.S. waters and the lipid content of consumed
aquatic organisms). Baseline BAFf s are not used directly in the derivation of the 304(a) criteria
because they do not reflect the conditions that affect bioavailability in U.S. waters.
Section 5.4.2 below provides detailed guidance for selecting the appropriate BAF
derivation procedure (Step 1 of the process). Guidance on calculating individual baseline
BAFf s, selecting the final baseline BAF, and calculating the national BAF (Steps 2 through 4 of
the process) is provided in separate sections under each of the four BAF derivation procedures.
5-16
-------
5-17
-------
Figure 5-2. BAF Derivation for Nonionic Organic Chemicals
f CHEMICAL
PROPERTY &
\^METABOLISM DATl
f LIPID CONTENT OF
( STUDY ORGANISM
DATA PREFERENCE
HIERARCHY
Stepl.
Select BAF
Procedure
BIOMAGNIFICATION }
DATA j
Step 2.
Calculate
Individual
Baseline BAFs
FREELY DISSOLVED\
FRACTION IN STUDY
WATER J
Step 3.
Select Final
Baseline BAF(s)
/ LIPID CONTENT OF
I CONSUMED
\ ORGANISMS j
UNCERTAINTYIN }
BASELINE BAFs J
Step 4.
Calculate
National BAF(s)
FREELY DISSOLVED\
FRACTION IN AWQC
WATERS J
5-18
-------
5.4.2 Selecting the BAF Derivation Procedure
This section describes the decisions that should be made to select one of the four
available hierarchical procedures for deriving a national BAF for nonionic organic chemicals
(Procedures #1 through #4 of Figure 5-1). As shown in Figure 5-1, two decision points exist in
selecting the BAF derivation procedure. The first decision point requires knowledge of the
chemical's hydrophobicity (i.e., the Kow of the chemical). Guidance for selecting the Kow for a
chemical is provided in the Bioaccumulation TSD. The Kow provides an initial basis for
assessing whether biomagnification may be a concern for nonionic organic chemicals. The
second decision point is based on the rate of metabolism for the chemical in the target organism.
Guidance for assessing whether a high or low rate of metabolism is likely for a chemical of
concern is provided below in Section 5.4.2.3. With the appropriate information for these two
decision points, the BAF derivation procedure should be selected using the following guidelines.
5.4.2.1 Chemicals with Moderate to High Hydrophobicity
1. For the purposes of the 2000 Human Health Methodology, nonionic organic chemicals
with log Kow values equal to or greater than 4.0 should be classified as moderately to
highly hydrophobic. For moderately to highly hydrophobic nonionic organic chemicals,
available data indicate that exposure through the diet and other non-aqueous routes can
become important in determining chemical residues in aquatic organisms (e.g., Russell et
al., 1999; Fisk et al., 1998; Oliver andNiimi, 1983; Oliver andNiimi, 1988; Niimi, 1985;
Swackhammer and Kites, 1988). Dietary and other non-aqueous exposure can become
extremely important for those nonionic organic chemicals that are poorly metabolized by
aquatic biota (e.g., certain PCB congeners, chlorinated pesticides, and poly chlorinated
dibenzo-p-dioxins and furans).
2. Procedure #1 should be used to derive national BAFs for moderately to highly
hydrophobic nonionic organic chemicals in cases where:
(a) the rate of chemical metabolism by target aquatic organisms is expected to be
sufficiently low such that biomagnification is of concern, or
(b) the rate of chemical metabolism by target aquatic organisms is not sufficiently
known.
Procedure #1 accounts for non-aqueous exposure and the potential for biomagnification
in aquatic food webs through the use of field-measured values for bioaccumulation (i.e.,
field measured BAF or BSAF) and FCMs when appropriate field data are unavailable.
Guidance on deriving national BAFs using Procedure #1 is found below in Section 5.4.3.
3. Procedure #2 should be used to derive the national BAFs for moderately to highly
hydrophobic nonionic organic chemicals in cases where:
(a) the rate of chemical metabolism by target aquatic organisms is expected to be
sufficiently high such that biomagnification is not of concern.
5-19
-------
Procedure #2 relaxes the requirement of using FCMs and eliminates the use of Kow-based
estimates of the BAF, two procedures that are most appropriate for poorly metabolized
nonionic organic chemicals. Guidance on deriving national BAFs using Procedure #2 is
found below in Section 5.4.4.
5.4.2.2 Chemicals with Low Hydrophobicity
1. For the purposes of these guidelines, nonionic organic chemicals with log Kow values less
than 4.0 should be classified as exhibiting low hydrophobicity. For nonionic organic
chemicals that exhibit low hydrophobicity (i.e., log Kow < 4.0), available information
indicates that non-aqueous exposure to these chemicals is not likely to be important in
determining chemical residues in aquatic organisms (e.g., Fisk et al., 1998; Gobas et al.,
1993; Connolly andPedersen, 1988; Thomann, 1989). For this group of chemicals,
laboratory-measured BCFs and Kow-predicted BCFs do not require adjustment with
FCMs for determining the national BAF (Procedures #3 and #4), unless other appropriate
data indicate differently.
Other appropriate data include studies clearly indicating that non-aqueous exposure is
important such that use of a BCF would substantially underestimate residues in aquatic
organisms. In these cases, Procedure #1 should be used to derive the BAF for nonionic
organic chemicals with log Kow < 4.0. Furthermore, the data supporting the Kow
determination should be carefully reviewed for accuracy and appropriate interpretation,
since the apparent discrepancy may be due to errors in determining Kow.
2. Procedure #3 should be used to derive national BAFs for nonionic organic chemicals of
low hydrophobicity in cases where:
(a) the rate of chemical metabolism by target aquatic organisms is expected to be
negligible, such that tissue residues of the chemical of concern are not
substantially reduced compared to an assumption of no metabolism, or
(b) the rate of chemical metabolism by target aquatic organisms is not sufficiently
known.
Procedure #3 includes the use of Kow-based estimates of the BCF to be used when lab or
field data are absent. Guidance on deriving national BAFs using Procedure #3 is found
below in Section 5.4.5.
3. Procedure #4 should be used to derive national BAFs for nonionic organic chemicals of
low hydrophobicity in cases where:
(a) the rate of chemical metabolism by target aquatic organisms is expected to be
sufficiently high, such that tissue residues of the chemical of concern are
substantially reduced compared with an assumption of no metabolism.
5-20
-------
Procedure #4 eliminates the option of using Kow-based estimates of the BAF because the
Kow may over-predict accumulation when a chemical is metabolized substantially by an
aquatic organism. Guidance on deriving national BAFs using Procedure #4 is found
below in Section 5.4.6.
5.4.2.3 Assessing Metabolism
Currently, assessing the degree to which a chemical is metabolized by aquatic organisms
is confounded by a variety of factors. First, conclusive data on chemical metabolism in aquatic
biota are largely lacking. Such data include whole organism studies where the metabolic rates
and breakdown products are quantified in fish and other aquatic organisms relevant to human
consumption. However, the majority of information on metabolism is derived from in vitro liver
microsomal preparations in which primary and secondary metabolites may be identified and their
rates of formation may or may not be quantified. Extrapolating results from in vitro studies to
the whole organism involves considerable uncertainty. Second, there are no generally accepted
procedures for reliably predicting chemical metabolism by aquatic organisms in the absence of
measured data. Third, the rate at which a chemical is metabolized by aquatic organisms can be
species and temperature dependent. For example, PAHs are known to be metabolized readily by
vertebrate aquatic species (primarily fish), although at rates much less than those observed for
mammals. However, the degree of metabolism in invertebrate species is generally much less
than the degree in vertebrate species (James, 1989). One hypothesis for this difference is that the
invertebrate species lack the detoxifying enzymes and pathways that are present in many
vertebrate species.
Given the current limitations on assessing the degree of chemical metabolism by aquatic
organisms, the assessment of metabolism should be made on a case-by-case basis using a
weight-of-evidence approach. When assessing a chemical's likelihood to undergo substantial
metabolism in a target aquatic organism, the following data should be carefully evaluated:
(1) in vivo chemical metabolism data,
(2) bioconcentration and bioaccumulation data,
(3) data on chemical occurrence in target aquatic biota, and
(4) in vitro chemical metabolism data.
1. In vivo Data. In vivo data on metabolism in aquatic organisms are from studies of
chemical metabolism using whole organisms. These studies are usually conducted using
large fish from which blood, bile, urine, and individual tissues can be collected for the
identification and quantification of metabolites formed over time. In vivo studies are
considered the most useful for evaluating a chemical's degree of metabolism in an
organism because both oxidative (Phase I) and conjugative (Phase II) metabolism can be
assessed in these studies. Mass-balance studies, in which parent compound elimination is
quantified separately from biotransformation and elimination of metabolites, allow
calculation of conversion rate of parent to metabolite as well as metabolite elimination.
This information might be used to estimate loss due to metabolism separately from that
due to elimination of the parent compound for adjustment of Kow-predicted BAFs.
However, due to the analytical and experimental challenges these studies pose, data of
5-21
-------
this type are limited. Less rigorous in vivo metabolism studies might include the use of
metabolic blockers to demonstrate the influence of metabolism on parent compound
kinetics. However, caution should be used in interpretation of absolute rates from these
data due to the lack of specificity of mammalian derived blockers in aquatic species
(Miranda et al., 1998).
2. Bioconcentration or Bioaccumulation Data. Data on chemical bioconcentration or
bioaccumulation in aquatic organisms can be used indirectly for assessing metabolism.
This assessment involves comparing acceptable lab-measured BCFs or field-measured
BAFs (after converting to baseline values using procedures below) with the chemical's
predicted value based on Kow. The theoretical basis of bioconcentration and
bioaccumulation for nonionic organic chemicals indicates that a chemical's baseline BCF
should be similar to its Kow-predicted value if metabolism is not occurring or is minimal
(see the Bioaccumulation TSD). This theory also indicates that baseline BAFs should be
similar to or higher than the Kow for poorly metabolized organic chemicals, with highly
hydrophobic chemicals often exhibiting higher baseline BAFs than Kow values. Thus, if a
chemical's baseline BCF or BAF is substantially lower than its Kow, this may be an
indication that the chemical is being metabolized by the aquatic organism of concern.
Note, however, that this difference may also indicate problems in the experimental design
or analytical chemistry, and that it may be difficult to discern the difference.
3. Chemical Occurrence Data. Although by no means definitive, data on the occurrence
of chemicals in aquatic biota (i.e., residue studies) may offer another useful line of
evidence for evaluating a chemical's likelihood to undergo substantial metabolism. Such
studies are most useful if they have been conducted repeatedly over time and over wide
geographical areas. Such studies might indicate a chemical is poorly metabolized if data
show that the chemical is being biomagnified in the aquatic food web (i.e., higher lipid-
normalized residues in successive trophic levels). Conversely, such studies might
indicate a chemical is being metabolized substantially if residue data show a decline in
residues with increasing trophic level. Again, other reasons for increases or decreases in
concentrations with increasing trophic level might exist and should be carefully evaluated
(e.g., incorrect food web assumptions, differences in exposure concentrations).
4. In vitro Data. In vitro metabolism data include data from studies where specific sub-
cellular fractions (e.g., microsomal, cytosolic), cells, or tissues from an organism are
tested outside the body (i.e., in test-tubes, cell- or tissue-culture). Compared with in vivo
studies of chemical metabolism in aquatic organisms, in vitro studies are much more
plentiful in the literature, with the majority of studies characterizing oxidative (Phase I)
reactions de-coupled from conjugative (Phase II) metabolism. Cell, tissue, or organ level
in vitro studies are less common but provide a more complete assessment of metabolism.
While such studies are particularly useful for identifying the pathways, rates of
formation, and metabolites formed, as well as the enzymes involved and differences in
the temperature dependence of metabolism across aquatic species, they suffer from
uncertainty when results are extrapolated to the whole organism. This uncertainty results
from the fact that dosimetry (i.e., delivery of the toxicant to, and removal of metabolite
5-22
-------
from, the target tissue) cannot currently be adequately reproduced in the laboratory or
easily modeled.
When assessing chemical metabolism using the above information, the following
guidelines apply.
a. A finding of substantial metabolism should be supported by two or more lines of
evidence identified using the data described above.
b. At least one of the lines of evidence should be supported by either in vivo metabolism
data or acceptable bioconcentration or bioaccumulation data.
c. A finding of substantial metabolism in one organism should not be extrapolated to
another organism or another group of organisms unless data indicate similar metabolic
pathways exist (or are very likely to exist) in both organisms. In vitro data may be
particularly useful in cross-species extrapolations.
d. Finally, in situations where sufficient data are not available to properly assess the
likelihood of significant metabolism in aquatic biota of concern, the chemical should be
assumed to undergo little or no metabolism. This assumptions reflects a policy decision
by EPA to err on the side of public health protection when sufficient information on
metabolism is lacking.
5.4.3 Deriving National BAFs Using Procedure #1
This section contains guidance for calculating national BAFs for nonionic organic
chemicals using Procedure #1 shown in Figure 5-1. The types of nonionic organic chemicals for
which Procedure #1 is most appropriate are those that are classified as moderately to highly
hydrophobic and subject to low (or unknown) rates of metabolism by aquatic biota (see Section
5.4.2 above). Non-aqueous contaminant exposure and subsequent biomagnification in aquatic
food webs are of concern for chemicals that are classified in this category. Some examples of
nonionic organic chemicals for which Procedure #1 is considered appropriate include:
tetra-, penta- & hexachlorobenzenes;
PCBs;
octachlorostyrene;
hexachl orobutadi ene;
endrin, dieldrin, aldrin;
mirex, photomirex;
DDT, DDE, ODD; and
heptachlor, chlordane, nonachlor.
Under Procedure #1, the following four methods may be used in deriving a national BAF:
using a BAF from an acceptable field study (i.e., a field-measured BAF);
predicting a BAF from an acceptable field-measured BSAF;
5-23
-------
• predicting a BAF from an acceptable laboratory-measured BCF and FCM; and
• predicting a BAF from an acceptable Kow and FCM.
As shown in Figure 5-2, once the derivation procedure has been selected, the next steps
in deriving a national BAF for a given trophic level include: calculating individual baseline
BAFf s (step 2), selecting the final baseline BAFf (step 3), and calculating the national BAF
from the final baseline BAFf (step 4). Each of these three steps is discussed separately below.
5.4.3.1 Calculating Individual Baseline BAF[ds
Calculating an individual baseline BAFf involves normalizing the field-measured BAF^
(or laboratory-measured BCF£) which are based on total concentrations in tissue and water by
the lipid content of the study organisms and the freely dissolved concentration in the study water.
Both the lipid content in the organism and the freely dissolved concentration (as influenced by
organic carbon in water) have been shown to be important factors that influence the
bioaccumulation of nonionic organic chemicals (e.g., Mackay, 1982; Connolly and Pederson,
1988; Thomann, 1989, Suffet et al., 1994). Therefore, baseline BAFf s (which are expressed on
a freely dissolved and lipid-normalized basis) are considered more amenable to extrapolating
between different species and bodies of water compared to BAFs expressed using the total
concentration in the tissue and water. Because bioaccumulation can be strongly influenced by
the trophic position of aquatic organisms (either due to biomagnification or physiological
differences), extrapolation of baseline BAFf s should not be performed between species of
different trophic levels.
1. For each species for which acceptable data are available, calculate all possible baseline
BAFf s using each of the four methods shown above for Procedure #1.
2. Individual baseline BAFf s should be calculated from field-measured BAFjs, field-
measured BSAFs, laboratory BCF^s, and the Kow according to the following procedures.
A. Baseline BAF^sfrom Field-Measured BAFs
A baseline BAFf should be calculated from each field-measured BAF^ using information
on the lipid fraction in the tissue of concern for the study organism and the fraction of the total
chemical that is freely dissolved in the study water.
1. Baseline BAF[d Equation. For each acceptable field-measured BAF^, calculate a
baseline BAFf using the following equation:
Baseline BAFt =
Measured
T
ffd
- 1
£
(Equation 5-10)
where:
5-24
-------
Baseline BAFf = BAF expressed on a freely dissolved and lipid-normalized
basis
Measured BAF] = BAF based on total concentration in tissue and water
fc = Fraction of the tissue that is lipid
ffd = Fraction of the total chemical that is freely dissolved in the
ambient water
The technical basis of Equation 5-10 is provided in the Bioaccumulation TSD. Guidance for
determining each component of Equation 5-10 is provided below.
2. Determining the Measured BAF].. The field-measured BAF] shown in Equation 5-10
should be calculated based on the total concentration of the chemical in the appropriate
tissue of the aquatic organism and the total concentration of the chemical in ambient
water at the site of sampling. The equation to derive a measured BAF] is:
t ct
Measured BAFT = — (Equation 5-11)
where:
Ct = Total concentration of the chemical in the specified wet tissue
Cw = Total concentration of chemical in water
The data used to calculate a field-measured BAF] should be reviewed thoroughly to
assess the quality of the data and the overall uncertainty in the BAF value. The following
general criteria apply in determining the acceptability of field-measured B AFs that are
being considered for deriving national BAFs using Procedure #1.
a. Aquatic organisms used to calculate a field-measured BAF] should be
representative of aquatic organisms that are commonly consumed in the United
States. An aquatic organism that is not commonly consumed in the United States
can be used to calculate an acceptable field-measured BAF] provided that the
organism is considered to be a reasonable surrogate for a commonly consumed
organism. Information on the ecology, physiology, and biology of the organism
should be reviewed when assessing whether an organism is a reasonable surrogate
of a commonly consumed organism.
b. The trophic level of the study organism should be determined by taking into
account its life stage, diet, size, and the food web structure at the study location.
Information from the study site (or similar sites) is preferred when evaluating
trophic status. If such information is lacking, general information for assessing
trophic status of aquatic organisms can be found in USEPA (2000a,b,c).
5-25
-------
c. The percent lipid of the tissue used to determine the field-measured BAFj should
be either measured or reliably estimated to permit lipid-normalization of the
chemical's tissue concentration.
d. The study from which the field-measured BAFj is derived should contain
sufficient supporting information from which to determine that tissue and water
samples were collected and analyzed using appropriate, sensitive, accurate, and
precise analytical methods.
e. The site of the field study should not be so unique that the BAF cannot be
reasonably extrapolated to other locations where the BAF and resulting criteria
will apply.
f. The water concentration(s) used to derive the BAF should reflect the average
exposure of the aquatic organism that corresponds to the concentration measured
in its tissue of concern. For nonionic organic chemicals, greater temporal and
spatial averaging of chemical concentrations is required as the Kow increases. In
addition, as variability in water concentrations increase, greater temporal and
spatial averaging is also generally required. Greater spatial averaging is also
generally required for more mobile organisms.
g. The concentrations of particulate organic carbon and dissolved organic carbon in
the study water should be measured or reliably estimated.
EPA is currently developing guidance for designing and conducting field studies for
determining field-measured BAFjs, including recommendations for minimum data
requirements. A more detailed discussion of factors that should be considered when
determining field-measured BAFjs is provided in the Bioaccumulation TSD.
Determining the Fraction Freely Dissolved (ffd). As illustrated by Equation 5-10, the
fraction of the nonionic organic chemical that is freely dissolved in the study water is
required for calculating a baseline BAFf from a field-measured BAF^. The freely
dissolved fraction is the portion of the nonionic organic chemical that is not bound to
particulate organic carbon or dissolved organic carbon. Together, the concentration of a
nonionic organic chemical that is freely dissolved, bound to dissolved organic carbon,
and bound to particulate organic carbon constitute its total concentration in water. As
discussed further in the Bioaccumulation TSD, the freely dissolved fraction of a chemical
is considered to be the best expression of the bioavailable form of nonionic organic
chemicals to aquatic organisms (e.g., Suffet et al., 1994; USEPA, 1995b). Because the
fraction of a nonionic organic chemical that is freely dissolved may vary among different
bodies of water as a result of differences in dissolved and particulate organic carbon in
the water, the bioavailability of the total chemical concentration in water is expected to
vary from one body of water to another. Therefore, BAFs which are based on the freely
dissolved concentration in water (rather than the total concentration in water) are
considered to be more reliable for extrapolating and aggregating BAFs among different
bodies of water. Currently, availability of BAFs based on measured freely dissolved
5-26
-------
concentrations is very limited, partly because of difficulties in analytically measuring the
freely dissolved concentration. Thus, if a BAF based on the total water concentration is
reported in a given study, the fraction of the chemical that is freely dissolved should be
predicted using information on the organic carbon content in the study water.
a. Equation for Determining the Freely Dissolved Fraction. If reliable measured
data are unavailable to directly determine the freely dissolved fraction of the
chemical in water, the freely dissolved fraction should be estimated using the
following equation.
ffd = [1 + (POC • Kow) + (DOC • 0.08 • Kow)] (Equation 5-12)
where:
POC = concentration of particulate organic carbon (kg/L)
DOC = concentration of dissolved organic carbon (kg/L)
Kow = n-octanol water partition coefficient for the chemical
In Equation 5-12, Kow is being used to estimate the partition coefficient to POC
(i.e., Kpocin L/kg) and 0.08-KOW is being used to estimate the partition coefficient
to DOC (i.e., the KDOC in L/kg). A discussion of the technical basis, assumptions,
and uncertainty associated with the derivation and application of Equation 5-12 is
provided in the Bioaccumulation TSD.
POC and DOC Values. When converting from the total concentration of a
chemical to a freely dissolved concentration using Equation 5-12 above, the POC
and DOC concentrations should be obtained from the original study from which
the field-measured BAF is determined. If POC and DOC concentrations are not
reported in the BAF study, reliable estimates of POC and DOC might be obtained
from other studies of the same site used in the BAF study or closely related site(s)
within the same water body. When using POC/DOC data from other studies of
the same water body, care should be taken to ensure that environmental and
hydrological conditions that might affect POC or DOC concentrations (i.e., runoff
events, proximity to ground water or surface water inputs, sampling season) are
reasonably similar to those in the BAF study. Additional information related to
selecting POC and DOC values is provided in the Bioaccumulation TSD.
In some cases, BAFs are reported using the concentration of the chemical in
filtered or centrifuged water. When converting these BAFs to a freely dissolved
basis, the concentration of POC should be set equal to zero when using Equation
5-12. Particulates are removed from water samples by filtering or centrifuging
the sample.
5-27
-------
c. Selecting K,,w Values. A variety of techniques are available to measure or predict
Kow values. The reliability of these techniques depends to a large extent on the
Kow of the chemical. Because Kow is an important input parameter for calculating
the freely dissolved concentration of nonionic organic chemicals and for deriving
BAFs using the other three methods of Procedure #1, care should be taken in
selecting the most reliable Kow value. The value of Kow for use in estimating the
freely dissolved fraction and other procedures used to derive national BAFs
should be selected based on the guidance presented in the Bioaccumulation TSD.
4. Determining the Fraction Lipid (ft). Calculating a baseline BAFf for a nonionic
organic chemical using Equation 5-10 also requires that the total chemical concentration
measured in the tissue used to determine the field-measured BAFj be normalized by the
lipid fraction (Q in that same tissue. Lipid normalization of tissue concentrations reflects
the assumption that BAFs (and BCFs) for nonionic organic chemicals are directly
proportional to the percent lipid in the tissue upon which they are based. This
assumption means that an organism with a two percent lipid content would be expected
to accumulate twice the amount of a chemical at steady state compared with an organism
with one percent lipid content, all else being equal. The assumption that aquatic
organisms accumulate nonionic organic chemicals in proportion to their lipid content has
been extensively evaluated in the literature (Mackay, 1982; Connell, 1988; Barren, 1990)
and is generally accepted. Because the lipid content in aquatic organisms can vary both
within and across species, BAFs that are expressed using the lipid-normalized
concentration (rather than the total concentration in tissue) are considered to be the most
reliable for aggregating multiple BAF values for a given species. Additional discussion
of technical basis, assumptions, and uncertainties involved in lipid normalization is
provided in the Bioaccumulation TSD.
a. The lipid fraction fc, is routinely reported in bioaccumulation studies involving
nonionic organic chemicals. If the lipid fraction is not reported in the BAF study,
it can be calculated using the following equation if the appropriate data are
reported:
(Equation 5-13)
t
where:
Mj = Mass of lipid in specified tissue
Mt = Mass of specified tissue (wet weight)
b. Because lipid content can vary within an aquatic organism (and among tissues
within that organism) due to several factors including the age and sex of the
organism, changes in dietary composition, season of sampling and reproductive
status, the lipid fraction used to calculate a baseline BAFf should be measured in
5-28
-------
the same tissue and organisms used to determine the field-measured BAFJ, unless
comparability is demonstrated across organisms.
c. Experience has shown that different solvent systems used to extract lipids for
analytical measurement can result in different quantities of lipids being extracted
and measured in aquatic organisms (e.g., Randall et al.,1991, 1998). As a result,
lipid measurements determined using different solvent systems might lead to
apparent differences in lipid-normalized concentrations and lipid-normalized
BAFs. The extent to which different solvent systems might affect lipid
extractions (and lipid-normalized concentrations) is thought to vary depending on
the solvent, chemical of concern, and lipid composition of the tissue being
extracted. Guidance on measurement of lipid content, including the choice of
solvent system and how different solvent systems may affect lipid content, is
provided in the Bioaccumulation TSD.
B. Baseline EAfff Derived from BSAFs
The second method of determining a baseline B AFf for the chemical of concern in
Procedure #1 involves the use of BSAFs. Although BSAFs may be used for measuring and
predicting bioaccumulation directly from concentrations of chemicals in surface sediment, they
may also be used to estimate BAFs (USEPA, 1995b; Cook and Burkhard, 1998). Since BSAFs
are based on field data and incorporate effects of chemical bioavailability, food web structure,
metabolism, biomagnification, growth, and other factors, BAFs estimated from BSAFs will
incorporate the net effect of all these factors. The BSAF approach is particularly beneficial for
developing water quality criteria for chemicals which are detectable in fish tissues and
sediments, but are difficult to detect or measure precisely in the water column.
As shown by Equation 5-14 below, predicting baseline BAFf s using BSAFs requires that
certain types of data be used for the chemicals of interest (for which BAFs are to be determined)
and reference chemicals (for which BAFs are measured) from a common sediment-water-
organism data set. Differences between BSAFs for different organic chemicals are good
measures of the relative bioaccumulation potentials of the chemicals. When calculated from a
common organism-sediment sample set, chemical-specific differences in BSAFs reflect the net
effect of biomagnification, metabolism, food chain, bioenergetics, and bioavailability factors on
the degree of each chemical's equilibrium/disequilibrium between sediment and biota. At
equilibrium, BSAFs are expected to be approximately 1.0. However, deviations from 1.0
(reflecting disequilibrium) are common due to: conditions where water is not at equilibrium with
surface sediment; differences in organic carbon content of water and sediment; kinetic
limitations for chemical transfer between sediments and water associated with specific biota;
biomagnification; or biological processes such as growth or biotransformation. BSAFs are most
useful (i.e., most predictable from one site to another) when measured under steady-state (or near
steady-state) conditions. The use of non-steady-state BSAFs, such as found with new chemical
loadings or rapid increases in loadings, increases uncertainty in this method for the relative
degree of disequilibrium between the reference chemicals and the chemicals of interest. In
general, the fact that concentrations of hydrophobic chemicals in sediment are less sensitive than
concentrations in water to fluctuations in chemical loading and distribution makes the BSAF
5-29
-------
method robust for estimating BAFs. Results from validation of the BAF procedure in Lake
Ontario, the Fox River and Green Bay, Wisconsin, and the Hudson River, New York,
demonstrate good agreement between observed and BSAF-predicted BAFs in the vast majority
of comparisons made. Detailed results of the validation studies for the BSAF procedure are
provided in the Bioaccumulation TSD.
Baseline BAFf s should be calculated using acceptable BSAFs for chemicals of interest
and appropriate sediment-to-water fugacity (disequilibrium) ratios (nsocw)r/(Kow)r for reference
chemicals under the following guidelines.
1. Baseline BAF[d Equation. For each species with an acceptable field measured (BSAF)j,
a baseline BAFf for the chemical of interest may be calculated using the following
equation with an appropriate value of (nsocw)r/(Kow)r:
sn i- Tt AT-ifd\ /Ttn j T~>\ ^ HT' ^-Li-socw'r ^ ow'i
(Baseline BAF^ ). = (BSAF). — (Equation 5-14)
where:
(Baseline BAFf )j = BAF expressed on a freely dissolved and lipid-
normalized basis for chemical of interest "I"
(BSAF)j = Biota-sediment accumulation factor for chemical of
interest "I"
(ILocwX = sediment organic carbon to water freely dissolved
concentration ratio of reference chemical "r"
(Kow)j = octanol-water partition coefficient for chemical of
interest "I"
(Kow)r = octanol-water partition coefficient for the reference
chemical "r"
Di/r = ratio between Hsocw/ Kow for chemicals "I" and "r"
(normally chosen so that Di/r =1)
The technical basis, assumptions, and uncertainties associated with Equation 5-14 are provided
in the Bioaccumulation TSD. Guidance for determining each component of Equation 5-14 is
provided below.
2. Determining Field-Measured BSAFs. BSAFs should be determined by relating lipid-
normalized concentrations of chemicals in an organism (Cc) to organic carbon-normalized
concentrations of the chemicals in surface sediment samples (Csoc) using the following
equation:
C.
BSAF = —— (Equation 5-15)
soc
5-30
-------
Lipid-Normalized Concentration. The lipid-normalized concentration of a
chemical in an organism should be determined by:
Ct
(Equation 5-16)
where:
Ct = Concentration of the chemical in the wet tissue (either
whole organism or specified tissue) (|ig/g)
f,, = Fraction lipid content in the tissue
b. Organic Carbon-Normalized Concentration. The organic carbon-normalized
concentration of a chemical in sediment should be determined by:
Cs
CSDC ~ ~r~ (Equation 5-17)
oc
where:
Cs = Concentration of chemical in sediment (|ig/g sediment)
foc = Fraction organic carbon in sediment
The organic carbon-normalized concentrations of the chemicals in surface
sediment samples should be associated with the average exposure environment of
the organism.
3. Sediment-to-Water Partition Coefficient (JJsocw)r. Sediment-to-water partition
coefficients for reference chemicals should be determined by:
(Equation 5-18)
(r
where:
(Csoc)r = Concentration of a reference chemical in sediment normalized to
sediment organic carbon
( C")r = Concentration of the reference chemical freely dissolved in water
4. Selecting Reference Chemicals. Reference chemicals with (Qsocw) / (Kow) similar to that
of the chemical of interest are preferred for this method. Theoretically, knowledge of the
5-31
-------
difference between sediment-to-water fugacity ratios for two chemicals, "I" and "r" (Di/r),
could be used when reliable reference chemicals that meet the fugacity equivalence
condition are not available. Similarity of (Qsocw) / (Kow) f°r two chemicals can be
indicated on the basis of similar physical-chemical behavior in water (persistence,
volatilization), similar mass loading histories, and similar concentration profiles in
sediment cores.
Validation studies have demonstrated that choosing reference chemicals with well
quantified concentrations in water is important because the uncertainty associated with
measurement of barely detected chemicals is large (see the Bioaccumulation TSD).
Similarity between Kow values of the reference and target chemicals is generally
desirable, although recent validation studies indicate that the accuracy of the method is
not substantially decreased through use of reference chemicals with large differences in
Kow, as long as the chemicals are structurally similar and have similar persistence
behavior in water and sediments.
5. The following data, procedural, and quality assurance requirements should be met for
predicting baseline BAFf s using field-measured BSAFs:
a. Data on the reference chemicals and chemicals of interest should come from a
common organism-water-sediment data set at a particular site.
b. The chemicals of interest and reference chemicals should have similar
physicochemical properties and persistence in water and sediment.
c. The loadings history of the reference chemicals and chemicals of interest should
be similar such that their expected sediment-water disequilibrium ratios
(ILocw/Kow) would not be expected to be substantially different (i.e., Di/r ~ 1).
d. The use of multiple reference chemicals is generally preferred for determining the
value of (HsocwX so l°n§ as the concentrations are well quantified and the
aforementioned conditions for selecting reference chemicals are met. In some
cases, use of a single reference chemical may be necessary because of limited
data.
e. Samples of surface sediments (0-1 cm is ideal) should be from locations in which
sediment is regularly deposited and is representative of average surface sediment
in the vicinity of the organism.
f. The Kow value for the target and reference chemicals should be selected as
described in the Bioaccumulation TSD.
g. All other data quality and procedural guidelines described earlier for determining
field-measured BAFs in Section 5.4.3.1(A) should be met.
5-32
-------
Further details on the requirements for predicting BAFs from BSAF measurements,
including the data, assumptions, and limitations of this approach are provided in the
Bioaccumulation TSD.
C. Baseline BAF^from a Laboratory-Measured BCF'T and FCM
The third method in Procedure #1 consists of using a laboratory-measured BCF^ (i.e., a
BCF based on total concentrations in tissue and water) and FCMs to predict a baseline BAFf for
the chemical of concern. The BCF] is used in conjunction with an FCM because non-aqueous
routes of exposure and subsequent biomagnification is of concern for the types of chemicals
applicable to Procedure #1. A laboratory-measured BCF inherently accounts for the effects of
chemical metabolism that occurs in the organism used to calculate the BCF, but does not account
for metabolism which may occur in other organisms of the aquatic food web.
1. Baseline BAF[d Equation. For each acceptable laboratory-measured BCF], calculate a
baseline BAFf using the following equation:
fd
Measured BCFT
Baseline BAFCIQ = (FCM) •
lfd
where:
T
- 1
(Equation 5-19)
Baseline BAFf = BAF expressed on a freely dissolved and lipid-
normalized basis
Measured BCF] = BCF based on total concentration in tissue and
water
ft = Fraction of the tissue that is lipid
ffd = Fraction of the total chemical in the test water that
is freely dissolved
FCM = The food chain multiplier either obtained from
Table 5-1 by linear interpolation for the appropriate
trophic level, or from appropriate field data
The technical basis for Equation 5-19 is provided in the Bioaccumulation TSD.
Guidance for determining each component of Equation 5-19 is provided below.
2. Determining the Measured BCF]. The laboratory-measured BCF] shown in Equation
5-19 should be calculated using information on the total concentration of the chemical in
the tissue of the organism and the total concentration of the chemical in the laboratory
test water. The equation to derive a measured BCF] is:
t ct
Measured BCFT = — (Equation 5-20)
5-33
-------
where:
Ct = Total concentration of the chemical in the specified wet tissue
Cw = Total concentration of chemical in the laboratory test water
The data used to calculate a laboratory-measured BCFj should be reviewed thoroughly to
assess the quality of the data and the overall uncertainty in the BCF value. The following
general criteria apply in determining the acceptability of laboratory-measured BCFj.
a. The test organism should not be diseased, unhealthy, or adversely affected by the
concentration of the chemical because these attributes may alter accumulation of
chemicals compared with healthy organisms.
b. The total concentration of the chemical in the water should be measured and
should be relatively constant during the exposure period.
c. The organisms should be exposed to the chemical using a flow-through or
renewal procedure.
d. The percent lipid of the tissue used to normalize the BCFj should be either
measured or reliably estimated to permit lipid normalization of chemical
concentrations.
e. The concentrations of particulate organic carbon and dissolved organic carbon in
the study water should be measured or reliably estimated.
f. Aquatic organisms used to calculate a laboratory-measured BCF^ should be
representative of those aquatic organisms that are commonly consumed in the
United States. An aquatic organism which is not commonly consumed in the
United States can be used to calculate an acceptable laboratory-measured BCF^
provided that the organism is considered to be a reasonable surrogate for a
commonly consumed organism. Information on the ecology, physiology, and
biology of the organism should be reviewed when assessing whether an organism
is a reasonable surrogate of a commonly consumed organism.
g. BCFs may be based on measurement of radioactivity from radiolabeled parent
compounds only when the BCF is intended to include metabolites, when there is
confidence that there is no interference due to metabolites of the parent
compounds, or when studies are conducted to determine the extent of metabolism,
thus allowing for a proper correction.
h. The calculation of the BCFj should appropriately address growth dilution, which
can be particularly important in affecting BCF j determinations for poorly
depurated chemicals.
5-34
-------
I. Other aspects of the methodology used should be similar to those described by the
American Society of Testing and Materials (ASTM, 1999) and USEPA
Ecological Effects Test Guidelines (USEPA, 1996).
j. In addition, the magnitude of the Kow and the availability of corroborating BCF
data should be considered. For example, if the steady-state method is used for the
BCF^ determination, exposure periods longer than 28 days will generally be
required for highly hydrophobic chemicals to reach steady state between the
water and the organism.
k. If a baseline BCFf derived from a laboratory-measured BCF j consistently
increases or decreases as the chemical concentration increases in the test solutions
for the test organisms, the BCF^ should be selected from the test concentration(s)
that would most closely correspond to the 304(a) criterion. Note: a BCF^ should
not be calculated from a control treatment.
Selecting Food Chain Multipliers. An FCM reflects a chemical's tendency to
biomagnify in the aquatic food web. Values of FCMs greater than 1.0 are indicative of
biomagnification and typically apply to organic chemicals with log Kow values between
4.0 and 9.0. For a given chemical, FCMs tend to be greater at higher trophic levels,
although FCMs for trophic level three can be higher than those for trophic level four.
Food chain multipliers used to derive baseline BAFf s using Procedure #1 can be selected
from model-derived or field-derived estimates.
a. Model-Derived FCMs. For nonionic organic chemicals appropriate for
Procedure #1, EPA has calculated FCMs for various Kow values and trophic levels
using the bioaccumulation model of Gobas (1993). The FCMs shown in
Table 5-1 were calculated using the Gobas model as the ratio of the baseline
BAFf s for trophic levels 2, 3, and 4 to the baseline BCFf.
EPA recommends using the biomagnification model by Gobas (1993) to derive
FCMs for nonionic organic chemicals for several reasons. First, the Gobas model
includes both benthic and pelagic food chains, thereby incorporating exposure of
organisms to chemicals from both the sediment and the water column. Second,
the input data needed to run the model can be readily defined. Third, the
predicted BAFs using the model are in agreement with field-measured BAFs for
chemicals, even those with very high log Kows. Finally, the model predicts
chemical residues in benthic organisms using equilibrium partitioning theory,
which is consistent with EPA's equilibrium partitioning sediment guidelines
(USEPA, 2000d).
The Gobas model requires input of specific data on the structure of the food chain
and the water quality characteristics of the water body of interest. For calculating
national BAFs, a mixed pelagic/benthic food web structure consisting of four
trophic levels is assumed. Trophic level 1 is phytoplankton, trophic level 2 is
5-35
-------
zooplankton, trophic level 3 is forage fish (e.g., sculpin and smelt), and trophic
level 4 are predatory fish (e.g., salmonids). Additional assumptions are made
regarding the composition of the aquatic species' diets (e.g., salmonids consume
10 percent sculpin, 50 percent alewives, and 40 percent smelt), the physical
parameters of the aquatic species (e.g., lipid values), and the water quality
characteristics (e.g., water temperature, sediment organic carbon).
A mixed pelagic/benthic food web structure has been assumed for the purpose of
calculating FCMs because it is considered to be most representative of the types
of food webs that occur in aquatic ecosystems. FCMs derived using the mixed
pelagic/benthic structure are also about mid-range in magnitude between a 100%
pelagic and 100% benthic driven food web (see the Bioaccumulation TSD). The
validity of FCMs derived using the mixed pelagic/benthic food web structure has
5-36
-------
Table 5-1
Food-Chain Multipliers for Trophic Levels 2, 3 and 4
(Mixed Pelagic and Benthic Food Web Structure and Hsocw / KQW = 23)
Log
KQW
4.0
4.1
4.2
4.3
4.4
4.5
4.6
4.7
4.8
4.9
5.0
5.1
5.2
5.3
5.4
5.5
5.6
5.7
5.8
5.9
6.0
6.1
6.2
6.3
6.4
6.5
Trophic
Level 2
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
Trophic
Level 3
1.23
1.29
1.36
1.45
1.56
1.70
1.87
2.08
2.33
2.64
3.00
3.43
3.93
4.50
5.14
5.85
6.60
7.40
8.21
9.01
9.79
10.5
11.2
11.7
12.2
12.6
Trophic
Level 4
1.07
1.09
1.13
1.17
1.23
1.32
1.44
1.60
1.82
2.12
2.51
3.02
3.68
4.49
5.48
6.65
8.01
9.54
11.2
13.0
14.9
16.7
18.5
20.1
21.6
22.8
Log
KQW
6.6
6.7
6.8
6.9
7.0
7.1
7.2
7.3
7.4
7.5
7.6
7.7
7.8
7.9
8.0
8.1
8.2
8.3
8.4
8.5
8.6
8.7
8.8
8.9
9.0
Trophic
Level 2
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
Trophic
Level 3
12.9
13.2
13.3
13.3
13.2
13.1
12.8
12.5
12.0
11.5
10.8
10.1
9.31
8.46
7.60
6.73
5.88
5.07
4.33
3.65
3.05
2.52
2.08
1.70
1.38
Trophic
Level 4
23.8
24.4
24.7
24.7
24.3
23.6
22.5
21.2
19.5
17.6
15.5
13.3
11.2
9.11
7.23
5.58
4.19
3.07
2.20
1.54
1.06
0.721
0.483
0.320
0.210
been evaluated in several different ecosystems including Lake Ontario, the tidally
influenced Bayou D'Inde in Louisiana, the Fox River and Green Bay, Wisconsin,
and the Hudson River in New York. Additional details of the validation of EPA's
national default FCMs and the assumptions, uncertainties, and input parameters
for the model are provided in the Bioaccumulation TSD.
5-37
-------
Although EPA uses the FCMs in Table 5-1 to derive its national 304(a) criteria,
EPA recognizes that food webs of other waterbodies might differ from the
assumptions used to calculate national BAFs. In these situations, States and
authorized Tribes may wish to use alternate food web structures for calculating
FCMs for use in setting State or Tribal water quality criteria. Additional guidance
on the use of alternate food web structures for calculating State, Tribal, or site-
specific criteria is provided in the Bioaccumulation TSD.
Field-Derived FCMs. In addition to model-derived estimates of FCMs, field
data may also be used to derive FCMs. Currently, the use of field-derived FCMs
is the only method recommended for estimating FCMs for inorganic and
organometalic chemicals because appropriate model-derived estimates are not yet
available (see Section 5.6). In contrast to the model-based FCMs described
previously, field-derived FCMs account for any metabolism of the chemical of
concern by the aquatic organisms used to calculate the FCM.
Field-derived FCMs should be calculated using lipid-normalized concentrations
of the nonionic organic chemical in appropriate predator and prey species using
the following equations.
FCM ^ = BMFTL2 (Equation 5-21)
FCM ^3 = (BMF^j) (BMF ^2) (Equation 5-22)
FCM ^4 = (BMF ^4) (BMF ^3) (BMF ^2) (Equation 5-23)
where:
FCM = Food chain multiplier for designated trophic level (TL2, TL3,
or TL4)
BMF = Biomagnification factor for designated trophic level (TL2,
TL3, or TL4)
The basic difference between FCMs and BMFs is that FCMs relate back to
trophic level one (or trophic level two as assumed by the Gobas (1993) model),
whereas BMFs always relate back to the next lowest trophic level. For nonionic
organic chemicals, BMFs can be calculated from tissue residue concentrations
determined in biota at a site according to the following equations.
BMF ^2 = (Cfi TL2) / (C,_ TL1) (Equation 5-24)
BMF ^3 = (C,, TL3) / (C, ^2) (Equation 5-25)
BMF ^4 = (C,, TL4) / (C,, ^3) (Equation 5-26)
where:
5-38
-------
C/ = Lipid-normalized concentration of chemical in tissue of
appropriate biota that occupy the specified trophic level
(TL2, TL3, or TL4)
In addition to the acceptability guidelines pertaining to field-measured BAFs, the
following procedural and quality assurance requirements apply to field-measured
FCMs.
(1) Information should be available to identify the appropriate trophic levels
for the aquatic organisms and appropriate predator-prey relationships for
the site from which FCMs are being determined. General information on
determining trophic levels of aquatic organisms can be found in USEPA
2000a,b,c.
(2) The aquatic organisms sampled from each trophic level should reflect the
most important exposure pathways leading to human exposure via
consumption of aquatic organisms. For higher trophic levels (e.g., 3 and
4), aquatic species should also reflect those that are commonly consumed
by humans.
(3) The studies from which the FCMs are derived should contain sufficient
supporting information from which to determine that tissue samples were
collected and analyzed using appropriate, sensitive, accurate, and precise
methods.
(4) The percent lipid should be either measured or reliably estimated for the
tissue used to determine the FCM.
(5) The tissue concentrations should reflect average exposure over the
approximate time required to achieve steady-state in the target species.
D. Baseline BAF^from a Kow and FCM
The fourth method in Procedure #1 consists of using a Kow and an appropriate FCM for
estimating the baseline BAFf. In this method, the Kow is assumed to be equal to the baseline
BCFf. Numerous investigations have demonstrated a linear relationship between the logarithm
of the BCF and the logarithm of the octanol-water partition coefficient (Kow) for organic
chemicals for fish and other aquatic organisms. Isnard and Lambert (1988) list various
regression equations that illustrate this linear relationship. When the regression equations are
constructed using lipid-normalized BCFs, the slopes and intercepts are not significantly different
from one and zero, respectively (e.g., de Wolf, et al., 1992). The underlying assumption for the
linear relationship between the BCF and Kow is that the bioconcentration process can be viewed
as the partitioning of a chemical between the lipid of the aquatic organisms and water and that
the Kow is a useful surrogate for this partitioning process (Mackay, 1982). To account for
biomagnification, Procedure #1 requires the Kow value be used in conjunction with an
appropriate FCM.
5-39
-------
1. Baseline BAF[d Equation. For each acceptable Kow value and FCM for the chemical of
concern, calculate a baseline BAFf using the following equation.
Baseline BAF/d = (FCM) • (Kow) (Equation 5-27)
where:
Baseline BAFf = BAF expressed on a freely dissolved and lipid-normalized
basis for a given trophic level
FCM = The food chain multiplier for the appropriate trophic level
obtained from Table 5-1 by linear interpolation or from
appropriate field data (used with Procedure #1 only)
Kow = Octanol-water partition coefficient
The BCF-KOW relationship has been developed primarily for nonionic organic chemicals
that are not readily metabolized by aquatic organisms and thus is most appropriate for
poorly-metabolized nonionic organic chemicals (i.e., Procedures #1 and #3 as depicted in
Figure 5-1). For poorly-metabolized nonionic organic chemicals with large log Kows (i.e.,
> 6), reported log BCFs are often not equal to log Kow. EPA believes that this
nonlinearity is primarily due to not accounting for several factors which affect the BCF
determination. These factors include not basing BCFs on the freely dissolved
concentration in water, not accounting for growth dilution, not assessing BCFs at steady-
state, inaccuracies in measurements of uptake and elimination rate constants, and
complications from the use of solvent carriers in the exposure. Application of Equation 5-
27 for predicting BAFs has been conducted in several different ecosystems including
Lake Ontario, the tidally influenced Bayou D'Inde in Louisiana, the Fox River and Green
Bay, Wisconsin, and the Hudson River in New York. Additional detail on the validation,
technical basis, assumptions, and uncertainty associated with Equation 5-27 and is
provided in the Bioaccumulation TSD.
2. FCMs and Kows. Food chain multipliers and Kow values should be selected as described
previously in Procedure #1.
5.4.3.2 Selecting Final Baseline BAF™s
After calculating individual baseline BAFf s using as many of the methods in Procedure
#1 as possible, the next step is to determine a final baseline BAFf for each trophic level from the
individual baseline BAFf s (see Figures 5-1 and 5-2). The final baseline BAFf will be used in
the last step to determine the national BAF for each trophic level. The final baseline BAFf for
each trophic level should be determined from the individual baseline B AFf s by considering the
data preference hierarchy defined by Procedure #1 and uncertainty in the data. The data
preference hierarchy for Procedure #1 is (in order of preference):
1. a baseline BAFf from an acceptable field-measured BAF (method 1)
5-40
-------
2. a baseline BAFf predicted from an acceptable field-measured BSAF (method 2),
3. a baseline BAFf predicted from an acceptable BCF and FCM (method 3), or
4. a baseline BAFf predicted from an acceptable Kow and FCM (method 4).
This data preference hierarchy reflects EPA's preference for BAFs based on field-measurements
of bioaccumulation (methods 1 and 2) over those based on laboratory-measurements and/or
predictions of bioaccumulation (methods 3 and 4). However, this data preference hierarchy
should not be considered inflexible. Rather, it should be used as a guide for selecting the final
baseline BAFf s when the uncertainty is similar among two or more baseline BAFf s derived
using different methods. The following steps and guidelines should be followed for selecting the
final baseline BAFf s using Procedure #1.
1. Calculate Species-Mean Baseline BAFfs. For each B AF method where more than one
acceptable baseline BAFf is available for a given species, calculate a species-mean
baseline BAFf as the geometric mean of all available individual baseline BAFf s. When
calculating a species-mean baseline BAFf, individual baseline BAFf s should be
reviewed carefully to assess the uncertainty in the BAF values. For highly hydrophobic
chemicals applicable to Procedure #1, particular attention should be paid to whether
sufficient spatial and temporal averaging of water and tissue concentrations was likely
achieved in the BAF, BSAF, or BCF study. Highly uncertain baseline BAFf s should not
be used. Large differences in individual baseline BAFf s for a given species (e.g., greater
than a factor of 10) should be investigated further. In such cases, some or all of the
baseline BAFf s for a given species might not be used. Additional discussion on
evaluating acceptability of BAF values is provided in the Bioaccumulation TSD.
2. Calculate Trophic-Level-Mean Baseline BAFfs. For each BAF method where more
than one acceptable species-mean baseline BAFf is available within a given trophic
level, calculate a trophic-level-mean baseline BAFf as the geometric mean of acceptable
species-mean baseline BAFf s in that trophic level. Trophic-level-mean baseline BAFf s
should be calculated for trophic levels two, three, and four because available data on U.S.
consumers offish and shellfish indicate significant consumption of organisms in these
trophic levels.
3. Select a Final Baseline BAF[d for Each Trophic Level. For each trophic level, select
the final baseline BAFf using best professional judgment by considering: (1) the data
preference hierarchy shown previously, (2) the relative uncertainty in the trophic-level-
mean baseline BAFf s derived using different methods, and (3) the weight of evidence
among the four methods.
a. In general, when more than one trophic-level-mean baseline BAFf is available for
a given trophic level, the final trophic-level-mean baseline BAFf should be
selected from the most preferred BAF method defined by the data preference
hierarchy for Procedure #1.
b. If uncertainty in a trophic-level-mean baseline BAF based on a higher tier (more
preferred) method is judged to be substantially greater than a trophic-level-mean
5-41
-------
baseline BAF from a lower tier method, and the weight of evidence among the
various methods suggests that a BAF value from lower tier method is likely to be
more accurate, then the final baseline BAFf should be selected using a trophic
level-mean baseline BAFf from a lower tier method.
c. When considering the weight of evidence among the various BAF methods,
greater confidence in the final baseline BAFf is generally assigned when BAFs
from a greater number of methods are in agreement for a given trophic level.
However, lack of agreement among methods does not necessarily indicate less
confidence if such disagreements can be adequately explained. For example, if
the chemical of concern is metabolized by aquatic organisms represented by a
BAF value, one would expect disagreement between a field-measured BAF (the
highest priority data) and a predicted BAF using a Kow and model-derived FCM.
Thus, field-measured BAFs should generally be given the greatest weight among
methods because they reflect direct measures of bioaccumulation and incorporate
any metabolism which might occur in the organism and its food web.
d. The above steps should be performed for each trophic level until a final baseline
BAFf is selected for trophic levels two, three, and four.
5.4.3.3 Calculating National BAFs
The last step in deriving a national BAF for each trophic level is to convert the final
baseline BAFf determined in the previous step to a BAF that reflects conditions to which the
national 304(a) criteria will apply (Figure 5-2). Since a baseline BAFf is by definition
normalized by lipid content and expressed on a freely dissolved basis, it needs to be adjusted to
reflect the lipid fraction of aquatic organisms commonly consumed in the U.S. and the freely
dissolved fraction expected in U.S. bodies of water. Converting a final baseline BAFf to a
national BAF requires information on: (1) the percent lipid of the aquatic organisms commonly
consumed by humans, and (2) the freely dissolved fraction of the chemical of concern that would
be expected in the ambient waters of interest. For each trophic level, a national BAF should be
determined from a final baseline BAFf according to the following guidelines.
1. National BAF Equation. For each trophic level, calculate a national BAF using the
following equation.
National BAF(1L n) = [(Final Baseline BAF/d )^ n • (Q^ n + 1] • (ffd) (Equation 5-28)
where:
Final Baseline BAFf = Final trophic-level-mean baseline BAF expressed
on a freely dissolved and lipid-normalized basis for
trophic level "n"
5-42
-------
fc(TLn) = Lipid fraction of aquatic species consumed at
trophic level "n"
ffd = Fraction of the total chemical in water that is freely
dissolved
The technical basis of Equation 5-28 is provided in the Bioaccumulation TSD. Guidance
for determining each component of Equation 5-28 is provided below.
2. Determining the Final Baseline BAF[d. The final trophic-level-mean baseline BAFf s
used in this equation are those which have been determined using the guidance presented
in Section 5.4.3.2 for selecting the final baseline BAFf s.
3. Lipid Content of Commonly Consumed Aquatic Species. As illustrated by Equation
5-28, the percent lipid of the aquatic species consumed by humans is needed to
accurately characterize the potential exposure to a chemical from ingestion of aquatic
organisms.
a. National Default Lipid Values. For the purposes of calculating a national
304(a) criterion, the following national default values for lipid fraction should be
used: 1.9% (for trophic level two organisms), 2.6% (for trophic level three
organisms), and 3.0% (for trophic level four organisms).
These national default values for lipid content reflect national per capita average
patterns offish consumption in the United States. Specifically, they were
calculated using the consumption-weighted mean lipid content of commonly
consumed fish and shellfish as identified by the USDA Continuing Survey of
Food Intake by Individuals (CSFII) for 1994 through 1996. This same national
survey data was used to derive national default values offish consumption. To
maintain consistency with the fish consumption assumptions, only freshwater and
estuarine organisms were included in the derivation of the national default lipid
values. Additional details on the technical basis, assumptions, and uncertainty in
the national default values of lipid fraction are provided in the Bioaccumulation
TSD.
Although national default lipid values are used by EPA to set national 304(a)
criteria, EPA encourages States and authorized Tribes to use local or regional data
on lipid content of consumed aquatic species when adopting criteria into their
water quality standards because local or regional consumption patterns (and lipid
content) can differ from national consumption patterns. Additional guidance on
developing site-specific values of lipid content, including a database of lipid
content for many commonly consumed aquatic organisms, is found in the
Bioaccumulation TSD.
4. Freely Dissolved Fraction. The third piece of information required for deriving a
national BAF is the freely dissolved fraction of the chemical of concern that is expected
5-43
-------
in waters of the United States. As noted previously, expressing BAFs on the freely
dissolved concentration in water allows a common basis for averaging BAFs from
several studies. However, for use in criteria development, these BAFs should be
converted back to values based on the total concentration in the water to be consistent
with monitored water column and effluent concentrations, which are typically based on
total concentrations of chemicals in the water. This should be done by multiplying the
freely dissolved baseline BAFf by the fraction of the freely dissolved chemical expected
in water bodies of the United States where criteria are to be applied, as shown in
Equation 5-29.
f _ _
fd " [1 + (POC • K J + (DOC • 0.08 • K J] (Equation 5-29)
where:
POC = national default value for the particulate organic carbon
concentration (kg/L)
DOC = national default value for the dissolved organic carbon
concentration (kg/L)
Kow = n-octanol water partition coefficient for the chemical
Equation 5-29 is identical to Equation 5-12, which was used to determine the freely
dissolved fraction for deriving baseline BAFf s from field-measured BAFs. However, the
POC and DOC concentrations used in Equation 5-29 reflect those values that are
expected in U.S. bodies of water, not the POC and DOC values in the study water used to
derive the BAF. Guidance for determining each component of Equation 5-29 follows.
a. National Default Values of POC and DOC. For estimating the freely dissolved
fraction of the chemical of concern that is expected in U.S. water bodies, national
default values of 0.5 mg/L (5 x lO'7 kg/L) for POC and 2.9 mg/L (2.9 x lO'6 kg/L)
for DOC should be used. These values are 50th percentile values (medians) based
on an analysis of over 110,000 DOC values and 85,000 POC values contained in
EPA's STORET database from 1980 through 1999. These default values reflect a
combination of values for streams, lakes and estuaries across the United States.
Additional details on the technical basis, assumptions, and uncertainty in the
derivation and application of the national default values of POC and DOC are
provided in the Bioaccumulation TSD.
Although national default values of POC and DOC concentrations are used by
EPA to set national 304(a) criteria as described by this document, EPA
encourages States and authorized Tribes to use local or regional data on POC and
DOC when adopting criteria into their water quality standards. EPA encourages
States and Tribes to consider local or regional data on POC and DOC because
local or regional conditions may result in differences in POC or DOC
5-44
-------
concentrations compared with the values used as national defaults. Additional
guidance on developing local or regional values of POC and DOC, including a
database of POC and DOC values segregated by waterbody type, is found in the
Bioaccumulation TSD.
b. KowValue. The value selected for the Kow of the chemical of concern should be
the same value used in earlier calculations (e.g., for calculating baseline BAFf s
and FCMs). Guidance for selecting the Kow value is found in the
Bioaccumulation TSD.
5.4.4 Deriving National BAFs Using Procedure #2
This section provides guidance for calculating national BAFs for nonionic organic
chemicals using Procedure #2 shown in Figure 5-1. The types of nonionic organic chemicals for
which Procedure #2 is most appropriate are those that are classified as moderately to highly
hydrophobic and subject to high rates of metabolism by aquatic biota (see Section 5.4.2 above).
Non-aqueous contaminant exposure and subsequent biomagnification in aquatic food webs are
not generally of concern for chemicals that are classified in this category. As a result, FCMs are
not used in this procedure. In addition, Kow -based predictions of bioconcentration are not used
in this procedure since the Kow /BCF relationship is primarily based on poorly metabolized
chemicals. Some nonionic organic chemicals for which Procedure #2 is probably appropriate
include certain PAHs which are believed to be metabolized substantially by fish (e.g.,
benzo[a]pyrene, phenanthrene, fluoranthene, pyrene, benzo [a] anthracene and
chrysene/triphenylene; USEPA, 1980; Burkhard and Lukasewycz, 2000).
According to Procedure #2, the following three methods can be used in deriving a
national BAF:
• using a BAF from an acceptable field study (i.e., a field-measured BAF) (method 1),
• predicting a BAF from an acceptable BSAF (method 2), and
• predicting a BAF from an acceptable BCF (method 3).
Each of these three methods relies on measured data for assessing bioaccumulation and
therefore, includes the effects of chemical metabolism by the study organism in the BAF
estimate. The field-measured BAF and BSAF methods also incorporate any metabolism which
occurs in the aquatic food web.
As shown in Figure 5-2, the next steps in deriving a national BAF after selecting the
derivation procedure are: (1) calculating individual baseline BAFf s, (2) selecting the final
baseline BAFf s, and (3) calculating the national BAFs. Each of these three steps is discussed
separately below.
5.4.4.1 Calculating Individual Baseline BAF[ds
As described previously in Procedure #1, calculating individual baseline BAFf s involves
normalizing the measured BAF^ or BCF^ (which are based on the total chemical in water and
5-45
-------
tissue) by the lipid content of the study organisms and the freely dissolved fraction of the
chemical in the study water. Converting measured BAFj (or BCFj) values to baseline BAFf (or
BCFf) values is designed to account for variation in measured BAF^s that is caused by
differences in lipid content of study organisms and differences in the freely dissolved fraction of
chemical in study waters. Therefore, baseline BAFf s are considered more amenable for
extrapolating and averaging BAFs across different species and different study waters compared
with total BAF^s.
1. For each species where acceptable data are available, calculate all possible baseline
BAFf s using each of the three methods shown above for Procedure #2.
2. Individual baseline BAFf s should be calculated from field-measured BAFjs, field-
measured BSAFs, and laboratory BCFjs according to the following procedures.
A. Baseline BAF^from Field-Measured BAFs
1. Except where noted below, a baseline BAFf should be calculated from a field-measured
BAF^ using the guidance and equations outlined in Section 5.4.3.1(A) for determining
baseline BAFf s from field-measured BAFs in Procedure #1.
2. Because nonionic organic chemicals applicable to Procedure #2 have relatively high rates
of metabolism in aquatic organisms, they will tend to reach steady state more quickly
than nonionic organic chemicals with similar Kow values but which undergo little or no
metabolism. Therefore, less temporal averaging of chemical concentrations would
generally be required for determining field-measured BAFjs with highly metabolizable
chemicals compared with chemicals that are poorly metabolized by aquatic biota.
5-46
-------
B. Baseline BAF^ Derived from Field-measured BSAFs
1 . A baseline BAFf should be calculated from a field-measured BSAF using the guidance
and equations outlined in Section 5.4.3.1(B) for determining baseline BAFfs from field-
measured BSAFs in Procedure #1.
C. Baseline BAF^from a Laboratory-Measured BCF
1 . Except where noted below, a baseline BAFf should be calculated from a laboratory-
measured BCF^ using the guidance and equations outlined in Section 5.4.3.1(c) for
determining baseline BAFfs from a laboratory-measured BCF and FCM in Procedure #1.
2. Because biomagnification is not an overriding concern for nonionic organic chemicals
applicable to Procedure #2, food chain multipliers are not used in the derivation of a
baseline BAFf from a laboratory-measured
5.4.4.2 Selecting Final Baseline BAF™s
After calculating individual, baseline BAFfs using as many of the methods in Procedure
#2 as possible, the next step is to determine a final baseline BAFf for each trophic level from the
individual baseline BAFfs. The final baseline BAFf will be used in the last step to determine
the national BAF for each trophic level. A final baseline BAFf for each trophic level should be
determined from the individual baseline BAFfs by considering the data preference hierarchy
defined by Procedure #2 and uncertainty in the data. The data preference hierarchy for
Procedure #2 is (in order of preference):
1 . a baseline BAFf from an acceptable field-measured BAF (method 1),
2. a baseline BAFf from an acceptable field-measured BSAF (method 2), or
3. a baseline BAFf from an acceptable laboratory-measured BCF (method 3).
This data preference hierarchy reflects EPA's preference for BAFs based on field-
measurements of bioaccumulation (methods 1 and 2) over those based on laboratory-
measurements (method 3). However, as explained in Procedure #1, this data preference
hierarchy should not be considered inflexible. Rather, it should be used as a guide for selecting
the final baseline BAFfs when the underlying uncertainty is similar among two or more baseline
BAFfs derived using different methods. Although biomagnification is not generally a concern
for chemicals subject to Procedure #2, trophic level differences in bioaccumulation might be
substantial to the extent that the rate of chemical metabolism by organisms in different trophic
levels differs. For example, certain PAHs have been shown to be metabolized to a much greater
extent by some fish compared with some invertebrate species (James, 1989). Therefore, final
baseline BAFfs for chemicals applicable to Procedure #2 should be determined on a trophic-
level-specific basis according to the following guidelines.
1 . The final baseline BAFfs in Procedure #2 should be selected according to the same steps
described in Procedure #1 but with the substitution of the data preference hierarchy
described above for Procedure #2. Specifically, the species-mean baseline BAFfs,
5-47
-------
trophic-level-mean baseline BAFf s, and the final baseline B AFf s should be determined
according to the guidelines presented in Procedure #1 (Section 5.4.3.2, Steps 1, 2, and 3).
5.4.4.3 Calculating the National BAFs
As described in Procedure #1, the last step in deriving national BAFs for nonionic
organic chemicals is to convert the final baseline BAFf s determined in the previous step to
BAFs which reflect conditions to which the national 304(a) criteria will apply (Figure 5-2).
1. For trophic levels two, three, and four, national BAFs should be calculated from the final
baseline BAFf s using the same equation and procedures described previously in
Procedure #1 (see Section 5.4.3.3 entitled "Calculating the National BAFs").
5.4.5 Deriving National BAFs Using Procedure #3
This section provides guidance for calculating national BAFs for nonionic organic
chemicals using Procedure #3 shown in Figure 5-1. The types of nonionic organic chemicals for
which Procedure #3 is most appropriate are those that are classified as low in hydrophobicity
(i.e., log Kow values less than 4.0) and subject to low (or unknown) rates of metabolism by
aquatic biota (see Section 5.4.2 above). Non-aqueous contaminant exposure and subsequent
biomagnification in aquatic food webs are not generally of concern for chemicals that are
classified in this category (Fisk et al., 1998; Gobas et al., 1993; Connolly and Pedersen, 1988;
Thomann, 1989). As a result, FCMs are not used in this procedure.
According to Procedure #3, the following three methods can be used in deriving a
national BAF:
• using a BAF from an acceptable field study (i.e., a field-measured BAF),
• predicting a BAF from an acceptable laboratory-measured BCF, and
• predicting a BAF from an acceptable Kow.
After selecting the derivation procedure, the next steps in deriving a national BAF at a
given trophic level for nonionic organic chemicals are: (1) calculating individual baseline
BAFf s, (2) selecting the final baseline BAFf, and (3) calculating the national BAF (Figure 5-2).
Each of these three steps is discussed separately below.
5.4.5.1 Calculating Individual Baseline BAFfs
Calculating individual baseline BAFf s involves normalizing each measured BAFj or
BCF^ (which are based on the total chemical in water and tissue) by the lipid content of the study
organism and the freely dissolved fraction of the chemical in the study water. For additional
discussion of the technical basis for calculating baseline BAFfs, see Section 5.4.3.1 in Procedure
#1.
5-48
-------
1 . For each species where acceptable data are available, calculate all possible baseline
BAFf s using each of the three methods shown above for Procedure #3.
2. An individual baseline BAFf should be calculated from field-measured BAFjs,
laboratory-measured BCF^s, and Kow values according to the following procedures.
A. Baseline BAP*? from Field-Measured BAFs
1 . Except where noted below, a baseline BAFf should be calculated from a field-measured
^ using the guidance and equations outlined in Section 5.4.3.1(A) in Procedure #1.
2. Freely Dissolved Fraction. Due to their low hydrophobicity (i.e., log Kow < 4.0),
nonionic organic chemicals applicable to Procedure #3 are expected to remain almost
entirely in the freely dissolved form in natural waters with dissolved and particulate
organic carbon concentrations typical of most field BAF studies. Therefore, the freely
dissolved fraction should be assumed to be equal to 1.0, unless the concentrations of
DOC and POC are very high in the field BAF study. For studies with very high DOC or
POC concentrations, (e.g., about 100 mg/L or higher for DOC or 10 mg/L or higher for
POC), the freely dissolved fraction may be substantially lower than 1.0 and therefore
should be calculated using Equation 5-12.
3 . Temporal Averaging of Concentrations. Also due to their low hydrophobicity,
nonionic organic chemicals appropriate to Procedure #3 will also tend to reach steady
state quickly compared with those chemicals to which Procedure #1 applies. Therefore,
the extent of temporal averaging of tissue and water concentrations is typically much less
than that required for highly hydrophobic chemicals to which Procedure #1 is applied. In
addition, field studies used to calculate BAFs for these chemicals should have sampled
water and tissue at similar points in time because tissue concentrations respond more
rapidly to changes in water concentrations. EPA will be providing additional guidance
on appropriate BAF study designs for nonionic organic chemicals (including those
appropriate to Procedure #3) in its forthcoming guidance document on conducting field
BAF and B SAP studies.
B. Baseline BAF^from a Laboratory-Measured BCF
1 . Except where noted below, a baseline BAFf should be calculated from a laboratory-
measured BCF^ using the guidance and equations outlined in Section 5.4.3.1(c) of
Procedure #1.
2. Food Chain Multipliers. Because biomagnification is not an overriding concern for the
minimally hydrophobic chemicals applicable to Procedure #3, FCMs are not used in the
derivation of a baseline BAFf from a laboratory -measured BCFj.
3. Freely Dissolved Fraction. Due to their low hydrophobicity (i.e., log Kow < 4.0),
nonionic organic chemicals to which Procedure #3 is applied are expected to remain
5-49
-------
almost entirely in the freely dissolved form in waters containing dissolved and participate
organic carbon concentrations typical of laboratory BCF studies. Therefore, the freely
dissolved fraction should usually be assumed equal to 1.0. The freely dissolved fraction
will be substantially less than 1.0 only in situations where unusually high concentrations
of DOC and POC are present in the laboratory BCF study (e.g., above about 100 mg/L
for DOC or about 10 mg/L for POC). In this situation, the freely dissolved fraction
should be calculated according to Equation 5-12.
C. Baseline EAfff from a Kow
1. Except where noted below, a baseline BAFf should be calculated from an acceptable Kow
using the guidance and equations outlined in Section 5.4.3.1(D) in Procedure #1.
2. Because biomagnification is not an overriding concern for nonionic organic chemicals
with low hydrophobicity (i.e., log Kow < 4.0), food chain multipliers are not used in
Procedure #3 for deriving the baseline BAFf from a Kow.
5.4.5.2 Selecting Final Baseline BAF™s
After calculating individual baseline BAFf s using as many of the methods in Procedure
#3 as possible, the next step is to determine a final baseline BAFf for each trophic level from the
individual baseline BAFf s (Figure 5-2). The final baseline BAFf will be used in the last step to
determine the national BAF for each trophic level. The final baseline BAFf for each trophic
level should be determined from the individual baseline BAFf s by considering the data
preference hierarchy defined by Procedure #3 and uncertainty in the data. The data preference
hierarchy for Procedure #3 is (in order of preference):
1. a baseline BAFf from an acceptable field-measured BAF or laboratory-measured
BCF, or
2. a baseline BAFf predicted from an acceptable Kow value.
This data preference hierarchy reflects EPA's preference for BAFs that are based on
measured data (field-measured BAFs and laboratory-measured BCFs) over BAFs based on
predictive methods (Kow). This data preference hierarchy should be used as a guide for selecting
the final baseline BAFf s when the uncertainty is similar among two or more baseline BAFf s
derived using different methods. Since bioaccumulation via dietary uptake and subsequent
biomagnification generally are not of concern for chemicals subject to Procedure #3, field-
measured BAFs and laboratory-measured BCFs are considered equally in determining the
national BAF.
Final baseline BAFf s should be selected for each trophic level using the following steps
and guidelines.
1. Calculate Species-Mean Baseline BAFfs. For each BAF method (i.e., field-measured
BAF, BAF from a lab-measured BCF, or BAF from a Kow) where more than one
5-50
-------
acceptable baseline BAFf is available for a given species, calculate a species-mean
baseline BAFf according to the guidance described previously in Procedure #1.
2. Calculate Trophic-Level-Mean Baseline BAFfs. For each BAF method where more
than one acceptable species-mean baseline BAFf is available within a given trophic
level, calculate the trophic-level-mean baseline BAFf as the geometric mean of
acceptable species-mean baseline BAFf s in that trophic level.
3 Select a Final Baseline BAF[d for Each Trophic Level. For each trophic level, select
the final baseline BAFf using best professional judgment by considering: (1) the data
preference hierarchy, (2) the relative uncertainties among trophic-level-mean baseline
BAFf s derived using different methods, and (3) the weight of evidence among the three
methods.
a. In general, when more than one trophic-level-mean baseline BAFf is available
within a given trophic level, the final baseline BAFf should be selected from the
most preferred BAF method defined by the data preference hierarchy for
Procedure #3. Within the first data preference tier, field-measured BAFs and
laboratory-measured BCFs are considered equally desirable for deriving a final
trophic-level-mean baseline BAFf using Procedure #3. If a trophic-level-mean
baseline BAFf is available from both a field-measured BAF and a laboratory-
measured BCF, the final baseline BAFf should be selected using the trophic-
level-mean baseline BAFf or BCFf with the least overall uncertainty.
b. If uncertainty in a trophic-level-mean baseline BAFf based on a higher tier (more
preferred) method is judged to be substantially greater than a trophic-level-mean
baseline BAFf from a lower tier method, then the final baseline BAFf should be
selected using a trophic-level-mean baseline BAFf from a lower tier method.
c. The above steps should be performed for each trophic level until a final baseline
BAFf is selected for trophic level two, three, and four.
5.4.5.3 Calculating the National BAFs
As described in Procedure #1, the last step in deriving a national BAF for a given trophic
level for nonionic organic chemicals is to convert the final baseline BAFf determined in the
previous step to a BAF that reflect conditions to which the national 304(a) criterion will apply
(Figure 5-2). Each national BAF should be determined from a final baseline BAFf according to
the following guidelines.
1. National BAF Equation. Except where noted below, national BAFs for trophic levels
two, three, and four should be calculated from the final, trophic-level-mean baseline
BAFf s using Equation 5-28 and associated guidance described in Procedure #1 (see
Section 5.4.3.3).
5-51
-------
2. Freely Dissolved Fraction. Due to their low hydrophobicity (i.e., log Kow < 4.0), a
freely dissolved fraction of 1.0 should be assumed for calculating national BAFs for
nonionic organic chemicals using Procedure #3. A freely dissolved fraction of 1.0 should
be assumed because at a log Kow of less than 4.0, nonionic organic chemicals are
expected to remain over 99 percent in the freely dissolved form at POC and DOC
concentrations corresponding to national default values for U.S. bodies of water (i.e., 0.5
mg/L and 2.9 mg/L, respectively).
5.4.6 Deriving National BAFs Using Procedure #4
This section provides guidance for calculating national BAFs for nonionic organic
chemicals using Procedure #4 shown in Figure 5-1. The types of nonionic organic chemicals for
which Procedure #4 is most appropriate are those that are classified as having low
hydrophobicity and subject to high rates of metabolism by aquatic biota (see Section 5.4.2
above). Non-aqueous contaminant exposure and subsequent biomagnification in aquatic food
webs are not generally of concern for chemicals that are classified in this category. As a result,
FCMs are not used in this procedure. In addition, Kow -based predictions of bioconcentration are
not used in this procedure since the Kow /BCF relationship is primarily based on poorly
metabolized chemicals. One example of a nonionic organic chemical for which Procedure #4
appears appropriate is butyl benzyl phthalate in fish. Using radiolabeling techniques with
confirmation by chromatographic analysis, Carr et al. (1997) present evidence that indicates
butyl benzyl phthalate is extensively metabolized in sunfish. Carr et al. (1997) also report
measured BCFs (and subsequently lipid-normalized BCFs) which are substantially below
predicted BCFs based on log Kow In a study of chlorinated anilines (which would be essentially
un-ionized at ambient pH), de Wolf et al. (1992) reported measured BCFs substantially lower
than those predicted based on Kow. The authors suggested that biotransformation (metabolism)
involving the amine (NH2) was responsible for the lower measured BCFs.
According to Procedure #4, the following two methods can be used in deriving a national
BAF:
• using a BAF from an acceptable field study (i.e., a field-measured BAF), and
• predicting a BAF from an acceptable BCF.
After selecting the derivation procedure, the next steps in deriving a national BAF for a
given trophic level for nonionic organic chemicals are: (1) calculating individual baseline
BAFf s, (2) selecting the final baseline BAFf, and (3) calculating the national BAF (Figure 5-2).
Each of these three steps is discussed separately below.
5.4.6.1 Calculating Individual Baseline BAFfs
Calculating individual baseline BAFf s involves normalizing the measured BAF^ or BCF j
(which are based on the total chemical in water and tissue) by the lipid content of the study
organism and the freely dissolved fraction of the chemical in the study water. For additional
discussion of the technical basis for calculating baseline BAFfs, see Section 5.4.3.1 in Procedure
#1.
5-52
-------
1. For each species where acceptable data are available, calculate all possible baseline
BAFf s using each of the two methods shown above for Procedure #4.
2. Individual baseline BAFf s should be calculated from field-measured BAFjs and
laboratory-measured BCF^s according to the following procedures.
A. Baseline BAP*? from Field-Measured BAFs
1. A baseline BAFf should be calculated from a field-measured BAFj using the guidance
and equations outlined in Section 5.4.3.1(A) in Procedure #1.
2. Freely Dissolved Fraction. Due to their low hydrophobicity (i.e., log Kow < 4.0),
nonionic organic chemicals applicable to Procedure #4 are expected to remain almost
entirely in the freely dissolved form in natural waters with dissolved and particulate
organic carbon concentrations typical of most field BAF studies. Therefore, the freely
dissolved fraction should be assumed equal to 1.0 unless the concentrations of DOC and
POC are very high in the field BAF study. For studies with very high DOC or POC
concentrations, (e.g., about 100 mg/L or higher for DOC or 10 mg/L or higher for POC),
the freely dissolved fraction may be substantially lower than 1.0 and therefore should be
calculated using Equation 5-12.
3. Temporal Averaging of Concentrations. Also due to their low hydrophobicity,
nonionic organic chemicals appropriate to Procedure #4 will also tend to reach steady-
state quickly compared with those chemicals to which Procedure #1 applies. Therefore,
the extent of temporal averaging of tissue and water concentrations is typically much less
than that required for highly hydrophobic chemicals to which Procedure #1 is applied. In
addition, field studies used to calculate BAFs for these chemicals should have sampled
water and tissue at similar points in time because tissue concentrations should respond
rapidly to changes in water concentrations. EPA will be providing additional guidance
on appropriate BAF study designs for nonionic organic chemicals (including those
appropriate to Procedure #4) in its forthcoming guidance document on conducting field
BAF and BSAP studies.
B. Baseline BAF^/from a Laboratory-Measured BCF
1. Except where noted below, a baseline BAFf should be calculated from a laboratory-
measured BCF^ using the guidance and equations outlined in Section 5.4.3.1(c) of
Procedure #1.
2. Food Chain Multipliers. Because biomagnification is not an important concern for the
minimally hydrophobic chemicals applicable to Procedure #4, FCMs are not used in the
derivation of a baseline BAFf from a laboratory-measured BCFj.
3. Freely Dissolved Fraction. Due to their low hydrophobicity (i.e., log Kow < 4.0),
nonionic organic chemicals to which Procedure #4 is applied are expected to remain
5-53
-------
almost entirely in the freely dissolved form in waters containing dissolved and participate
organic carbon concentrations typical of laboratory BCF studies. Therefore, the freely
dissolved fraction should usually be assumed to be equal to 1.0. The freely dissolved
fraction will be substantially less than 1.0 only in situations where unusually high
concentrations of DOC and POC are present in the lab BCF study (e.g., above about 100
mg/L for DOC or about 10 mg/L for POC). In this situation, the freely dissolved fraction
should be calculated according to Equation 5-12.
5.4.6.2 Selecting Final Baseline BAF™s
After calculating individual baseline BAFf s using as many of the methods in Procedure
#4 as possible, the next step is to determine a final baseline BAFf for a given trophic level from
the individual baseline BAFf s (Figure 5-2). The final baseline BAFf will be used in the last step
to determine the national BAF for each trophic level. A final baseline BAFf should be
determined for each trophic level from the individual baseline BAFf s by considering the data
preference hierarchy defined by Procedure #4 and uncertainty in the data. The data preference
hierarchy for Procedure #4 is:
1. a baseline BAFf from an acceptable field-measured BAF or predicted from an
acceptable laboratory-measured BCF.
Since bioaccumulation via dietary uptake and subsequent biomagnification generally are
not of concern for chemicals subject to Procedure #4, field-measured BAFs and laboratory-
measured BCFs are considered equally in determining the national BAF.
Final baseline BAFf s should be selected for each trophic level using the following steps
and guidelines.
1. Calculate Species-Mean Baseline BAFfs. For each BAF method (i.e., field-measured
BAF or a BAF from a lab-measured BCF) where more than one acceptable baseline
BAFf is available for a given species, calculate a species-mean baseline BAFf according
to the guidance described previously in Procedure #1.
2. Calculate Trophic-Level-Mean Baseline BAFfs. For each BAF method where more
than one acceptable species-mean baseline BAFf is available within a given trophic
level, calculate the trophic-level-mean baseline BAFf as the geometric mean of
acceptable species-mean baseline BAFf s for that trophic level.
3. Select a Final Baseline BAF[d for Each Trophic Level. For each trophic level, select
the final baseline BAFf using best professional judgment by considering: (1) the data
preference hierarchy, and (2) the relative uncertainties among trophic-level-mean BAFs
derived using different methods.
a. As discussed above, field-measured BAFs and laboratory-measured BCFs are
considered equally desirable for deriving a final trophic-level-mean baseline
5-54
-------
BAFf using Procedure #4. If a trophic-level-mean baseline BAFf is available
from both a field-measured BAF and a laboratory-measured BCF, the final
baseline BAFf should be selected using the trophic-level-mean baseline BAFf or
BCFf with the least overall uncertainty.
b. The above steps should be performed for each trophic level until a final baseline
BAFf is selected for trophic levels two, three, and four.
5.4.6.3 Calculating National BAFs
As described in Procedure #1, the last step in deriving a national BAF for a given trophic
level for nonionic organic chemicals is to convert the final baseline BAFf determined in the
previous step to a BAF that reflects conditions to which the national 304(a) criterion will apply
(Figure 5-2). Each national BAF should be determined from a final baseline BAFf according to
the following guidelines.
1. National BAF Equation. Except where noted below, national BAFs for trophic-levels
two, three, and four should be calculated from the final, trophic-level-mean baseline
BAFf s using the same equation and procedures described previously in Procedure #1
(see Section 5.4.3.3 in Procedure #1).
2. Freely Dissolved Fraction. Due to their low hydrophobicity (i.e., log Kow < 4.0), a
freely dissolved fraction of 1.0 should be assumed for calculating national BAFs for
nonionic organic chemicals using Procedure #4. A freely dissolved fraction of 1.0
should be assumed because at a log Kow value of less than 4.0, nonionic organic
chemicals are expected to remain over 99 percent in the freely dissolved form at POC and
DOC concentrations corresponding to national default values for U.S. bodies of water
(i.e., 0.5 mg/L and 2.9 mg/L, respectively).
5.5 NATIONAL BIO ACCUMULATION FACTORS FOR IONIC ORGANIC
CHEMICALS
This section contains guidelines for deriving national BAFs for ionic organic chemicals
(i.e., organic chemicals which undergo significant ionization in water). As defined in Section
5.3.5, ionic organic chemicals contain functional groups which can either readily donate protons
(e.g., organic acids with hydroxyl, carboxylic, and sulfonic groups) or readily accept protons
(e.g., organic bases with amino and aromatic heterocyclic nitrogen groups). Some examples of
ionic organic compounds include:
• chlorinated phenols (e.g., 2,4,6-trichlorophenol, pentachlorophenol),
• chlorinated phenoxyalkanoic acids (e.g., 2,4-dichlorophenoxyacetic acid [2,4-D]),
• nitrophenols (e.g., 2-nitrophenol, 2,4,6-trinitrophenol),
• cresols (e.g., 2,4-dinitro-o-cresol [DNOC]),
• pyridines (e.g., 2,4-dimethypyidine),
• aliphatic and aromatic amines (e.g., trimethylamine, aniline), and
5-55
-------
• linear alkylbenzenesulfonate (LAS) surfactants.
Ionic organic chemicals are considered separately for deriving national BAFs because the
anionic or cationic species of these chemicals behave much differently in the aquatic
environment compared with their neutral (un-ionized) counterparts. The neutral species of ionic
organic chemicals are thought to behave in a similar manner as nonionic organic compounds
(e.g., partitioning to lipids and organic carbon as a function of hydrophobicity). However, the
ionized (cationic, anionic) species exhibit a considerably more complex behavior involving
multiple environmental partitioning mechanisms (e.g., ion exchange, electrostatic, and
hydrophobic interactions) and a dependency on pH and other factors including ionic strength and
ionic composition (Jafvert et al., 1990; Jafvert 1990; Schwarzenbach, et al., 1993). As a
consequence, methods to predict the environmental partitioning of organic cations and anions are
less developed and validated compared with methods for nonionic organic chemicals (Spacie,
1994;Suffetetal., 1994).
Given the current limitations in the state of the science for predicting the partitioning and
bioaccumulation of the ionized species of ionic organic chemicals, procedures for deriving
national BAFs for these chemicals differ depending on the extent to which the fraction of the
total chemical is likely to be represented by the ionized (cationic, anionic) species in U.S.
surface waters. When a significant fraction of the total chemical concentration is expected to be
present as the ionized species in water, procedures for deriving the national BAF rely on
empirical (measured) methods (i.e., Procedures #5 and 6 in Section 5.6). When an insignificant
fraction of the total chemical is expected to be present as the ionized species (i.e., the chemical
exists essentially in the neutral form), procedures for deriving the national BAF will follow those
established for nonionic organic chemicals (e.g., Procedures #1 through #4 in Section 5.4). The
following guidelines apply for assessing the occurrence of cationic and anionic forms at typical
environmental pH ranges.
1. For the ionic organic chemical of concern, the dissociation constant, pKa, should be
compared to the range of pH values expected in fresh and estuarine waters of the U.S. At
pH equal to the pKa, 50% of the organic acid or base is expected to be present in the
ionized species. The pH values for U.S. fresh and estuarine waters typically range
between 6 and 9, although somewhat higher and lower values can occur in some bodies
of water (e.g., acidic bogs and lakes, highly alkaline and eutrophic systems, etc.).
2. For organic acids, the chemical will exist almost entirely in its un-ionized form when pH
is about 2 or more units below the pKa. For organic bases, the chemical will exist almost
entirely in its un-ionized form when pH is about 2 or more units above the pKa. In these
cases, the aqueous behavior of the chemical would be expected to be similar to nonionic
organic chemicals. Therefore, national BAF should usually be derived using Procedures
#1 through #4 in Section 5.4.
3. When pH is greater than the pKa minus 2 for organic acids (or less than the pKa plus 2
for organic bases), the fraction of the total chemical that is expected to exist in its ionized
form can become significant (i.e., > 1% in the ionized). In these cases, the national BAF
should usually be derived using Procedures #5 and #6 in Section 5.6.
5-56
-------
4. In general, most organic acids (e.g., pentachlorophenol and silvex), exist primarily in the
ionized form in ambient waters because their pKa's (4.75 and 3.07, respectively) are
much smaller than the pH of the ambient waters. Conversely, most organic bases, (e.g.,
aniline) exist mostly in the un-ionized form in ambient waters because their pKa's (4.63
for aniline) are much smaller than the pH of the ambient waters.
5. The above guidelines are intended to be a general guide for deriving national BAFs for
ionic organic chemicals, not an inflexible rule. Modifications to these guidelines should
be considered on a case-by-case basis, particularly when such modifications are strongly
supported by measured bioaccumulation or bioconcentration data. For example, initial
models have been developed for predicting the solid and organic-phase partitioning of
certain organic acids (e.g., Jafvert 1990, Jafvert et al., 1990). As these or other models
become more fully developed and appropriately validated in the future, they should be
considered in the development of national BAFs. In addition, since pH is a controlling
factor for dissociation and subsequent partitioning of ionic organic chemicals,
consideration should be given to expressing BAFs or BCFs as a function of pH (or other
factors) where sufficient data exist to reliably establish such relationships.
5-57
-------
5.6 NATIONAL BIO ACCUMULATION FACTORS FOR INORGANIC AND
ORGANOMETALLIC CHEMICALS
This section contains guidelines for deriving national BAFs for inorganic and
organometallic chemicals as defined in Section 5.3.5. The derivation of BAFs for inorganic and
organometallic chemicals differs in several ways from procedures for nonionic organic
chemicals. First, lipid normalization of chemical concentrations in tissues does not generally
apply for inorganic and organometallic chemicals. Thus, BAFs and BCFs cannot be extrapolated
from one tissue to another based on lipid-normalized concentrations as is done for nonionic
organic chemicals. Second, the bioavailability of inorganics and organometallics in water tends
to be chemical-specific and thus, the techniques for expressing concentrations of nonionic
organic chemicals based on the freely dissolved form do not generally apply. Third, at the
present time there are no generic bioaccumulation models that can be used to predict BAFs for
inorganic and organometallic chemicals as a whole, unlike the existence of Kow-based models for
nonionic organic chemicals. While some chemical-specific bioaccumulation models have been
developed for inorganic and organometallic chemicals (e.g., Mercury Cycling Model by Hudson
et. al, 1994), those models currently tend to require site-specific data for input to the model and
are restricted to site-specific applications. As the models become more fully developed and
validated in the future, they should be considered on a case-by-case basis in conjunction with the
following procedures for deriving national BAFs.
5.6.1 Selecting the BAF Derivation Procedure
As shown in Figure 5-1, national BAFs can be derived using two procedures for
inorganic and organometallic chemicals (Procedures #5 and #6). The choice of the BAF
derivation procedure depends on whether or not the chemical undergoes biomagnification in
aquatic food webs.
1. For many inorganic and organometallic chemicals, biomagnification does not occur and
the BCF will be equal to the BAF. For these types of chemicals, Procedure #5 should be
used to derive the national BAF. Procedure #5 considers BAFs and BCFs to be of equal
value in determining the national BAF and does not require the use of FCMs with BCF
measurements. Guidance for deriving BAFs using Procedure #5 is provided in Section
5.6.3.
2. For some inorganic and organometallic chemicals (e.g., methylmercury),
biomagnification does occur and Procedure #6 should be used to determine the national
BAF. Procedure #6 gives general preference to the use of field-measured BAFs over
laboratory-measured BCFs and requires FCMs to be used with BCF measurements for
predicting BAFs. Guidance for deriving BAFs using Procedure #6 is provided in Section
5.6.4.
3. Determining whether or not biomagnification occurs for inorganic and organometallic
chemicals requires chemical-specific data on measured concentrations of the chemical in
aquatic organisms and their prey. Concentrations in aquatic organisms that increase
substantially at successive trophic levels of a food web suggest that biomagnification is
5-58
-------
occurring. Concentrations in aquatic organisms that remain about the same or decrease at
successive trophic levels of a food web suggest that biomagnification is not occurring.
When comparing tissue concentrations for assessing biomagnification, care should be
taken to ensure that the aquatic organisms chosen actually represent functional predator-
prey relationships and that all major prey species are considered in the comparisons.
5.6.2 Bioavailability
The chemical-specific nature of inorganic and organometallic bioavailability is likely due
in part to chemical-specific differences in several factors which affect bioavailability and
bioaccumulation. These factors include differences in the mechanisms for chemical uptake by
aquatic organisms (e.g., passive diffusion, facilitated transport, active transport), differences in
sorption affinities to biotic and abiotic ligands, and differences in chemical speciation in water.
Some inorganic and organometallic chemicals exist in multiple forms and valence states in
aquatic ecosystems that can differ in their bioavailability to aquatic organisms and undergo
conversions between forms. For example, selenium can exist in various forms in aquatic
ecosystems, including inorganic selenite(+4) and selenate(+6) oxyanions, elemental selenium (°)
under reducing conditions (primarily in sediments), and organoselenium compounds of selenide
("2). Dominant forms of mercury in natural, oxic waters include inorganic (+2) mercury
compounds and methylmercury; the latter is generally considered to be substantially more
bioavailable than inorganic mercury compounds to higher trophic level organisms. Although a
generic analogue to the "freely dissolved" conversion for nonionic organic chemicals does not
presently exist for inorganic and organometallic chemicals as a whole, the occurrence and
bioavailability of different forms of these chemicals should be carefully considered when
deriving national BAFs.
1. If data indicate that: (1) a particular form (or multiple forms) of the chemical of concern
largely governs its bioavailability to target aquatic organisms, and (2) BAFs are more
reliable when derived using the bioavailable form(s) compared with using other form(s)
of the chemical of concern, then BAFs and BCFs should be based on the appropriate
bioavailable form(s).
2. Because different forms of many inorganic and organometallic chemicals may
interconvert once released to the aquatic environment, regulatory and mass balance
considerations typically require an accounting of the total concentration in water. In
these cases, sufficient data should be available to enable conversion between total
concentrations and the other (presumably more bioavailable) forms in water.
5.6.3 Deriving BAFs Using Procedure #5
This section contains guidance for calculating national BAFs for inorganic and
organometallic chemicals using Procedure #5 as shown in Figure 5-1. The types of inorganic
and organometallic chemicals for which Procedure #5 is appropriate are those that are not likely
to biomagnify in aquatic food webs (see Section 5.1 above). In Procedure #5, two methods are
available to derive the national BAF for a given trophic level:
5-59
-------
using a BAF from an acceptable field study (i.e., field-measured BAF), or
predicting a BAF from an acceptable laboratory-measured BCF.
Individual BAFs should be determined from field-measured BAFs or laboratory-measured BCFs
according to the following guidelines.
5.6.3.1 Determining Field-Measured BAFs
1. Except where noted below, field-measured BAFs should be determined using the
guidance provided in Section 5.4.3.1(A) of Procedure #1.
2. As described previously, conversion of field-measured BAFs to baseline BAFf s based on
lipid-normalized and freely-dissolved concentrations does not apply for inorganic and
organometallic chemicals. Therefore, the guidance and equations provided in Procedure
#1 which pertain to converting field-measured BAFs to baseline BAFf s and subsequently
to national BAFs do not generally apply to inorganic chemicals. As discussed in Section
5.6.2 above, an analogous procedure in concept might be required for converting total
BAFs to BAFs based on the most bioavailable form(s) for some inorganic and
organometallic chemicals of concern. Such procedures should be applied on a chemical-
specific basis.
3. BAFs should be expressed on a wet-weight basis; BAFs reported on a dry-weight basis
can be used only if they are converted to a wet-weight basis using a conversion factor
that is measured or reliably estimated for the tissue used in the determination of the BAF.
4. BAFs should be based on concentrations in the edible tissue(s) of the biota unless it is
demonstrated that whole-body BAFs are similar to edible tissue BAFs. For some finfish
and shellfish species, whole body is considered to be the edible tissue.
5. The concentrations of an inorganic or organometallic chemical in a bioaccumulation
study should be greater than normal background levels and greater than levels required
for normal nutrition of the test species if the chemical is a micronutrient, but below levels
that adversely affect the species. Bioaccumulation of an inorganic or organometallic
chemical that is essential to the nutrition of aquatic organisms might be overestimated if
concentrations are at or below normal background levels due to selective accumulation
by the organisms to meet their nutritional requirements.
5-60
-------
5.6.3.2 Determining Laboratory-Measured BCFs
1. Except where noted below, BAFs should be predicted from laboratory-measured BCFs
using the guidance provided in Section 5.4.3.1(c) of Procedure #1.
2. As described previously, conversion of laboratory-measured BCFs to baseline BCFf s
based on lipid-normalized and freely dissolved concentrations does not apply for
inorganic and organometallic chemicals. Therefore, the guidance and equations provided
in Procedure #1 which pertain to converting laboratory-measured BCFs to baseline
BCFf s and subsequently to national BCFs do not generally apply to inorganic and
organometallic chemicals. As discussed in Section 5.6.2 above, an analogous procedure
in concept might be required for converting total BCFs to BCFs based on the most
bioavailable form(s) of some inorganic and organometallic chemicals of concern. Such
procedures should be applied on a chemical-specific basis. In addition, the use of FCMs
with BCFs does not apply to chemicals applicable to Procedure #5.
3. BCFs should be expressed on a wet-weight basis; BCFs reported on a dry-weight basis
can be used only if they are converted to a wet-weight basis using a conversion factor
that is measured or reliably estimated for the tissue used in the determination of the BCF.
4. BCFs should be based on concentrations in the edible tissue(s) of the biota unless it is
demonstrated that whole-body BCFs are similar to edible tissue BCFs. For some finfish
and shellfish species, whole body is considered to be the edible tissue.
5. The concentrations of an inorganic or organometallic chemical in a bioconcentration test
should be greater than normal background levels and greater than levels required for
normal nutrition of the test species if the chemical is a micronutrient, but below levels
that adversely affect the species. Bioaccumulation of an inorganic or organometallic
chemical that is essential to the nutrition of aquatic organisms might be overestimated if
concentrations are at or below normal background levels due to selective accumulation
by the organisms to meet their nutritional requirements.
5.6.3.3 Determining the National BAFs
After calculating individual BAFs using as many of the methods in Procedure #5 as
possible, the next step is to determine national BAFs for each trophic level from the individual
BAFs. The national BAFs will be used to determine the national 304(a) criteria. The national
BAFs should be determined from the individual BAFs by considering the data preference
hierarchy defined for Procedure #5 and uncertainty in the data. The data preference hierarchy
for Procedure #5 is:
1. a BAF from an acceptable field-measured BAF or predicted from an acceptable
laboratory-measured BCF.
Since bioaccumulation via dietary uptake and subsequent biomagnification are not of
concern for chemicals subject to Procedure #5, field-measured BAFs and laboratory-measured
5-61
-------
BCFs are considered equally in determining the national BAFs. The national BAFs should be
selected for each trophic level using the following steps and guidelines.
1. Calculate Species-Mean BAFs. For each BAF method where more than one acceptable
field-measured BAF (or a BAF predicted from a BCF) is available for a given species,
calculate the species-mean BAF as the geometric mean of all acceptable individual
measured or BCF-predicted BAFs. When calculating species-mean BAFs, individual
measured or BCF-predicted BAFs should be reviewed carefully to assess uncertainties in
the BAF values. Highly uncertain BAFs should not be used. Large differences in
individual BAFs for a given species (e.g., greater than a factor of 10) should be
investigated further and in such cases, some or all of the BAFs for a given species might
not be used. Additional discussion on evaluating the acceptability of BAF and BCF
values is provided in the Bioaccumulation TSD.
2. Calculate Trophic-Level-Mean BAFs. For each BAF method where more than one
acceptable species-mean BAF is available within a given trophic level, calculate the
trophic-level-mean BAF as the geometric mean of acceptable species-mean BAFs in that
trophic level. Trophic-level-mean BAFs should be calculated for trophic levels two,
three and four because available data on U.S. consumers offish and shellfish indicate
significant consumption of organisms in these trophic levels.
3 Select a Final National BAF for Each Trophic Level. For each trophic level, select the
final national BAF using best professional judgment by considering: (1) the data
preference hierarchy in Procedure #5, and (2) the relative uncertainties among trophic
level-mean BAFs derived using different methods.
a. As discussed above, field-measured BAFs and laboratory-measured BCFs are
considered equally desirable for deriving a final national BAF using Procedure
#5. If a trophic-level-mean BAF is available from both a field-measured BAF
and a laboratory-measured BCF, the final national BAF should be selected using
the trophic-level-mean BAF with the least overall uncertainty.
b. The above steps should be performed for each trophic level until a national BAF
is selected for trophic levels two, three, and four.
5.6.4 Deriving BAFs Using Procedure #6
This section contains guidance for calculating national BAFs for inorganic and
organometallic chemicals using Procedure #6 as shown in Figure 5-1. The types of inorganic
and organometallic chemicals for which Procedure #6 is appropriate are those that are
considered likely to biomagnify in aquatic food webs (see Section 5.6.1 above). Methylmercury
is an example of an organometallic chemical to which Procedure #6 applies. In Procedure #6,
two methods are available to derive the national BAF:
• using a BAF from an acceptable field study (i.e., field-measured BAF), or
5-62
-------
• predicting a BAF from an acceptable laboratory-measured BCF and a FCM.
Individual BAFs should be determined from field-measured BAFs or laboratory-measured BCFs
and FCMs according to the following guidelines.
5.6.4.1 Determining Field-Measured BAFs
1. Field-measured BAFs should be determined using the guidance provided in Section
5.6.3.1 of Procedure #5.
5.6.4.2 Determining Laboratory-Measured BCFs
1. Except where noted below, BAFs should be predicted from laboratory-measured BCFs
using the guidance provided in Section 5.6.3.2 of Procedure #5.
2. Because biomagnification is of concern for chemicals applicable to Procedure #6, BAFs
should be predicted from laboratory-measured BCF using FCMs. Currently, there are no
generic models from which to predict FCMs for inorganic or organometallic chemicals.
Therefore, FCMs should be determined using field data as described in the section
entitled: "Field-Derived FCMs" in Section 5.4.3.1(c) of Procedure #1. Unlike nonionic
organic chemicals, field-derived FCMs for inorganic and organometallic chemicals are
not based on lipid-normalized concentrations in tissues. For calculating FCMs for
inorganic and organometallic chemicals, concentrations in tissues should be based on the
consistent use of either wet-weight or dry-weight concentrations in edible tissues. FCMs
should be derived for trophic levels two, three, and four.
5.6.4.3 Determining the National BAF
After calculating individual BAFs using as many of the methods in Procedure #6 as
possible, the next step is to determine national BAFs for each trophic level from the individual
BAFs. The national BAFs will be used to determine the national 304(a) criteria. The national
BAFs should be determined from the individual BAFs by considering the data preference
hierarchy defined for Procedure #6 and uncertainty in the data. The data preference hierarchy
for Procedure #6 is (in order of preference):
1. a BAF from an acceptable field-measured BAF, or
2. a predicted BAF from an acceptable laboratory-measured BCF and FCM.
This data preference hierarchy reflects EPA's preference for field-measured BAFs over
BAFs predicted from a laboratory-measured BCF and FCM, because field-measured BAFs are
direct measures of bioaccumulation and biomagnification in aquatic food webs. BAFs predicted
from laboratory-measured BCFs and FCMs indirectly account for biomagnification through the
use of the FCM. For each trophic level, the national BAFs should be determined using the
following steps and guidelines.
5-63
-------
1. Calculate Species-Mean BAFs. For each BAF method where more than one acceptable
field-measured BAF or BAF predicted using a BCF and FCM is available, calculate a
species-mean BAF according to the guidance described previously in Procedure #5.
2. Calculate Trophic Level-Mean BAFs. For each BAF method where more than one
acceptable species-mean BAF is available within a given trophic level, calculate the
trophic level-mean BAF according to guidance described previously in Procedure #5.
3. Select a Final National BAF for Each Trophic Level. For each trophic level, select the
final national BAF using best professional judgment by considering: (1) the data
preference hierarchy in Procedure #6, and (2) the relative uncertainties among trophic
level-mean BAFs derived using different methods.
a. When a trophic-level mean BAF is available using both methods for a given
trophic level (i.e., a field-measured BAF and a BAF predicted from a BCF and
FCM), the national BAF should usually be selected using the field-measured BAF
which is the preferred BAF method in the data preference hierarchy in Procedure
#6.
b. If uncertainty in the trophic-level mean BAF derived using field-measured BAFs
is considered to be substantially greater than a trophic-level mean BAF derived
using a BCF and FCM, the national BAF for that trophic level should be selected
from the second tier (BCF • FCM) method.
c. The above steps should be performed for each trophic level until a national BAF
is selected for trophic levels two, three, and four.
5.7 REFERENCES
ASTM (American Society of Testing and Materials). 1999. Standard practice for conducting
bioconcentration tests with fishes and saltwater bivalve molluscs. Designation E 1022 -
94. In: Annual Book of ASTM standards. Volume 11.05. Pp. 333-350.
Barren, M.G. 1990. Bioconcentration: Will water-borne organic chemicals accumulate in
aquatic animals? Environ. Sci. Technol. 24:1612-1618.
Burkhard, L.P. and M.T. Lukasewycz. 2000. Some bioaccumulation factors and biota-sediment
accumulation factors for polycyclic aromatic hydrocarbons in lake trout. Environ.
Toxicol. Chem. 19:1427-1429.
Carr, K.H., G.T. Coyle and R.A. Kimerle. 1997. Bioconcentration of [14C]butyl benzyl phthalate
inbluegill sunfish (LepomisMacrochirus). Environ. Toxicol. Chem. 16:2200-2203.
Connell, D.W. 1988. Bioaccumulation behavior of persistent organic chemicals with aquatic
organisms. Rev. Environ. Contain. Toxicol. 101:117-159.
5-64
-------
Connolly, J.P. and C.G. Pedersen. 1988. A thermodynamic-based evaluation of organic
chemical accumulation in aquatic organisms. Environ. Sci. Technol. 22:99-103.
Cook, P.M. and L.P. Burkhard. 1998. Development of bioaccumulation factors for protection of
fish and wildlife in the Great Lakes. In: National Sediment Bioaccumulation Conference
Proceedings. U.S. Environmental Protection Agency, Office of Water. Washington, DC.
EPA 823-R-002.
de Wolf, W., J.H.M. de Bruijn, W. Seinen and J.L.M. Hermans. 1992. Influence of
biotransformation on the relationship between bioconcentration factors and octanol-water
partition coefficients. Environ. Sci. Technol. 26:1197-1201.
Fisk, A.T., RJ. Norstrom, C.C. Cymbalisty and D.C.B. Muir. 1998. Dietary accumulation and
depuration of hydrophobic organochlorines: Bioaccumulation parameters and their
relationship with the octanol/water partition coefficient. Environ. Toxicol. Chem. 17:
951-961.
Gobas, F.A.P.C., J.R. McCorquodale and G.D. Haffner. 1993. Intestinal absorption and
biomagnification of organochlorines. Environ. Toxicol. Chem. 12:567-576.
Gobas, F.A.P.C. 1993. A model for predicting the bioaccumulation of hydrophobic organic
chemicals in aquatic food-webs: application to Lake Ontario. Ecol. Mod. 69:1-17
Hudson, R.J.M., A.S. Gherini, CJ. Watras and D.B. Porcella. 1994. Modeling the
biogeochemical cycle of mercury in lakes: the Mercury Cycling Model (MCM) and its
application to the MTL Study Lakes, In: CJ. Watras and J.W. Huckabee (eds.), Mercury
Pollution: Integration and Synthesis. Lewis Publishers. Boca Raton, FL. Pp. 473-523.
Isnard, P. and S. Lambert. 1988. Estimating bioconcentration partition coefficients and aqueous
solubility. Chemosphere 17:21-34.
Jafvert, C.T. 1990. Sorption of organic acid compounds to sediments: initial model
development. Environ. Toxicol. Chem. 9:1259-1268.
Jafvert, C.T., J.C. Westell, E. Grieder and R.P. Schwarzenbach. 1990. Distribution of
hydrophobic ionogenic organic compounds between octanol and water: Organic acids.
Environ. Sci. Technol. 24:1795-1803.
James, M.O. 1989. Biotransformation and disposition of PAH in aquatic invertebrates, In: U.
Varanasi (ed.). Metabolism ofPolycyclic Aromatic Hydrocarbons in the Aquatic
Environment. CRC Press, Inc. Boca Raton, FL. Pp. 69-92.
Mackay, D. 1982. Correlation of bioconcentration factors. Environ. Sci. Technol. 16:274-278.
Miranda, C.L., M.C. Henderson, and D.R. Buhler. 1998. Evaluation of chemicals as inhibitors
of trout cytochrome p450s. Toxicol. Appl. Pharmacol. 148:327-244.
5-65
-------
Niimi, AJ. 1985. Use of laboratory studies in assessing the behavior of contaminants in fish
inhabiting natural ecosystems. Wat. Poll. Res. J. Can. 20:79-88.
Oliver, E.G. and AJ. Niimi. 1983. Bioconcentration of chlorobenzenes from water by rainbow
trout: Correlations with partition coefficients and environmental residues. Environ. Sci.
Technol. 17:287-291.
Oliver, E.G. and AJ. Niimi. 1988. Trophodynamic analysis of poly chlorinated biphenyl
congeners and other chlorinated hydrocarbons in the Lake Ontario ecosystem. Environ.
Sci. Technol. 22:388-397.
Randall, R.C., H. Lee II, RJ. Ozretich, J.L. Lake and RJ. Pruell. 1991. Evaluation of Selected
Lipid Methods for Normalizing Pollutant Bioaccumulation. Environ. Toxicol. Chem. 10:
1431-1436.
Randall, R.C., D.R. Young, H. Lee, and S.F. Echols. 1998. Lipid methodology and pollutant
normalization relationships for neutral nonpolar organic pollutants. Environ. Toxicol.
Chem. 17:788-791.
Russell, R.W., F.A.P.C. Gobas, and G.D. Haffner. 1999. Role of chemical and ecological
factors in trophic transfer of organic chemicals in aquatic food webs. Environ. Toxicol.
Chem. 18:1250-1257.
Schwarzenbach, R.P., P.M. Gschwend, and D.M. Imboden. 1993. Environmental Organic
Chemistry. John Wiley and Sons, Inc. New York, NY.
Spacie, L.L. 1994. Interactions of organic pollutants with inorganic solid phases: are they
important to bioavailability? In: Bioavailability: Physical, Chemical and Biological
Interactions. Hamelink, J.L., Landrum, P.P., Bergman, H.L., and W.H. Benson (eds.).
Proceedings of the Thirteenth Pellston Workshop, Pellston, MI. August 17-22, 1992.
SETAC Special Publication Series. CRC Press, Inc. Boca Raton, FL. Pp. 73-82.
Suffet, I.H., C.T. Jafvert, J. Kukkonen, M.R. Servos, A. Spacie, L.L. Williams, and J.A. Noblet.
1994. Synopsis of discussion sessions: influence of particulate and dissolved material on
the bioavailability of organic compounds, In: Bioavailability: Physical, Chemical and
Biological Interactions. Hamelink, J.L., Landrum, P.P., Bergman, H.L., and W.H.
Benson (Eds.). Proceedings of the Thirteenth Pellston Workshop, Pellston, MI. August
17-22, 1992. SETAC Special Publication Series. CRC Press, Inc. Boca Raton, FL. Pp.
93-108.
Swackhamer, D.L. and R.A. Kites. 1988. Occurrence and bioaccumulation of organochlorine
compounds in fishes from Siskiwit Lake, Isle Royale, Lake Superior. Environ. Sci.
Technol. 22:543-548.
Thomann, R.V. 1989. Bioaccumulation model of organic chemical distribution in aquatic food
chains. Environ. Sci. Technol. 23:699-707.
5-66
-------
USEPA (U.S. Environmental Protection Agency). 1980. Appendix C-Guidelines and
methodology used in the preparation of health effect assessment chapters of the consent
decree water criteria documents. Federal Register 45:79347-79357. November 28.
USEPA (U.S. Environmental Protection Agency). 1991. Technical Support Document for
Water Quality-Based Toxics Control. Office of Water. Washington, DC. EPA/505/2-
90/001.
USEPA (U.S. Environmental Protection Agency). 1993. Assessment and control of
bioconcentratable contaminants in surface water. Federal Register 56:13150.
USEPA (U.S. Environmental Protection Agency). 1995a. Final water quality guidance for the
Great Lakes system; Final Rule. Federal Register 60:15366-15425. March 23.
USEPA (U.S. Environmental Protection Agency). 1995b. Great Lakes Water Quality Initiative
Technical Support Document for the Procedure to Determine Bioaccumulation Factors.
Office of Water. Washington, DC. EPA/820/B-95/005.
USEPA (U.S. Environmental Protection Agency). 1996. Ecological Effects Test Guidelines.
OPPTS 850.1730 Fish BCF. Public Draft. Office of Prevention, Pesticides and Toxic
Substances. Washington, DC. EPA/712/C-96/129. April.
USEPA (U.S. Environmental Protection Agency). 1997. Revisions to the polychlorinated
biphenyl criteria for human health and wildlife for the water quality guidance for the
Great Lakes system; Final Rule. Federal Register 62:11723-11731. March 12.
USEPA (U.S. Environmental Protection Agency). 2000a. Trophic Level and Exposure Analyses
for Selected Piscivorous Birds and Mammals. Volume I: Analyses of Species for the
Great Lakes. Draft. Office of Water. Washington, DC. August.
USEPA (U.S. Environmental Protection Agency). 2000b. Trophic Level and Exposure Analyses
for Selected Piscivorous Birds and Mammals. Volume II: Analyses of Species in the
Conterminous United States. Draft. Office of Water. Washington, DC. August.
USEPA (U.S. Environmental Protection Agency). 2000c. Trophic Level and Exposure Analyses
for Selected Piscivorous Birds and Mammals. Volume III: Appendices. Draft. Office of
Water. Washington, DC. August.
USEPA (U.S. Environmental Protection Agency). 2000d. Technical Basis for the Derivation of
Equilibrium Partitioning Sediment Guidelines (ESGs) for the Protection ofBenthic
Organisms: Nonionic Organics. Office of Science and Technology, Office of Research
and Development. Washington, DC. June Draft.
Veith, G.D., D.F.L. DeFoe and B.V. Bergstedt. 1979. Measuring and estimating the
bioconcentration factor in fish. J. Fish. Res. Board Can. 36:1040-1045.
5-67
-------
5-68
------- |