Guidance for Developing
Ecological Soil Screening Levels
OSWER Directive 9285.7-55
Soil Screening
^ PROtซ
U.S. Environmental Protection Agency
Office of Solid Waste and Emergency Response
1200 Pennsylvania Avenue, N.W.
Washington, DC 20460
November 2003
Revised February 2005
-------
This Page Intentionally Left Blank
-------
EXECUTIVE SUMMARY
This document describes the process used to derive a set of risk-based ecological soil screening
levels (Eco-SSLs) for many of the soil contaminants that are frequently of ecological concern for
plants and animals at hazardous waste sites and provides guidance for their use. The Eco-SSL
derivation process represents the group effort of a multi-stakeholder workgroup consisting of
federal, state, consulting, industry, and academic participants led by the U.S. Environmental
Protection Agency Office of Superfund Remediation and Technology Innovation (OSRTI). The
Eco-SSLs are concentrations of contaminants in soil that are protective of ecological receptors
that commonly come into contact with soil or ingest biota that live in or on soil. These values
can be used to identify those contaminants of potential concern in soils requiring further
evaluation in a baseline ecological risk assessment. The Eco-SSLs should be used during Step 2
of the Superfund Ecological Risk Assessment process, the screening-level risk calculation. The
Eco-SSLs are not designed to be used as cleanup levels and EPA emphasizes that it is
inappropriate to adopt or modify these Eco-SSLs as cleanup standards.
EPA derived the Eco-SSLs in order to conserve resources by limiting the need for EPA and other
risk assessors to perform repetitious toxicity data literature searches and data evaluations for the
same contaminants at every site. This effort should also allow risk assessors to focus their
resources on key site-specific studies needed for critical decision-making. EPA also expects that
the Eco-SSLs will increase consistency among screening risk analyses and decrease the
possibility that potential risks from soil contamination to ecological receptors will be
overlooked.
EPA prepared a list of twenty-four (24) contaminants to be addressed initially by the Eco-SSL
guidance. This list was based on a review of the contaminants of concern reported in recent
Record of Decisions at Superfund National Priority List sites. The Eco-SSL contaminant list
also included contaminants nominated by the EPA regional Biological Technical Assistance
Group Coordinators. The list of 24 Eco-SSL contaminants contained 17 metals including
aluminum, antimony, arsenic, barium, beryllium, cadmium, chromium, cobalt, copper, iron, lead,
manganese, nickel, selenium, silver, vanadium, and zinc. The organic contaminants on the list
were dieldrin, Hexahydro -l,3,5-trinitro-l,3,5-triazine (RDX), trinitrotoluene (TNT),
l,l,l-Trichloro-2,2-bis (p-chlorophenyl)ethane (DDT) and metabolites (DDE and ODD),
pentachlorophenol, polycyclic aromatic hydrocarbons (PAHs), and polychlorinated biphenyls
(PCBs).
The omission of other contaminants, such as phthalates, cyanides, dioxins and mercury, does not
imply that these contaminants can be excluded from the ERA screening process for soil
contamination. The processes and procedures established here for developing the Eco-SSLs
were intended to be sufficiently transparent to allow others to derive values for additional
contaminants, as needed. PCBs were included by the workgroup in the original Eco-SSL
contaminant list. However, it became apparent early in the process, that development of a PCB
soil screening value was not warranted. Because of the known high persistence and toxicity of
PCBs, and the conservative nature of the Eco-SSLs, it was acknowledged that soil screening
ES- 1
-------
levels derived for PCBs would normally be lower than the soil analyses detection limits. EPA
believes that if PCBs are detected in soil above background levels, the PCBs are probably site-
related and therefore should be included as a contaminant of potential concern in the baseline
risk assessment.
The approach developed for deriving the Eco-SSLs for plants and soil invertebrates was similar
to the approach taken for deriving the wildlife Eco-SSLs (specifically the toxicity reference
values). The general approach included four steps: (1) conduct literature searches, (2) screen
identified literature with exclusion and acceptability criteria, (3) extract, evaluate, and score test
results for applicability in deriving an Eco-SSL, and (4) derive the value. These procedures were
finalized as standard operating procedures prior to initiating any work to derive the actual values.
Chapter 3 provides a description of the procedures used for deriving plant and soil invertebrate
Eco-SSL values. The values were derived directly after an evaluation of all available plant and
soil invertebrate chronic toxicity test data (measured toxicity related to soil contaminant
concentrations). Chapter 4 provides a description of the procedures for deriving the wildlife
Eco-SSLs. The wildlife Eco-SSLs were the result of back-calculations from a hazard quotient of
1.0. The hazard quotient is equal to the estimated exposure dose divided by the toxicity
reference value (TRV). An HQ of 1.0 is the condition where the exposure and the dose
associated with no adverse chronic effects are equal, indicating adverse effects at or below this
soil concentration are unlikely. A generic food-chain model was used to estimate the
relationship between the concentration of the contaminant in soil and the dose for the receptor
(mg per kg body weight per day). The TRV represents a receptor-class specific estimate of a
no-observed adverse effect level (dose) for the respective contaminant for chronic exposure.
The Eco-SSLs apply to sites where terrestrial receptors may be exposed directly or indirectly to
contaminated soil. Seven groups of ecological receptors were initially considered in the
development of the Eco-SSLs. These included mammals, birds, reptiles, amphibians, soil
invertebrates, plants, and soil microbes and their processes. After investigation, the toxicity data
for amphibians and reptiles were deemed insufficient to derive Eco-SSLs. EPA recognizes that
the Eco-SSL may not be protective of these receptor groups. Eco-SSLs protective of microbes
and soil microbial processes were also not derived. Like amphibians and reptiles, EPA
recognizes their importance within terrestrial systems, but concurs with the workgroup that data
are insufficient and the interpretation of test results too uncertain for establishing risk-based
thresholds.
Eco-SSLs are appropriate to all sites where key soil parameters fall within a certain range of
chemical and physical parameters. The Eco-SSLs for plants and soil invertebrates were derived
to apply to soils where the pH is greater than or equal to 4.0 and less than or equal to 8.5, and the
organic matter content is less than or equal to 10%. Based on these stated parameters, it is
expected that there are certain soils and situations to which Eco-SSLs do not apply. These
situations include (but may not be limited to) wetland soils that are regularly flooded (i.e.,
sediments), sewage sludge amended soils where the organic matter content is > 10%, and waste
types where the pH is < 4.0.
ES-2
-------
Because of the diversity of the workgroup scientists, the process developed to derive the Eco-
SSLs underwent constant peer review. There were also two external peer reviews performed
during the development process. The first was a consultation requested by EPA's Office of Solid
Waste and Emergency Response of EPA's Science Advisory Board (SAB). This consultation
was held on April 6, 1999. At this meeting, the SAB provided verbal comments to the presenters
which were subsequently addressed, as appropriate, by the workgroup as they prepared the
guidance. A peer review of the draft guidance document was also performed. The peer review
workshop was held on July 26 and 27, 2000 and was open to the public. Each of the comments
received was carefully considered by the workgroup Steering Committee and appropriate
changes were made to both the procedures for establishing Eco-SSLs and to the guidance
document.
After developing the procedures and completing the peer review process, the workgroup focused
primarily on deriving Eco-SSL values for the list of contaminants. The results of the application
of the derivation procedures reported in this document are provided as separate contaminant-
specific documents. In cases where data were limited or not available and Eco-SSL values could
not be derived for specific contaminants and receptor groups, EPA may at some point in the
future revise the contaminant-specific documents or add contaminants as appropriate.
ES-3
-------
This Page Intentionally Left Blank
-------
ACKNOWLEDGMENTS
The development of this guidance was a team effort led by the U.S. Environmental Protection Agency Office of
Superfund Remediation and Technology Innovation (OSRTI). A Steering Committee coordinated the activities of
four task groups. A team of scientists from Syracuse Research Corporation (SRC) managed by Janet Burris
provided technical and administrative support in the development of this guidance. Listed below are the members of
the Steering Committee and each task group.
Steering Committee
Steve Ells, U.S. EPA OSRTI, Co-Chair
Ralph Stahl, DuPont, Co-chair
Randy Wentsel, U.S. EPA ORD, Co-chair
Bill Adams Rio Tinto
Doris Anders U.S. Air Force Center for Environmental Excellence (AFCEE)
John Bascietto U.S. Department of Energy (DOE)
David Charters U.S. EPA OSRTI
Ron Checkai U.S. Army Edgewood Chemical Biological Center (ECBC)
Stiven Foster U.S. EPA, Office of Research and Development (ORD)
Dale Hoff U.S. EPA, Region 8
Charlie Menzie Menzie-Cura & Associates
Chris Russom U.S. EPA, ORD
Brad Sample CH2MHill
Jason Speicher North Division of the Naval Facilities Engineering Command
Mike Swindell Exxon Mobile Biomedical Sciences, Inc.
Task Group on Soil Chemistry
Randy Wentsel, U.S. EPA, Co-Chair
Charlie Menzie, Menzie-Cura & Associates, Co-Chair
Bill Berti DuPont
Ron Checkai U.S. Army Edgewood Chemical Biological Center (ECBC)
Mary Goldade U.S. EPA, Region 8
Roman Lanno Oklahoma State University
Charles R. Lee U.S. Army Corps of Engineers, Waterways Experiment Station
Linda Lee Purdue University
Mike Ruby Exponent
John Samuelian OGDEN Environmental and Energy Services
Task Group on Eco-SSLs for Soil Invertebrates and Plants
Ron Checkai, U.S. Army ECBC, Co-Chair
Mike Swindell, Exxon Mobile Biomedical Sciences, Inc., Co-Chair
David Barclift North Division of the Naval Facilities Engineering Command
Ron Checkai U.S. Army ECBC
William Desmond Tetra Tech EM, Inc.
Steve Ells U.S. EPA, OSRTI
Charlene Falco Illinois EPA
Stiven Foster U.S. EPA, ORD.
Dave Gannon Zeneca Corp.
Andrew Green International Lead Zinc Research Org. (ILZRO)
Larry Kapustka ecological, planning & toxicology, Inc.
Roman Kuperman U.S. Army ECBC
Les Morrow Illinois EPA
Daniel Mazur U.S. EPA, Region 5
Chris Russom U.S. EPA, ORD
Jason Speicher North Division of the Naval Facilities Engineering Command
Gladys Stephenson ESG International, Ltd. / University of Guelph
-------
Task Group on Wildlife Toxicitv Reference Values
Dale Hoff, U.S. EPA, Co-Chair
Doris Anders, U.S. Air Force, Co-Chair
Nelson Beyer U.S. Geological Survey
David Charters U.S. EPA OSRTI
David Cozzie U.S. EPA Office of Solid Waste
Dave Cragin Elf-Atochem North America
Steve Dole Exponent
Anne Fairbrother U.S. EPA, ORD, Corvallis, Oregon
Bob Fares Environmental Standards, Inc.
Gary Friday Westinghouse Savannah River Co.
Kinzie Gordon Parsons Engineering Science, Inc.
Gerry Henningsen U.S. EPA, Region 8 (retired)
Mark Johnson U.S. Army Center for Health Promotion and Preventative Medicine
Paul Kuhlmeier Dames & Moore
Jackie Little TN & Assoc.
Patricia Newell Environmental Health Associates, Inc.
Drew Rak U.S. Army Corps of Engineers, Baltimore District
Linda Schmeising Exponent
Task Group on Exposure Models for Wildlife Species
John Bascietto, U.S. Department of Energy (DOE), Co-Chair
Brad Sample, CH2MH111, Co-Chair
Amber Brenzikofer Parsons Engineering Science
Bridgette Deshields Harding Lawson Associates
Will Gala Chevron
Tracy Hammon Colorado Department of Health and Environmental Protection
Rob Pastorok Exponent
Phillip Rury Arthur D. Little, Inc.
Randy Ryti Neptune and Co.
William Schew Environmental Standards, Inc.
Ralph Stahl DuPont
Jeff Yurk U.S. EPA, Region 6
-------
TABLE OF CONTENTS
EXECUTIVE SUMMARY
LIST OF FIGURES
LIST OF TABLES
LIST OF ATTACHMENTS
LIST OF ACRONYMS AND ABBREVIATIONS
1.0 INTRODUCTION 1-1
1.1 Scope of the Eco-SSLs 1-4
1.2 The General Process for Establishing Eco-SSLs 1-7
1.3 Peer Review Process 1-11
1.4 Quality, Objectivity, Utility, and Integrity of Information 1-12
1.5 Using Eco-SSLs to Screen Contaminated Soils 1-12
2.0 SOIL PROPERTIES 2-1
2.1 Introduction 2-1
2.2 Soil Properties Influencing Contaminant Bioavailability 2-1
2.3 Using Soil Properties to Guide Eco-SSL Derivation 2-8
3.0 DERIVATION OF PLANT AND SOIL INVERTEBRATE ECO-SSLs 3-1
3.1 Literature Search, Acquisition, and Screening 3-2
3.2 Identification of Potentially-Acceptable Literature 3-3
3.3 Extraction of Data and Scoring Studies 3-3
3.4 Derivation of Eco-SSLs 3-6
3.5 Soil Toxicity Test Methods 3-8
4.0 DERIVATION OF WILDLIFE ECO-SSLs 4-1
4.1 The Wildlife Risk Model for Eco-SSLs 4-1
4.2 Selection of Surrogate Wildlife Species 4-2
4.3 The Exposure Dose 4-4
4.4 Derivation of Toxicity Reference Values (TRVs) 4-11
4.5 Calculation of Wildlife Eco-SSLs 4-18
5.0 ECO-SSL DOCUMENTS 5-1
6.0 REFERENCES 6-1
Guidance for Developing Eco-SSLs i November 2003
-------
LIST OF FIGURES
Title Page
Figure 1.1 Eight Step Process Recommended in Ecological Risk Assessment Guidance for
Superfund (ERAGS) 1-2
Figure 1.2 Eco-SSL Contaminants 1-4
Figure 1.3 Comparison of Eco-SSL Derivation Procedures for Wildlife and Plant and Soil
Invertebrates 1-9
Figure 3.1 The Four-Step Process for Deriving Eco-SSLs for Plants and Soil Invertebrates3-l
Figure 3.2 Literature Exclusion Criteria 3-2
Figure 3.3 Eleven Study Acceptance Criteria 3-3
Figure 4.1 The Wildlife Risk Model for Eco-SSLs (Equation 4-1) 4-2
Figure 4.2 Summary of Method Used for Estimation of Contaminant Concentrations
in Biota Types (B;) 4-7
Figure 4.3 Comparison of Soil-to-Foilage and Soil-to-Seed Bioaccumulation for Cadmium,
Copper, Lead and Zinc 4-8
Figure 4.4 Comparison of Mean Concentrations in Multiple Species near a Smelter ... 4-10
Figure 4.5 Steps of the Wildlife TRV Derivation Process 4-11
Figure 4.6 Ten Attributes Scored as Part of the Wildlife Toxicological Data
Evaluation 4-12
Figure 4.7 Example of a Toxicological Plot for the TRV Derivation Process (Cobalt) .. 4-15
Figure 4.8 Procedure for Deriving the Wildlife Toxicity Reference Value (TRV) 4-16
Figure 4.9 Example of Wildlife TRV Derivation (Cobalt) 4-17
Guidance for Developing Eco-SSLs ii November 2003
-------
Table 2.2
Table 4.:
LIST OF TABLES
Title
Page
Table 1.1 Comparison of Eco-SSL Derivation Procedures 1-10
Table 2.1 General Contaminant Classification 2-3
Log Kow Values for Organic Contaminants 2-5
Table 2.3 Mean Reported Soil Metal Background Concentrations (ppm dry weight) by
State 2-9
Table 2.4a Qualitative Bioavailability of Metal Cations in Natural Soils to Plants 2-10
Table 2.4b Qualitative Bioavailability of Metal Cations in Natural Soils to Soil
Invertebrates 2-10
Table 2.5 Qualitative Bioavailability of Non-Ionizing Organic Contaminants in Natural
Soils 2-11
Table 2.6 Qualitative Bioavailability of Metal Anions in Natural Soils 2-11
Table 3.1 Ecologically Relevant Endpoints (EREs) for Soil Invertebrate Eco-SSLs .... 3-4
Table 3.2 Ecologically Relevant Endpoints (EREs) for PI ant Eco-SSLs 3-5
Table 3.3 Summary of Nine Study Evaluation Criteria for Plant and Soil Invertebrate
Eco-SSLs 3-7
Table 3.4 Standard Methods Appropriate for use in Generating Data for the Derivation of
Plant or Soil Invertebrate Eco-SSLs 3-9
Table 4.1 Parameterization of the Eco-SSL Wildlife Exposure Model 4-5
Table 4.2 Cases where the Median of the BAF Distribution for Metals is Greater
or Less than One 4-10
Example of Extracted and Scored Toxicity Data for Wildlife TRY (Cobalt) .4-14
Table 4.4 Summary of Conservatism Associated with the Wildlife Eco-SSL Risk Model
Parameters 4-20
Guidance for Developing Eco-SSLs
in
November 2003
-------
Attachment 1-1
Attachment 1-2
Attachment 1-3
Attachment 1-4
Attachment 3-1
Attachment 3-2
Attachment 4-1
Attachment 4-2
Attachment 4-3
Attachment 4-4
Attachment 4-5
LIST OF ATTACHMENTS
Review of Existing Soil Screening Guidelines
Discussion Concerning Soil Microbial Processes
Review of Dermal and Inhalation Exposure Pathway for Wildlife
Review of Background Concentrations for Metals
Eco-SSL Standard Operating Procedure #1: Plant and Soil
Invertebrate Literature Search and Acquisition
Eco-SSL Standard Operating Procedure #2: Plant and Soil
Invertebrate Literature Evaluation and Data Extraction, Eco-SSL
Derivation, Quality Assurance Review, and Technical Write-up
Exposure Factors and Bioaccumulation Models for Derivation of
Wildlife Eco-SSLs
Eco-SSL Standard Operating Procedure #3: Wildlife TRY:
Literature Search and Retrieval
Eco-SSL Standard Operating Procedure # 4: Wildlife TRY
Literature Review, Data Extraction and Coding
Eco-SSL Standard Operating Procedure # 5: Wildlife TRY Data
Evaluation
Eco-SSL Standard Operating Procedure #6: Wildlife TRY
Derivation
Due to the length of many of these Attachments, they are only available on the EPA Superfund
web site atwww.epa.gov/oswer/riskassessment/ecorisk/ecossl.htm.
Guidance for Developing Eco-SSLs
IV
November 2003
-------
LIST OF ACRONYMS AND ABBREVIATIONS
AEC Anion exchange capacity
ASTM American Society of Testing and Materials
AUF Area use factor
Absorbed fraction of contaminant (j) from biota type (i)
Absorbed fraction of contaminant (j) from soil (s)
B; Contaminant concentration in biota type (i)
BOy Intercept from log-linear bioaccumulation model for contaminant (j) for biota type (i)
Ely Slope from log-linear bioaccumulation model for contaminant (j) for biota type (i)
BAF Bioaccumulation factor
BEH Behavior
BIO Biochemical
BTAG Biological Technical Assistance Group
BW Body weight
Csoil Concentration of contaminant in soil
CCME Canadian Council of Ministers of the Environment
CEC Cation exchange capacity
CI Confidence interval
COC Contaminant of concern
COPC Contaminant of potential concern
DDT 1,1,1 -Trichloro-2,2-bis(p-chlorophenyl)ethane
DQO Data quality objective
dw Dry weight
ECXX Effect concentration for the xx% of the test population
Eco-SSL Ecological soil screening level
EPA U.S. Environmental Protection Agency
ERA Ecological risk assessment
ERAGS Ecological Risk Assessment Guidance for Superfund
ERE Ecologically relevant endpoint
foc fraction organic carbon content
FBRC Federal Biology Research Cooperative
FIR Food ingestion rate
g Grams
GRO Growth
HQ Hazard quotient
HQj Hazard quotient for contaminant (j)
HSDB Hazardous Substance Data Bank
ISO International Standards Organization
kg Kilogram
Koc Organic carbon-normalized partition coefficient
Kow Octanol water partition coefficient
LC50 Concentration lethal to 50 percent of test population
LOAEC Lowest-observed adverse effect concentration
LOAEL Lowest-observed adverse effect level
Guidance for Developing Eco-SSLs
-------
MATC
mg
mmol
MOR
MPA
MFC
NC
NOAEL
NOAEC
NOC
OC
OECD
OM
ORNL
OSRTI
P;
Ps
pKa
PAH
PCB
PCP
PHY
PTH
QA
QSAR
RDX
REP
RIVM
ROD
SAB
SMDP
Soilj
SOP
SQG
T
u
T
Aver
TNT
TRY
TRVj
UCL
U.S.
LIST OF ACRONYMS AND ABBREVIATIONS
(Continued)
Maximum acceptable toxicant concentration
Milligram
Millimole
Mortality
Maximum permissible addition
Maximum permissible concentration
Negligible concentration
No-observed adverse effect level
No-observed adverse effect concentration
Non ionic organic compounds
Organic carbon
Organization for Economic Cooperation and Development
Organic matter
Oak Ridge National Laboratory
Office of Superfund Remediation and Technology Innovation
Proportion of biota type (i) in diet
Soil ingestion as proportion of diet
Acid dissociation constant
Polycyclic aromatic hydrocarbon
Polychlorinated biphenyl
Pentachlorophenol
Physiology
Pathology
Quality assurance
Quantitative Structure Activity Relationship
Hexahy dro-1,3,5 -trinitro-1,3,5 -triazine
Reproduction
Dutch National Institute of Public Health and the Environment
Record of decision
Science Advisory Board
Scientific Management Decision Point
the Eco-SSL for contaminant j for wildlife
Standard operating procedure
Soil quality guideline
Soil-to-biota bioaccumulation factor for contaminant (j) for biota type (i)
diet-to-biota B AF
Trinitrotoluene
Toxicity reference value
Toxicity reference value for contaminant (j)
Upper confidence limit
United States
micrometer
Guidance for Developing Eco-SSLs
VI
-------
LIST OF ACRONYMS AND ABBREVIATIONS
(Continued)
VOC Volatile organic compound
ww Wet weight
Guidance for Developing Eco-SSLs vii
-------
This Page Intentionally Left Blank
-------
1.0 INTRODUCTION
This guidance describes the process used to derive a set of risk-based ecological soil screening
levels (Eco-SSLs) for many of the soil contaminants that are frequently of ecological concern for
plants and animals at hazardous waste sites and further provides guidance on using Eco-SSLs. The
specific Eco-SSL values and the data upon which they were derived are described in separate
contaminant specific Eco-SSL documents. The Eco-SSL derivation process represents the group
effort of a multi-stakeholder workgroup consisting of federal, state, consulting, industry, and
academic participants led by the U.S. Environmental Protection Agency (EPA) Office of Superfund
Remediation and Technology Innovation (OSRTI). The workgroup developed the following
mission statement at the initiation of the Eco-SSL project:
Develop a set of generic, scientifically sound, ecologically based, soil screening levels that are
protective of the terrestrial environment for up to 24 contaminants of concern; and methodologies
and models that use site-specific exposure data to modify these screening levels. The screening
levels and methodologies should be sufficiently specific and transparent to allow for consistent
implementation by EPA and other Federal Agencies, States, and private parties at all Superfund
sites.
The Eco-SSLs are screening values that can be used routinely to identify those contaminants of
potential concern (COPCs) in soils requiring further evaluation in a baseline ecological risk
assessment (ERA). Although these screening levels were developed specifically to be used during
Step 2 of the Superfund ecological risk assessment process, EPA envisions that any federal, state,
or private environment assessment or cleanup program could use these values to screen soil
contaminants and sites in order to determine if additional ecological site study was warranted. The
Eco-SSLs were not designed to be used as cleanup levels and EPA emphasizes that it would be
inappropriate to adopt or modify these Eco-SSLs as cleanup standards.
This document provides guidance and is designed to communicate national policy on identifying
contaminants in soil that may present an unacceptable ecological risk to terrestrial receptors. The
document does not, however, substitute for EPA's statutes or regulations, nor is it a regulation itself.
Thus, it does not impose legally-binding requirements on EPA, states, or the regulated community,
and may not apply to a particular situation based upon the circumstances of the site. EPA and state
personnel may use and accept other technically sound approaches, either on their own initiative, or
at the suggestion of potentially responsible parties, or other interested parties. Therefore, interested
parties are free to raise questions and objections about the substance of this guidance and the
appropriateness of the application of this document to a particular situation. EPA welcomes public
comments on this guidance at any time and may consider such comments in future revisions of this
guidance.
What are Eco-SSLs?
Eco-SSLs are concentrations of contaminants in soil that are protective of ecological receptors that
commonly come into contact with soil or ingest biota that live in or on soil. Eco-SSLs are derived
separately for four groups of ecological receptors: plants, soil invertebrates, birds, and mammals.
Guidance for Developing Eco-SSLs 1-1 November 2003
-------
W o
i
O 5
o
JLJ
"o
O
STEP 1: SCREENING LEVEL
Site Visit
Problem Formulation
Toxicity Assessment
STEP 2: SCREENING LEVEL
Exposure Estimate
Risk Characterization
STEP 3: PROBLEM FORMULATION
Toxicity Evaluation
Assessment
Endpoints
Conceptual Model
Exposure Pathways
I
Questions/ Hypotheses
STEP 4: STUDY DESIGN AND DQO PROCESS
Lines of Evidence
Measurement Endpoint
Workplan and Sampling and Analyses Plan
STEP 5: VERIFICATION OF FIELD SAMPLING DESIGN
STEP 6: SITE INVESTIGATION AND DATA ANALYSES
Risk Assessor
and Risk
Manager
Agreement
STEP 7: RISK CHARACTERIZATION
STEP 8:
RISK MANAGEMENT
fe Qivrnp
SMDP = Sample Management Decision Point
DQO = Data Quality Objective
Figure 1.1 Eight Step Process Recommended in Ecological Risk Assessment
Guidance for Superfund (ERAGs) (U.S. EPA, 1997)
Guidance for Developing Eco-SSLs
1-2
November 2003
-------
Protection of Rare, Endangered and
Threatened Species
Eco-SSLs should be protective of rare,
endangered, and threatened species. However,
the final decision should be made on a site-
specific basis in consultation with the United
States Fish and Wildlife Service and other
natural resource trustees.
As such, these values are presumed to provide adequate
protection of terrestrial ecosystems. Eco-SSLs for
wildlife are derived to be protective of the
representative of the conservative end of the
distribution in order to make estimates for local
populations. The Eco-SSLs are conservative and are
intended to be applied at the screening stage of the
assessment. These screening levels should be used in
the ERA process to identify the COPCs that require
further evaluation in the site-specific baseline risk
assessment. This Eco-SSL guidance was written with
the assumption that the reader is familiar with Superfund's guidance on performing ERAs
(Ecological Risk Assessment Guidance for Superfund (ERAGS)), U.S. EPA, 1997, Figure 1.1),
with Superfund's ecological risk assessment and risk management principles (U.S. EPA, 1999) and
with EPA's risk assessment guidelines (U.S. EPA, 1998).
The Eco-SSLs should be used during Step 2 of the Superfund ERA process, the screening-level risk
calculation. This step normally is completed at a time when limited soil concentration data are
available, and other site-specific data (e.g., contaminant bioavailability information, area use
factors) are not available. It is expected that the Eco-SSLs will be used to screen the site soil data
to identify those contaminants that are not of potential ecological concern and do not need to be
considered in the subsequent baseline ERA. The Eco-SSLs are intentionally conservative in order
to provide confidence that contaminants which could present an unacceptable risk are not screened
out early in the ERA process. EPA believes that the Eco-SSLs generally provide an appropriate
balance of protectiveness and reasonableness.
Why are Eco-SSLs Needed?
EPA derived the Eco-SSLs with the intent to
conserve resources by limiting the need for
EPA, state, contractor, and other federal risk
assessors to perform repetitious toxicity data
literature searches and toxicity data evaluations
for the same contaminants at every site. These
Eco-SSLs are also intended to increase
consistency among screening risk analyses,
decrease the possibility that potential risks from
soil contamination to ecological receptors will
be overlooked, and allow risk assessors to focus
their resources on identifying key site studies
needed for critical decision-making.
In the process of deriving the Eco-SSLs, the
workgroup examined currently available soil
screening guidelines (see text box) for their
potential use within the Superfund process.
Some Other Available Soil Screening Guidelines
Canadian Council of Ministers of the Environment
(CCME) Canadian Soil Quality Guidelines (SQGs).
The CCME guidelines are numerical limits for
contaminants intended to maintain, improve, or protect
environmental quality and human health. They are
intended for use in the assessment and remediation of
contaminants at sites in Canada (CCME, 1997).
The Dutch National Institute of Public Health and the
Environment (RIVM). Maximum permissible
concentrations (MFCs), maximum permissible additions
(MPAs), and negligible concentrations (NCs) were
developed in a series of reports for soils, sediments, and
water for metals and pesticides (RIVM, 1997a and 1997b).
Oak Ridge National Laboratory (ORNL). A series of
reports have been issued from ORNL that provide
screening levels for plants (Efroymson et al, 1997a), soil
invertebrates and microbial processes (Efroymson et al.,
1997b), wildlife (Sample et al., 1996), and sediments
(Jones etal, 1997).
Guidance for Developing Eco-SSLs
1-3
November 2003
-------
Because these existing guidelines were developed in response to country-specific legislation and/or
included policies not totally consistent with current EPA policies, EPA believes use of these
guidelines may not be appropriate. A summary and evaluation of the available guidelines was
completed and provided as Attachment 1-1.
How Are the Eco-SSLs Derived?
Eco-SSLs were derived using standardized procedures for literature review, toxicity data selection,
and data evaluation. Where acceptable data were judged to be adequate, four Eco-SSLs were
derived for each contaminant: one each for plants, soil invertebrates, birds, and mammals.
Plant and soil invertebrate Eco-SSL values were derived directly from an evaluation of available
plant and soil invertebrate toxicity test data (measured toxicity related to soil contaminant
concentrations), as described in Chapter 3. The process for deriving mammalian and avian
Eco-SSLs is described in Chapter 4. The wildlife Eco-SSLs were the result of back-calculations
from a Hazard Quotient (HQ) of 1.0. The HQ is equal to the estimated exposure dose divided by a
toxicity reference value (TRV). An HQ of 1.0 is the condition where the exposure and the dose
associated with no adverse effects are equal, indicating adverse effects at or below this soil
concentration are unlikely. A generic food-chain model was used to estimate the relationship
between the concentration of the contaminant in soil and the dose for the receptor (milligrams per
kilogram body weight per day (mg/kg bw/d)). The TRV represents a receptor-class specific
estimate of a no-observed adverse effect level (NOAEL) (dose) for the respective contaminant.
Figure 1.2 Eco-SSL Contaminants
Dieldrin
Hexahydro-l,3,5-trinitro-l,3,5-triazine(RDX)
Trinitrotoluene (TNT)
U,l-Trichloro-2,2-bis (p-chlorophenyl)ethane (DDT)
and metabolites DDE and ODD.
Pentachlorophenol (PCP)
Poly cyclic Aromatic Hydrocarbons (PAHs)
Poly chlorinated biphenyls (PCBs)
1.1 Scope of the Eco-SSLs
Contaminants Considered
EPA prepared a list of twenty-four (24)
contaminants to be addressed initially by the
Eco-SSL guidance. This list was based on a
review of the contaminants of concern reported
to be the subject of soil remediation in recent
Record of Decisions (ROD) at Superfund
National Priority List sites. The Eco-SSL
contaminant list also included contaminants
nominated by the EPA regional Biological
Technical Assistance Group (BTAG)
Coordinators. The list of 24 Eco-SSL
contaminants contained seven organics and 17
metals (see Figure 1.2).
The omission of other contaminants, such as
phthalates, cyanides, dioxin, and mercury, does
not imply that these contaminants can be
excluded from the ERA screening process for soil contamination, only that these 24 contaminants
have historically been of greatest ecological concern in soil. The process and procedures
Metals
Aluminum
Antimony
Arsenic
Barium
Beryllium
Cadmium
Chromium
Cobalt
Copper
Iron
Lead
Manganese
Nickel
Selenium
Silver
Vanadium
Zinc
Guidance for Developing Eco-SSLs
1-4
November 2003
-------
established here for developing the Eco-SSLs are intended to be sufficiently transparent to allow
others to derive values for additional contaminants, as needed. Poly chlorinated biphenyls (PCBs)
were included by the workgroup in the original 24 identified Eco-SSL contaminants. However, it
became apparent early in the process that development of a screening value would not be
appropriate. Because of the known persistence and toxicity of PCBs, and the conservative nature of
the Eco-SSLs, any soil screening level derived for PCBs would often be lower than the detection
limits used in a screen. EPA recommends that if PCBs are detected in soil above background
levels, then they should be considered site-related and therefore should be included as a COPC in a
baseline ERA.
Ecological Receptors of Concern
The Eco-SSLs generally apply to sites where terrestrial receptors may be exposed directly or
indirectly to contaminated soil. Seven groups of ecological receptors were initially considered in
the development of the Eco-SSLs. These initial groups included mammals, birds, reptiles,
amphibians, soil invertebrates, plants, and soil microbes and their processes. After further
investigation, the toxicity data for amphibians and reptiles were deemed insufficient to derive
Eco-SSLs. EPA recognizes that the Eco-SSL values do not address possible risks for reptiles and
amphibians and may not be protective of these receptor groups. The user should consider including
these receptors in the site-conceptual model for the site-specific risk assessment.
Eco-SSLs protective of microbes and soil microbial processes were also not derived. Like
amphibians and reptiles, EPA recognized their importance within terrestrial systems, but believes
that data are insufficient and the interpretation of test results too uncertain for establishing risk-
based thresholds for risk screening purposes. While Eco-SSLs for microbes and soil microbial
processes were not established at this time, they may be considered in the future as the science
develops and appropriate studies are completed. A summary of the task group discussion
concerning establishing Eco-SSLs for soil microbes and their processes was documented as
Attachment 1-2.
Eco-SSLs were derived for four general groups of ecological receptors: mammals, birds, plants,
and soil invertebrates. By deriving conservative soil screening values considered protective of
these groups, it is assumed that these receptor groups are protected from possible adverse effects
associated with soil contamination. This assumption is consistent with the use of "generic
assessment endpoints" as discussed in Section 1.2.5 of ERAGS.
Exposure Pathways for Ecological Receptors
A complete exposure pathway is defined in ERAGS as "one in which the contaminant can be traced
or expected to travel from the source to a receptor that can be affected by the contaminant."
Exposure pathways can be classified as incomplete, complete, or potentially complete. An
exposure pathway is not considered complete if natural habitat for ecological receptors is not
present and is not expected to be present in the future.
Guidance for Developing Eco-SSLs 1-5 November 2003
-------
The Eco-SSLs for plants considered direct
contact of contaminants in soils. The Eco-SSLs
for soil invertebrates considered ingestion of soil
and direct contact exposures with a preference
for conditions of high bioavailability (refer to
Chapters 2 and 3).
The Eco-SSLs for birds and mammals considered
two potentially complete exposure pathways: 1)
incidental ingestion of soils during feeding,
grooming and preening; and 2) ingestion of food
contaminated as a result of the uptake of soil
contaminants. Two potentially complete
exposure pathways (dermal contact and
inhalation) were not considered in the derivation
of wildlife Eco-SSLs. The rationale for this
decision included the following:
Burrowing animals could be exposed to
relatively high concentrations of volatile
organic compounds (VOCs) in their burrows via inhalation. However, with the exception of
some of the polynuclear aromatic hydrocarbons (PAHs), none of the Eco-SSL contaminants
were VOCs. At sites with high VOC and/or certain PAH concentrations in soils with
burrowing mammals present, the inhalation exposure pathway should be considered in the
baseline ERA. In this case, the contaminants would not be excluded in the screening step.
What is a Complete Ecological Exposure Pathway
for Contaminants in Soil?
For an exposure pathway to be complete, a contaminant
must be able to travel from the source to ecological
receptors and be taken up by the receptors via one or more
exposure routes (U.S. EPA, 1997).
Exposure pathways may not be complete for ecological
receptors if:
S Soil contamination exists only below the root
zone, and deep burrowing mammalian species
are not identified as potential receptors in the
site conceptual model.
S The site is within urban and/or industrialized
areas where natural habitat and receptors are
absent.*
*Urban settings may in some cases be used by protected
species. The appropriate trustees should be consulted.
Exposure Pathways Considered
in Eco-SSLs
Birds and Mammals
Ingestion of soils during grooming,
feeding, and preening
Ingestion of food contaminated as a
result of uptake of soil contaminant
Plants
Direct contact
Soil Invertebrates
Direct contact
Soil ingestion
Soil particles containing non-VOC contaminants
(by either adsorption or absorption) could also be
inhaled by wildlife. Respirable particles (greater
than five |im) are, however, most likely ingested
as a result of mucocilliary clearance rather than
being inhaled (Witschi and Last, 1996). At equal
exposure concentrations, inhalation of
contaminants associated with dust particles is
expected to contribute less than 0.1 % of total risk
compared to oral exposures (Attachment 1-3).
Birds and mammals may also be exposed to
contaminants in soils via dermal contact. Studies
investigating dermal exposures to birds from the
application of pesticides by spray to tree branches
have shown this exposure route to be significant
relative to oral exposures for some substances; e.g.
organophosphate pesticides, (Abou-Donia and
Graham 1978, Driver et al. 1991, and Henderson
etal. 1994). However, current information is
Guidance for Developing Eco-SSLs
1-6
November 2003
-------
insufficient to evaluate dermal exposure for the 24 selected Eco-SSL contaminants in
various soil matrices, or to predict possible rates of absorption for many species. For most
contaminants, the dermal exposure is expected to contribute less than one percent to 11 % of
the total risk (Attachment 1-3) compared to oral exposures.
This approach is consistent with Section 9.2.4 of ERAGS (U.S. EPA, 1997), which states that the
ingestion route is most important for terrestrial animals and that "although other exposure routes
can be important, more assumptions are needed to estimate exposure levels for these routes, and the
results are less certain". Exclusion of dermal and inhalation exposure routes for the Eco-SSLs does
not preclude their inclusion in the site-specific baseline ERA. If it is expected that receptors may
be more exposed to contaminant(s) via dermal and/or inhalation exposures relative to oral
exposures due to site-specific conditions, these exposure routes should be evaluated as part of the
baseline ERA.
Soils for Which Eco-SSLs are Applicable
It is recommended that Eco-SSLs be considered at sites where key soil parameters fall within a
certain range of chemical and physical parameters. The Eco-SSLs for plants and soil invertebrates
are usually appropriate for application to soils where: the pH is greater than or equal to 4.0 and less
than or equal to 8.5 and the organic matter content is less than or equal to 10 %.
The Eco-SSLs are intended for use in upland soils. However, they may also be useful for screening
wetland soils. The wildlife Eco-SSLs are derived for several general receptor groups that are likely
to be representative of wildlife found in wetlands. A major caveat, however, is the omission of the
amphibians and reptiles from derivation of the wildlife Eco-SSLs. These groups could be
especially important in wetlands. The Eco-SSLs for plants and soil invertebrates are expected to be
broadly applicable (i.e., conservative enough for most soils) as preference was given to studies with
high bioavailability of the contaminants in soils. For this reason, the Eco-SSLs for plants and soil
invertebrates may be useful for screening for contaminants in wetland soils. In general, wetland
soils are expected to exhibit a lower bioavailability (compared to those used to derive Eco-SSLs) as
a result of the high organic content.
Based on these stated parameters, it is expected that there are certain soils and situations to which
Eco-SSLs may not be appropriate. These situations include (but may not be limited to):
Wetland soils that are regularly flooded (i.e., sediments).
Sewage sludge amended soils where the organic matter (OM) content is > 10 %.
Waste types where the pH is < 4.0.
1.2 The General Process for Establishing Eco-SSLs
Four separate task groups were formed to develop the procedures for establishing Eco-SSLs and to
accomplish the work necessary to derive the Eco-SSL values for the 24 listed contaminants. The
task groups were composed of a cross-section of the stakeholders represented in the larger work
Guidance for Developing Eco-SSLs 1-7 November 2003
-------
Eco-SSL Task Groups
0 ., . ^ TT.- * i -A A f * *i, information related to the factors that
Soil Chemistry. This task group provided information on the
factors that influence bioavailability of contaminants from soils.
The task group developed the soils matrix used to select studies
for establishing plant and soil invertebrate Eco-SSLs, the
background soils database, and Chapter 2 of this guidance
document.
Plant and Soil Invertebrate Eco-SSLs. This task group
developed the procedures for establishing plant and soil
invertebrate Eco-SSLs (Chapter 3). The task group derived the
plant and soil invertebrate Eco-SSL values according to these
procedures.
Wildlife Eco-SSLs. The wildlife Eco-SSLs were developed by
the combined efforts of two separate task groups:
group developed the procedures for establishing the
wildlife toxicity reference values as presented in
Chapter 4.
Although the task groups worked
Wildlife Exposure Model. This task group developed
the equations, assumptions, and factors for establishing
the exposure model for wildlife. The model is presented
in Chapter 4. The model estimates the soil contaminant
concentration associated with no adverse effects to
representative wildlife species.
group with efforts directed by a steering
committee. One task group focused on
influence the bioavailiblity of
contaminants from soils. A second task
group focused on deriving the Eco-SSLs
protective of plants and soil invertebrates.
The third and fourth task groups focused
on deriving Eco-SSLs protective of
wildlife. One group developed the
procedures for deriving wildlife TRVs
(mammalian and avian receptors) while
the other group developed the models
necessary for estimating contaminant
transfer from soils to the terrestrial food
Wildlife Toxicity Reference Values (TRVs). This task , . j r- j f
chain and tor deriving exposures tor
wildlife.
independently, the approach taken for
deriving the Eco-SSLs for plants and soil
invertebrates was similar (see Figure 1.3)
to the approach taken for deriving the
wildlife Eco-SSLs (specifically the
TRVs). The general approach included
four steps (1) conduct literature searches,
(2) screen identified literature with exclusion and acceptability criteria, (3) extract, evaluate, and
score test results for applicability in deriving an Eco-SSL, and (4) derive the value.
Step 1: Literature Search
For all receptors, potentially relevant publications were identified through literature searches of
computerized abstracting databases combined with the examination of citations associated with
published literature review articles. The literature searches for avian and mammalian species
included all publication years, while those for plants and soil invertebrates included only
publications after 1987 (Table 1.1). The task group theorized that most of the pre-1988
publications would be identified through the review bibliographies, and this was confirmed in
subsequent analysis of the literature search results with approximately 40 % of potentially
applicable papers identified through non-computerized search techniques. In cases where
contaminant/receptor pairings had less than 20 potentially applicable articles, searches were
completed for all publication years.
Step 2: Determine Acceptability of Study for Use in Deriving Eco-SSL
Both task groups used similar exclusion criteria to assess the potential applicability of publications
identified through the literature searches. Publications were excluded for similar reasons as
Guidance for Developing Eco-SSLs 1-8 November 2003
-------
Wildlife TRV
Plant and
Soil Invertebrate Eco-SSL
Stepl
Conduct Electronic
Literature Search
Complete according to
established process
(Attachment 4-3)
Complete according to
established process
(Attachment 3-1)
Step 2
Determine
Acceptability of
Study for use in
Deriving an Eco-
SSL
Apply Literature Rejection
Criteria (Attachment 4-3)
Apply Literature Rejection
Criteria (Attachment 3-1)
Apply 11 Literature
Acceptance Criteria
(Attachment 3-1)
Step3
Extract, Evaluate
and Score Data
from Accepted
Studies
Data extracted and scored
according to established
procedures (Attachments 4-
3 and 4-4)
Data extracted and scored
according to established
procedures (Attachment 3-2)
Studies that score 66 or
higher (66%) can be used to
derive TRV used to calculate
Eco-SSL
Studies that score 11 or
higher (61%) can be used to
derive Eco-SSL
Step 4
Derive Value
Derive according to
established procedures
(Attachment 4-5)
Geometric mean of NOAEL
values for growth and
reproduction. In some cases,
the highest bounded NOAEL
below the lowest bounded
LOAEL for growth,
reproduction or survival
Derive according to
established procedures
(Attachment 3-2)
Geometric mean of EC
MATC or EC10 values
20'
Figure 1.3
NOAEL = No-Observed Adverse Effect Level
LOAEL = Lowest-Observed Adverse Effect Level
EC20 = Effect Concentration 20%
EC10 = Effect Concentration 10%
MATC = Maximum Acceptable Toxicant Concentration
Comparison of Eco-SSL Derivation Procedures for Wildlife and Plant and Soil
Invertebrates
Guidance for Developing Eco-SSLs
1-9
November 2003
-------
Table 1.1. Comparison of Eco-SSL Derivation Procedures
Step in Eco-
SSL Process
Common to Both Approaches
Unique to Wildlife Approach
Unique to Plant and Soil Invertebrate Approach
Step 1:
Conduct
electronic
literature
search
Searches were conducted with appropriate abstracting
databases finding the intersection of contaminant,
species and toxicological terms. The final electronic
search result was reviewed to determine potentially
applicable studies. Bibliographies of review articles
were skimmed to identify additional publications.
All publication years were searched.
Searches were limited to 1988 to present. For limited data sets,
searches included literature published prior to 1988. Holdings of
U. S. EPA MED-Duluth library were searched.
Step 2:
Determine
acceptability of
study for use
in deriving an
Eco-SSL
Must be the primary source of the data
Exposure to single contaminant
Control must be included
Duration of exposure must be reported
Effects must be reported for relevant endpoints
Dose or concentration must be reported
Laboratory or field studies accepted
Beneficial effects not considered
Species must be reported
Percent metal or purity used to calculate nominal
concentrations
Oral route of exposure
At least two exposures: 1 control and 1 contaminant
Endpoints are behavioral, biochemical, growth,
mortality, pathology, population, physiology, and
reproduction
Chronic studies only ( > 3 days exposure duration)
Application rates not used
Natural or artificial soils
At least three exposures: 1 control and 2 contaminant
Test media must have: % organic matter (OM) content < 10
%; pH > 4 and < 8.5
Endpoints are growth , physiology (plants only), population,
and reproduction
Priority given to chronic studies, but acute studies with
sublethal effects or plant emergence as an endpoint were used
Step 3:
Extract,
evaluate and
score data from
accepted
studies
Papers were reviewed and the results were extracted
according to established guidelines.
Separate results were used if any of the following varied:
test organism (species or strains), contaminant form, test
location, control type, doses, application frequency, or
route of exposure.
Separate results were used if any of the following parameters
varied: test species (not strain), contaminant (not form), soil
(natural vs. artificial), pH, or % OM content.
Each result was scored according to established
guidelines.
All endpoints were scored in order of preference based on:
source of data; chemical form, measurement or no
measurement in substrate; ability to calculate a dose;
bounded vs unbounded NOAEL/LOAEL, route of
exposure, endpoint type, duration of exposure, statistical
power, and adherence to test guidelines.
Studies that scoreed > 65 % were used to derive a TRV.
Within a study with multiple endpoints only one was used in the
following order of preference: reproduction > population > growth
> physiology. Score was based on soil bioavailability score;
experimental design (i.e., adherence to ideal test guidelines);
reporting of chemical concentration; use of appropriate controls;
duration of exposure; NOAEC and LOAEC or EC10,EC20 reported
(or can be estimated); statistical analysis; and origin of test
organism.
Studies that scored > 1 1 out of 18 (
Eco-SSL.
61 %) were used to derive an
Step 4:
Calculation of
value
Results that scored above a cutoff were used in and
established approach to determine value.
The TRV was equal to the geometric mean of NOAEL
values for growth and reproduction or the highest bounded
NOAEL below the lowest bounded LOAEL for growth,
reproduction or survival. The process considered NOAEL
and LOAEL endpoints and included unbounded NOAEL
values, but not unbounded LOAEL values.
The Eco-SSL was equal to geometric mean of EC20, MATC or EC10
values. Only EC20, MATC and EC10 values were considered.
NOAEC and LOAEC values were used to calculate the MATC.
Unbounded NOAECs and LOAECs were not used.
When a study reported multiple endpoints, the selection followed
the following order of preference: EC20 > MATC > EC10
Guidance for Developing Eco-SSLs
1 - 10
November 2003
-------
specified later in Chapters 3 and 4 and summarized in Table 1.1. Acceptable plant and soil
invertebrate studies were further restricted to those that followed state-of-the-art soil testing
requirements (e.g., 4.0 < pH < 8.5, organic matter content < 10 %, use of natural or artificial soil).
Both task groups agreed that three or more treatment levels (including the control) were preferred,
but a substantial portion of the avian and mammalian studies would have been excluded if this
requirement was applied. Therefore only two treatment levels were required for mammalian and
avian receptors and five were required for plants and soil invertebrates. Only chronic toxicity
studies (greater than a three-day exposure) were accepted for mammalian and avian studies.
Although acute studies were not excluded for plants and soil invertebrates, the exposure duration
was considered later in the process for selecting the most appropriate test results for deriving the
Eco-SSL.
Step 3: Extract, Evaluate and Score Data
The extraction of toxicity data from the acceptable literature, evaluation of test methods and results,
and scoring of each test result also followed similar processes. When the methods diverged, it was
due mainly to inherent differences in study designs (e.g., direct exposures to soils versus exposures
in the diet or drinking water). Both task groups determined a cutoff for acceptable results for use in
deriving the wildlife TRY or plant or soil invertebrate Eco-SSL by plotting data for various data
sets and identifying logical breaks. For mammalian and avian species this break occurred at an
evaluation score of 66 out of a possible 100 (i.e., studies with a score greater than or equal to 66 %
were used to derive the TRY), while the break for plants and soil invertebrates occurred with a
score of 11 out of 18.
Step 4: Select Value
For both task groups, only the results from studies scoring above the respective cutoffs were used to
determine the value. For mammals and birds the TRY was equal to the geometric mean of NOAEL
values for growth and reproduction, or the highest bounded NOAEL (NOAEL with paired LOAEL)
below the lowest bounded lowest-observed adverse effect level (LOAEL) for growth, reproduction
or survival (whichever was lower). For contaminants with no growth, reproduction, or survival
data, the TRY was derived from biochemical, behavioral, pathology and physiology results. The
methodology considered both NOAEL and LOAEL endpoints including unbounded NOAEL values
but not unbounded LOAEL values.
For plants and soil invertebrates, the Eco-SSL was equal to the geometric mean of values for the
maximum acceptable toxicant concentration (MATC), the effective concentration that affects 20 %
of the test population (EC20), or the effective concentration that affects 10 % of the test population
(EC10). When a study reported multiple endpoints, the selection of endpoints for use in deriving the
Eco-SSL followed an order of preference of the EC20, then the MATC, then the EC10. The MATC
was either reported in the study or was calculated from the geometric mean of the no-observed
adverse effect concentration (NOAEC) and the lowest-observed adverse effect concentration
(LOAEC).
Guidance for Developing Eco-SSLs 1-11 November 2003
-------
1.3 Peer Review Process
The Eco-SSL procedures received both internal and external peer review. Because of the diversity
of the workgroup scientists, the process developed to derive the Eco-SSLs underwent constant peer
review. There were also two external peer reviews performed during the development of the
Eco-SSLs. The first was a consultation requested by EPA's Office of Solid Waste and Emergency
Response of EPA's Science Advisory Board (SAB). This consultation was held April 6, 1999. At
this meeting the SAB provided verbal comments to presenters at the SAB review, which were
passed on to the workgroup and incorporated into the guidance as appropriate. A formal external
peer review of the draft guidance document was also performed. The peer review workshop was
held on July 26 and 27, 2000 and was open to the public. The results of this peer review are
provided in separate documentation on the Eco-SSL website. Each of the comments received was
carefully considered by the Steering Committee and the individual work groups and appropriate
changes were made to both the procedures for establishing Eco-SSLs and the associated guidance
document.
1.4 Quality, Objectivity, Utility, and Integrity of Information
The EPA is committed to a policy of ensuring that the information it provides to the public, and
uses to make its decisions, maintain a basic standard of quality, which includes objectivity, utility,
and integrity. The Eco-SSLs are primarily derived from information the EPA obtained from
external sources that may not have used the same standards, guidelines, and controls that EPA
imposes on itself, and those who gather data on behalf of the agency. The agency has recently
proposed five assessment factors for evaluating the quality of information it obtains from external
sources (67 FR 57225, September 9, 2002):
Soundness The extent to which the procedures, measures, methods, or models
employed to generate the information are reasonable for, and consistent with, the
intended application and are scientifically/technically appropriate.
Applicability and Utility The extent to which the information is applicable and
appropriate for the Agency's intended use.
Clarity and Completeness The degree of clarity and completeness with which the
data, assumptions, methods, quality controls, and analyses employed to generate the
information are documented.
Uncertainty and Variability The extent to which the variability and uncertainty in
the information or in the procedures, measures, methods, or models are evaluated or
characterized.
Evaluation and Review The extent of independent application, replication,
evaluation, validation, and peer review of the information or of the procedures,
measures, methods, or models.
Guidance for Developing Eco-SSLs 1-12 November 2003
-------
Because these five assessment factors were developed after this guidance document was prepared,
they are not addressed by name. However, all five of these principles were embedded in the
standardized procedures for literature review, toxicity data selection, and data evaluation that were
used to derive the Eco-SSLs.
1.5 Using Eco-SSLs to Screen Contaminated Soils
The Eco-SSLs are intended for use in identifying soil contaminants ( i.e., COPCs) and/or areas of
soil contamination that warrant further consideration in a baseline ERA. Screening is typically
completed during Step 2 of the 8-step Superfund ERA process, as depicted in Figure 1.1. Prior to
using the Eco-SSLs, it is assumed that the risk assessor has completed Step 1, including the site
visit and initial problem formulation. With the information gathered in Step 1, it is recommended
that the risk assessor completes a screening of soils data using the Eco-SSLs in the risk calculation
performed during Step 2.
Comparing the Site Conceptual Model to the General Eco-SSL Model
The user should compare the preliminary site conceptual model developed for their site during Step
1, with the assumptions and limitations inherent in the Eco-SSLs to determine if additional or more
detailed assessments are needed for any exposure pathways or contaminants. Early identification of
areas, conditions, or receptors where Eco-SSLs are not appropriate is important for adequate
planning and sampling strategies for the ERA.
Are There Soil Exposure Pathways for Ecological Receptors?
The Eco-SSLs are designed for use at sites where terrestrial receptors may be exposed directly or
indirectly to contaminated soil. Therefore, generally the first step in determining whether use of the
Eco-SSLs is appropriate is to identify all complete and potentially complete soil pathways present
at the site. The following are the receptor group-specific pathways of exposure to soil contaminants
considered in deriving the Eco-SSLs:
Mammals and Birds
Incidental ingestion of soil
Ingestion of food (soil invertebrates and plants)
Soil Invertebrates and Plants
Direct contact
Ingestion of soil (by soil invertebrates)
Uptake (by plants)
For surface soils (i.e., those soils within the root zone at the specific site), all the above pathways
were considered. Ecological risks from potential exposure to contaminated subsurface soils were
generally not considered. In some site-specific cases, however, there may be risks to animals that
Guidance for Developing Eco-SSLs 1-13 November 2003
-------
burrow beneath the root zone. It should also be noted that, for some plants, the root zone can
extend several feet.
As part of Step 1 of the ERAGS process, the site manager and risk assessor generally need to know
enough about the site to answer at least the following questions:
What contaminants are known or suspected to exist at the site?
What complete exposure pathways might exist at the site?
What habitat types located on or near the site are potentially contaminated?
If it is determined that there are no complete soil exposure pathways (e.g., the current and future
land-use is industrial and there are no terrestrial habitats, or the only soil contamination is well
below both the root zone and the burrow zone at the site), then additional screening for soil effects
on ecological receptors is generally not needed.
Are There Exposure Pathways at the Site Not Addressed by the Eco-SSLs?
In some cases, the site-specific conceptual model may identify complete or potentially complete
ecological soil exposure pathways that are not considered in the derivation of the Eco-SSLs. In
these instances, the additional pathways should be considered in a separate screening analysis or as
part of the baseline ERA. Examples of such instances include:
The contaminated soil is near a surface water body or wetland where there is
potential for contamination of surface water and/or sediments by transport during
rain events.
There are other likely ecological exposure routes not considered in the derivation of
the Eco-SSLs. For example, inhalation of VOCs may be of concern for burrowing
animals.
Some site conditions may be a source of contamination to groundwater. For
example, contaminants from soils may leach to groundwater, which could result in
exposures for ecological receptors upon discharge to surface waters.
Comparing Site Soil Concentrations to the Eco-SSLs
Comparisons of site soil concentrations to the Eco-SSLs during Step 2 of the ERAGS process may
be used to answer the following questions:
Are there any potential ecological risks associated with soil contamination, and is it
necessary to proceed with a baseline ERA (Steps 3 to 8 of ERAGS)?
Which contaminants in soil can be dropped from further consideration and which
ones should be the focus of the baseline ERA?
Guidance for Developing Eco-SSLs 1-14 November 2003
-------
Which geographic areas of soil contamination may result in ecological risks?
Which receptors/functional groups (e.g., birds or soil invertebrates) appear to be at
most risk and should be the focus of the baseline ERA?
Are the Existing Site Soil Contaminant Data Adequate?
At this point of the process, the user should make a decision concerning the adequacy of the
available contaminant concentration data for soils for use in completing a screening level analysis.
This decision, typically made by the site manager and risk assessor, usually considers the
following:
Are all expected areas of soil contamination sampled, or are there other areas of
potential exposure for ecological receptors for which soil data are not available?
Are the parameters of the soil analyses sufficient to identify the possible
contaminants deposited as part of known waste disposal processes and practices?
For example, if DDT is suspected as part of the deposited waste, are soil analyses
available for DDT? Or are data only available for metals?
Are the quantification limits adequate to measure the contaminants at the Eco-SSL
levels?
How do you Calculate the Concentration Term for Comparison to the Eco-SSLs?
The appropriate soil contaminant concentration for comparison to the Eco-SSL is dependent on a
number of factors, including the size of the site, the nature and extent of the contamination, and the
level of confidence in the site sampling data. In most cases, there are limited soil data available at
Step 2 of the ERAGS process; therefore, it is recommended that the maximum soil contaminant
concentrations be compared to the Eco-SSLs. However, if the data set is large, the 95 % upper
confidence limit (UCL) of the arithmetic mean may be the appropriate value to use. Decisions
concerning concentration terms used for comparisons should be made in consultation with the site
manager, site risk assessor, and the regional BTAG (as applied in Superfund).
Which Eco-SSL Should be Used?
Assuming there is a complete exposure pathway, the lowest of the four reported Eco-SSLs should
generally be used to compare to the site soil concentrations. The ERA process assumes that
complete exposure pathways exist for each of the four receptor groups; i.e., every terrestrial habitat
at or near a hazardous waste site is, or should be, suitable for mammals, birds, plants, and soil
invertebrates.
Guidance for Developing Eco-SSLs 1-15 November 2003
-------
What if Soil Contaminant Concentrations Exceed Eco-SSLs?
If the appropriate site soil contaminant concentration exceeds an Eco-SSL, then the user should
retain that contaminant as a COPC for further consideration in the baseline ERA. If soil
concentrations exceed some receptor-specific Eco-SSLs and not others, then it is recommended that
the contaminant be retained as a COPC only for those receptor groups where Eco-SSLs are
exceeded.
What if Soil Contaminant Concentrations Do Not Exceed Eco-SSLs?
Contaminants in soils with concentrations lower than Eco-SSLs can be excluded as COPCs in the
subsequent ERA. However, the user should recognize that new information may become available
during the baseline ERA which may show that initial assumptions are no longer valid (e.g., site
contaminant levels are higher than reported earlier). In this case, contaminants may be placed back
on the list of COPCs. If there are no soil contaminant concentrations that exceed the Eco-SSLs, a
baseline ERA for soils is generally not needed for that site.
What if There is No Eco-SSL?
At this time, Eco-SSLs are not available for all four receptor groups and for all 24 soil
contaminants. For some of the Eco-SSL contaminants, there was an insufficient number of
acceptable toxicity studies to establish an Eco-SSL. For some contaminants, EPA had not
completed the review of toxicity data for derivation of Eco-SSLs. The current Eco-SSLs are
available on the EPA website http://www.epa.gov/oswer/riskassessment/ecorisk/ecossl.htm. The
user should consult this source for the current values.
For those contaminants with an insufficient number of acceptable toxicity studies to establish an
Eco-SSL, a summary of the toxicity studies evaluated in the Eco-SSL process was made available
in the contaminant specific Eco-SSL documents. The information from these studies can be used
according to the process described in Section 1.3.1 of ERAGS to derive screening values. As more
toxicity information becomes available, EPA may use these processes to revise existing Eco-SSLs
or to develop new ones.
Can I Use Site-specific Data to Modify an Eco-SSL or Should I Proceed to a Baseline Risk
Assessment?
If one or more Eco-SSL values are exceeded it is recommended that the user proceed to a baseline
ERA. The Eco-SSLs are intended to be conservative values used to identify those contaminants
that should be the focus of a baseline ERA and as such they should not be modified as part of the
screening step. During the baseline ERA, it will be possible for the user to collect and use site-
specific data as part of development of the site-specific exposure and toxicity assessments in the
baseline ERA.
Guidance for Developing Eco-SSLs 1-16 November 2003
-------
Consideration of Background Soil Concentrations
Due to conservative modeling assumptions (e.g., metal exists in most toxic form or highly
bioavailable form, high food ingestion rate, high soil ingestion rate) which are common to
screening processes, several Eco-SSLs are derived below the average background soil
concentration for a particular contaminant as presented in Attachment 1-4. Since Eco-SSLs are not
designed to be used as clean-up levels, EPA is not promoting clean-up to below background
concentrations. It is EPA's policy to not screen against background levels. Background
concentrations, the speciation of metals, and the effects of conservative modeling assumptions are
generally taken into account in the initial steps of the baseline risk assessment. Specific procedures
for comparing risk information to site specific background levels is addressed in the Office of Solid
Waste and Emergency Response document Role of Background in the CERCLA Cleanup Program
(OSWER 9285.6-07P) (U.S. EPA, 2002). However, if a specific exposure parameter or toxicity
endpoint is found to consistently produce an Eco-SSL that is below background concentrations
(taking into account natural species of metals), EPA may reevaluate specific Eco-SSL values as
they are updated.
Information on background concentrations of
contaminants in soils was collected and
reviewed during the Eco-SSL derivation
process to examine how the Eco-SSL values
compare to natural soil conditions. These
comparisons were used to guide the process
and are presented as Attachment 1-4. The
review indicated that there are regions of the
country where natural background levels for
some metals exceed Eco-SSLs. For these
regions and for specific local areas, the
acquisition of adequate data on background
soil concentrations is an important step
toward evaluating, on a site-specific basis, if
observed concentrations are related to
releases or are naturally occurring.
Background concentrations are further
discussed in each of the contaminant-specific
Eco-SSL documents.
Definitions
Contaminants of concern (COCs) are the contaminants that, at
the completion of the risk assessment, are found to pose
unacceptable human or ecological risks. The COCs drive the
need for a remedial action.
Contaminants of potential concern (COPCs) generally
comprise the contaminants that are investigated during the
baseline risk assessment that may or may not pose unacceptable
risks.
Screening is a common approach used by risk assessors to refine
the list of COPCs to those hazardous substances, pollutants and
contaminants that may pose substantial risks to health and the
environment.
Background refers to constituents or locations that are not
influenced by the releases from a site, and is usually described as
naturally occurring or anthropogenic (US EPA, 2002)
Anthropogenic -natural and human-made
substances present in the environment as a
result of human activities (not specifically
related to the release in question); and,
Naturally occurring - substances present in
the environment in forms that have not been
influenced by human activity.
Guidance for Developing Eco-SSLs
1- 17
November 2003
-------
This Page Intentionally Left Blank
-------
2.0 SOIL PROPERTIES
2.1 Introduction
Soil properties influence the toxicity of contaminants to invertebrates, plants, and wildlife.
Therefore, they are generally important to consider in the development of Eco-SSLs and provide
a basis for guiding site-specific evaluations that may follow application of Eco-SSLs. This
chapter discusses the primary soil parameters that influence bioavailability of contaminants from
soils. This soil property information provides the rationale for defining a set of soil parameters
used in the selection of the most appropriate studies for deriving Eco-SSLs for plants and soil
invertebrates and specific recommendations for screening soils for aluminum and iron.
This chapter focuses primarily on the relationship between soil chemistry factors that influence
the toxicity to and accumulation of contaminants in soils to plants and soil invertebrates. The
absorption of contaminants bound to incidentally ingested soil particles in the animal gut is
influenced by other parameters including gut residence time as well as toxicokinetic and
physiological factors that may affect the uptake of contaminants in wildlife. Because wildlife
species vary in the anatomy and physiology of their digestive systems (e.g., herbivores vs.
carnivores), generalizations can not and should not be made concerning bioavailability of
contaminants on soil incidentally ingested by wildlife at the screening level.
2.2 Soil Properties Influencing Contaminant Bioavailability
Bioavailability is a measure of the potential for entry of the contaminant into ecological or
human receptors and is specific to the receptor, the route of entry, time of exposure, and the soil
matrix containing the contaminant (Anderson et al., 1999). In order to ensure that Eco-SSLs are
adequately conservative for a broad range of soils, an effort is made in the procedures developed
to select studies that favor the bioavailability of the selected contaminants. To accomplish this,
it is first necessary to develop a basic understanding of how various soil properties may
influence bioavailability. Several authors stress the importance of physical and chemical
properties of contaminants that influence the bioavailability in soils and thus exposure and
toxicity (Alexander, 1995; Allen et al., 1999; Linz and Nakles, 1997; and Loehr and Webster,
1996). The behavior and bioavailability of contaminants are greatly influenced by their
interactions with soil parameters, such that not all contaminants are equally available to biota.
However, estimating the availability of metals and organic contaminants in soil to soil biota and
plant toxicity is not a straightforward process.
The bioavailability of contaminants depends on their chemical properties and the specific
physical and geochemical binding mechanisms that vary among contaminants and soil types.
Contaminants interact with soil through interactions with the surface of particulate material in
soils (adsorption), by penetration through the particulate surfaces where the contaminant
becomes associated with the internal material (absorption or partitioning), and through specific
contaminant reactions sometimes referred to as chemisorption. Also some contaminants, in
particular metals, can associate with inorganic ligands and precipitate. The affinity of a
Guidance for Developing Eco-SSLs 2-1 November 2003
-------
contaminant to be removed from solution and become associated with soil participates
irrespective of mechanism, is generally referred to as "sorption". The exception are precipitation
reactions, which are often discussed independently from sorption processes. Contaminants are
generally considered to be bioavailable when they are released from interactions with the soil
and soil constituents and released into the soil pore-water. The exception to this rule is the direct
ingestion of soil by terrestrial wildlife.
Identifying and quantifying soil properties that control the distribution of a contaminant in
soil/water systems at equilibrium are useful for exposure situations where time is sufficient for
equilibrium conditions to develop. For exposure situations that are dominated by discrete events
often of short duration (e.g., incidental ingestion of soil), the kinetics of contaminant release
from soils into another medium (i.e., the amount released per unit time) and residence time (i.e.,
time allowed for transfer to occur) controls the fraction of a contaminant that would be labile to
target biota. Both adsorption and absorption partitioning processes are considered reversible,
although mass transfer from the particle to the pore-water can be constrained. In the case of
interactions within a particle, a contaminant can become sequestered or trapped through various
physical and contaminant alterations that occur over time, such that contaminant release is
completely constrained. The decline of the availability of many organic contaminants in soil
over months or years has been well-documented (Alexander, 1995; Loehr and Webster, 1996).
For chemisorption, the binding mechanism is considered irreversible under most environmental
conditions. For precipitation reactions, release to pore-water is controlled by the factors
affecting the stability or solubility of the contaminant precipitate. Overall, bioavailability of a
contaminant in soil strongly depends on its physical and chemical properties, the characteristics
of the soil, the interactions between the contaminant and the medium, including time of
exposure, and the physiological and biochemical conditions of the receptor.
Contaminant Characteristics Impacting Lability
The soil parameters important in affecting sorption and precipitation reactions and the extent of
their influence and thus contaminant bioavailability, are dependent on the intrinsic properties of
the contaminants. The 24 contaminants considered in this guidance include both metals and
organic contaminants. Metals can exist as either cations or anions in the soil environment, which
significantly affects their sorption, mobility, and solubility in soils. For example, soil is
primarily negatively charged, thus, metal cations have a higher propensity to be sorbed by soil
particles relative to metal anions. For organics, lipophilicity and persistence alter their
availability, as well as ionic potential in the case of organic contaminants with ionizable
functional groups. Collectively, the 24 contaminants are classified into the four groups (Table
2.1).
Metals. Metals occur naturally in soils primarily as amorphous oxides and hydroxides, and to a
lesser extent as carbonates, phosphates, sulfates, and sulfides, which are relatively insoluble.
The same is generally true for metal-contaminated soils, because metals quickly undergo
precipitation and coprecipitation reactions forming relatively insoluble solid phases, and/or are
strongly complexed by soil minerals or organic matter (Lindsay, 1979). Toxicity testing on the
Guidance for Developing Eco-SSLs 2-2 November 2003
-------
other hand usually employs very soluble metals not commonly found in any appreciable amounts
in soils relative to total metal concentrations.
Table 2.1 General Contaminant Classification
Contaminant Class
Metal Cations
Metal Anions
Nonionic Organics
lonizable Organics
Eco-SSL Contaminant
aluminum, antimony, barium, beryllium, cadmium, cobalt, copper,
iron, lead, manganese, nickel, silver, and zinc
arsenic, chromium, selenium, and vanadium
PCBs, DDT and metabolites, dieldrin, PAHs, TNT, and RDX
PCP
As identified in Table 2.1, most of the 24 contaminants considered for Eco-SSLs are metals that
typically exist as cationic species. These metals can complex with inorganic soil constituents,
e.g., carbonates, sulfates, hydroxides, sulfides, to form either precipitates or positively charged
complexes. Both complexation and precipitation reactions are pH dependant. Therefore,
although these metals can form complexes with a net negative charge, under most
environmentally relevant scenarios (pH = 4 to 8.5), these metals either precipitate or exist as
cations.
Arsenic, chromium, selenium, and vanadium complex with oxygen and typically exist as anionic
species under most environmentally relevant scenarios (Bohn et al., 1985; Lindsay, 1979). The
most common forms of arsenic are arsenate (arsenic V) and arsenite (arsenic III), which are
present in soil solution in the form of AsO43" and AsO2", respectively. The chemistry of arsenic
resembles that of phosphate (Barber, 1995; Bohn et al., 1985). Chromium can exist as chromate
(chromium VI or CrO42"), which is usually considered more soluble, mobile and bioavailable
than the sparingly soluble chromite (chromium (III)), which is normally present in soil as the
precipitate Cr(OH)3 (Barnhart, 1997; James et al., 1997). Similarly, selenium can be present as
selenates (SeO42") and selenites (SeO32"). For vanadium, vanadate (VO43") is the most common
form.
Metals in their various forms can exist in the pore-water as charged species, as soluble
complexes, or precipitate out of solution. Retention by soil is usually electrostatic with cationic
species and anionic species being associated with negatively and positively charged sites on the
soil, respectively. For most soils in the United States, negatively charged sites are more plentiful
with less than 5% of the total available charge on the soil surface being positively charged.
Therefore, metals existing as cationic species have a greater propensity to associate with the soil
and are less bioavailable, whereas, distribution of anionic metals is generally more towards the
pore-water for most soil/water systems. The soil pH and availability of charged sites on soil
surfaces are the primary soil factors controlling their release to the pore-water, and subsequently,
bioavailability.
Guidance for Developing Eco-SSLs
2-3
November 2003
-------
Organic Contaminants. Of the seven organic contaminants identified in Table 2.1, DDT and
metabolites, dieldrin and PCBs are very hydrophobic, highly lipophilic, and persistent nonionic
organic contaminants. These contaminants are highly sorbed to soil surfaces and organic matter
and tend to bioaccumulate and biomagnify in the food chain. The structure and degree of
chlorination of PCBs and associated congeners directly impacts their behavior, persistence, and
bioavailability (e.g., see citations in Hansen et al., 1999). As solubility decreases, sorption
increases, and bioavailability generally decreases with increasing chlorination. However,
uptake, degradability, and toxicity are also impacted by placement of the chlorines in the
biphenyl structure. The remaining nonionic organic contaminants, PAHs and explosives (TNT
and RDX) are generally considered less persistent and more bioavailable than pesticides or PCBs
under identical soil conditions. PAHs are compounds with two or more aromatic rings in their
structure and consist of only carbon and hydrogen. PAHs can be highly retained by soil in a
similar manner as pesticides or PCBs, but are considered less persistent due to their higher
affinity to be degraded microbially. TNT and RDX, a trinitro aromatic and trinitro
nitrogen-heterocyclic respectively, are explosive materials and are more polar than PCBs or
PAHs. The only ionizable organic contaminant being considered at this time in the development
of Eco-SSLs, is the organic acid pentachlorophenol (PCP). Organic acids can exist as either a
nonionic species or as an organic anion, which is dependent on the acid dissociation constant
(pKa) and pH. In the pH range relevant to most environmental scenarios, PCP can exist as both
a neutral species and as an anionic species; however, the majority will exist as the organic anion
(Leeetal., 1990).
For all nonionic organic compounds (NOC) and the neutral form of PCP, sorption by soil is
primarily related to their hydrophobicity and the amount of organic matter present in the soil
(Lagrega, 1994; Lee et al., 1990), with the exception of the more polar, nitro-substituted organic
contaminants (i.e., the explosives). Differences in the distribution of several NOCs in diverse
soil-water and sediment-water systems have been minimized by normalization to organic matter
or more specifically organic carbon (OC) with OC-normalized distribution coefficients, referred
to as Koc values (e.g., Gertsl, 1990; Lyman et al., 1990;). The greater the affinity of a
contaminant for organic matter, the larger the organic carbon-normalized partition coefficient
(Koc), and a soil with higher amounts of organic matter has a higher propensity to sorb NOCs.
The hydrophobicity and Koc, of organic compounds increases with the size of the compound and
with increasing chlorine content, as in the case of chlorinated organics. Therefore, sorption of
PAHs by soils increases with the number of aromatic rings. For compounds like PCBs, sorption
increases with increasing chlorination. Increasing compound hydrophobicity also reflects
increasing lipophilicity, which will result in a greater propensity to bioaccumulate in the lipid
fraction of biota. For PCP, an ionic contaminant, the anionic species has a greater tendency
relative to the neutral PCP to remain in the pore-water similar to metal anions. Therefore,
pH-dependent speciation drastically modifies the solubility, sorption, transport, and
bioavailability of PCP. Although organic matter is the primary sorption domain in soils, all
contaminants have some affinity to be associated with any surface through weak physical forces
(Schwarzenbach et al., 1993). In addition, the nitro-substituted NOCs are known to have
specific interactions with clay surfaces that are impacted by the inorganic cations present and
Guidance for Developing Eco-SSLs 2-4 November 2003
-------
clay charge density, and less so by the amount of organic matter present (Weissmahr et al., 1998;
1999).
A common contaminant index representing the degree of hydrophobicity and lipophilicity of an
organic contaminant is the octanol-water partition coefficient (Kow), which is the contaminant
distribution between octanol and water phases. Kow values are positively correlated to both Koc
values and bioconcentration factors (Lyman et al., 1990). For reference, log Kow values for
selected organic contaminants are summarized in Table 2.2.
Table 2.2. Log K,,w Values for Organic Contaminants
Analyte
RDX
TNT
DDT
ODD
DDE
Dieldrin
Pentachlorophenol (PCP)
PCBs
CAS no.
121824
118967
50293
72548
72559
60571
87865
logK,,w
0.87
1.6
6.53
6.1
6.76
5.37
5.09
4.5 (1 chlorine)
>8 (10 chlorines)
Source
Syracuse Research Corporation (SRC)
SRC
U.S. EPA (1996)
U.S. EPA (1996)
U.S. EPA (1996)
U.S. EPA (1996)
U.S. EPA (1996)
Verschueren(1996)
Schwarzenbach et al. (1993)
PAHs
Naphthalene (2 rings)
Acenaphthene (3 rings)
Phenanthrene (3 rings)
Anthracene (3 rings)
Chrysene (4 rings)
Benzo(a)anthracene (4 rings)
Benzo(a)pyrene (5 rings)
Dibenzo(ah)anthracene (5 rings)
Benzo(b)fluoranthene (5 rings)
Benzo(k)fluoranthene (5 rings)
Benzo(ghi)perylene (6 rings)
91203
83329
85018
120127
218019
56553
50328
53703
92240
207089
191242
3.36
3.92
4.55
4.55
5.7
5.7
6.11
6.69
6.2
6.2
6.7
U.S. EPA (1996)
U.S. EPA (1996)
U.S. EPA (1995)
U.S. EPA (1996)
U.S. EPA (1996)
U.S. EPA (1996)
U.S. EPA (1996)
U.S. EPA (1996)
U.S. EPA (1996)
U.S. EPA (1996)
U.S. EPA (1995)
Key Soil Parameters Affecting Contaminant Bioavailability in Soils
From the preceding overview of how contaminants interact with soil constituents, it is clear that
soil plays a very significant role in reducing the potential bioavailability of contaminants in the
environment. Given the types of contaminant-soil interactions presented, the primary soil factors
controlling the potential bioavailability of all contaminants are identified as soil pH, available
charged sites on soil surfaces, clay content, and soil organic matter. Below is a discussion
Guidance for Developing Eco-SSLs
2-5
November 2003
-------
briefly detailing the key soil parameters affecting the various contaminants' availability to the
pore-water, thus bioavailability.
Soil pH. Soil pH is often termed the master soil variable because it controls virtually all aspects
of contaminant and biological processes in soil. These processes include solubility,
precipitation, speciation, and sorption processes as well as microbial activity. Soil pH controls
the speciation of both ionizable organic contaminants such as PCP, and metals. For metals, the
net charge of the metal complexes and their precipitation/dissolution reactions are directly
impacted by soil pH. For organic acids such as PCP, the fraction of contaminant existing as an
anion increases with increasing pH. The anion has a lower affinity for the soil relative to the
neutral species. Increasing soil pH also results in an increase in the number of negatively
charged soil sites with a concomitant decrease in the positively charged sites. Therefore,
increasing the soil pH directly impacts the sorption and removal from the pore-water of metal or
organic ions (Bohn et al., 1985). The impact of pH on the behavior and bioavailability of
nonionic organic contaminants is less marked and is generally achieved through its influence on
organic matter and on microbial activity.
Cation and Anion Exchange Capacities. The available charges on soil surfaces are quantified
in the soil parameters known as cation exchange capacity (CEC) and anion exchange capacity
(AEC). CEC is a measure of the soil's ability to adsorb and release cations, which is directly
proportional to the number of available, negatively charged sites. Likewise, AEC is a measure
of the soil's ability to adsorb and release anions. As a result, the AEC is a measure of available
positively-charged surface sites. CEC is directly related to the clay mineral content and type,
organic matter and soil pH. CEC is greater for 2:1 clays such as montmorillonite (600 to 1,000
millimole (mmol)/kg) compared to 1:1 clays such as kaolinite (20 to 160 mmol/kg). CEC in
organic matter ranges from 2,000 to 4,000 mmol/kg; however, the organic matter fraction of a
soil is usually much less than the clay fraction. CEC arising from pH-dependent charge, which
includes organic matter contributions to CEC, increases with increasing pH. CEC in soil ranges
from values as low as 10 millimole (mmol) per kg for extremely coarse-textured soil to as much
as 600 mmol/kg for fine textured soil, containing large amounts of 2:1 clays and organic matter
(Bohn et al., 1985). AEC, which is primarily associated with amorphous oxides, decreases with
increasing soil pH. As previously mentioned, the number of positively charged sites (i.e., AEC)
on the majority of soil types is very small, and in environmentally-relevant pH ranges, is usually
negligible. Therefore, AEC is not generally considered an important parameter in assessing
contaminant availability at most sites in the United States.
Clay Minerals Clays, by definition, are soil particles less than 2 microns in size (Miller and
Gardiner, 1998); therefore, high clay soils have higher surface areas relative to sandy soils (sand
particle size ranges from: 20 microns to 2 millimeters (mm)). For nonionic organic
contaminants, the primary sorption domain is organic matter; however, soils with high surface
area will result in enhanced sorption of organic contaminants through weak physical interactions,
as well. Much of the CEC of a soil comes from the negatively charges sites on clay surfaces.
Therefore, high clay soils will have a higher affinity to sorb cationic species whether organic or
Guidance for Developing Eco-SSLs 2-6 November 2003
-------
inorganic due to CEC, and to sorb nonionic organic contaminants due to high surface areas, thus
making contaminants less bioavailable relative to sandy soils. In addition to charged sites
available in clays, siloxane oxygens present in clays can interact specifically with contaminants
such as the nitro-substituted explosives. Metals can form precipitates with inorganic soil
constituents, such as carbonate and phosphate minerals under certain soil conditions. Carbonate-
and phosphate-metal complexes have varying degrees of solubility and reactivity depending on
the metal, its oxidation state, the ligand to which it is bound, and pH. Precipitation removes a
contaminant from the pore-water, thus decreasing bioavailability.
Organic Matter (Organic Carbon) Content. Organic matter includes plant and animal
remains in various stages of decomposition (e.g., cells and tissues, of soil organisms and
substances from plant roots and soil microbes) (Sumner, 2000). Organic matter is primarily
composed of carbon, oxygen, and nitrogen. Organic matter is often reported or analytically
determined on a carbon basis. On average, approximately 58% of organic matter is organic
carbon. Soils encompass a range in organic matter from <1% for a sandy soil to almost 100%
for a peat soil, with most soils having organic matter contents <10% (Bohn et al., 1985). Also,
organic matter content is usually higher in surface soils or in the root zone and decreases with
depth in the soil profile.
Organic matter has a high affinity to bind organic compounds as well as some metals in soils
thereby, reducing their availability. Organic contaminants preferentially partition to organic
matter relative to the polar aqueous phase, while the organic acid functional groups typically
present in organic matter have a high affinity to attract metal cations. For nonpolar or neutral
organic contaminants at equilibrium, sorption is positively correlated to the amount of organic
matter, usually reported as the fraction of organic carbon (foc), and inversely proportional to
aqueous solubility. Sorption of organic contaminants increases with increasing amounts of soil
organic matter. The greater the hydrophobicity or lipophilicity of an organic contaminant, the
greater potential it has to be sorbed onto organic matter. The latter has led to the use of Koc for
estimating contaminant sorption with the soil-specific distribution coefficient estimated by Koc
multiplied by foc. Another indirect effect of soil organic matter is its role on limiting
contaminant mass-transfer. The rate of mass-transfer of an organic contaminant from soil
particles to the surrounding pore-water is inversely proportional to the contaminant's soil-water
distribution coefficient (Pignatello, 2000). Therefore, with increasing organic matter content,
retention of an organic contaminant increases and rates of release decrease, thereby, decreasing
overall contaminant bioavailability.
Other Factors
Background. Background refers to constituents or locations that are not influenced by the
releases from a site, and is usually described as naturally occurring or anthropogenic background
(US EPA, 2002). Background contaminant concentrations can vary due to soil type, depth, and
region of the country. Due to this variation, background metal levels in soils are addressed on a
Guidance for Developing Eco-SSLs 2-7 November 2003
-------
site-specific basis in the baseline ERA. As part of the development of Eco-SSLs, background
metal concentrations from the literature were compiled to provide a general comparison of the
screening levels with current available information. Background metal concentrations were
reviewed and compiled in Attachment 1-4. For reference, background levels by state are
provided in Table 2.3. In the contaminant specific Eco-SSL documents, the Eco-SSL values are
compared to a summary box-and-whisker chart depicting background concentrations in eastern
and western United States soils. A discussion on the quality of available toxicity data and the
reasonableness of the Eco-SSL values when compared to background concentrations is
presented. Background data is site-specific and should be derived for each site investigated.
Aging. The issue of adsorption, complexation, lability of contaminants in soils, and the
corresponding reduction in toxicity over time is an important issue in understanding the fate of
contaminants in soils. In the evaluation of the available toxicological literature for plants and
soil biota, few studies incorporated a step to age or weather spiked contaminants in soils. The
use of contaminated soils from the field in laboratory tests is a viable option; however, due to the
common presence of mixtures and the specific soil chemistry parameters, the use of field
contaminated soils is primarily useful only for site-specific studies. While no standard setting
organization has established methods for aging contaminants in soils, an aging step has been
added by some investigators in plant and soil invertebrate test methods that involves several
cycles of wetting and drying of a freshly spiked soil (Kuperman et al., 2002 and Phillips et al.,
2002).
2.3 Using Soil Properties to Guide Eco-SSL Derivation
In identifying a set of soil parameters for use in selecting studies for deriving Eco-SSLs for
plants and soil invertebrates, four soil parameters were selected: soil pH, CEC, clay content, and
organic matter. However, when the plant and soil invertebrate task group evaluated the current
literature, they observed that CEC and clay content were not routinely reported. Thus, these
parameters were not used and the matrices were constructed using only pH and organic matter
content as the primary soil parameters affecting bio-availability and toxicity. For these soil
parameters, ranges were within that typically found in soils. Tests that used soils with
characteristics that fell outside the selected ranges were not considered. Although other soil
factors can be significant, combinations of these two soil parameters are sufficient for use in this
screening process as a qualitative guide to address how most soils across the United States may
influence bioavailability of contaminants. Qualitative rankings of high, medium, and low
bioavailability were used to categorize each combination of the soil parameters and their ranges.
Information on bioavailability was used to help select and score studies to include in the
derivation of the Eco-SSL values. Greater weight was given to those studies that had higher
bioavailability. Using the selected soil parameters and defining ranges that correspond
qualitatively to the soil's affinity for the contaminant and thus for bioavailability, Tables 2.4a and
2.4b, 2.5, and 2.6 are provided for metal cations, nonionic organics, and anionic species,
respectively. For each of the soil parameters, the values typically found in soils were divided
into three ranges. For example, most environmentally relevant scenarios fall between pH values
Guidance for Developing Eco-SSLs 2-8 November 2003
-------
Table 2.3. Mean Reported Soil Metal Background Concentrations (mg/kg dry weight) by State*
Alabama
Arkansas
Arizona
California
Colorado
Connecticut
Delaware
Florida
Georgia
Iowa
Idaho
Illinois
Indiana
Kansas
Kentucky
Louisiana
Massachusetts
Maryland
Maine
Michigan
Minnesota
Missouri
Mississippi
Montana
Nebraska
North Carolina
North Dakota
New Hampshire
New Jersey
New Mexico
Nevada
New York
Ohio
Oklahoma
Oregon
Pennsylvania
Rhode Island
South Carolina
South Dakota
Tennessee
Texas
Utah
Virginia
Vermont
Washington
Wisconsin
West Virginia
Wyoming
Aluminum
23100
33429
32933
75633
61557
85000
22500
9944
38250
64667
58500
48714
50000
61818
54123
42188
34083
39167
65385
10964
49457
42094
45368
70938
59474
60105
62857
66667
10075
54423
66078
58800
54615
39200
94412
63438
100000
39143
74333
31894
41958
45638
60438
56667
66834
48000
67000
56125
Antimony
3.6
1.2
1.4
0.8
1.1
1.0
0.9
1.0
1.0
1.0
1.1
1.0
1.1
1.0
1.0
1.0
1.2
1.0
1.3
1.0
1.0
1.0
1.1
1.0
1.0
1.0
1.4
1.0
1.0
1.0
1.0
1.2
1.0
1.3
0.7
1.1
1.1
1.2
1.0
1.0
1.3
1.1
Arsenic
4.7
9.7
9.6
5.1
6.7
4.1
1.4
3.0
5.0
7.3
6.4
7.1
7.5
6.8
7.8
7.6
8.6
3.8
9.4
4.2
5.5
10
8.8
8.8
5.5
4.8
7.0
4.4
7.0
5.9
9.0
6.4
12
7.0
5.1
13
3.5
3.9
8.5
16
6.4
8.0
5.1
3.6
4.5
4.4
8.6
6.5
|
cS
03
200
336
364
598
662
400
400
48
232
617
757
551
500
694
349
441
203
393
319
127
571
499
390
739
711
356
682
500
54
727
822
666
469
430
682
366
500
151
1043
193
404
493
436
333
606
543
360
756
Beryllium
0.6
0.9
1.0
1.1
1.4
0.5
0.5
0.6
0.6
1.3
1.1
0.7
0.7
1.0
1.1
0.6
1.3
1.3
1.6
0.7
0.7
1.0
0.9
1.1
1.1
0.6
0.9
2.3
0.3
1.0
1.3
1.4
1.0
1.1
0.9
1.4
0.5
1.4
1.4
0.8
0.9
0.9
0.9
1.7
0.9
2.0
1.0
0.7
Cadmium
0.4
0.4
0.1
0.2
0.9
0.3
0.6
0.3
0.2
0.2
0.8
Chromium
30.6
53.1
37.3
119.9
41.7
40.0
30.0
15.4
32.4
64.7
52.1
48.4
46.8
49.0
79.8
60.8
39.5
47.9
71.2
13.8
25.4
50.0
53.2
63.3
32.5
64.8
53.2
18.4
13.9
55.5
36.8
66.9
55.0
46.0
121.6
52.8
50.0
21.4
58.7
40.3
39.6
45.6
54.3
66.7
49.9
40.3
46.0
47.9
1
0
v
4.4
12
9.9
14
6.8
7.5
3.3
1.6
6.9
11
12
9.8
10
8.9
11
8.6
7.8
7.5
10
4.6
7.2
12
12
7.5
5.9
15
6.9
5.3
1.7
8.8
8.4
9.1
13
7.1
16
15
10
3.5
7.7
14
5.3
6.6
9.7
12
18
7.7
14
8.3
b
a
0
v
9.6
17
23
39
21
15
5.0
5.6
21
31
28
24
27
25
17
33
16
20
28
12
20
19
20
29
15
34
23
12
14
21
25
36
28
16
53
37
15
16
29
17
15
26
33
18
31
12
22
21
a
8
11950
19857
20787
36867
23048
17500
7500
3705
16976
23278
32000
19159
21364
18788
30432
19688
19000
28571
45385
10520
19581
24733
19684
27766
16000
37053
25357
33333
11632
20898
22725
38900
27308
19320
50147
36063
30000
12500
25667
28479
16328
18830
27750
30000
42635
15667
28500
25250
Manganese
420
731
447
640
343
450
85
86
252
603
580
646
518
452
483
470
439
291
581
230
583
940
471
366
306
563
530
633
221
367
481
418
550
465
725
609
500
87.1
1013
1112
303
371
441
800
760
365
770
416
1
o
iz
11
18
23
48
13
13
6.0
8.5
17
26
22
19
18
17
23
33
13
13
30
12
14
20
21
20
15
24
20
10
3.8
28
15
21
25
15
23
24
15
7.8
28
18
12
13
17
25
23
14
23
16
"S
,_J
9.3
21
16
26
31
5.0
15
12
19
19
22
39
18
32
16
16
13
22
19
9.2
9.9
23
18
14
16
17
13
28
35
18
25
20
23
18
15
23
15
5.0
16
23
14
35
36
20
14
12
17
17
Selenium
0.3
0.7
0.4
0.2
0.4
0.8
0.3
0.3
0.4
0.4
0.3
0.5
0.4
0.4
0.5
0.7
1.9
0.2
0.7
0.3
0.3
0.5
0.5
0.4
0.4
0.4
0.4
0.3
0.9
0.3
0.3
0.3
0.6
0.3
0.3
0.5
0.9
0.3
0.5
0.6
0.3
0.3
0.4
0.4
0.3
0.3
0.5
0.5
tH
1)
>
CW
0.5
0.8
0.5
0.5
1.2
0.7
Vanadium
38
52
42
118
74
60
20
11
43
97
90
62
74
77
66
76
87
63
98
44
72
72
68
101
62
107
83
57
30
72
78
132
88
50
168
80
70
45
108
49
52
70
77
70
160
48
65
84
3
si
26
39
51
113
87
40
23
12
47
57
83
67
56
67
35
55
54
39
80
33
38
53
45
69
54
56
64
23
22
44
69
82
69
50
70
81
30
25
75
57
39
96
233
43
78
44
60
57
* Summary of background soil concentration data provided as Attachment 1-4.
Guidance for Developing Eco-SSLs
2-9
November 2003
-------
of 4.0 and 8.5. This pH range was divided into the following sub-ranges: 4.0 to 5.5, 5.5 to 7.0,
and 7.0 to 8.5. Qualitative bioavailability indices of very high, high, medium, low, and very low
were assigned for each combination of soil parameters within each class of the contaminants
(Tables 2.4a, 2.4b, 2.5, and 2.6). For example, a soil with a pH between 5.5 to 7.0, and organic
matter content between 2 and 6%, would bind metal cations to a moderate extent. Therefore, a
bioavailability index of 'medium' for metal cations was assigned (see Table 2.4a and 2.4b).
These tables simplified and facilitated the use of soil chemistry information in the derivation of
Eco-SSLs for plants and soil invertebrates. The ranges given in these tables were used in
selecting the most appropriate plant and soil invertebrates toxicity data for deriving Eco-SSLs
(Chapter 3).
Table 2.4a.
Qualitative Bioavailability of Metal Cations in Natural Soils
to Plants
Soil Type
4 < SoilpH< 5.5
5.5
-------
Table 2.5
Qualitative Unavailability of Non-Ionizing Organic Contaminants in Natural Soils
Soil Type
4 < SoilpH<5.5
5.53.5
LogKow<3.5
LogKow>3.5
LogKow<3.5
LogKow>3.5
LogKow<3.5
Organic Matter (%)
<2
High
Very High
Medium
High
Low
Medium
2to<6
Medium
High
Low
Medium
Low
Low
6 to 10
Low
Medium
Low
Low
Low
Low
Table 2.6
Qualitative Unavailability of Metal Anions in Natural Soils
Soil Type
4< SoilpH<5.5
5.5
-------
This Page Intentionally Left Blank
-------
3.0 DERIVATION OF PLANT AND SOIL INVERTEBRATE ECO-SSLs
Eco-SSLs for plants and soil invertebrates were derived using a four-step process. These
procedures built upon previous efforts (CCME, 1997; Efroymson et a/., 1997 a,b), and
established a process for evaluating published studies and selecting relevant data from the
literature. The process (Figure 3.1) was guided by a set of standard operating procedures (SOPs)
that specified the steps for identifying and evaluating data for appropriateness in deriving an
Eco-SSL as well as the actual calculation of the screening values (Attachments 3-1 and 3-2).
The process was intended to ensure that the Eco-SSLs for plants and soil invertebrates were
based on sound science. The process required data from ecotoxicity studies that met prescribed
requisites including but not limited to a thorough experimental design, appropriate quality
control, and specific soil boundary conditions.
When appropriate and sufficient data were available for a specific contaminant, Eco-SSLs for
plants and soil invertebrates generally could be derived using the prescribed process.
Alternatively, if sufficient data does not exist, then toxicity testing could be completed under the
preferred Eco-SSL experimental conditions (refer to Section 3.5) to generate the data needed to
derive an Eco-SSL for the contaminant of interest.
Eco-SSLs for plants and soil invertebrates were
derived using data from tests performed within
soil boundary conditions favoring relatively
high bioavailability for upland aerobic soils
(refer to Tables 2.4 - 2.6 of Chapter 2). The soil
chemistry conditions of relatively high
bioavailability were defined by low soil pH and
organic matter. These parameters were
frequently found to be the predominant factors
affecting contaminant bioavailability to plants
and soil invertebrates in aerobic soils. Extreme
pH values (< 4 and > 8.5) will substantially
affect the solubility, precipitation, speciation,
and sorption processes of contaminants, and
therefore were not appropriate for use in Eco-
SSL derivation. Similarly, organic matter,
composed primarily of carbon, oxygen, and
nitrogen, at elevated levels in the soils (i e.
>10%) has a high affinity to bind organic
compounds as well as some metals, thereby
reducing their bioavailability.
Figure 3.1 The Four-Step Process for Deriving
Eco-SSLs for Plants and Soil Invertebrates
Step 1. Literature search, acquisition, and
screening (Attachment 3-1). Apply 22
Literature Exclusion Criteria.
Step 2. Identify acceptable literature by
applying eleven Study Acceptance
Criteria to retrieved papers (Attachment
3-1).
Step 3. Extract and score data from acceptable
literature according to nine Study
Evaluation Criteria (Attachment 3-2).
Step 4. Derive soil invertebrate and plant Eco-
SSLs according to specified procedures
(Attachment 3-2).
Sort study data by bioavailability score
Complete QA review
Calculate value
Guidance for Developing Eco-SSLs
3- 1
November 2003
-------
Figure 3.2 Literature Exclusion Criteria
3.1 Literature Search, Acquisition, and Screening
The first step in the process of developing Eco-SSLs for plants and soil invertebrates was to
identify, retrieve, and screen published papers that reported data for soil toxicity to terrestrial
plants or soil invertebrates. The procedures used to complete the literature search, retrieval and
review are provided as Attachment 3-1.
The literature search included both paper-based searches and searches of computerized
abstracting databases. The paper-based literature search process consisted primarily of the
manual review of bibliographies, guidance documents, and review articles. This manual search
was not limited by publication year. Searches of computerized abstracting databases included
the use of DIALOG, SilverPlatter
and Ovid commercial database
vendors. Within DIALOG, the
targeted databases included
AGRICOLA, BIOSIS and
ChemAbstracts. The searches were
supplemented with other electronic
databases including Toxline,
PolToxl, Toxnet, and Current
Contents. Searches of computerized
abstracting databases were limited
to papers published after 1988,
except in cases when fewer than 20
publications were identified for a
contaminant- receptor pairing (e.g.,
cadmium- plants). In these cases,
the electronic search was expanded
to include all publication years in
the databases. It was assumed that
relevant studies prior to 1988 would
be identified in the bibliographies of
review articles. A list of 22
Literature Exclusion Criteria (see
Figure 3.2) was then used to screen
out those studies not appropriate for
use in deriving Eco-SSLs. These
Literature Exclusion Criteria were
applied to retrieved abstracts, or to
the entire publication if the needed
information was not available in the
abstract.
Biological Product
Chemical Methods
Drug
Effluent
Contaminant Fate
Human Health
In Vitro
Methods
Mixture
Modeling
No Cone
No Duration
No Effect
No Species
No Toxicant
No Tox Data
Nutrient
Oil
Publ As
QSAR
Review
Survey
Biological toxins (venoms, etc.)
Methods for measuring contaminants
Testing for drug effects
Effluent, sewage, polluted run-off
Fate and transport of substance in the
environment (only)
Human or primate subjects
In vitro studies, including cell cultures and
excised tissues
Methods reported but no usable specific
toxicity test results
Combinations of chemicals in laboratory
testing
Only modeling results reported
No dose or concentration reported, or not
able to calculate from information given
No exposure duration reported
No effect reported for a biological test
species
No viable plant or animal present or tested
No toxicant used
Toxicant used, but no results reported that
had a negative impact
Nutrition studies reporting no concentration-
related negative impact
Oil and petroleum products
Author states information is published in
another source
Data developed only from
quantitative-structure activity relationships
Data reported are not primary data
Assessment of toxicity in the field over a
period of time
QSAR = Quantitative Structure Activity Relationship
Guidance for Developing Eco-SSLs
3-2
November 2003
-------
For the 24 contaminants listed in the Chapter 1, EPA completed the literature searches according
to the SOP (Attachment 3-1). The searches identified more than 7,600 papers. These
publication abstracts and titles were screened to determine if they were likely to meet the Eco-
SSL requirements using the 22 Literature Exclusion Criteria. This process resulted in the
acquisition of more than 5,200 papers.
Figure 3.3 Eleven Study Acceptance Criteria
1. The document is the primary source of the test
result.
2. Adverse effects are caused by an identified
chemical stressor (i.e., no mixture testing in
laboratory studies).
3. The chemical form (e.g., metal salt used) and
concentration are reported by the author(s).
4. The test medium used in the study is a natural
or artificial soil.
5. The study reports the organic matter content
and it is < 10 % of the composition of the soil;
or equivalent concentration reported on the
basis of organic carbon.
6. Except for studies on non-ionizing substances
(e.g., PCP), the study reports the pH of the soil,
and the soil pH is within the range 4.0 < soil
pH < 8.5.
7. The study includes at least one control
treatment.
8. The duration of the exposure is reported, or a
standard study method with a defined duration
is used.
9. For studies conducted in a laboratory setting, at
least three treatment levels are used (i.e.,
control plus two chemical exposures).
10. Biological effects are reported for ecologically
relevant endpoints (ERE).
11. Either the test species' scientific name,
common name, variety, or strain is reported.
3.2 Identification of Potentially-
Acceptable Literature
The second step of the process identified
which publications acquired through Step 1
included at least the minimum information
necessary for deriving an Eco-SSL. Eleven
Study Acceptance Criteria (see Figure 3.3)
were applied to each of the acquired
publications. Publications that meet all 11
Study Acceptance Criteria were further
evaluated in Step 3. Detailed descriptions of
the Study Acceptance Criteria are presented
in Attachment 3.1. Approximately seven
percent of the plant and soil invertebrate
publications reviewed by EPA (identified
during Step 1) met all 11 Study Acceptance
Criteria.
3.3 Extraction of Data and Scoring
Studies
In Step 3, each of the publications that met
all 11 Study Acceptance Criteria were
reviewed and study data extracted and
scored. If a publication contained more than
one study, each study was evaluated and
scored separately. For each study reviewed,
a set of critical notes were recorded in a
spreadsheet. A separate "study" was defined
if any of the following parameters varied:
test species (not strain), contaminant (not
form), soil (natural vs. artificial), pH or
percent organic matter (% OM) content.
The specific procedures used to extract and
record study data are provided as
Attachment 3-2.
Guidance for Developing Eco-SSLs
3-3
November 2003
-------
Ecologically Relevant Endpoints
For each of the studies reviewed, the measures of toxic effects to either plants or soil
invertebrates were grouped into one of four ecologically relevant endpoints (EREs) (see Table
3.1). If a study reported toxic effects for multiple EREs, only the preferred ERE and
corresponding contaminant concentration were recorded in the critical notes. For soil
invertebrates, the preferred ERE followed the order: reproduction > population > growth. For
soil invertebrates, reproduction was the preferred ERE because it is necessary for sustaining
populations, and reproductive endpoints are good indicators of longer term (i.e., chronic)
exposure. Although population endpoints were considered less robust than reproductive
endpoints, they were next in preference because screening ecological risk assessments focus on
protection at the population level. Growth measurements on individuals were the traditional
measurement endpoints, and are frequently extrapolated to represent impact at the population
level.
For plants, the preferred ERE was biomass production as it is normally the most sensitive
measurement. If no measurement of biomass production was reported, some physiological
endpoints were accepted as there were established linkages between certain measures of
photosynthesis and productivity (refer to Table 3.1).
Table 3.1 Ecologically Relevant Endpoints (EREs) for Soil Invertebrate Eco-SSLs
Ecologically Relevant Endpoint
Definition
Reproduction
Measures of the effect of toxicants on offspring production. Examples of
EREs associated with reproduction included changes in fecundity, number
of progeny produced (eggs, cocoons, etc.), rate of reproduction (hatching
rates, etc.), rate of maturation, sexual development, change in sex
expression, and sterility number or proportion of abnormal progeny.
Population
Measurements and endpoints regarding a group of soil invertebrates of the
same species occupying the same area at a given time. Measurement
included population dynamics. Examples of EREs associated with
population included changes in size and age class structures, changes in sex
ratio, intrinsic population growth rate, survivability of subsequent
generations, diversity, evenness, index to population size (count, number,
abundance), life table data, population density (number/area).
Growth
Broad category which encompassed measures of weight/mass and length.
EREs associated with growth and development included responses such as a
change in body weight.
Guidance for Developing Eco-SSLs
3-4
November 2003
-------
Table 3.2 Ecologically Relevant Endpoints (EREs) for Plant Eco-SSLs
Ecologically Relevant Endpoint
Definition
Growth (Biomass)
Measurement of plant products including standing crop biomass, seedling
emergence, shoot length/growth, root elongation/growth, fresh or dry mass,
yield or production (e.g., seed production).
Physiology
For the purposes of developing Eco-SSLs, plant studies reporting EREs
associated with physiological responses were used. Physiological endpoints
for plants included net photosynthesis (CO2 uptake, oxygen release),
decrease in chlorophyll content or chlorophyll fluorescence, increased
deformation, membrane damage, desiccation/decrease in water content,
detrimental changes in dormancy measures, decreased flowering, and
increased senescence.
Toxicity Parameters
For each study, a toxicity parameter was recorded. The toxicity parameter was the
concentration-response measurement associating the adverse effect with the contaminant
exposure. Toxicity parameters considered acceptable for deriving Eco-SSLs were the EC20
(effective concentration that affects 20% of the test population), the MATC and the EC10
(effective concentration that affects 10% of the test population). The MATC was equal to the
geometric mean of the NOAEC and the LOAEC. Some toxicity data were not used to derive
Eco-SSLs including acute toxicity data such as the concentration lethal to 50% of the test
population (LC50) or the effective concentrations affecting more than 50% of the test population
(EC50). Effect concentrations affecting less than 5% of the test population (ECX <5) were also
not considered acceptable for deriving Eco-SSLs. The LC50 and EC50 values were not considered
sufficiently protective of ecological resources, while EC5 values have low levels of confidence
due to natural variability.
When NOAEC and LOAEC values were reported, these data were used to calculate an MATC.
A bounded value was a study result with both a NOAEC and LOAEC reported (other than
control). Unbounded NOAEC or LOAEC values do not describe a dose-response curve and thus
were not used due to the high uncertainty over where the real threshold of toxicity lies. If a
study reported multiple toxicity parameters, a single preferred toxicity parameter was selected.
Selection of the preferred toxicity parameter followed the order:
EC20 > MATC > EC10
If a study reported more than one adverse effect concentration with the "preferred" ERE and
toxicity parameter, the lowest effect concentration was then selected. If a publication did not
report an EC20, MATC or EC10, but sufficient data were provided, the reviewer calculated and
recorded the toxicity value under the appropriate toxicity parameter. Toxicity data were reported
as mg/kg of the chemical on the critical notes form (Attachment 3-2). Data not reported in these
units were converted to mg/kg or converted from formulation to active ingredient. When metal
Guidance for Developing Eco-SSLs 3-5 November 2003
-------
concentrations were reported, these were converted to an elemental basis. The only data
extracted were for results where an adverse effect was significantly different from the control.
Scoring Each Study
Each study was scored according to nine specific Study Evaluation Criteria (Table 3.3) and the
results documented in the critical notes. If a single publication contained data for multiple
studies (e.g., reports toxicity data for more than one species or soil type, etc.), each study within
the publication was scored separately. The information recorded for each study included test
species, soil characteristics (e.g., pH, %OM), relative bioavailability score, the ERE, toxicity
parameter, and toxicity value (Attachment 3.2). The specific procedures used to score studies
and document the results are provided as Attachment 3-2.
Scoring for each Study Evaluation Criterion used a three-point scale (i.e., 0, 1, or 2). The
maximum score for a study was 18. Studies were deemed inappropriate for deriving Eco-SSLs
for plants and soil invertebrates if they did not score above ten. Studies with an evaluation score
of 10 or less were presumed to lack sufficient detail about the study to enable reviewers to assess
the quality of the data. These studies results were not included in the pool of data considered for
deriving the Eco-SSL values.
3.4 Derivation of Eco-SSLs
Sorting Data by Bioavailability Score
The first step in deriving an Eco-SSL was to sort the studies accepted at Step 3 (receiving a
Study Evaluation Score greater than ten out of 18 possible points) by their bioavailability score.
The bioavailability score was determined as part of the scoring process in Step 3 (Table 3.3).
This score was based on the soil matrix tables presented in Chapter 2 (Tables 2.4 to 2.6), the type
of soil (natural versus artificial), pH and OM content. A score of two was applied to natural
soils with relatively high or very high bioavailability, a score of one was applied to natural soil
with medium bioavailability and to standard artificial soil, and a score of zero was applied for
natural soil with low or very low relative bioavailability.
Quality Control Review
A quality assurance review of the sorted data was preformed by a panel of experts. The
reviewers verified that all studies used to derive the Eco-SSLs were correctly evaluated and
scored. All studies were reviewed by at least two individuals (other than the original reviewer).
The quality control review provided a forum for confirming that the appropriate data were
identified and documented, resolving any comments or concerns, and ensuring consensus on data
selection. For example, in cases where a study reported data for multiple test species and for
several endpoints, the quality control review provided a forum to reach a consensus that the most
Guidance for Developing Eco-SSLs 3-6 November 2003
-------
Table 3.3 Summary of Nine Study Evaluation Criteria for Plant and Soil Invertebrate Eco-SSLs
Criteria Title
Rationale
Scoring
#1: Testing was Done Under
Conditions of High
Bioavailability
Bioavailability of metals and polar organic compounds is
influenced by pH and soil organic matter, cationic exchange
capacity, and clay content. The scoring is intended to favor
relatively high bioavailability.
Scores were based on the bioavailability matrix
(see Chapter 2). Scored 2 if bioavailability of
natural soil was high or very high. Scored 1 for
natural soil with medium bioavailability or
standard artificial soil. Scored 0 for natural soil
with low and very low bioavailability.
#2A (laboratory) and #2B
(field): Experimental Designs
for Studies are Documented and
Appropriate
Experimental design can significantly influence the quality
of a study. Higher quality studies will use an experimental
design sufficiently robust to allow analysis of the test
variables and discriminate non-treatment effects.
Scored based on experimental design and
methods used for statistical analyses. Scored 2,
1 or 0. Specific criteria used provided in
Attachment 3-2.
#3: Concentration of Test
Substance in Soil is Reported
The concentration of the contaminant tested must be
reported unambiguously.
Scored 2 if measured concentrations were
reported. Scored 1 for nominal concentrations
and scored 0 in all other cases.
#4: Control Responses are
Acceptable
Negative controls are critical to distinguish treatment effects
from non-treatment effects.
Scored 2 if a standardized procedure were used
and control values were within procedural
guidelines or acceptable range (if non-standard
procedure used). Scored 1 if results of control
were not reported or were ambiguous. Score dO
if control results were not within an acceptable
range.
#5: Chronic or Life Cycle Test
was Used
Chronic toxicity tests assessing long-term adverse sub-lethal
impacts on the life-cycle phases of an organism are
considered superior to acute toxicity tests.
Scored 2 if chronic exposures were used.
Scored 1 if acute tests were used. Scored 0 if
very short term exposures were used.
#6: Contaminant Dosing
Procedure is Reported and
Appropriate for Contaminant
and Test
Contaminant dosing procedure may affect the outcome of a
test. Dosing procedure should include: (A) the form of the
contaminant; (B) the carrier or vehicle (e.g., solvent, water,
etc.); (C) how the carrier was dealt with following dosing
(i.e., allowed to volatilize, controls, etc.); (D) procedure for
mixing of soil with contaminant (homogeneity).
Score applied based on how well the study
reports the four contaminant dosing procedures
(A to D). Scored 2 if study reported all. Scored
1 if information for items A and B, but not C or
D; Scored 0 if details were not provided and
could not be inferred.
#7: A Dose-Response
Relationship is Reported or can
be Established from Reported
Data
Two methodologies can be used to identify this benchmark
concentration. The first method generates a no observed
adverse effect concentration (NOAEC) and a lowest
observed adverse effect concentration (LOAEC). The
second method uses a statistical model to calculate a dose
response curve and estimate an effect concentration for
some percentage of the population (ECX), usually between
an EC, and an EC,n.
Scored 2 if an EC10- EC20; or a NOEC and
LOEC were within a factor of 3. Scored 1 if
the difference between the NOEC and LOEC
was > 3x but < lOx. Scored 0 if an ECX was not
reported or the difference between the NOEC
and LOEC was > 10, or only a NOEC or LOEC
was reported.
#8: The Statistical Tests used to
Calculate the Benchmark and
the Level of Significance were
Described
Statistical tests and results reported in the study should be
sufficient to determine the significance of the results.
Scored 2 if ANOVA or statistical method were
based on a P = 0.05; or the 95% CI of the ECX.
Scored 1 if an ANOVA was completed but P
level not provided or > 0.05; or if EC data did
not include the 95% CI or used a 90% CI.
Scored 0 if a NOEC, LOEC, or EC/LCX were
not reported, or were reported without a
description of the method used to calculate the
values.
#9: The Origin of the Test
Organisms is Described
The results of a toxicity test can be influenced by the
condition of the test organisms. Culture conditions should
be maintained such that the organisms are healthy and have
had no exposure above background to contamination prior
to testing (inverts) or detailed information is provided about
the seed stock (plants).
Scored 2 if the source and condition of the test
organisms were known and described.
Scored 1 for a non-commercial source not
adequately described, or if insufficient
information was provided about a commercial
source. Scored 0 if organisms were from a
known contaminated site, or insufficient
information was provided on the a commercial
source.
Guidance for Developing Eco-SSLs
3-7
November 2003
-------
appropriate data were used to derive the Eco-SSL. The quality assurance review consisted of
verifying the following information:
Adherence to all eleven Study Acceptance Criteria
Accuracy of study evaluation scores
Accuracy of soil pH, OM, and relative bioavailability score for each study
Selection of preferred EREs, toxicity parameter(s), and toxicity results
Units of the toxicity result (wet weight/dry weight, contaminant form)
Calculation of the Eco-SSL Values
The Eco-SSL was calculated as the geometric mean of all the toxicity values at the highest
relative bioavailability score for which sufficient data existed (i.e., > three data points). If less
than three data values were available at the highest relative bioavailability level, data from the
next highest bioavailability level were included in that Eco-SSL data set. This process
proceeded until a combined data set of three or more data values were identified for calculating
the Eco-SSL. For example, if there were only two toxicity values from studies using soils with
relatively high bioavailability (i.e., score = two), but there were data from four medium
bioavailability studies (i.e., score = one), then an Eco-SSL was calculated using the combined
six data values. If there were less than three acceptable studies (score > ten) an Eco-SSL was not
calculated.
3.5 Soil Toxicitv Test Methods
If sufficient data for deriving an Eco-SSL were not available from the published literature,
additional plant or soil invertebrate data could be generated through completion of appropriately
designed soil toxicity studies. The Agency recommends that such studies be designed with
consideration of the nine Study Evaluation Criteria (Table 3.3) to generate the highest quality
data to derive an Eco-SSL. For example, the ideal studies would be conducted using natural
soils within the specified soil chemistry conditions, have the highest relative bioavailability (i.e.,
score = two), and satisfy the other evaluation criteria.
Several standardized soil test procedures are available that can be used to generate the data for
deriving Eco-SSLs. Alternative test methods or designs could also be used if they use the
appropriate EREs. Whether standardized or other methods are used, experimental design
options should be selected that best fit the Study Evaluation Criteria outlined in Table 3.3. To
obtain data that best represent natural conditions, investigators are encouraged to incorporate
aging/weathering of contaminated soil into their study design. Examples of standard procedures
that may be used are described in the following subsections and are listed in Table 3.4.
Soil Invertebrate Toxicity Testing
Several internationally standardized soil invertebrate toxicity tests may be used to generate data
for Eco-SSLs. Specifically, three soil toxicity tests are identified as generally appropriate: 1) a
21-day chronic earthworm reproduction (cocoon production) toxicity test (ISO/11268-2:1998),
Guidance for Developing Eco-SSLs 3-8 November 2003
-------
2) the enchytraeid reproduction test (ISO/16387:2001), and 3) the collembolan reproduction test
(ISO/11267:1998) (Table 3.4). These specific tests are recommended as they measure
contaminant toxicity during chronic assays, and include at least one reproductive component
among the measurement endpoints. Guidelines for these methods have been approved by the
International Standards Organization (ISO), and similar efforts are in the final stages of review
or approval by one of several national and international organizations (e.g., the Organization for
Economic Cooperation and Development (OECD), the American Society for Testing and
Materials (ASTM), the European Community, and the Federal Biology Research Cooperative
(FBRC)).
Plant Toxicity Testing
There are relatively few standardized plant toxicity test procedures that EPA expects would
generate data acceptable for deriving Eco-SSLs (Table 3.4). The EPA guidelines for the early
seedling growth and vigor plant test may be used, as well as similar ASTM methods (e.g.,
ASTM E 1963-98). The use of these procedures may require a modified design to best meet the
nine Study Evaluation Criteria.
Table 3.4 Standard Methods Appropriate for use in Generating Data for the Derivation of
Plant or Soil Invertebrate Eco-SSLs
Species
Eisenia fetida
Enchytraeus sp.
Folsomia Candida
Terrestrial Plants
Terrestrial plants
Terrestrial plants
Terrestrial plants
Citation
ISO (International Standardization Organization). 1998. Soil Quality - Effects of Pollutants
on Earthworms (Eisenia fetida) - Part 2: Determination of Effects on Reproduction. ISO
11268-2:1998
ISO (International Standardization Organization) (2001). Soil Quality - Effects of
Pollutants on Enchytraeidae (Enchytraeus sp.) - Determinations of effects on reproduction
and survival. ISO/CD 16387.
ISO (International Standardization Organization) (1999). Soil Quality - Inhibition of
Reproduction of Collembola (Folsomia Candida) by Soil. ISO 1 1267: 1999.
ASTM (American Society for Testing and Materials). 2002. Standard Guide for
Conducting Terrestrial Plant Toxicity Tests. Designation: E 1963-98 Annual Book of
Standards. American Society for Testing and Materials. West Conshohocken, PA.
U.S. EPA (U.S. Environmental Protection Agency). 1996. Ecological effects test
guidelines. OPPTS 850.4230: Early seedling growth toxicity test. Report Number EPA
712-C-96-347. Office of Prevention, Pesticides and Toxic Substances, Washington, DC.
U.S. EPA (U.S. Environmental Protection Agency). 1996. Ecological effects test
guidelines. OPPTS 850.4250. Vegetative vigor, Tier II. Report Number EPA
712-C-96-364. Office of Prevention, Pesticides and Toxic Substances, Washington, DC.
U.S. EPA (U.S. Environmental Protection Agency). 1996. Ecological effects test
guidelines. OPPTS 850. 4150. Terrestrial Plant Toxicity, Tier 1: Vegetative vigor. Report
Number EPA 712-C-96-163. Office of Prevention, Pesticides and Toxic Substances,
Washington, DC.
Guidance for Developing Eco-SSLs
3-9
November 2003
-------
This Page Intentionally Left Blank
-------
4.0 DERIVATION OF WILDLIFE ECO-SSLs
Eco-SSLs for wildlife were derived using a five-part process that included: selecting a wildlife
risk model, selecting surrogate species, estimating an exposure dose, deriving the TRVs, and
calculating the Eco-SSL. Wildlife Eco-SSLs were derived for two groups of receptors: mammals
and birds. Eco-SSLs were not derived for amphibians or reptiles at this time, as described in
Chapter 1.
4.1 The Wildlife Risk Model for Eco-SSLs
The basic equation that was used for estimating potential risks to wildlife was:
Hazard Quotient (HQ) =
Exposure Dose (mg / kg bw / day)
Toxicity Reference Value (mg / kg bw / day)
Contaminant exposure for terrestrial wildlife was expressed as an Exposure Dose in mg/kg
bw/day, and the Toxicity Reference Value (TRV) was expressed in the same units.
2.
The Eco-SSL is the soil concentration that
results in an HQ=1, that is, when the TRV
and the Exposure Dose are equal. The
Exposure Dose for wildlife was equal to the
amount of contaminant in the diet that is
taken up or transferred from the soil.
Therefore, it was necessary to model the
contaminant specific soil concentration that
would result in a dietary concentration equal
to the Exposure Dose that was equal to the
TRV. Estimation of the Exposure Dose is
described in Section 4.3. Derivation of the
TRV is described in Section 4.4. Calculation
of the Eco-SSLs to solve for an HQ = 1 is
described in Section 4.5. The full HQ
equation is provided in Figure 4.1. In this
model, incidental oral soil exposure was
added to the total dietary (food-based)
exposure, making the total oral exposure
greater than 100%. This equation also
included terms for the absorbed fraction (AF) of the contaminant from soil and the diet as well as
an area use factor (AUF) (representing the fraction of time an animal would be exposed). For
the purposes of establishing Eco-SSLs, which are conservative screening values, these terms (AF
and AUF) were set equal to one.
4.
Steps for Establishing a Wildlife Eco-SSL
Identify the Wildlife Risk Model - Equation
relates the contaminant soil concentration to an
acceptable threshold based on a food-chain exposure
model.
Select Surrogate Wildlife Species - Specific
indicator species were identified for parameterization
of the exposure model.
Estimate Exposure Dose - Parameterization of the
exposure dose model for the estimation of exposure
doses for each contaminant.
Derive the Effects Dose or TRV- Identification of
an acceptable dose.
Calculate the Eco-SSL - Calculation of the Eco-
SSLs by solving equation for an HQ=1.
Guidance for Developing Eco-SSLs
4- 1
Revised February 2005
-------
Figure 4.1. The Wildlife Risk Model for Eco-SSLs (Equation 4-1)
N
[Soil * Ps * FIR * AFJ+ [ z Bti* Pt* FIR * AFJ \ *AUF
1=1 '
1W;
where:
//gj = Hazard quotient for contaminant (j) (unitless),
Soilj = Concentration of contaminant (j) in soil (mg/kg dry weight),
jV = Number of different biota types in diet,
Bi} = Concentration of contaminant (j) in biota type (i) (mg/kg dry weight),
Pt = Proportion of biota type (i) in diet,
FIR = Food ingestion rate (kg food [dry weight]/ kg BW [wet weight] /day),
AFi} = Absorbed fraction of contaminant (j) from biota type (i) (for screening purposes set equal to 1),
AFS] = Absorbed fraction of contaminant (j) from soil (s) (for screening purposes set equal to 1),
TRVj = Toxicity reference value (mg/kg BW/day) (Section 4.4),
Ps = Soil ingestion as proportion of diet,
AUF = Area use factor (for screening purposes set equal to 1).
4.2 Selection of Surrogate Wildlife Species
It was neither feasible nor necessary to derive an Eco-SSL for each and every wildlife species
potentially present at a hazardous waste site; therefore, surrogate species were used to derive
wildlife Eco-SSLs. In this approach, specific species were selected as "representatives" for
other species within the same class (mammalian or avian) with similar diets. The advantages of
focusing Eco-SSLs on representatives within generic trophic groups included, but were not
limited to, the following:
This approach provided generic screening values that could be applied to any site,
regardless of the presence or absence of a particular species. The trophic groups
selected were expected to be present or potentially present at all sites across the
nation.
This approach provided results that could be used to examine comparative risks
associated with different exposure routes (e.g., ingestion of food versus ingestion
of soil) representing different contaminant transport pathways (e.g., soil to
herbivore, soil to ground insectivore, soil to soil invertebrate, and soil to plant)
versus direct soil ingestion.
This approach was consistent with ERAGS which states: "for the screening-level
ERA, assessment endpoints are any adverse effects on ecological receptors, where
receptors are plant and animal populations and communities, habitats, and
sensitive environments." (p. 1-7; U.S. EPA, 1997)
Guidance for Developing Eco-SSLs 4-2 Revised February 2005
-------
Criteria for Selection of Surrogate Taxa
Three general trophic groups (e.g., herbivore, ground insectivore, and carnivore) for both
mammals and birds were used for the Eco-SSL wildlife exposure model. Within each of these
trophic groups, a specific species was identified as a "surrogate" species. Surrogate species were
selected to provide a conservative representation of their respective trophic groups. Selected
species were generally small in size relative to other species within their respective trophic
groups (e.g., weasels and voles vs. foxes and coyotes or rabbits and deer). Because small size
was considered to be associated with higher metabolic rates (Nagy et al., 1999) and smaller
home ranges (McNab, 1963), exposure for small receptors was assumed to be high. Eco-SSLs
based on these species were therefore likely to be protective of other, larger species in their
trophic group.
Selection of specific surrogate species was necessary for parameterizing the Eco-SSL wildlife
model, which required estimates of body weights, food ingestion rates, and soil ingestion rates.
The following criteria were used to guide the selection of surrogate species for each trophic
group:
1) Exposure pathway link to soil. Each surrogate species had to have a clear direct or
indirect exposure pathway link to soil. Direct exposure pathways to soil included
ingestion of soil dwelling biota (e.g., plants or soil invertebrates) and incidental ingestion
of soil as a result of foraging at the soil surface (as opposed to from plants). Species with
direct exposure pathways to soil were assumed to be the most highly exposed to soil
contamination with the exception of contaminants that biomagnify. Indirect exposure
included ingestion by carnivores of prey that have direct contact with soil.
2) Diet Composition. The selected,
surrogate species must forage in
terrestrial, upland habitats. This
criteria ensured that only potential
exposures related to soil
contamination were considered and
consumption of aquatic prey items
(exposures to the aquatic
environment) were not considered.
3) Diet composition can be
simplistically classified. The
dietary composition of each
surrogate species had to be easily
classified into one of the three
selected trophic groups (herbivore,
ground insectivore, carnivore).
Clear classification of diet served to
What Wildlife Groups are not Considered
Appropriate for Eco-SSLs?
Some specific wildlife groups were not considered suitable
for deriving wildlife Eco-SSLs. These groups include:
Generalist species (e.g., raccoons, jays) were
excluded due to difficulty in defining diet and,
therefore, exposure. These species forage
opportunistically and are likely to consume different
foods in different parts of their range.
Piscivores (e.g., herons, otter) were excluded due to
the lack of a direct exposure pathway to soil.
Aerial Insectivores (e.g., swallows) and Arboreal
Insectivores (e.g., warblers) were excluded as they do
not forage primarily from terrestrial environments.
Guidance for Developing Eco-SSLs
4-3
Revised February 2005
-------
simplify the exposure assumptions related to dietary composition into three classes:
plants, invertebrates and animals. Further, the dietary composition for the surrogate
species needed to be realistically assumed to consist of a single food type. This
assumption allowed for the evaluation of the potential maximum exposure and risk from
that dietary pathway. Evaluation of the maximum risk that may be presented by a given
pathway (i.e., plants, invertebrates, or vertebrates) produced results that are protective of
species with more varied diets. Omnivorous species were assumed to be more likely to
consume foods with differing contaminant concentrations. As a result, their total
exposure would be less than that for species whose diets that consisted of the single most
contaminated food type. By selecting surrogate species that foraged exclusively on a
specific food type (plants, invertebrates, or vertebrates), maximum risks were expressed
for any given contaminant. This helped to ensure protectiveness of all other species.
4) Mammalian and avian species identified. Because toxic responses for the same
contaminant could differ among wildlife taxa, surrogate species were selected for both
mammals and birds. Based upon the above factors, three mammalian and three avian
species (listed in Table 4.1) were selected to represent some of the most highly exposed
species. It was assumed that the use of these six species also protected other herbivores,
ground insectivores, and carnivores.
4.3 The Exposure Dose
Estimation of the exposure dose associated with contaminant concentrations in soil required
parameterization of the general model provided as Equation 4-1 (see Figure 4.1).
Reduced Wildlife Risk Model for Screening
The Eco-SSLs are intended to be conservative screening values that can be used to eliminate
contaminants clearly not associated with unacceptable risks. Therefore, several simplifying,
conservative assumptions were made in the parameterization of the general wildlife risk model
(Equation 4-1). These assumptions included:
Surrogate species were assumed to reside and forage exclusively on and within
the contaminated site. Therefore, the area use factor (AUF) was set equal to 1.
Bioavailability of the contaminant in both soil and food was assumed to be
comparable to the bioavailability of the contaminant in the laboratory studies used
to establish the TRVs. Therefore, the absorbed fraction from soil (AFsj) and
absorbed fraction from biota type 'i' (AFy) were both equal to 1.
The surrogate species' diet consisted of just one food type. Therefore, the
proportion of biota type 'i' in the diet (P;) was equal to 1 and the number of biota
types (N) in the diet was equal to 1.
Guidance for Developing Eco-SSLs 4-4 Revised February 2005
-------
Parameterizing the Model for Estimating the Exposure Dose
Parameterization of the model included using exposure factors related to the surrogate species
(see Table 4.1) and estimating the contaminant concentrations in biota items (By) consumed in
the diet. The exposure factors identified and derived for surrogate species-specific exposure
factors are described in Attachment 4-1. The food and soil ingestion rates used in the exposure
model were represented by high-end values. For the food ingestion rate a high end point
estimate was selected based on measured data. For soil ingestion, the 90th percentile from the
distribution was selected. Use of exposure parameter values from the upper tails of the
distributions ensures the protectiveness of the Eco-SSLs for other wildlife species.
Table 4.1. Parameterization of the Eco-SSL Wildlife Exposure Model
Receptor Group
(Surrogate Species)
Mammalian Herbivore
(Meadow Vole)
Mammalian Ground Insectivore
(Short- tailed shrew)
Mammalian Carnivore
(Long-tailed weasel)
Avian Grainivore
(Mourning dove)
Avian Ground Insectivore
(American woodcock)
Avian Carnivore
(Red-tailed hawk)
Food Ingestion Rate
(kg dw/kg bw day) 1
0.0875
0.209
0.130
0.190
0.214
0.0353
Soil Ingestion
(py
0.032
0.030
0.043
0.139
0.164
0.057
Assumed Diet
100% foliage
100% earthworms
100% small mammals
100% seeds
100% earthworms
100% small mammals
that consume 100%
earthworms
Parameterization details provided in Attachment 4-1 .
1 High end point estimate based on measured data. Derivation presented in Attachment 4-1 .
2 Soil ingestion as proportion of diet. 90th percentile. Derivation presented in Attachment 4-1 .
dw = dry weight
Estimating Contaminant Concentrations in Biota (Dietary Items)
The contaminant concentrations in biota types (By) composing the wildlife diets (plants,
earthworms, or small mammals) were estimated by assuming that the concentration of the
contaminant 'j' in the food type 'i' could be predicted from the concentration of the contaminant
in the soil (Soilj).
Guidance for Developing Eco-SSLs
4-5
Revised February 2005
-------
The concentration of contaminant (j) in biota or food type type (i) (By) is related to the
concentration in soil (Soilj) by an uptake model which has one of the following forms:
Case 1) By = BAF;j * Soilj (constant)
Case 2) ln(B;j) = I;j + S;j * In(Soilj) (log-linear)
Case 3) B;j = I;j + S;j * Soilj (linear)
where:
By = Concentration of contaminant (j) in food type (i) (where i = plant,
earthworm, or small mammal)
BAFy = Soil-to-biota bioaccumulation factor (BAF) for contaminant (j) for biota
type (i)
Iy = Intercept from bioaccumulation model for contaminant (j) for food type (i)
Sy = Slope from bioaccumulation model for contaminant (j) for food type (i)
Soilj = Concentration of contaminant (j) in soil
For the predatory surrogate species that are assumed to consume 100% small mammals, BAFs or
regression equations were not available for all contaminants relating soil concentration to
concentration in small mammal tissue. In these instances, it was necessary to estimate small
mammal tissue concentrations (By) from dietary-based BAFs or regressions:
Case 4) B, = Cdiet * BAFdm
Case 5) ln(B,) = I;j + S;j * In^,)
Case 6) B;j = I;j + S;j * Cdiet
where:
By = Concentration of contaminant (j) in food type (i) (where i = small
mammal)
Cdiet = Concentration of contaminant (j) in diet where diet is 100% earthworms
estimated as in Case 1, 2 or 3, above
BAFdm = Diet-to-biota BAF for contaminant (j) in mammal or bird tissue
A hierarchy was established for decision-making concerning the use of available data to estimate
contaminant concentrations in biota types (By). The following values were used in order of
preference:
1) Existing Regression (R). If regression equations were available from the literature and
the r-square values were > 0.2, then these were preferentially used. The primary sources
of existing regression equations were: Sample et al. (1999) for earthworms; Sample et al.
(1998a) for small mammals; and Bechtel-Jacobs (1998) for plants.
2) New Regression (R). If regressions were not available from the literature, then paired
data (contaminant concentration in soil versus concentration in plant foliage;
concentration in soil versus soil invertebrate; or concentration in diet versus mammal or
Guidance for Developing Eco-SSLs 4-6 Revised February 2005
-------
bird tissue) were identified from the literature. If the paired data were sufficient to
establish a regression and the regression was significant with r-square values > 0.2, then
the regression was used to estimate uptake.
3) Ratio (BAF). If regressions were not significant then BAFs (or ratios of the contaminant
in soil to the contaminant in the food item) were used to estimate uptake. BAFs were
identified either from the literature or from the paired data compiled in the attempt to
establish a regression in Step 2. The median BAF value was selected to estimate uptake.
4) Modeled BAFs or B;j (M for
modeled). If BAFs were not
available in the literature or
the paired data were not
available to derive the BAF,
then By was estimated using
relationships established
between physical or chemical
parameters of the contaminant
'j' and By. Existing models
from the literature were used
where possible. In some
cases, new models were
derived. The models are
provided in Attachment 4-1.
5) Assumptions (A). In
instances where data were not
available to complete any of
the previously listed options in
the hierarchy (1 to 4), then it
was necessary to make
assumptions concerning the
bioaccumulation of
contaminants for soil into biota
types. These assumptions are
discussed in following
subsections.
Figure 4.2 Summary of Method Used for Estimation of
Contaminant Concentrations in Biota Types (B,)
COC
Antimony
Arsenic
Barium
Beryllium
Cadmium
Chromium
Cobalt
Copper
Lead
Manganese
Nickel
Selenium
Silver
Vanadium
Zinc
Dieldrin
DDT
ODD
DDE
PCP
PAHs
TNT
RDX
Soil to
Plant
R
BAF
BAF
R
R
BAF
BAF
R
R
BAF
R
R
BAF
BAF
R
BAF
R
R
R
R
R or BAF
BAF
BAF
Soil to
Earthworm
A
R
BAF
BAF
R
BAF
BAF
BAF
R
R
BAF
R
BAF
BAF
R
R
R
R
R
R
M
TBD
TBD
Diet to Soil to
Mammal Mammal
BAF
NA
BAF
BAF
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
R
R
R
R
R
A
A
A
NA
R
NA
NA
R
R
R
R
R
R
R
R
BAF
BAF
R
NA
NA
NA
NA
NA
A
A
A
A = Assumption .
BAF = Bioaccumulation Factor.
M = Estimated based on equation relating physical-chemical factor to
bioaccumulation (model).
NA = Not applicable. Only one method is needed for estimating tissue
concentrations in mammals based on either diet or soil concentrations.
R = Either existing or new regression equation (Attachment 4-1)
TBD = To be determined in the chemical specific Eco-SSL document.
Figure 4.2 summarizes the method (based on the hierarchy) used for estimating contaminant
concentrations in biota types (dietary items) for each of the Eco-SSL contaminants. The
following subsections describe how contaminant concentrations were estimated in each of the
dietary items including plants, soil invertebrates and small mammals.
Guidance for Developing Eco-SSLs
4-7
Revised February 2005
-------
How Contaminant Concentrations Are Estimated for Plants and Soil Invertebrates as Dietary
Items
The specific information concerning how contaminant concentrations were estimated for the
plant, earthworm and small mammal biota types (By) of the diets of the surrogate species is
provided as Attachment 4-1. Regressions (R), BAFs, models (M) and assumptions (A) were all
used (Figure 4.2).
The existing models for plants produce estimated contaminant concentrations in foliage.
Although these data are suitable for receptors that consume foliage (e.g., meadow vole), their
suitability for receptors whose diets consist of seeds (e.g., mourning dove) is unknown. To
address this issue, a literature search was performed to locate studies in which contaminant
concentrations in seeds and in soil were evaluated. The few data located (limited to metals) were
plotted over reported soil-to-plant foliage bioaccumulation data (Figure 4.3).
0)
1000
100
10
1
0.1
Cadmium
o ...-'o
ฃ
"c 0.01
"5.
c 0.001
0.001 0.01
C 10000
o
tS 1000
ง 100
o
i 10
o
1
0.1
0.01
0.001
Lead
0.1
O Foliage
95% prediction
limit-foliage
Seeds
1000
100
10
0.1
Copper
0.1 1 10 100 1000 01
100000.0
10000.0
1000.0
100.0
10.0
1.0
0.1
10 100
1000
10000
Zinc
10 100 1000 10000 1 10 100 1000
Concentration in soil (mg/kg dry wt.)
10000
100000
Figure 4.3 Comparison of Soil-to-Foilage and Soil-to-Seed Bioaccumulation
for Cadmium, Copper, Lead and Zinc
Guidance for Developing Eco-SSLs
4-8
Revised February 2005
-------
For cadmium, copper, lead, and zinc soil-to-seed bioaccumulation mirrors the pattern observed
for soil-to-foliage bioaccumulation. The degree of spread in the seed bioaccumulation data was
comparable to that for foliage data. Based on these data, the models estimating contaminant
concentrations in foliage were assumed to be suitable surrogates (at least for metals) for
estimating contaminant concentrations in seeds.
How Contaminant Concentrations Are Estimated for Mammals as Dietary Items
Empirical soil-to-whole body log-linear regression models and BAFs were available from
Sample et al. (1998a) for 11 of the 24 contaminants. For the remaining contaminants for which
empirical regression models or BAFs were not available, bioaccumulation was estimated using
the methods presented in Attachment 4-1.
Although many species of predatory wildlife consume both birds and mammals as prey, few data
were available to estimate bioaccumulation of contaminants into birds. As a consequence, the
bioaccumulation models for mammals were assumed to produce estimates that adequately
represent concentrations in birds. The validity of this assumption was supported by data
presented in Beyer et al. (1985). Birds (representing multiple species), white-footed mice, and
short-tailed shrews were collected from two locations in the vicinity of a zinc smelter in
Pennsylvania. Analyses were available for carcasses (tissue remaining after removal of the
gastrointestinal tract, skin, feet, and beaks) for lead, zinc, cadmium, and copper. Mean analyte
concentrations (and 95% confidence limits) in birds and mammals from both locations are
presented in Figure 4.4. Based on these data, concentrations in birds appear to be approximately
equivalent to or less than those found in omnivorous or insectivorous small mammals.
What if Data are not Available to Estimate Concentrations in Dietary Items?
For some contaminants and biota types (i.e., earthworms for antimony), data were not available
to derive BAFs (as described in Attachment 4-1). For these contaminants, an assumption or
default BAF of 1 was used. This assumption was supported by analyses of BAFs for plants,
earthworms, and small mammals from Bechtel Jacobs (1998), Sample et al. (1998b), and Sample
et al. (1998a), respectively (refer to Table 4.2). Approximately 90% of plant BAFs, 68% of
earthworm BAFs and 83% of small mammal BAFs were less than one. Assuming BAFs for
inorganic contaminants when data are absent generally over-estimate uptake and are, therefore,
suitable for screening purposes only. A default BAF of 1 was also used for organic compounds.
Guidance for Developing Eco-SSLs 4-9 Revised February 2005
-------
ฃ1 1 UUU
o
&
~s>
E_
ซ 100-
8
re
O
c
| 1ฐ"
2
c
0)
u
c
o 1
O '
">^ 100 -i
o) :
1
TO :
o :
3
c
1 1l
'a :
c
0)
o
c
/\ n 1 -
Lead
f
,,
* f f
Birdsl Birds2 Micel Mice2 Shrewl Shrew2
Cadmium
i i
i * *
IUUU -
100 :
14 -
12 -
10-
8-
6 -
A -
Zinc
i I
f * !
Birdsl Birds2 Micel Mice2 Shrewl Shrew2
Copper T
T
I "
M * '
Birdsl Birds2 Micel Mice2 Shrewl Shrew2
Birdsl Birds2 Micel Mice2 Shrewl Shrew2
Figure 4.4 Comparison of Mean Concentrations in Multiple Species near a Smelter
Table 4.2 Cases where the Median of the BAF Distribution for Metals
is Greater or Less than One
Biota Type
Plants
Earthworms
Small Mammals
Total Number of
Contaminants
21
31
24
BAFs < 1
19
21
20
BAFs > 1
2
10
4
Guidance for Developing Eco-SSLs
4- 10
Revised February 2005
-------
4.4 Derivation of Toxicity Reference Values (TRVs)
The wildlife TRV is defined as:
Dose above which ecologically relevant effects might occur to wildlife species
following chronic dietary exposure and below which it is reasonably expected
that such effects will not occur.
As presented in Figure 4.5, a four-step process was used to select TRVs appropriate for
calculating wildlife Eco-SSLs. The four steps included: 1) conduct a literature search, 2)
complete a review of the literature and extract data, 3) complete an evaluation of the extracted
data and score data, and 4) derive a TRV.
Figure 4.5. Steps of the Wildlife TRV Derivation
Process
The wildlife TRV derivation process is composed of four
general steps that are documented as separate standard operating
procedures (SOPs):
Literature Search and Retrieval
Wildlife TRV Literature Search and Retrieval (Attachment
4-2). A literature search identifies dose-response literature
for retrieval.
Literature Review and Data Extraction
Wildlife TRV Literature Review, Data Extraction and
Coding (Attachment 4-3). The retrieved literature studies
are reviewed and data are extracted according to an
established coding system. Data are entered into an
electronic data base.
Data Evaluation
Wildlife TRV Data Evaluation (Attachment 4-4). Each of
the results identified in the reviewed literature is scored for
quality and applicability for TRV derivation.
TRV Derivation
Wildlife TRV Derivation (Attachment 4-5). This procedure
plots the collective dose-response information and
establishes the process for estimating the TRV.
Literature Search and Retrieval
The first step in deriving the TRVs was to
conduct a literature search to identify
toxicological studies for mammals and
birds for retrieval and review. The search
procedure was described in detail in
Attachment 4-2. EPA completed a
literature search for 23 of the Eco-SSL
contaminants (PCBs were excluded). The
search process identified over 44,000
records for review. The literature search
process was documented in sufficient
detail that others may use it to identify
relevant data for additional contaminants
not completed by EPA.
Literature Review and Data Extraction
The toxicological literature identified
from the literature search was next
reviewed for usefulness in establishing
wildlife TRVs. Literature exclusion
criteria (similar to those discussed in
Chapter 3 for plants and soil
invertebrates) and listed in detail in Attachment 4-3 were applied to the identified literature.
Additional types of studies excluded specifically for wildlife included those that only report
genotoxic, mutagenic, or carcinogenic effects, acute or non-oral exposures (inhalation, injection,
dermal, etc.), or studies unrelated to the contaminant and receptor groups of interest. Where
possible, the literature exclusion criteria were applied to identified titles and abstracts prior to
retrieval of the paper. For retrieved studies that passed the literature exclusion criteria, the
Guidance for Developing Eco-SSLs
4- 11
Revised February 2005
-------
Figure 4.6 Ten Attributes Scored as Part of
the Wildlife Toxicological Data Evaluation
1.
2.
relevant toxicological data were extracted and
entered into an electronic database according to
established extraction and coding procedures
detailed as Attachment 4-3. These extraction and
coding procedures were consistent with EPA's
EcoTox database (USEPA, 2003).
The primary purpose of the data extraction process
was to identify two values associated with each
study result:
A no-observed adverse effect level
(NOAEL), which is the highest dose that
does not cause a statistically significant
adverse effect; and
A lowest-observed adverse effect level
(LOAEL), which is the lowest dose that
caused a statistically significant adverse
effect.
In theory, the threshold for the particular adverse
effect lies between the NOAEL and the LOAEL.
Data Evaluation
Each test result extracted during the literature
review process was scored for quality and
applicability for TRY derivation. The data
evaluation process was provided as Attachment
4-4. In instances where more than one
"experiment" (i.e., different combinations of test
organism (species or strains), contaminant form,
test location, control type, doses, application
frequency, or route of exposure) were reported in a
paper, the individual "experiments" were scored
separately. In cases of more than one experiment,
the scoring system was applied independently to
each experimental result. The scoring system was
based on the evaluation often attributes of the toxicological study (Figure 4.6) assigning a score
for each attribute, ranging from zero (no merit in setting a TRV) to 10 (extremely valuable and
relevant to setting a TRV). Note that a low score does not necessarily imply the study itself is
poor, only that the study design is not the most appropriate for deriving an oral TRV. The total
score was calculated by adding the results of the evaluation of each attribute. The use of total
7.
10.
Data Source
Primary sources only considered.
Dose Route
Dietary studies scored higher than gavage,
capsule and liquid. Non oral exposures were
excluded.
Test Substance Concentrations
Studies with measured exposures scored higher
than nominal exposures.
Contaminant Form
Contaminant forms similar to soil forms scored
higher compared to dissimilar forms.
Dose Quantification
Exposures reported as doses scored higher than
those reported as concentrations.
Endpoint
Reproductive effects scored higher than
lethality and growth. Physiology, behavioral,
biochemical and pathology changes were scored
lower and biomarkers scored lowest.
Dose Range
Studies with both NOAEL and LOAEL values
scored higher than studies which reported only
one value. Narrower ranges between NOAEL
and LOAEL scored higher.
Statistical Power
The statistical power of a NOAEL was scored.
Exposure Duration
Exposure durations encompassing multiple
generations and critical life stages scored higher
than chronic, subchronic, and acute.
Test Conditions
Studies that reported standard exposure
conditions scored higher then those that
reported fewer or none.
Guidance for Developing Eco-SSLs
4- 12
Revised February 2005
-------
Data Evaluation Score was interpreted as follows:
80 to 100 High confidence
71 to 79 Medium confidence
66 to 70 Low confidence
0 to 65 Not used in Eco-SSL derivation
A web-based data entry system and database was created by EPA as a tool to facilitate efficient,
accurate, and consistent data extraction from individual reviewed toxicological studies as well as
data evaluation scoring. Extraction of the data directly into an electronic database facilitated
necessary sorting, searching, and presentation of the data for the purposes of TRV derivation.
The TRV database was focused on extracting the NOAEL and LOAEL doses from each of the
toxicological studies.
TR V Derivation
The dose-response information for mammals and birds was plotted separately, and a TRV was
derived for each class using an established procedure. The process used was documented in
Attachment 4-5. The following general steps were completed to derive the TRVs:
Data Sorted. The toxicity data (effect doses) were downloaded from the database into
spreadsheet files for each contaminant using a consistent tabular format. One table was
constructed for avian data and a second for mammalian data. The tables provided the
essential information concerning each of the toxicity testing results. Table 4.4 provides
an example using the results for mammals and cobalt. The results were numbered
sequentially and sorted by general effect group, then by effect measure.
Data Plotted. Summary plots were constructed depicting the NOAELs and LOAELs for
each contaminant. Separate plots were completed for mammalian and avian data. The
data plots (example provided as Figure 4.7) were organized by General Effect Group in
order from left to right as: biochemical (BIO), behavior (BEH), physiology (PHY),
pathology (PTH), reproduction (REP), growth (GRO), and mortality (MOR). Within
each toxicological study there may be several effect measures reported that have the same
NOAEL and/or LOAEL values. Inclusion of the NOAEL and LOAEL values for all
endpoint measures would result in repetitive values on the plots. To avoid the inclusion
of repetitive and duplicative data, the results for only one Effect Measure per Effect
Group were recorded on the plots. For example, a study may report more the one adverse
reproductive effect such as reduced progeny weight as well as progeny survival and
reduced number of litters. In this case, only one effect was coded for the most
conservative "worst case" result.
Exclusion of Data with Limited Utility in Establishing an Eco-SSL Each NOAEL
and LOAEL result was evaluated according to the Data Evaluation process (Attachment
Guidance for Developing Eco-SSLs 4-13 Revised February 2005
-------
Table 4.3. Example of Extracted and Sorted Data for Wildlife TRY Derivation (Cobalt)
Result #
Reference
Ref
No.
Test Organism
# of Cone/ Doses
Method of Analyses
Route of Exposure
Exposure Duration
Duration Units
ex
Age Units
Lifestage
X
General Effect Group
Effect Measure
Response Site
NOAEL Dose
(mg/kg/day)
LOAEL Dose
(mg/kg/day)
Data Evaluation Score
Biochemical Effects
1
2
3
4
Maroetal., 1980
Chettyetal., 1979
Kadiiska et al., 1985
Derretal., 1970
171
115
19290
129
Cow (Bos taurus )
Rat (Rattus norvegicus )
Rat (R. norvegicus )
Rat (R. norvegicus )
2
6
2
2
M
U
U
U
FD
FD
DR
DR
45
4
30
35
d
w
d
d
7
NR
NR
NR
mo
NR
NR
NR
JV
NR
JV
JV
F
B
M
M
BIO
BIO
BIO
BIO
HMGL
HMGL
P450
HMCT
BL
BL
LI
BL
0.30
19
29
20
118
70
75
69
65
Behavioral Effects
5
6
7
Gershbein et al., 1983
Huck and Claws on, 1976
Bourgetal., 1985
136
86
111
Rat (R. norvegicus )
Pig (Sus scrofa)
Rat (R. norvegicus )
2
4
2
U
U
M
FD
FD
DR
80
28
57
d
d
d
44
NR
80
d
NR
d
JV
NR
JV
M
NR
M
BEH
BEH
BEH
NMVM
FCNS
ACTP
WO
WO
wo
1.5
7.1
20
66
69
77
Physiology Effects
8
Hagaetal., 1996
105 |Rat (R. norvegicus )
2
U
FD
16
w
NR
NR
NR
M
PHY
Other
HE
80
.0
77
Pathology Effects
9
10
11
12
13
14
Gershbein et al., 1983
Chettyetal., 1979
Hagaetal., 1996
Van Vleet et al., 1981
Seidenberg et al., 1986
Derretal., 1970
136
116
105
149
113
129
Rat (R. norvegicus )
Rat (R. norvegicus )
Rat (R. norvegicus )
Pig (S. scrofa)
Mouse (Mus musculus)
Rat (R. norvegicus )
2
6
2
2
2
2
U
U
U
U
U
U
FD
FD
FD
FD
GV
DR
80
4
16
10
5
35
d
w
w
w
d
d
44
NR
NR
NR
NR
NR
d
NR
NR
NR
NR
NR
JV
NR
NR
JV
GE
JV
M
B
M
M
F
M
PTH
PTH
PTH
PTH
PTH
PTH
GHIS
SMIX
BDWT
GLSN
BDWT
SMIX
NR
TS
WO
HE
WO
HE
1.5
4.8
9.6
8.8
19
82
118
73
78
77
73
75
67
Reproductive Effects
15
16
17
18
19
20
21
22
23
24
Nation et al., 1983
Domingo et al., 1985
Paternain et al., 1988
Seidenberg et al., 1986
Pedigoetal., 1988
Anderson et al., 1992
Corner et al., 1985
Mollenhauer et al., 1985
Anderson et al., 1993
Pedigoetal., 1993
126
124
109
113
121
120
123
119
139
187
Rat (R. norvegicus )
Rat (R. norvegicus )
Rat (R. norvegicus )
Mouse (M. musculus)
Mouse (M. musculus)
Mouse (M. musculus)
Rat (R. norvegicus )
Rat (R. norvegicus )
Mouse (M. musculus)
Mouse (M. musculus)
3
4
4
2
4
2
2
2
2
2
U
U
U
U
U
U
U
U
U
U
FD
GV
GV
GV
DR
DR
FD
FD
DR
DR
69
28
9
5
13
9
70
98
13
10
d
d
d
d
w
w
d
d
w
w
80
NR
NR
NR
10
12
100
100
12
8 to 10
d
NR
NR
NR
w
w
d
d
w
w
MA
MA
GE
GE
MA
MA
SM
MA
MA
JV
M
F
F
F
B
M
M
M
M
M
REP
REP
REP
REP
REP
REP
REP
REP
REP
REP
TEWT
PRWT
PRWT
PROG
RSUC
TEWT
TEDG
TEWT
TEWT
PRFM
TE
WO
WO
WO
WO
TE
TE
TE
TE
WO
5.0
5.4
25
82
20
11
10
14
20
24
43
56
83
87
81
72
78
73
77
73
78
73
Growth Effects
25
26
27
28
29
30
31
32
33
34
35
Maroetal., 1980
Gershbein et al., 1983
Huck and Claws on, 1976
Pedigoetal., 1988
Mohiuddin et al., 1970
Bourgetal., 1985
Chettyetal., 1979
Paternain et al., 1988
Van Vleet et al., 1981
Anderson et al., 1993
Derretal., 1970
171
136
86
121
132
111
116
109
149
139
129
Cow (Bos taurus )
Rat (R. norvegicus )
Pig (S. scrofa)
Mouse (M. musculus)
Guinea pig (Caviaporcellus)
Rat (R. norvegicus )
Rat (R. norvegicus )
Rat (R. norvegicus )
Pig (Sus scrofa)
Mouse (M. musculus)
Rat (R. norvegicus )
2
2
4
4
2
2
6
4
2
2
2
M
U
U
U
U
M
U
U
U
U
U
FD
FD
FD
DR
OR
DR
FD
GV
FD
DR
DR
45
80
16
5
5
57
4
9
5
13
24
d
d
w
w
w
d
w
d
w
w
d
7
44
NR
6 to 7
NR
80
NR
NR
NR
12
NR
mo
d
NR
w
NR
d
NR
NR
NR
w
NR
JV
JV
NR
SM
MA
JV
NR
GE
JV
MA
JV
F
M
NR
M
M
M
B
F
M
M
M
GRO
GRO
GRO
GRO
GRO
GRO
GRO
GRO
GRO
GRO
GRO
BDWT
BDWT
BDWT
BDWT
BDWT
BDWT
BDWT
BDWT
BDWT
BDWT
BDWT
WO
WO
WO
WO
WO
WO
WO
WO
WO
WO
WO
0.30
1.5
2.4
19
20
20
33
0.96
6.2
20
43
122
77
68
74
82
72
72
77
79
77
76
72
Mortality Data
36
37
38
Van Vleet et al., 1981
Seidenberg et al., 1986
Mohiuddin et al., 1970
149
113
132
Pig (S. scrofa)
Mouse (M. musculus)
Guinea pig (Caviaporcellus)
2
2
2
U
U
U
FD
GV
OR
10
5
5
w
d
w
NR
NR
NR
NR
NR
NR
JV
GE
MA
M
F
M
MOR
MOR
MOR
MORT
MORT
SURV
NR
NR
WO
19
82
20
78
80
73
ACTP = activity level; B = both; BIO = biochemical; BL = blood; d = days; BDWT = body weight changes; BEH = behavior; DR = Drinking water; F =
female; FCNS = food consumption; FD = food; GE = gestation; GRO = growth; GLSN = gross lesions; GV = gavage; HE = heart; HMCT = hematocrit;
HMGL = hemoglobin; JV = juvenile; LI = liver; MA = mature; M = male; M = measured; m = months; MOR = mortality, NMVM = number of movements;
NR = Not reported; OR = other oral; PHY = physiology; PTH = pathology; PRFM = sexual performance; REP = reproduction; SM = sexually mature;
SMIX = weight relative to body weight; SURV = survival; TE = testes; TEDG = testes degeneration; TEWT = testes weight; U = unmeasured; w = weeks;
WO = whole organism
Guidance for Developing Eco-SSLs
4-14
Revised February 2005
-------
Figure 4.7 Example of a Toxicological Plot for the TRV Derivation Process (Cobalt)
o/kgBW/day)
C
5 C
0
Q
1 j
1 j
i ~T ' 75 Q;|
^ " "1 '
2 i i ! ^5^
3> ro ' x^^1 ' 77 ; 'ฐ 77 ^ !
^ a ; Qyp ; | s "
i 2. i ฃo ! ซb ~ i
S" ' i ^I ^~ J
4Qt I / N o"
1 ^m 1 . ( 73) -r- 1
i i ! i
ฎ ! ! i i
i i ! >
tiป i < j<
Biochemical (BIO) Behavior (BEH) Pathology (PTH)
BIO-NOAEL OBIO-LOAEL ปBEH-NOAEL
PTH-LOAEL REP-NO AEL OREP-LOAEL
Result number Test Species Key
~^y ฐ - ? R - rat Pg pig
^>^*"^ T M = mouse Gp = guinea p
Reference Number ^^ Test Species ~ _
! 72 !
^^ x^^^ ' 82 ~ s " i
x-v ^m 5 ^C"J^ " ' x~y~y^\ <* i s~~\
83 J ^^ f 775"^ซ 5- ! C82V2X72) 77 $2 ' (78) 73
JWp '^^ ! 5 ~ ^ | ! 5 a
ct ^ ฐ i ^- i
1 Q. m 77 1
; ซb P i
1 CO CM '
*- cc j
Reproduction (REP) Growth (GRO) Mortality (MOR)
OBEH-LOAEL OPHY-NOAEL PHY-LOAEL OPTH-NOAEL
OGRO-NOAEL GRO-LOAEL OMOR-NOAEL MOR-LOAEL
X-N
lg ^^^^
Guidance for Developing Eco-SSLs
4-15
Revised February 2005
-------
4-4) and scored within a range of 0 to 100 (worst to best) for usefulness in establishing
an oral TRV. Data with limited utility were defined as study endpoints receiving a Total
Data Evaluation Score of 65 or less. These data points were excluded from the plots.
The purpose of the exclusion was to ensure that the TRV derivation used the most
suitable data.
TRV Selected. The general steps and conditional statements of the derivation process
are outlined in Figure 4.8. These steps are an a priori framework for selection of the
TRV value based on the results of the toxicological plots. The flow chart (Figure 4.8) is
used with the toxicity data plots to derive the TRV according to the described steps. The
TRV is equal to the geometric mean of the NOAEL values for growth (GRO) and
reproductive (REP) effects. In cases where the geometric mean NOAEL is higher than
the lowest bounded LOAEL (bounded refers to a LOAEL that has a paired NOAEL) for
either, GRO, REP or mortality (MOR), the TRV is equal to the highest bounded NOAEL
below the lowest bounded LOAEL. An example is provided using the mammalian cobalt
plot depicted as Figure 4.9.
NO
Are there at least 3
toxicity values for 2
species for REP, GRO or
MOR?
YES
Step 2:
Are there 3 or more
NOAELs in REP
and GRO?
TRV = lowest
LOAEL /10
YES
4
Step 5:
At least 3
LOAELs for
GRO & REP?
NO
NO
YES
NO
At Least one NOAEL
for REP and GRO?
At least 6 NOAELs 01
LOAELs for all
endpoints?
Calculate the geometric mean of
NOAELs for REP and GRO
YES
YES
TRV = Highest NOAEL
below lowest LOAEL
for all endpoints
Is NOAEL <
lowest bounded
LOAEL for REP,
GRO or MOR?
NO
NO
TRV = Highest bounded
NOAEL below lowest
bounded LOAEL for
REP, GRO or MOR
TRV = Lowest
NOAEL for GRO, REP
or MOR
Is geometric mean
NOAEL < lowest
bounded LOAEL for
REP, GRO or MOR?
NO
TRV = Highest bounded
NOAEL below lowest
bounded LOAEL for
appropriate effect group
YES
Is Mechanism of
Toxicity Addressed?
YES
TRV = geometric mean
of NOAELs for REP &
GRO
Figure 4.8 Procedure for Deriving the Wildlife Toxicity Reference Value (TRV)
Guidance for Developing Eco-SSLs
4- 16
Revised February 2005
-------
o/kgBW/day)
C
.5 C
0
Q
Figure 4.9 Example of a TRV Derivation (Cobalt)
1 i !
1 j 1 1
1 ! !
i 1 i i
J jj KS d ;ii|:
ฃ gj ! /~\~ 77 ' 'ฐ 77 PT ',ซ_1_&-F V-^ - R ' ^ S ' 2 ss
s" Fii ff iff Fyl ^^ JB r "
I ^^ I f 73 J -^ I E" ^ Geometric Mean of (^^) ^ ^ ^ j
i i ! ^-^ i ^ NOAELs for REP and ^^ Jฐ j
(3 j & j I 0^ j GRO = 7.3 ! Rr p "
|T ; S I ฃ8 j ' co CM
~ ; K- : ' ~ 1 ! CD ci
*- I <" I . ro 1 i CM ^ ,
ฉ i i ! i !ฉฃ, ^ i
' i ' ' ' CN i
. .'< i '. Jซ . . ; . . i
Biochemical (BIO) Behavior (BEH) Pathology (PTH) Reproduction (REP) Growth (GRO) Mortality (MOR)
BIO-NOAEL OBIO-LOAEL ปBEH-NOAEL OBEH-LOAEL OPHY-NOAEL PHY-LOAEL OPTH-NOAEL
PTH-LOAEL REP-NO AEL OREP-LOAEL OGRO-NOAEL GRO-LOAEL OMOR-NOAEL MOR-LOAEL
Result number Test Species Key ^
Reference Ni
i '^^\tv- M = mouse Gp = guineapig Xr^ . .
iVildlife TRV Derivation Process
1) There are at least three results available for two test species within the GRO, REP and MOR effect groups.
There is enough data to derive TRV.
2) There are are at least three NOAEL results available for calculation of a geometric mean.
3) The geometric mean of the NOAEL values for GRO and REP equals 7.3 mg Co/kg BW/day.
4) The geometric mean NOAEL value is less than the lowest bounded LOAEL for REP, GRO or MOR.
5) The mammalian wildlife TRV for cobalt is equal to 7.3 mg Co/kg BW/day.
Guidance for Developing Eco-SSLs
4-17
Revised February 2005
-------
4.5 Calculation of Wildlife Eco-SSLs
Based on the assumptions described in Section 4.3, the Eco-SSL wildlife risk model (Equation
4-1) was reduced to the following after removal of the parameters set equal to 1:
[FIR* (Soilj. * Ps+ B(j)\
Equation 4-2
TRVj
where:
Soilj = Concentration for contaminant (j) in soil (mg/kg [dry weight]),
FIR = Food ingestion rate (kg food [dry weight]/ kg bw [wet weight] / d),
Ps = Proportion of diet that is soil,
TRVj = Toxicity reference value for contaminant (j) (mg [dry weight]/kg bw [wet
weight] /d),
By = Concentration of contaminant (j) in food type (i) mg/kg [dry weight].
As described previously, the concentration of contaminant (j) in biota or food type type (i) (By)
was related to the concentration in soil (Soilj) by an uptake model which has one of the following
forms:
Case 1) By = BAFy * Soilj (constant)
Case 2) ln(By) = I;j + S;j * In(Soilj) (log-linear)
Case 3) B;j = I;j + S;j * Soilj (linear)
where:
BAFy = Soil-to-biota Bioaccumulation factor (BAF) for contaminant (j) for biota
type (i),
Iy = Intercept from bioaccumulation model for contaminant (j) for food type (i)
Sy = Slope from bioaccumulation model for contaminant (j) for food type (i)
In instances where it was necessary to estimate small mammal tissue concentrations (By) based
on dietary based BAFs or regressions, the uptake model may have one of the following forms:
Case 4) B;j = Cdiet * BAF
Case 5) ln(By) = I;j + S;j *
Case 6) B, = I;j + S;j * Cdiet
where:
By = Concentration of contaminant (j) in food type (i) (where i = small
mammal)
Guidance for Developing Eco-SSLs 4-18 Revised February 2005
-------
Cdiet = Concentration of contaminant (j) in diet where diet is 100% earthworms
estimated as in Case 1, 2 or 3, above
BAFdm = Diet-to-biota BAF for contaminant (j) in mammal or bird tissue
The general procedure for calculating the wildlife Eco-SSL for a contaminant (j) was to solve
Equation 4-2 to determine the concentration in soil (Soilj) equivalent to an HQj equal to 1. In
most cases, solution of Equation 4-2 was accomplished by use of a computer spreadsheet
program.
In cases where By was estimated using a simple BAF from soil to food type, Equation 4-2 was
rearranged resulting in the following simplified equation:
TRY
Eco-SSL = Soil. = (Eauation 4-3)
1 FIR*[Ps+ BAF(j\ ^quanonu)
Conservatism of Model Parameterization
The purpose of the Eco-SSLs is to identify concentrations of contaminants in soil that may
present a risk from those concentrations that clearly do not. As a screening tool, the Eco-SSLs
should be conservative, but not so much so that no concentrations pass the screen. To this end,
the wildlife Eco-SSL model was parameterized using a combination of conservatively skewed
and non-conservatively skewed parameter values. The level of conservatism associated with the
model parameter values is summarized in Table 4.4. Whereas the TRVs, food and soil ingestion
rates, diet composition, area use, and bioavailability were all conservatively skewed, all
bioaccumulation values were based on central tendency estimates.
Guidance for Developing Eco-SSLs 4-19 Revised February 2005
-------
Table 4.4 Summary of Conservatism Associated with the Wildlife Eco-SSL Risk Model Parameters.
Parameter
Toxicity Reference Value
(TRY)
Food Ingestion Rate (FIR)
Soil Ingestion as Fraction of
Diet (P.)
Soil to Biota
Bioaccumulation Factor
Diet to Biota
Bioaccumulation Factor
log-linear regression
equations for
bioaccumulation
Diet composition (Pj)
Area Use Factor (AUF)
Bioavailability (AF^ AF^)
Value selected for use in
calculation of Eco-SSL
the geometric mean of
NOAELs for REP or GRO
or the highest bounded
NOAEL below the lowest
bounded LOAEL for REP,
GRO or MOR
90th percentile
90th percentile
median
median
value predicted by
regression equation
100% plant, small mammal,
or soil invertebrate
(depending on trophic
group)
100%
Equivalent to that for the
chemical form used to
develop the TRY
Conservatism of
Value
moderate
high
high
neither conservative
nor anti-
conservative
neither conservative
nor anti-
conservative
neither conservative
nor anti-
conservative
high
high
high
Rationale
Represents highest dose that did not cause
any adverse effects in any test species
Conservatively-skewed value selected to
represent majority of individuals within
population without being over protective.
Conservatively-skewed value selected to
represent majority of individuals within
population without being over protective.
Lowest uncertainty - not expected to over-
predict concentrations
Lowest uncertainty - not expected to over-
predict concentrations
Lowest uncertainty - not expected to over-
predict concentrations
Maximum pathway-specific exposure to
allow screening of risks by pathway.
Consistent with ERAGs (U.S. EPA 1997)
Consistent with ERAGs (U.S. EPA 1997)
Consistent with ERAGs (U.S. EPA 1997)
Guidance for Developing Eco-SSLs
4-20
Revised February 2005
-------
5.0 ECO-SSL DOCUMENTS
Presented in this chapter is a description of the format and content of the Eco-SSL documents
which report the results of the Eco-SSL derivation procedures as described in the previous
chapters. Each document is structured with the following standard sections. However, the
structure is not followed for iron and aluminum where the documents specifically review the
chemistry, bioavailability, and toxicity in soils instead of providing numeric screening levels.
EPA is continuing to evaluate existing toxicity studies (primarily for birds and mammals) and
intends to issue additional Eco-SSL documents in the future. These documents will be posted on
the afore mentioned website as they become available. The following subsections describe the
basic components of each of the documents.
Introduction
The introduction to each contaminant summary provides a brief review of the contaminant
including environmental forms, sources, background concentrations, mechanisms of toxicity, and
essential elemental status if applicable. These summaries are intended to provide the reader with
a general introduction only and do not represent an exhaustive review of the environmental fate
and effects of the contaminant. It is recommended that the user of the Eco-SSL values review
site-specific data concerning contaminant sources and possible fate and transport processes to
evaluate as necessary the site-specific properties of the contaminant in soils. The general
introductory material is based on a review of literature obtained during the search process for
plants and soil invertebrates and wildlife. Some review material is based on information
available from the Hazardous Substances Databank (HSDB), a database of the National Library
of Medicine's TOXNET system (http://toxnet.nlm.nih.gov). Some Eco-SSLs for metals are
within the range of reported background concentrations that may occur at sites without any
contaminant release due to hazardous waste management activities. As part of the Eco-SSL
project, available data for the background concentrations of metals are summarized in Table 2.3
(see Chapter 2) and Attachment 1-4.
Summary Table
This subsection of each document contains a summary table of the Eco-SSL values calculated
for each receptor group expressed as mg contaminant per kg dw soil. If an Eco-SSL could not
be calculated for a receptor group then "NC" is noted for not calculated. In some cases, the Eco-
SSL is pending further review of toxicity information. In these cases, "pending" is noted. The
Eco-SSL values are rounded to two significant digits.
Eco-SSL for Plants
This subsection describes the results of the derivation of an Eco-SSL value for terrestrial plants
completed according to the procedures describe in Chapter 3 and appropriate Attachments.
Guidance for Developing Eco-SSLs 5-1 November 2003
-------
Eco-SSLfor Soil Invertebrates
This subsection describes the results of the derivation of an Eco-SSL value for soil invertebrates
completed according to the procedures described in Chapter 3 and Attachments.
Eco-SSL for Avian Wildlife
This subsection describes the results of the derivation of an Eco-SSL value for avian wildlife.
The subsection provides only the data used to derive the Eco-SSL according to the procedures
described in Chapter 4 and Attachments. The results are provided as two parts.
Toxicity Reference Value
Estimation of Dose and Calculation of the Eco-SSL
Eco-SSLfor Mammalian Wildlife
This subsection describes the results of the derivation of an Eco-SSL value for mammalian
wildlife. The subsection provides only the data used to derive the Eco-SSL according to the
procedures described in Chapter 4 and Attachments. The results are provided as two parts.
Toxicity Reference Value
Estimation of Dose and Calculation of the Eco-SSL.
References
The references are provided as the last subsection of each document. The references are
segregated by receptor with separate lists for plants and soil invertebrates and wildlife. The
references for each receptor are further segregated into two lists. The first list provides the
papers used to derive the Eco-SSL and the second list provides the papers that were rejected for
use. For each of the papers rejected the reason for the rejection is listed.
Appendices
A complete tabulation of the toxicity data used to derive the wildlife toxicity reference values
(TRVs) is provided as an appendix.
Guidance for Developing Eco-SSLs 5-2 November 2003
-------
6.0 REFERENCES
Abou-Donia, M.B. and D. G. Graham. 1978. Delayed neurotoxicity from long term low-level
topical administration of Leptophos to the comb of hens. Toxicology and Applied
Pharmacology. 46: 199-213.
Alexander, M. 1995. How toxic are toxic chemicals in soil? Environ. Sci. Technol. 29: 2713-
2717.
Allen, H., S. McGrath, M. McLaughlin, W. Peijnenburg, and S. Suave. 1999. Bioavailability of
Metals in Terrestrial Ecosystems. Draft Document.
Anderson, W.C., R. C. Loehr, and B. P. Smith. 1999. Environmental Availability of
Chlorinated Organics, Explosives, and Heavy Metals in Soils. American Academy of
Environmental Engineers.
Anderson, M. B., Lepak, K., Farinas, V., and George, W. J. 1993. Protective action of zinc
against cobalt-induced testicular damage inthe mouse. Reprod. Toxicol. 7(1): 49-54.
Anderson, M. B., Pedigo, N. G., Katz, R. P., and George, W. J. 1992. Histopathology of testes
from mice chronically treated with cobalt. Reprod. Toxicol. 6(1): 41-50.
Barber, S. A. 1995. Nutrient Bioavailability - A Mechanistic Approach. John Wiley and Sons.
Barnhart, J. 1997. Chromium Chemistry and Implications for Environmental Fate and Toxicity.
1997. Journal of Soil Contamination. 6(6): 561-568.
Betchel-Jacobs. 1998. Empirical Models for the Uptake of Inorganic Chemicals from Soil by
Plants. Oak Ridge National Laboratory BJC/OR-133. Bechtel Jacobs Company L.L.C.
Beyer, W. N., O. H. Pattee, L. Siteo, D. J. Hoffman, and B. M. Mulhern. 1985. Metal
contamination in wildlife living near two zinc smelters. Environ. Poll. Series A. 38(1):
63-86.
Bohn, H., B. McNeal, and G. O'Connor. 1985. Soil Chemistry. 2nd ed. John Wiley and Sons.
Bourg, W. J., Nation, J. R., and Clark, D. E. 1985. The effects of chronic cobalt exposure on
passive-avoidance performance in the adult rat. Bull. Psychon. Soc. 3(6): 527-530.
Canadian Council of Ministers of the Environment (CCME). 1997. Recommended Canadian
Soil Quality Guidelines. March 1997.
Chetty, K. N., Subba, R. A. O. D SV, Drummond, L., and Desaiah, D. 1979. Cobalt-induced
changes in immune response andatpase activities in rats. J Environ. Sci. Health PartB
Guidance for Developing Eco-SSLs 6-1 November 2003
-------
Pestic. Food. Contam. Agric. Wastes. 14(5): 525-544.
Corrier, D. E., Mollenhauer, H. H., Clark, D. E., Hare, M. F., and Elissalde, M. H. 1985.
Testicular degeneration and necrosis induced by dietary cobalt. VetPathol. 22(6):
610-6.
Cullen, A.C. andH.C. Frey. 1999. Probabilistic Techniques in Exposure Assessment. A
Handbook for Dealing with Variability and Uncertainty in Models and Inputs. New
York: Plenum Press.
Derr, R. F., Aaker, H., Alexander, C. S., and Nagasawa, H. T. 1970. Synergism between cobalt
and ethanol on rat growth rate. J. Nutr. 100(5): 521-524.
Domingo, J. L., Paternain, J. L., Llobet, J. M., and Corbella, J. 1985. Effects of cobalt on
postnatal development and late gestation in rats upon oral administration. RevEsp Fisiol.
Driver, C.J., M. W. Ligotke, P. V. Van Voris, B. D. McVeety, and B. J. Greenspan. 1991.
Routes of uptake and their relative contribution to the toxicologic response of northern
bobwhite (Colinus virginianus) to an organophosphate pesticide. Environ. Tox. and
Chem. 10:21-33.
Efroymson, R. A., M. E. Will, and G. W. Suter II. 1997a. Toxicological Benchmarks for
Contaminants of Potential Concern for Effects on Soil And Litter Invertebrates and
Heterotrophic Process: 1997 Revision. ES/ER/TM-126/R2. Prepared for the U.S.
Department of Energy, Office of Environmental Management by Lockheed Martin
Energy Systems, Inc. managing the activities at the Oak Ridge National Laboratory
(ORNL).
Efroymson, R.A., M.E. Will, G. W. Suter II and A.C. Wooten. 1997b. Toxicological
Benchmarks for Screening Contaminants of Potential Concern for Effects on Terrestrial
Plants: 1997 Revision. ES/ER/TM-85/R3. Prepared for the U.S. Department of Energy,
Office of Environmental Management by Lockheed Martin Energy Systems, Inc.
managing the activities at the Oak Ridge National Laboratory (ORNL).
Gershbein, L. L., Gershbein, J. D., and French, R. 1983. Behavior of male rats fed low levels of
metallic salts. Res Commun Chem PatholPharmacol. 39(3):507-510.
Gertsl, Z. 1990. Estimation of Organic Chemical Sorptionby Soils. Journal of Contaminant
Hydrology. 6: 357-375.
Haga, Y., Clyne, N., Hatroi, N., Hoffman-Bang, C., Pehrsson, S. K., and Ryden, L. 1996.
Impaired myocardial function following chronic cobalt exposure in an isolated rat heart
model. Trace Elements and Electrolytes. 13(2): 69-74.
Guidance for Developing Eco-SSLs 6-2 November 2003
-------
Hansen, E.G., A. B. Paya-Perez,, M. Rahman, and B. R. Larsen. 1999. QSARs for Kow and
Koc of PCB congeners: a critical examination of data, assumptions, and statistical
approaches. Chemosphere 39: 2209-2228.
Henderson, J. D., J. T. Yamamoto, D.M. Fry, J. N. Seiber, and B. W. Wilson. 1994. Oral and
dermal toxicity of organophosphate pesticides in the domestic pigeon (Columba livid).
Bulletin of Environ. Contain. Toxicol. 52: 633-640.
Huck, D. W. and Clawson, A. J. 1976. Excess dietary cobalt in pigs. J Anim Sci. 43(6):
1231-1246.
James, B.R., J.C. Petura, RJ. Vitale, and G.R. Mussoline. 1997. Oxidation-Reduction
Chemistry of Chromium: Relevance to the Regulation and Remediation of Chromate-
Contaminated Soils. Journal of Soil Contamination. 6(6): 569-580.
Jones, D.S., G. W. Suter, and R. N. Hull. 1997. Toxicological Benchmarks for Screening
Contaminants of Potential Concern for Effects on Sediment-Associated Biota: 1997
Revision. ES/ER/TM-95/R4. Prepared for the U.S. Department of Energy, Office of
Environmental Management by Lockheed Martin Energy Systems, Inc. managing the
activities at the Oak Ridge National Laboratory (ORNL).
Kuperman, R. G., Checkai, R. T., Phillips, C. T., and Simini, M. 2002. Toxicity Assessments of
Antimony, Barium, Beryllium, and Manganese for Development of Ecological Soil
Screening Levels (Eco-SSL) Using Enchytraeid Reproduction Benchmark Values.
Rep.Doc.#ECBC-TR-XXX (Draft) , 81 p.
Lagrega, M. 1994. Hazardous Waste Management. McGraw-Hill, Inc.
Lee, L. S., P. S. C. Rao, P. Nkedi-Kizza, and J. J. Delfmo. 1990. Influence of solvent and
sorbent characteristics on distribution of pentachlorophenol in octanol-water and soil-
water systems. Environ. Sci. Tech. 24: 654-661.
Lindsay, W. L. 1979. Chemical Equilibria in Soils. John Wiley & Sons, New York.
Linz, D. G. and D. V. Nakles. 1997. Environmentally Acceptable Endpoints in Soil: Risk-Based
Approach to Contaminated Site Management Based on Availability of Chemicals in Soil.
American Academy of Environmental Engineers. Annapolis, MD.
Loehr, R. C. and M. T. Webster. 1996. Effect of Treatment on Contaminant Availability,
Mobility and Toxicity. Chapter 2. In: Environmentally Acceptable Endpoints in Soils.
D. Linz, and D. Nakles (eds.), American Academy of Environmental Engineers.
Annapolis, MD.
Guidance for Developing Eco-SSLs 6-3 November 2003
-------
Lyman, W. I, W. F. Reehl, and D. H. Rosenblatt. 1990. Handbook of Chemical Property
Estimation Methods. McGraw-Hill, New York.
McNab, B. K. 1963. Bioenergetics and the determination of home range size. Am. Nat.
97:133-140.
Maro, J. K., Kategile, J. A., and Hvidsten, H. 1980. Studies on copper and cobalt in dairy
calves. British Journal of 'Nutrition. 44(1): 25-31.
Miller, R.W. and D. T. Gardiner. 1998. Soil in Our Environment, 8th edition, Prentice Hall,
Upper Saddle River, NJ.
Mohiuddin, S. M., Taskar, P. K., Rheault, M., Roy, P. E., Chenard, J., and Morin, Y. 1970.
Experimental cobalt cardiomyopathy. Am Heart J. 80(4): 532-543.
Mollenhauer, H. H., Corrier, D. E., Clark, D. E., Hare, M. F., and Elissalde, M. H. 1985. Effects
of dietary cobalt on testicular structure Virchow Arch [Cell Pathol]. 49(3): 241-248.
Nagy, K. A., I. A. Girard, and T. K. Brown. 1999. Energetics of free-ranging mammals, reptiles,
and birds. Ann. Rev. Nutr. 19: 247-277.
Nation, J. R., Bourgeois, A. E., Clark, D. E., and Hare, M. F. 1983. The effects of chronic
cobalt exposure on behavior and metallothionein levels in the adult rat. Neurobehav
Toxicol Teratol. 5(1): 9-15.
Paternain, J. L., Domingo, J. L., and Corbella, J. 1988. Developmental toxicity of cobalt in the
rat. J. Toxicol. Environm. Health. 24(2): 193-200.
Pedigo, N. G. and Vernon, M. W. 1993. Embryonic losses after 10-week administration of
cobalt to male mice. Reprod. Toxicol. 7: 111-116.
Pedigo, N. G., George, W. J., and Anderson, M. B. 1988. Effects of acute and chronic exposure
to cobalt on male reproduction in mice. Reprod Toxicol. 2(1): 45-53.
Pignatello, J. J. 2000. The measurement and interpretation of sorption and desorption rates for
organic compounds in soil media. In: Advances in Agronomy, Vol. 69, Academic Press,
San Diego, CA.
Phillips, C. T., Checkai, R. T., Kuperman, R. G., and Simini, M. 2002. Toxicity Assessment of
Antimony, Barium, Beryllium, and Manganese for Development of Ecological Soil
Screening Levels (Eco-SSL) Using Folsomia Reproduction Benchmark Values.
Rep.Doc.#ECBC-TR-XXX (Draft) , 98 p.
Guidance for Developing Eco-SSLs 6-4 November 2003
-------
Rijksinstituut Voor Volksgezondheid en Milieu (RIVM). 1997a. National Institute of Public
Health and Environment. Maximum Permissible Concentrations and Negligible
Concentrations for Pesticides. T. Crommentuijn, D.F. Kalf, M.D. Polder, R. Posthumus
and EJ. van de Plassche. October, 1997.
Rijksinstituut Voor Volksgezondheid en Milieu (RIVM). 1997b National Institute of Public
Health and Environment. Maximum Permissible Concentrations and Negligible
Concentrations for Metals. T. Crommentuijn, D.F. Kalf, M.D. Polder, R. Posthumus and
EJ. van de Plassche.
Sample, B. E., J. J. Beauchamp, R. Efroymson, and G. W. Suter, II. 1999. Literature-derived
bioaccumulation models for earthworms: development and validation. Environmental
Toxicology and Chemistry. 18: 2110-2120.
Sample, B. E., D. M. Opresko, and G. W. Suter II. 1996. Toxicological Benchmarks for
Wildlife: 1996Revision. ES/ER/TM-86/R3. Prepared for the U.S. Department of
Energy, Office of Environmental Management by Lockheed Martin Energy Systems, Inc.
managing the activities at the Oak Ridge National Laboratory (ORNL).
Sample, B. E, J. J. Beauchamp, R. A. Efroymson, and G. W. Suter, II. 1998a. Development and
Validation of Bioaccumulation Models for Small Mammals. Oak Ridge National
Laboratory ES/ER/TM-219. Lockheed Martin Energy Systems Environmental
Restoration Program.
Sample, B. E, J. J. Beauchamp, R. A. Efroymson, G. W. Suter, II and T.L. Ashwood. 1998b.
Development and Validation of Bioaccumulation Models for Earthworms. Oak Ridge
National Laboratory ES/ER/TM-220. Lockheed Martin Energy Systems Environmental
Restoration Program.
Schwarzenbach, R. P., P.M. Gschwend, and D. M. Imboden. 1993. Environmental Organic
Chemistry. In: Methods of Soil Analysis, Part 2, Microbiological and Biochemical
Properties, 2nd Edition.. R. W. Weaver et al (eds.), Agronomy Monograph 9,
Agronomy Science of America. Soil Science Society of America. Wiley -Inter science.
Madison, WI. p. 63.
Seidenberg, J. M., Anderson, D. G., and Becker, R. A. 1986. Validation of an in vivo
developmental toxicity screen in the mouse. Teratog CarcinogMutagen. 6:361-374.
Sumner, M.E. (ed). 2000. Handbook of Soil Science. CRC Press, Boca, Raton. FL.
Syracuse Research Corp. (SRC). Physical Properties Database, http://esc.syrres.com/interkow
/Phy sProp.htm
Guidance for Developing Eco-SSLs 6-5 November 2003
-------
United States Environmental Protection Agency (U.S. EPA). 2003. ECOTOX
(ECOTOXicology) database http://www.epa.gov/ecotox/.
United States Environmental Protection Agency (U.S. EPA). 2002. Role of Background in the
CERCLA Cleanup Program. U.S. EPA Office of Solid Waste and Emergency Response,
Office of Emergency and Remedial Response, April 26, 2002. OSWER 9285.6-07P.
United States Environmental Protection Agency (U.S. EPA). 1999. Ecological Risk Assessment
and RiskManagement Principles for Superfund Sites. Office of Emergency and
Remedial Response, Washington, DC. OSWER Directive 9285.7-28.P.
United States Environmental Protection Agency (U.S. EPA). 1998. Guidelines for Ecological
Risk Assessment. Risk Assessment Forum. U.S. Environmental Protection Agency,
Washington DC. EPA/630/R-95/002F. April. May 14, 1998 Federal Register 63(93):
26846-26924.
United States Environmental Protection Agency (U.S. EPA). 1997. Ecological Risk Assessment
Guidance for Superfund: Process for Designing and Conducting Ecological Risk
Assessments. Interim Final. U.S. Environmental Protection Agency, Environmental
Response Team (Edison, NJ). June 5, 1997.
United States Environmental Protection Agency (U.S. EPA). 1996. Soil Screening Guidance:
Technical Background Document. Office of Emergency and Remedial Response,
Washington, D.C., EPA/540/R-95/128. May.
United States Environmental Protection Agency (U.S. EPA). 1995. Internal Report on
Summary of Measured, Calculated, and Recommended Log Km Values. U.S.
Environmental Protection Agency, Office of Water, Washington, D.C. 38 pp.
United States Environmental Protection Agency (U.S. EPA). 1993. Wildlife Exposure Factors
Handbook, Volumes 1 andII. U.S. Environmental Protection Agency.
EPA/600/R-93/187a and b.
Van Vleet, J. F., Boon, G. D., and Ferrans, V. J. 1981. Induction of lesions of selenium-vitamin
E deficiency in weanling swine fed silver, cobalt, tellurium, zinc, cadmium, and
vanadium. Am. J. Vet. Res. 42(5): 789-799.
Verschueren, K. 1996. Handbook of Environmental Data on Organic Chemicals. 3rd ed. Van
Nostrand Reinhold, New York, NY, USA.
Weissmahr, K.W., S.B. Haderlein, and R.P. Schwarzenbach. 1998. Complex formation of soil
minerals with nitroaromatic explosives and other pi-acceptors. Soil Science Society of
America Journal. 62: 369-378.
Guidance for Developing Eco-SSLs 6-6 November 2003
-------
Weissmahr, K.W., M. Hildenbrand, R. P. Schwarzenbach, and S. B.Haderlein. 1999.
Laboratory and Field Scale Evaluation of Geochemical Controls on Groundwater
Transport of Nitroaromatic Ammunition Residues. Environ. Sci. Technol. 33 (15): 2593
-2600.
Witschi, H.R., and J.A. Last. 1996. Toxic responses of the respiratory system. In: Casarett and
Doull's Toxicology, 5th Edition, pp 443-462. C.D. Klaassen (ed.), McGraw-Hill, New
York.
Guidance for Developing Eco-SSLs 6-7 November 2003
------- |