DRAFT EPA/630/P-04/068B
DO NOT CITE OR QUOTE November 2004
Peer Review Draft
Framework for
Inorganic Metals Risk Assessment
NOTICE
THIS DOCUMENT IS A PRELIMINARY DRAFT. THIS INFORMATION IS
DISTRIBUTED SOLELY FOR THE PURPOSE OF PEER REVIEW UNDER
APPLICABLE INFORMATION QUALITY GUIDELINES. IT HAS NOT
BEEN FORMALLY DISSEMINATED BY THE EPA AND SHOULD NOT BE
CONSTRUED TO REPRESENT ANY AGENCY DETERMINATION OR
POLICY.
Risk Assessment Forum
. Environmental Protection Agency
Washington, DC 20460
-------
DISCLAIMER
This information is distributed solely for the purpose of peer review under applicable
information quality guidelines. It has not been formally disseminated by the EPA and should not
be construed to represent any Agency determination or policy. Mention of trade names or
commercial products does not constitute endorsement or recommendation for use.
11/24/2004
11
DRAFT—DO NOT CITE OR QUOTE
-------
CONTENTS
LISTOFTABLES vi
LIST OF FIGURES viii
LIST OF ABBREVIATIONS AND ACRONYMS ix
PREFACE .. xv
AUTHORS, CONTRIBUTORS, AND REVIEWERS xviii
EXECUTIVE SUMMARY xxv
1. INTRODUCTION 1-1
1.1. PURPOSE AND AUDIENCE 1-1
1.2. METALS FRAMEWORK SCOPE ' - 1-2
1.3. RISK ASSESSMENT FRAMEWORK 1-3
1.4. METALS ASSESSMENT CONTEXT . 1-5
1.4.1. National Ranking and Categorization 1-5
1.4.2. National-Level Assessments 1-7
1.4.3. Site-Specific Assessments 1-8
1.5. ORGANIZATION OF METALS FRAMEWORK 1-9
2. PROBLEM FORMULATION AND METALS PRINCIPLES 2-1
2.1. PRINCIPLES OF METALS RISK ASSESSMENT 2-2
2.1.1. Environmental Background Concentrations 2-2
2.1.2. Essentiality 2-3
2.1.3. Environmental Chemistry : 2-4
2.1.4. Bioavailability 2-6
2.1.5 Bioaccumulation and Bioconcentration 2-9
2.1.6 Acclimation, Adaptation, and Tolerance 2-10
2.1.7. Toxicity Testing 2-10
2.1.8. Mixtures 2-11
2.2. METALS CONCEPTUAL MODEL 2-12
2.3. NEXT STEPS 2-17
3. METALS RISK ASSESSMENT RECOMMENDATIONS 3-1
3.1. HUMAN HEALTH RISK ASSESSMENT RECOMMENDATIONS 3-1
3.1.1. Fate and Transport 3-2
3.1.2. Exposure Assessment 3-2
3.1.3. Effects Analysis 3-7
3.2. METALS RISK ASSESSMENT RECOMMENDATIONS FOR AQUATIC
ENVIRONMENT 3-10
3.2.1. Fate and Transport 3-10
3.2.2. Water Column Exposure, Bioavailability, and Effects 3-12
3.2.3. Background 3-15
3.2.4. Bioaccumulation 3-16
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
iii
-------
CONTENTS (continued)
3.2.5. Trophic Transfer, Biomagnification, and Dietary Toxicity 3-18
3.2.6. Sediment Exposure and Effects 3-20
3.2.7. Metals Mixtures 3-21
3.3. METALS RISK ASSESSMENT RECOMMENDATIONS FOR
TERRESTRIAL ECOSYSTEMS 3-21
3.3.1. Fate and Transport 3-22
3.3.2. Exposure Assessment 3-25
3.3.3. Toxicity Assessment 3-34
4. METAL-SPECIFIC TOPICS AND METHODS 4-1
4.1. ENVIRONMENTAL CHEMISTRY 4-1
4.1.1. Introduction and Terminology 4-1
4.1.2. Hard and Soft Acids and Bases: The Stability of Complexes 4-2
4.1.3. Aquatic Chemistry 4-4
4.1.4. Ground Water and Metal Mobility 4-10
4.1.5. Sediment Chemistry 4-17
4.1.6. Soil Chemistry ; 4-23
4.1.7. Atmospheric Behavior/Chemistry 4-32
4.1.8. Metal Speciation Techniques 4-34
4.1.9. Organo-Metals/Metalloids Transformation Processes 4-36
4.2. HUMAN HEALTH EXPOSURE PATHWAY ANALYSIS 4-44
4.2.1. Introduction 4-44
4.2.2. Human Exposure 4-44
4.2.3. Physiological Routes of Entry 4-52
4.2.4. Integrated Exposure Approaches 4-52
4.2.5. Toxicokinetics and Toxicodynamics 4-62
4.2.6. Pharmacokinetic/Pharmacodynamic Modeling of Metals 4-67
4.3. HUMAN HEALTH EFFECTS 4-71
4.3.1. Introduction 4-71
4.3.2. Essentiality Versus Toxicity 4-71
4.3.3. RDAs and RfDs/RfCs ' 4-73
4.3.4. Biology Relevant to Toxic and Essential Properties of Metals 4-74
4.3.5. Toxicity : 4-76
4.3.6. Metals Mixtures 4-77
4.3.7. Variations in Susceptibility 4-82
4.4. ECOLOGICAL EXPOSURE PATHWAY ANALYSIS 4-88
4.4.1. Aquatic and Terrestrial Transport Pathways for Metals 4-89
4.4.2. Routes of Exposure to Aquatic and Terrestrial Species 4-103
4.4.3. Food Chain Modeling for Wildlife 4-121
4.5. CHARACTERIZATION OF ECOLOGICAL EFFECTS 4-126
4.5.1. Essentiality 4-126
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
IV
-------
CONTENTS (continued)
4,5.2. Acclimation and Adaptation 4-129
4.5.3. Metals Mixtures 4-135
4.5.4. Background 4-140
4.5.5. Indirect Effects of Metals 4-144
4.5.6. Unavailability in Terrestrial Systems 4-147
4.5.7. Bioavailability in Aquatic Systems 4-152
4.5.8. Bioaccumulation and Bioconcentration in Aquatic Organisms 4-160
4.5.9. Bioaccumulation in Terrestrial Organisms 4-170
4.5.10. Sediment Toxicity: Equilibrium Partitioning Approach for Metals 4-174
4.5.11. Soil Toxicity 4-181
4.5.12. Food Chain (Wildlife) Toxicity 4-184
5. METALS RESEARCH NEEDS 5-1
5.1. U.S. EPA RESEARCH 5-1
5.2. EXTERNAL RESEARCH 5-2
5.3. SPECIFIC RECOMMENDATIONS 5-3
5.3.1. Environmental Chemistry 5-3
5.3.2. Bioaccumulation and Bioavailability 5-4
5.3.3. Exposure 5-5
5.3.4. Human Health Effects 5-6
5.3.5. Characterization of Ecological Effects 5-7
5.3.6. Unit World Model for Metals 5-8
GLOSSARY 6-1
REFERENCES 7-1
APPENDIX A-l
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
V
-------
LIST OF TABLES
Table 2-1. Classification of metals based on characteristics of health effects
Table 2-2. Factors that primarily control metal sorption to soils, aquifers, and sediments ..
Table 4-1. Hard and soft acids (metal cations) and bases (ligands)
Table 4-2. Important properties of trace metals as they occur in natural waters and mineral
systems: oxidation states, redox sensitivity, tendency to form sulfides at low Eh
dominant chemical species of metals in soils and natural waters not considering
most (especially weak) metal complexes •
Table 4-3. Comparison of model predictions and measured values of percent metals
associated with the suspended paniculate fraction of mine drainage waters
from selected sites 4-13
2-4
2-6
4-3
4-9
Table 4-4. Partition coefficients as a function of pH for several important elements of
potential concern 4
-16
Table 4-5. Time to achieve 95% of steady-state metal concentration in soil and total soil
metal concentrations after 100 years and at steady state 4-28
Table 4-6. Some stable organometallic compounds 4-37
Table 4-7. Metals/metalloids involved in methylation processes 4-37
Table 4-8. General trends of environmental factors affecting rates of
methylation/demethylation 4-40
Table 4-9. Metal speciation techniques for solids and their associated pore waters 4-43
•
Table 4-10. Metal (versus organic) compound properties affecting absorption, distribution,
metabolism, and elimination 4-63
Table 4-11. Kinetic factors to consider when evaluating the use of PBPK models or other
dosimetric adjustments in the risk assessment process for metals in humans .. 4-69
Table 4-12. Metals classified by their known essentiality 4-71
Table 4-13. List of fate and transport models 4-95
Table 4-14. Metals classified by their known essentiality 4-128
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
vi
-------
LIST OF TABLES (continued)
Table 4-15. Maximum tolerable levels of dietary minerals for domestic livestock in
comparison with levels in forages 4-132
Table 4-16. Examples of indirect effects of metal toxicity 4-145
Table 4-17. Qualitative bioavailability of metal cations in natural soils to plants and soil
invertebrates 4-150
Table 4-18. Qualitative bioavailability of metal anions in natural soils to plants and soil
invertebrates 4-150
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
vii
-------
LIST OF FIGURES
Figure 1-1. Risk assessment/risk management process for metals 1-4
Figure 2-1. Health effect curves for (a) essential elements and (b) nonessential metals .... 2-5
Figure 2-2. Conceptual diagram for evaluating bioavailability processes and
bioaccessibiliry for metals in soil, sediment, or aquatic systems 2-7
Figure 2-3. Generic conceptual model for metals risk assessments 2-14
Figure 4-1. Approximate positions of some natural environment in terms of Eh and pH ... 4-6
Figure 4-2. Sequence of microbially mediated oxidation-reduction reactions 4-7
Figure 4-3. Adsorption of various metal cations and oxyanions, each at 5 * 10"7 M,
by ferrihydrite (ZFe[III] = 10"3 M) as a function of pH at ionic strength of
0.1 mol/kg 4-26
Figure 4-4. Dose-response curves for essential elements 4-72
Figure 4-5. A generalized framework for chemical fate and transport in aquatic systems .. 4-92
Figure 4-6. Conceptual model for direct and indirect exposure of ecological receptors to
metals in soil zones 4-115
Figure 4-7. Relative contribution of incidental soil ingestion to oral dose for wildlife at
different soil ingestion rates and bioaccumulation factors, and a
bioavailability of 100% 4-120
Figure 4-8. Wildlife Oral Exposure Model 4-123
/
Figure 4-9. Linkages between dietary toxicity threshold, bioaccumulation in prey organisms
and waterbome exposure 4-169
Figure 5-1. Schematic representation of the Unit World Model for metals 5-10
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
viii
-------
1 LIST OF ABBREVIATIONS AND ACRONYMS
2
3 1CFOK One-Compartment, First-Order Kinetics Model
4 ABA Absolute bioavailability
5 aBLM Aquatic Biotic Ligand Model
6 ACF Accumulation factor
7 AEC Anion exchange capacity
8 AESOP Advanced Ecological Systems Operating Program
9 AI Adequate intake
10 ALA Aminolevulinic acid
11 ALM Adult lead methodology
12 ANZECC Australian and New Zealand Environment and Conservation Council
13 ARMC ANZ Agriculture and Resources Management Council of Australia and New Zealand
14 ATSDR Agency for Toxic Substances and Disease Registry
15 AUC Area under the curve '
16 AVS Acid-volatile sulfide
17 AWQC Ambient Water Quality Criteria
18 BAF Bioaccumulation factor
19 BCF Bioconcentration factor
W 20 BF Bioaccessible fraction
21 BLM Biotic Ligand Model
22 BMD Benchmark dose
23 BSAF Biota/sediment accumulation factor
24 CAA Clean Air Act
25 CATM Center for Air Toxic Metals
26 CBR Critical body residue
27 CCA Chromated copper arsenate
28 CCC Criterion continuous concentration
29 CEC Cation exchange capacity
30 CHESS Chemical Equilibria in Soils and Solutions
31 CHMTRNS Chemical Transport model
32 CHNTRN Channel Transport model
33 CMAQ Community Multi-scale Air Quality model
34 CMC Criterion maximum concentration
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
ix
-------
LIST OF ABBREVIATIONS AND ACRONYMS (continued)
1 CSF Cancer slope factor
2 CTAP Chemical Transport and Analysis Program
3 CWA Clean Water Act
4 DELFT3D Delft 3-Dimensional model
5 DEPM Dietary Exposure Potential Model
6 DHHS U.S. Department of Health and Human Services
7 DJOC Donald J. O'Connor model
8 DL Diffuse Layer model
9 DOC Dissolved organic carbon
10 DOM Dissolved organic matter
11 DYNTOX Dynamic Toxics model
12 EC Effect concentration
13 ECOFRAM Ecological Committee on FIFRA Risk Assessment Methods
14 ECOM Estuary, Coastal, Ocean Model
15 ECOMSED Estuary, Coastal, Ocean Model (ECOM) updated for sediment transport
16 EcoSSL Ecological soil screening level
17 EERC Energy and Environmental Research Center (University of North Dakota)
18 Eh Redox (reduction-oxidation) potential
19 EPMA-SEM Electron probe microanalysis-scanning electron microscopy
20 EqP Equilibrium partitioning
21 ESB Equilibrium partitioning sediment benchmark
22 EU European Union
23 EUSES European Union System for the Evaluation of Substances model
24 EUTRO Water Quality Analysis Simulation Program (WASP) for Eutrophication
25 EXAMS Exposure Analysis Modeling System
26 foc Fraction of organic carbon (mass, for sediment)
27 FAV Final acute value
28 FCV Final chronic value
29 FDA Food and Drug Administration
30 FETRA Sediment/Radionuclide Transport Model
31 FIAM Free Ion Activity Model
32 FIFRA Federal Insecticide, Fungicide, and Rodenticide Act
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
X
-------
LIST OF ABBREVIATIONS AND ACRONYMS (continued)
Food and Nutrition Board (NAS)
Gastrointestinal tract
General Motors III model
Genus mean acute value
Gill Surface Interaction Model
Generalized Two-Layer Model
Hydrous ferric oxide
Human leukocyte antigens
Hydrophobic organic compound
Hard and soft acid and base
Hydrologic Simulation Program (Fortran program also referenced as HSPF)
Hydrologic Simulation Fortran Program
Hazardous Substance Research Center
3-Dimensional Flow, Heat and Solute Transport model
Ion Balance Model
Inductively coupled plasma-mass spectrometry
Integrated Exposure Uptake Biokinetic model
Integrated Risk Information System
Interstitial Water Benchmark Unit
Partition distribution coefficient
Octanol-water partition coefficient
Lethal concentration (for x percent of the study population)
Lowest-observed-adverse-effect level
Metals Action Plan
Metal-Biotic Ligand model
Maximum contaminant level
Metal Exposure and Transformation Assessment model
Metals Exposure Analysis Modeling Systems
Metals in the Environment Research Network
Monitored natural attenuation
Metallothionein
Molecular weight
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
xi
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
FNB
GI
GM
GMAV
GSIM
GTLM
HFO
HLA
HOC
HSAB
HSP
HSPF
HSRC
HST3D
IBM
ICP-MS
IEUBK
IRIS
IWBU
K,
KOW
LCX
LOAEL
MAP
M:BL
MCL
META4
MEXAMS
MITE-RN
MNA
MT
MW
-------
LIST OF ABBREVIATIONS AND ACRONYMS (continued)
National Academy of Sciences
National Academy of Sciences/Institute of Medicine
National Stream Quality Accounting Network
National Institute of Environmental Health Sciences
No-observed-adverse-effect level
Natural organic matter
National Research Council
Office of Air and Radiation (EPA)
Organization for Economic Cooperation and Development
Organic matter
Organophosphorous
Organophosphate-induced delayed neurotoxicity
Office of Pesticide Programs (EPA)
Office of Research and Development (EPA)
Office of Water, Office of Science and Technology (EPA)
Office of Solid Waste and Emergency Response (EPA)
Pawtuxent Toxic model
Blood lead
Physiologically based pharmacodynamic
Physiologically based pharmacokmetic
Persistent bioaccumulative toxic
Physiologically based toxicokinetic
Probabilistic Dilution Model
Negative log of electron activity
Predicted environmental concentration
Predicted no-effect concentration
Pollution-induced community .tolerance
Particle induced x-ray emission (also uPIXE)
Particulate matter ' '
Particulate organic carbon
Quantitative ion character-activity relationship
Quantitative structure-activity relationship
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
xii
1
2
3
4'
5
•6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
2.1
22
23
24
25
26
27
28
29
30
3*1
32
NAS
NAS/IOM
NASQAN
NIEHS
NOAEL
NOM
NRC
OAR
OECD
OM
OP
OPIDN
OPP
ORD
OST
OSWER
Pawtoxic
PbL
PBPD
PBPK
PBT
PBTK
PDM
..pE
PEC
PNEC
PICT
PIXE
PM
POC
QICAR
QSAR
-------
LIST OF ABBREVIATIONS AND ACRONYMS (continued)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
QWASI
RAF
RAGS
RBA
RCATOX
RCRA
RDA
RfC
RfD
RIVEQLII
RIVRISK
RTC
SAB
SCAMP
SEM
SERATRA
SHEDS
SIMS
SLSA
SMAV
SPC
SPM
SMPTOX
SPQ
SQG
SRWG
SSD
STATSGO
STORET
tBLM
Quantitative Water Air Sediment Interaction model
Relative absorption factor
Risk Assessment Guidance for Superfund
Relative bioavailability
Row-Column Advanced ecological systems operating program for Toxics
model
Resource Conservation and Recovery Act
Recommended dietary allowance
Reference concentration
Reference dose
River Quality II model
River Risk model
Report to Congress
Science Advisory Board
Surface Chemistry Assemblage Model for Particles
Simultaneously extracted metals
Sediment Radionuclide Transport model
Stochastic Human Exposure and Dose Simulation model
Secondary ion mass spectrometry
Simplified Lake and Stream Analysis model
Species mean acute value
Science Policy Council (EPA)
Suspended particulate matter
Simplified Method-Program Variable-Complexity Stream Toxics
Hydrologic Simulation Program-FORTRAN model
Sediment quality guideline
Science and Research Working Group (SRWG) of the Non-Ferrous Metals
Consultative Forum on Sustainable Development
Species sensitivity distribution
State Soil Geographic Database
Storage and Retrieval data system
Terrestrial Biotic Ligand Model
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
xiii
-------
LIST OF ABBREVIATIONS AND ACRONYMS (continued)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
TD
TEF
TIE
TMDL
TRANSPEC
TRIM
TRIM.FaTE
TRV
TSCA
TTD
TTM
TU
U.S. EPA
USGS
VOC
WASP
WASTOX
WER
WHAM
WHO
WHO/IPCS
WOE
WQAM
WQC
WQG
XAS
XPS
XRD
Toxicodynamic
Toxicity equivalence factor
Toxicity identification evaluation
Total maximum daily load
Transport and Speciation model
Total Risk Integrated Methodology model
Total Risk Integrated Methodology Fate, Transport, and Ecological Exposure
model
Toxicity reference value
Toxic Substances Control Act
Target-organ toxicity dose
Total toxicity of mixture
Toxic unit
U.S. Environmental Protection Agency
U.S. Geological Survey
Volatile organic compound
Water Quality Analysis Simulation Program
Water Quality Analysis Simulation of Toxics
Water-effect ratio
Windermere Humic Aqueous Model
World Health Organization
World Health Organization/International Programme on Chemical Safety
Weight of evidence
Water Quality Assessment Methodology model
Water quality criterion
Water quality guideline
X-ray absorption spectroscopy
X-ray photoelectron spectroscopy
X-ray diffraction
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
xiv
-------
1 PREFACE
2
3 Many U.S. Environmental Protection Agency (EPA or the Agency) programs face
4 decisions on whether and how to regulate metals. These decisions range from setting standards
5 or permitting for environmental releases, to establishing safe levels in different environmental
6 media, to setting priorities for programmatic or voluntary efforts. A fundamental input to the
7 decision-making process for most EPA programs is an assessment of potential risks to human
8 health and the environment.
9 EPA's Science Policy Council recognizes that metals present unique risk assessment
10 issues and tasked an Agency workgroup, under the auspices of EPA's Risk Assessment Forum,
11 with devising a Metals Action Plan (MAP) to establish a process for ensuring the consistent
12 application of scientific principles to metals risk assessment. The MAP included brief
13 descriptions of the Agency's current activities on metals, identified critical scientific issues that
14 need addressing, and recommended the development of a Metals Risk Assessment Framework.
15 The MAP stated that the framework should offer general guidance to EPA programs for
16 considering the various properties of metals, such as environmental chemistry, bioavailability,
17 and bioaccumulation.
18 Because of the scientific complexity of metal-specific risk assessment, the Agency
19 recognized the need to include stakeholders and the public in the framework development
20 process and to involve experts throughout the Agency. A stepwise process was initiated,
21 beginning with the MAP and continuing with framework development and review. Workshops
22 and peer-review activities were conducted at multiple intervals during framework production to
23 ensure current and accurate science that supported program applications. To gain additional
24 information, the Agency contracted for the development of issue papers on important topics in
25 metals assessment. These activities, along with input from other federal agencies and review by
26 EPA's Science Advisory Board (SAB), provided additional improvements. Additional details on
27 these activities are provided below.
28 MAP Stakeholder Input. In February 2002, a meeting was convened to gather
29 stakeholder input to help EPA formulate the plan for developing this framework. EPA solicited
30 input on organization and content and received comments that were adopted to the extent
31 practicable. The meeting report and comments are available on EPA's Web site at
32 http://cfpub2.epa.gov/ncea/raf/recordisplay.cfm?deid=51737 and
33 http://cfpub2.epa.gov/ncea/raf/recordisplay.cfm?deid=51736. .
11/24/2004 Peer Review Draft
DISCLAIMER: This infbrmation is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
xv
-------
1 Science Advisory Board Review. In September 2002, EPA's SAB reviewed the MAP
2 and provided comments. Some of the panel's recommendations are summarized below, and all
3 are available in full at http://www.epa.gov/sab/pdf/ecl03001.pdf.
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
• The panel agreed that inorganic metals should be assessed differently from organic
pollutants in a number of contexts. Metals are elements and, although they do not
degrade, they have complex environmental chemistry. Moreover, some metals are
essential for living organisms, and metals occur naturally in the environment.
• The panel agreed that chemical speciation, bioavailability, bioaccumulation, and
toxicity are key issues in assessing the hazards of metals, with some qualifications.
• The panel recommended consideration of stability and environmental residence
times, as well as overall environmental chemistry, to determine temporal
characteristics of metal hazards.
• The panel recommended greater emphasis on the combined effects of metals,
including nutritional and toxicological considerations.
Issue Paper Topics and Science Questions. Issue papers were developed to discuss key
scientific topics pertaining to inorganic metals. The issue paper authors were asked to expand on
these topics,\vith focus on decison-making applications, framework-specific uses, and research
needs. The papers are available athttp://cfpub.epa.gov/ncea/raf/recordisplay.cfm?deid=86119.
The topics and primary questions addressed by the papers include the following:
• Environmental chemistry. How can environmental chemistry be better
incorporated into assessments for inorganic metals?
* Bioavailability and bioaccumulation of metals. What methods or tools can be used
now to reflect metal bioavailability? What scientifically based approaches can be
used to determine metal bioaccumulation?
• Metal exposure assessment. What are the relevant exposure pathways for inorganic
metals to humans and ecological endpoints?
• Human health effects. What populations are most susceptible to effects from
inorganic metals? How should toxicity tests be conducted and interpreted, including
issues of essential elements, dietary salts, and others?
• Ecological effects. What ecological system characteristics promote increased
toxicity from metals?
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
xvi
-------
1' Peer Consultation Workshop. A draft framework was completed in July 2004, and a
2 peer consultation workshop was held on July 27-28 to seek input from scientists expert in the
3 field of metals risk assessment. Scientists participating in the workshop were from academia;
4 industry, state, federal, and Canadian agencies; and various Offices within EPA. Stakeholder
5 comments were also received for consideration. Based on comments received at the workshop,
6 the Agency contracted with a few workshop participants to expand on several gaps and issues
7 identified in the human health and environmental chemistry discussions. The document was
8 revised, and the revised report was made available for inter-Agency review.
9 Inter-Agency Review. Based on comments received, the framework was revised.
10 SAB Review and Public Comment. Pending.
11
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
xvii
-------
1 AUTHORS, CONTRIBUTORS, AND REVIEWERS
2
3 CO-CHAIRS
4 Anne Fairbrother, U.S. EPA, Office of Research and Development, National Health and
5 Environmental Effects Laboratory, Corvallis, OR
6
7 Randy Wentsel, U.S. EPA, Office of Water, Washington, DC
8
9 STEERING COMMITTEE
10 Marilyn Brower, U.S. EPA, Office of Research and Development, National Center for
11 Environmental Assessment, Risk Assessment Forum, Washington, DC
12
13 Stephen Devito, U. S. EPA, Office of Environmental Information, Washington, DC
14
15 Alexander McBride, U.S. EPA, Office of Solid Waste and Emergency Response, Washington,
16 DC (Retired)
17
18 David Mount, U.S. EPA, Office of Research and Development, National Health and
19 Environmental Effects Laboratory, Duluth, MN
20
21 Pamela Noyes, U.S. EPA, Office of Research and Development, National Center for
22 Environmental Assessment, Risk Assessment Forum, Washington, DC
23
24 Keith Sappington, U.S. EPA, Office of Research and Development, National Center for
25 Environmental Assessment, Washington, DC v
26
27 William Wood, U.S. EPA, Office of Research and Development, National Center for
28 Environmental Assessment, Risk Assessment Forum, Washington, DC
29
30 CONTRIBUTORS
31 Environmental Chemistry:
32 Kim Anderson, Oregon State University, Corvallis, OR
33
34 Rufus Chaney, U. S. Department of Agriculture, Beltsville, MD1
35
36 Paul Chrostowski, CPF Associates, Inc., Takoma Park, MD1
37
38 Miriam L. Diamond, University of Toronto, Toronto, ON, Canada
39
40 Robert Ford, U.S. EPA, Office of Research and Development, National Risk Management
41 Research Laboratory, Ada, OK
42
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
xviii
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS (continued)
1 Donald Langmuir, Hydrochem Systems Corp., Colorado School of Mines, Golden, CO1
2
3 Deborah Luecken, U.S. EPA, Office of Research and Development, National Exposure Research
4 Laboratory, Research Triangle Park, NC
5
6 Michael Newman, The College of William and Mary, Gloucester Point, VA
7
8 Robert Puls, U.S. EPA, Office of Research and Development, National Risk Management
9 Research Laboratory, Ada, OK
10
11 Bernard Vigneault, CANMET Mining & Mineral Sciences Laboratories/Natural Resources,
12 Ottawa, ON, Canada1
13
14 Unavailability and Bioaccumulation:
15 Michael Beringer, U.S. EPA, Region VII, Kansas City, KS
16
17 John Drexler, University of Colorado, Boulder, CO1
18
19 Nicholas Fisher, State University of New York, Stony Brook, NY1
20
21 Gerry Henningsen, H&H Scientific Services, Centennial, CO1
22
23 Roman Lanno, Ohio State University, Columbus, OH1
24
25 Jim McGeer, Natural Resources Canada, Ottawa, ON, Canada1
26
27 Keith Sappington, U.S. EPA, Office of Research and Development, National Center for
28 Environmental Assessment, Washington, DC1
29
30 Exposure:
31 Gary Diamond, Syracuse Research Corp., Syracuse, NY1
32
33 Charlie Menzie, Menzie-Cura & Associates, Inc., Chelmsford, MA1
34
35 Jacqueline Moya, U.S. EPA, Office of Research and Development, National Center for
36 Environmental Assessment, Washington, DC1
37
38 Michael Newman, The College of William and Mary, Gloucester Point, V A1
39
40 Jerome Nriagu, University of Michigan, Ann Arbor, MI1
41
11/24/2004 Peer Review D raft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
xix
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
AUTHORS, CONTRIBUTORS, AND REVIEWERS (continued)
Human Health Effects:
Ed Bender, U.S. EPA, Office of Research and Development, Office of Science Policy,
Washington, DC
Michael Beringer, U.S. EPA, Region VII, Kansas City, KS
Harlal Choudhury, U.S. EPA, Office of Research and Development, National Center for
Environmental Assessment, Cincinnati, OH1
Dominic Di Toro, University of Delaware, Newark, DE
Mari Hirtzel Golub, University of California, Davis, CA1
Robert Goyer, University of Western Ontario, National Research Council Board Member/Chair1
Mike Hughes, U.S. EPA, Office of Research and Development, National Health and Ecological
Effects Research Laboratory, Washington, DC1
Richard Hertzberg, U.S. EPA, Office of Research and Development, National Center for
Environmental Assessment, Atlanta, GA
Elaina Kenyon, U.S. EPA, Office of Research and Development, National Health and Ecological
Effects Research Laboratory, Washington, DC1
Fran Kremer, U.S. EPA, Office of Research and Development, National Risk Management
Research Laboratory, Washington, DC
Deirdre Murphy, U.S. EPA, Office of Air, Research Triangle Park, NC
Mark Stifelman, U.S. EPA, Region X, Seattle, WA1
Ecological Effects:
William Clements, Colorado State University, Ft. Collins, CO1
Heidi Bell, U.S. EPA, Office of Water, Washington, DC
Peter Chapman, EVS Environment Consultants, N. Vancouver, BC, Canada
Dale Hoff, U.S. EPA, Region VIII, Denver, CO
Larry Kapustka, Ecological Planning and Toxicology, Inc., Corvallis, OR1
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
XX
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS (continued)
1 Michael Newman, The College of William and Mary, Gloucester Point, VA
2
3 Paul Paquin, HydroQual, Inc., Mahwah, NJ1
4
5 Cindy Roberts, U.S. EPA, Office of Water, Washington, DC
6
7 Mark Sprenger, U.S. EPA, Office of Solid Waste and Emergency Response, Washington, DC1
8
9 Dan Wall, U.S. Department of the Interior, U.S. Fish and Wildlife Service, Washington, DC
10
11 Linda Ziccardi, Exponent, Boulder, C0!
12
13 METALS ISSUE PAPER AUTHORS2
14 Environmental Chemistry of Metals:
15 Donald Langmuir, Hydrochem Systems Corp., Colorado School of Mines, Golden, CO
16
17 Paul Chrostowski, CPF Associates, Inc., Takoma Park, MD
18
19 Rufus Chaney, U.S. Department of Agriculture, Beltsville, MD
20
21 Bernard Vigneault, CANMET Mining and Mineral Sciences Laboratories/Natural Resources
22 Canada,'Ottawa, ON
23
24 Metal Exposure Assessment:
25 Michael Newman, The College of William and Mary/YIMS, Gloucester Point, VA
26 :
27 Gary Diamond, Syracuse Research Corp., Syracuse, NY
28
29 Charles Menzie, Menzie-Cura & Associates, Chelmsford, MA
30
31 Jerome Nriagu, University of Michigan, Ann Arbor, MI
32
33 Ecological Effects of Metals:
34 Lawrence Kapustka, Ecological Planning and Toxicology, Inc., Corvallis, OR
35
36 William Clements, Colorado State University, Fort Collins, CO
37
38 Linda Ziccardi, Exponent, Boulder, CO
39
40 Paul Paquin, HydroQual, Inc., Mahwah, NJ
41
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
xxi
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS (continued)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
Mark Sprenger, U.S. EPA, Edison, NJ
Daniel Wall, U.S. Fish and Wildlife Service, Denver, CO
Human Health Effects of Metals:
Robert Goyer, University of Western Ontario, ON
Man, Golub, University of California, Davis, CA
Harlal Choudhury, U.S. EPA, Cincinnati, OH (contributor)
Michael Hughes, U.S. EPA, Research Triangle Park, NC (contributor)
Elaina Kenyon, U.S. EPA, Research Triangle Park, NC (contributor)
Marc Stifelman, U.S. EPA, Seattle, WA (contributor)
Bioavailability and Bioaccumulation of Metals:
John Drexler, University of Colorado, Boulder, CO
Nicholas Fisher, State University of New York, Stony Brook, NY
Gerry Henningsen, H&H Scientific Services, Centennial, CO
Roman Lanno, Ohio State University, Columbus, OH
Jim McGeer, Natural Resources Canada, Ottawa, ON
Keith Sappington, U.S. EPA, Office of Research and Development, National Center for
Environmental Assessment, Washington, DC
Michael Beringer, U.S. EPA, Region VII, Kansas City, KS (contributor)
PEER CONSULTATION WORKSHOP PARTICIPANTS
Breakout Group 1 (Overall Framework):
Pat Doyle, Environment Canada, Ottawa, ON (breakout group lead)
Charles Menzie, Menzie-Cura & Associates, Chelmsford, MA (Workshop Chairman)
Bruce K. Hope, Oregon Department of Environmental Quality, Portland, OR
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
xxii
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS (continued)
1 Breakout Group 2 (Environmental Chemistry):
2 Rufus Chaney, U.S. Department of Agriculture, Beltsville, MD (breakout group lead)
3
4 Donald Langmuir, Hydrochem Systems Corp., Colorado School of Mines, Golden, CO
5
6 Mark Barnett, Department of Civil Engineering, Auburn University, AL
7
8 Roman Larino, Ohio State University, Columbus, OH
9
10 Jim Ryan, U.S. EPA, Office of Research and Development, National Risk Management
11 Research Laboratory, Cincinnati, OH
12
13 Breakout Group 3 (Terrestrial Exposure/Effects):
14 John Drexler, University of Colorado, Boulder, CO (breakout group lead)
15
16 Lawrence Kapustka, Ecological Planning and Toxicology, Inc., Corvallis, OR
17
18 Bradley Sample, CH2M Hill, Sacramento, CA
19
20 Beverly Hale, University of Guelph, Guelph, ON
21
22 Breakout Group 4 (Aquatic Exposure/Effects):
23 Keith Sappington, U.S. EPA, Office of Research and Development, National Center for
24 Environmental Assessment, Washington, DC (breakout group lead)
25
26 David Mount, U.S. EPA, Office of Research and Development, National Health and
27 Environmental Effects Laboratory, Duluth, MN
28
29 Jim McGeer, Natural Resources Canada, Ottawa, ON
30
31 Steve Klaine, Clemson University, Clemson, SC
32
33 Paul Paquin, HydroQual, Inc., Mahwah, NJ
34
35 Anne Smith-Reiser, Analytical Services Inc., Jessup, MD
36
37 Breakout Group 5 (Human Exposure/Health Effects):
38 Steve Devito, U.S. EPA, Office of Environmental Information, Washington, DC (breakout group
39 lead)
40
41 Mari Golub, University of California, Davis, CA
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
xxiii
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS (continued)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Herman Gibb, Sciences International, Alexandria, VA
>
Rosalind A. Schoof, Integral Consulting, Inc., Mercer Island, WA
Deirdre Murphy, U.S. EPA, Office of Research and Development, Office of Radiation, Office of
Air Quality Planning and Standards, Research Triangle Park, NC
Alexander McBride, U.S. EPA, Office of Solid Waste and Emergency Response, Washington,
DC
Margaret MacDonell, U.S. Department of Energy, Argonne National Laboratory, Argonne, IL
Bob Jones, U.S. EPA, Office of Prevention, Pesticides, and Toxic Substances, Washington, DC
Richard Troast, U.S. EPA, Office of Solid Waste and Emergency 'Response, Washington, DC
'Scientific expert commissioned to develop paper on issues and state-of-the-art
approaches in metals risk assessment.
2 Information from the metals issue papers was used in the development of this
framework.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
xxiv
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
EXECUTIVE SUMMARY
The Framework for Inorganic Metals Risk Assessment is a science-based document that
addresses the special attributes and behaviors of metals and metal compounds when assessing
their human health and ecological risks. The document describes basic principles to be
considered in assessing risks posed by metals and presents a consistent approach for use across
the Agency when conducting these assessments. Although the audience for the framework is
primarily intended to be Agency risk assessors, it will also communicate principles, tools, and
recommendations for metal risk assessment to stakeholders and the public. The Agency
developed this framework document to supplement previous guidance for use in site-specific
risk assessments, criteria derivation, ranking or categorization, and other similar Agency
activities related to metals.
One of the purposes of this document is to present key principles that contain specific
issues which differentiate inorganic metals from other chemicals when assessing their risk to
human health and the environment. While we recognize that organic compounds, for example,
undergo bioaccumulation, there are unique properties, issues, and processes within these
principles that assessors need to consider when evaluating metal compounds. For example, the
latest scientific data on bioaccumulation do not currently support the use of bioconcentration
factor (BCF) and bioaccumulation factor (BAF) data when applied as generic threshold criteria
for the hazard potential of inorganic metals (e.g., for classification as a "PBT" chemical). These
principles should be addressed and incorporated into metals risk assessment to the extent
practicable. They include the following:
Environmental background
concentrations
Essentiality
Environmental chemistry
Bioavailability
• Accumulation/bioaccumulation and
bioconcentration
• Acclimation, adaptation, and tolerance
• Toxicity testing
• Mixtures
This document discusses, in Section 2, why these principles are important in ecological
and human health risk assessments and presents conceptual models on metal-specific attributes
and bioavailability. In Section 3, the framework provides assessors with recommendations and
method applications, and supports these recommendations with technical discussions, in Section
4, on metal-specific topics, including environmental chemistry, human health exposure pathway
• and effects analysis, and ecological exposure pathway and effects analysis. Section 5 presents
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information.
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
XXV
-------
1 research needs to improve the science supporting the assessment of metals and metal
2 compounds. This framework document will not be a prescriptive guide on how any particular
3 type of assessment should be conducted within a U.S. EPA program office. Rather, it is intended
4 to provide recommendations and foster the consistent application of methods and data to metals
5 risk assessment in consideration of the unique properties of metals.
6 While this document discusses scientific issues and makes recommendations about
7 scientific approaches, this framework does not address the science policy questions and issues
8 which are raised. Rather, it is intended to make recommendations and foster the consistent
9 application of methods and data to metals risk assessment in consideration of the unique
10 properties of metals. Consistent with these recommendations, the Agency will be analyzing the
11 science policy implications and developing appropriate policy approaches which are protective
12 of human health and the environment.
13 The framework is the result of contributions from a variety of individuals inside and
14 outside the Agency. Their combined expertise and enthusiasm have improved the technical
15 quality of the document and its applicability for various risk assessment activities.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
xxvi
-------
1 1. INTRODUCTION
2
3 Metals and metal compounds have unique characteristics that should be considered when
4 assessing their risks. Some of these characteristics are typically not considered when assessing
5 the risks of organic substances. For example, although metals are neither created nor destroyed
6 by biological or chemical processes, they are transformed from one chemical form to another.
7 Native (zero valence) forms of metal and some inorganic metal compounds are not readily
8 soluble, and as a result, toxicity tests based on soluble salts may overestimate the bioavailability
9 and toxicity of these substances. Some metals are essential elements at low levels but toxic at
10 higher levels (e.g., copper, selenium, and zinc), whereas others have no known biological
11 functions (e.g., lead, arsenic, and mercury). Because metals are naturally occurring, many
12 organisms have evolved mechanisms to regulate accumulations, especially those of essential
13 metals. Because the majority of compounds assessed by the U.S. Environmental Protection
14 "Agency (EPA or the Agency) are organic substances, the various guidance documents provided
15 for risk assessments of either human health or ecological receptors lack specificity on how to
16 account for these and other metal attributes.
17
18 1.1. PURPOSE AND AUDIENCE
19 The Agency developed this framework document to supplement previous guidance for
20 use in site-specific risk assessments; criteria derivation, ranking, or categorization; and other
21 similar Agency activities related to metals. This framework document will not be a prescriptive
22 guide on how any particular type of assessment should be conducted within an EPA program
23 office. Rather, it is intended to provide recommendations and foster the consistent application of
24 methods and data to metals risk assessment in consideration of the unique properties of metals.
25 The inorganic metals risk assessment framework describes basic principles to be
26 considered in assessing risks posed by metals and presents a consistent approach for use across
27 the Agency when conducting these assessments. Although the primary audience will be Agency
28 risk assessors, the framework will also communicate principles, tools, and recommendations for
29 metals risk assessment to stakeholders and the public. In addition, the framework relies heavily
30 on issue papers developed, under EPA commission, on key scientific topics pertaining to
31 inorganic metals. The papers are available on EPA's website at
32 http://cfpub.epa.gov/ncea/rai7recordisplay.cfm?deid=86119.
33 The metals framework is intended for guidance only. It does not establish any
34 substantive "rules" under the Administrative Procedure Act or any other law and will have no
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
1-1
-------
1 binding effect on EPA or any regulated entity. Rather, it represents a nonbinding statement of
2 policy. EPA believes that the metals framework provides a sound, up-to-date presentation of
3 principles, tools, and recommendations for assessing the risk posed by metals and serves to
4 enhance the application of the best available science in Agency risk assessments. However, EPA
5 may conduct metals risk assessments using.approaches and tools mat differ from those described
6 in the framework for many reasons, including, but not limited to, new information, new
7 scientific understandings, and new science policy judgments. The science surrounding metals
8 risk assessment continues to be intensively studied and thus is rapidly evolving. Specific
9 principles, tools, or recommendations presented in the framework may become outdated or may
10 otherwise require modification to reflect the best available science. Application of this
11 framework in future metals risk assessments will be based on EPA decisions that its approaches
12 are suitable and appropriate. These judgments will be tested and examined through peer review,
13 and any risk analysis will be modified as deemed appropriate.
14
15 1.2. METALS FRAMEWORK SCOPE
16 The metals risk assessment framework is a science-based document that focuses on the
17 special attributes and behaviors of metals and metal compounds affecting human health and
18 ecological risk assessments. It does not set forth a step-by-step process to assess the risk of
19 metals to human health or the environment but, rather, focuses on principles, tools, and methods
20 coupled with recommendations to guide assessors in addressing the unique properties of
21 inorganic metals. It supplements existing guidance and does not cover elements of the risk
22 assessment process that are not unique to metals because these are adequately addressed in other
23 Agency guidelines and strategies (e.g., U.S. EPA, 2003a, 2000a, 1998a).
24 The Agency regulates metals and their
25 inorganic and organometallic compounds because
26 they have the potential to harm human health and the
27 environment. The Agency's Science Advisory Board
28 has stressed the importance of environmental
29 chemistry and its relevance to the assessment of both
30 inorganic and organometallic compounds. However,
31 the complexities of addressing all types of metal
32 compounds within a single document would result in
33 a framework that would be difficult to follow or to
34 apply in specific cases. Because organometallic
35 compounds exhibit properties common to both
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
1-2
Metals and Metalloids of Interest
Aluminum
Antimony
Arsenic
Barium
Beryllium
Boron
Cadmium
Chromium
Cobalt
Copper
Iron
Lead
Manganese
Mercury (inorganic)
Molybdenum
Nickel
Selenium
Silver
Strontium
Tin
Thallium
Vanadium
Zinc
-------
1 organic substances and inorganic metal compounds, the properties of both the organic moieties
2 of these compounds and their inorganic components would need to be addressed. Frameworks
3 and associated guidance documents for assessing properties of organic compounds have already
4 been developed by EPA and do not need to be discussed further here. Therefore, this document
5 addresses only those assessment issues associated with inorganic metal compounds. The
6 framework does discuss natural transformation pathways that form organometallic compounds
7 and refers the reader to appropriate Agency documentation or research efforts related to relevant
8 risk assessment issues.
9 In this document, the term "metals" generally refers to metals and metalloids that may
10 pose a toxic hazard and are currently of primary interest to EPA. However, the principles and
•11 approaches set forth in the framework are applicable to all metals. In some instances, metal-by-
12 metal considerations are included, either to serve as examples or to highlight particular
13 exceptions.
14
15 1.3. RISK ASSESSMENT FRAME WORK
16 Risk assessment provides a qualitative and quantitative comparison of the relationship
17 between environmental exposures and effects in exposed individuals and other organisms. In
18 1983, the National Research Council described four primary steps in the process of risk
19 assessment: hazard identification, dose-response assessment, exposure assessment, and risk
20 characterization. EPA has developed a similar framework for ecological risk assessment and
21 included a problem formulation step (U.S. EPA, 1998a). This framework document provides
22 recommendations, including applications and limitations of currently available tools and
23 methods, for conducting metals risk assessment. These recommendations are designed to be
24 incorporated into current principles and elements of human health and ecological risk assessment
25 guidance developed by the U.S. EPA (e.g., U.S. EPA, 2003a, 2000a, 1998a). Additional general
26 risk assessment information is also available on EPA's website at http://cfpub.epa.gov/ncea/ and
27 ' http://cfpub.epa.gov/ncea/rafindex.cfm.
28 Figure 1-1 broadly illustrates the overall risk assessment/risk management process, and
29 by way of example, identifies in the problem formulation and analysis steps some metals-
30 specific considerations. An effective risk assessment for metals will consider the unique aspects
31 of metals, differentiating them from other substances, early and throughout the risk assessment
32 process. These unique aspects are captured and formulated in this framework as metals concepts
33 and principles; they are summarized in Section 2 and discussed throughout the framework. They
34 include environmental background concentrations; essentiality; environmental chemistry;
35 bioavailability; bioaccumulation and bioconcentration; acclimation, adaptation, and tolerance;
36 toxicity testing; and mixtures.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
1-3
-------
1
2
3
4
5
6
7
8
9
10
11
12
,13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Figure 1-1. Risk assessment/risk management process for metals.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
1-4
-------
1 1.4. METALS ASSESSMENT CONTEXT
2 The context for the risk assessment is a major factor in determining the type of analysis
3 that is appropriate for any particular situation. To provide a context for discussion of the
4 framework principles for metals, EPA has defined three general categories of assessments:
5 national ranking and categorization, national-level assessments, and site-specific assessments.
6 Each type of assessment can vary in level of detail from simple screening analysis to highly
7 complex definitive assessments.
8
9 1.4.1. National Ranking and Categorization
10 In 1he first type of assessment, EPA
11 may rank or categorize chemicals on the basis
12 of their potential to cause risk. For many
13 chemicals, there are significant data gaps
14 regarding their chemistry, environmental fate,
15 toxicity, or exposure potential, notably with
16 regard to location-specific characteristics that
17 directly influence these factors and make
18 broad generalizations difficult. Nonetheless,
19 EPA is tasked with protecting human health
20 and the environment from the potentially
21 harmful effects of these chemicals and thus
22 had to develop methods to identify those most
23 likely to pose a significant threat.
24 With more than 80,000 chemicals
25 currently listed on the Toxic Substances
26 Control Act (TSCA) inventory that can
27 legally be used in commerce within the United States (not including pesticides or chemicals that
28 are created as byproducts during industrial processes), the Agency needs a way to prioritize
29 substances for review or action. Many of the statutes administered by EPA provide specific lists
30 of chemicals that require,consideration, but often those lists are based on information and
31 analyses previously developed by EPA. In addition, the statutes generally provide for adding or
32 deleting chemicals from the initial list on the basis of their potential threat to human health or
33 ecological receptors. Consequently, a need exists for methods that rapidly screen chemicals for
34 placement on lists or that prioritize potentially hazardous substances.
Hazardous Waste Listing Determination
Under the Resource Conservation and Recovery
Act, EPA is required to make formal decisions on
whether to designate certain specific industry waste
streams as hazardous. For waste streams that are listed
as hazardous, the generators and handlers of those
wastes must comply with a comprehensive set of
management and treatment standards.
In determining whether to list a waste as hazardous,
the Agency evaluates the ways in which that waste is
currently being managed or could plausibly be managed
by the generators and handlers of the waste. The
Agency also assesses the physical and chemical
composition of the waste. Based on the waste
characteristics and management practices, EPA then
conducts an analysis to determine whether potentially
harmful constituents in the waste might be released and
transported to human or ecological receptors. In
conducting these analyses, the Agency evaluates the
potential for constituents in the waste material to be
released to air, surface water, soil, and ground water. It
then models the fate and transport of those constituents
to potential receptors.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
1-5
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
Some of the ranking and categorization methods used by EPA involve identifying certain
attributes of chemicals that can then be used as indicators of potential risk. Example attributes
include toxicity, production volume, quantities released to the environment, persistence in the
environment, mobility in the environment as indicated by volatility or solubility, and potential to
accumulate in the food chain. Other
methods, which may be less quantitative,
rely more on a combination of expert
judgment, stakeholder input, and
availability of information to determine the
priority or categorization of chemicals for
decision making or other action. Examples
of programs where EPA identifies or
categorizes chemicals for priority action
include the following:
Selecting chemicals for the
Agency's Toxicity
Characteristic regulation that
defines hazardous wastes;
Establishing reporting
thresholds for spills of
hazardous materials under .
Superfund;
Setting priorities for revisions to
the Ambient Water Quality
Criteria (AWQC);
Listing chemicals under the
Toxics Release Inventory;
Determining priorities for
developing drinking water
standards;
Setting priorities for hazardous
air pollutant data collection and
assessment; and
Ambient Water Quality Criteria
EPA's Office of Water is charged with developing
Ambient Water Quality Criteria (AWQC) to support the
Clean Water Act goals of protecting and maintaining
physical, chemical, and biological integrity of waters of
the United States. Examples of chemical-specific
criteria include those designed to protect human health,
aquatic life, and wildlife. Although AWQC are typically
derived at a national level, there is a long history behind
the development of methods to accommodate site-
specific differences in metals bioavailability. For
example, since the 1980s aquatic life criteria for several
cationic metals have been expressed as a function of
water hardness to address the combined effect of certain
cations (principally calcium and magnesium) on toxicity.
Recognizing that water hardness adjustments did not
account for other important ions and ligands that can
alter metals bioavailability and toxicity, EPA developed
the Water Effect Ratio (WER) procedure as an empirical
approach for making site-specific bioavailability
adjustments to criteria (U.S. EPA, 1994a). This
approach relies on comparing toxicity measurements
made in site water with those made in laboratory water
to derive a WER. The WER is then used to adjust the
national criterion to reflect site-specific bioavailability.
More recently, the Office of Water has been
developing a mechanistic-based approach for addressing
metals bioavailability using the Biotic Ligand Model
(BLM)(U.S. EPA,2000b;DiToreetal.,2001; Santore
et a]., 2001). This model, which is described in further
detail in Section 4, predicts acute toxicity to aquatic
organisms on the basis of physical and chemical factors
affecting speciation, complexation, and competition of
metals for interaction at the biotic ligand (i.e., the gill in
the case offish). The BLM has been most extensively
' developed for copper and is being incorporated directly
into the national copper aquatic life criterion. The BLM
is also being developed for use with other metals,
including silver. Conceptually, the BLM has appeal
because metals criteria could be implemented to account
for predicted periods of enhanced bioavailability at a site
that may not be captured by purely empirical methods,
such as the WER.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
1-6
-------
1 • Setting priorities for reviewing existing chemicals under TSCA.
2
3 This list of needs for ranking or categorizing chemicals is not comprehensive but does
4 provide an indication of the kinds of activities that EPA conducts in this regard. In addition, the
5 Agency may set national standards and guidelines for specific chemicals, including metals, as
6 described in the next section.
7
8 1.4.2. National-Level Assessments
9 National-level assessments may be performed when the Agency is setting media
10 standards or guidelines for chemicals (e.g., Maximum Contaminant Levels [MCLs], National
11 Ambient Air Quality Standards, AWQC, Superfund soil-screening levels) or when the Agency is
12 using risk assessments to establish controls for environmental releases from industry or other
13 sources (e.g., hazardous waste listings under the Resource Conservation and Recovery Act,
14 residual risk determinations under the Clean Air Act, pesticide registrations). EPA also is
15 charged with establishing controls on environmental releases based on the best available
16 treatment technologies (e.g., maximum achievable control technology for air emissions, best
17 available treatment technology for surface water discharges and for hazardous wastes).
18 However, even though the standards are based on technological achievability, the Agency
19 typically performs risk assessments in support of these regulations to help inform management
20 decisions and for use in cosl/benefit analyses.
21 . Differing environmental conditions across the country affect the biogeochemistry of
22 metals, making it difficult to set single-value national criteria (national standards that apply at
23 the point of exposure, such as MCLs, are less affected by these factors). To conduct such
24 assessments, the Agency commonly undertakes several approaches. One is to define one or
25 more exposure scenarios and to conduct a relatively detailed analysis. The difficulty in this
26 approach is in selecting the appropriate scenario; typically, the Agency tries to ensure that the
27 scenario is sufficiently conservative to be protective of the population at highest risk (such as
28 populations exposed above the 90th percentile) without being so conservative that the standards
29 are protective of hypothetical individuals whose calculated risks are above the real risk
30 distribution. In selecting the appropriate scenario, the Agency needs to consider all of the factors
31 mat may affect potential risk, including environmental factors affecting the fate, transport,
32 exposure potential, and toxicity of the chemicals released.
33 Another common approach for a national assessment or criteria derivation is to conduct a
34 probabilistic analysis (such as a Monte Carlo analysis), wherein the variability of the key factors
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
1-7
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
is described by parameter distributions used as inputs to the probability analysis procedure. The
result is an integrated distribution of potential risk levels. The difficulties related to conducting
this kind of analysis are in developing appropriate distributions for each parameter and in
ensuring that adequate attention is paid to
potential correlations among key
parameters. These correlations often are
more complex and difficult to describe for
metals than for organic compounds.
Establishing Water Discharge Permit Conditions
The Clean Water Act establishes a two-tier process
for setting water discharge permit conditions. First, all
dischargers must meet the technology-based effluent
guidelines limitations requirements. Second, if those
limitations are not adequate to allow the receiving
stream to achieve its designated water quality standards,
then more stringent limits are developed to ensure that
those standards are met.
The water quality standards are established by the
, states and consist of a designated use for the waterbody
and a set of criteria for individual chemicals that allow
that use to be achieved. EPA has published national
water quality criteria values for the states to use as
guidance in setting their standards.
Once the standards that include the criteria have been
established and it has been determined that the effluent
guidelines alone will not be sufficient to allow those
criteria to be met, the state prepares a wasteload
allocation for all the dischargers to that stream segment,
including, where appropriate, the nonpoint source
discharges. The wasteload allocation generally consists
of modeling the potential impact on the stream from
each discharge of the chemicals of concern and then
setting the allowable discharges to ensure that the
criteria for the chemicals are met.
The modeling process can be quite complex,
potentially taking into account the interactions of the
ambient stream conditions with the chemicals in the
discharge, including dilution, chemical transformations,
degradation, settling, resuspension, and other processes.
For metals, stream characteristics such as pH, organic
content, suspended solids levels, and numerous other
factors can significantly affect how the metal will
behave and affect aquatic life in the stream segment.
Therefore, it is important to understand these processes
in conducting the wasteload allocation.
1.4.3. Site-Specific Assessments
Site-specific assessments are
conducted to inform a decision concerning a
particular location and may also support
some national regulatory decisions.
Examples include the following:
• Determining appropriate soil
cleanup levels at a Superfund
site,
• Establishing water discharge
permit conditions to meet
ambient water quality standards,
and
• Determining the need for
emission standards for sources of
hazardous air pollutants.
An accurate site-specific assessment
for a metal requires knowledge of the form
of the metal as it enters the environment, the
environmental conditions affecting the metal (climatological conditions, soil geochemistry,
water and sediment chemistry, etc.), the existence of plants and/or animals that might
accumulate the metal as well as the uptake factors for whatever form(s) the metal may be in,
plausible pathways and routes of exposures to the human or ecological receptors, and the effect
the metal will have on target organisms in whatever form in which it reaches that organism and
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
1-8
-------
1 its target organ/system. Although many of these same factors also affect the risk potential of
2 organic chemicals, models for predicting fate, transport, and toxic properties are generally better
3 defined for organic chemicals than for metals.
4 In summary, the Agency conducts a variety of assessments, from site-specific risk
5 assessments to national criteria setting and ranking. All of these assessments share common
6 elements and rely on accurate information and knowledge about how chemicals behave in the
1 7 environment and when they come in contact with humans or other organisms of concern. Metals
8 have unique environmental and toxicological properties that may confound such assessments if
9 they are not given consideration. This framework provides the basic tools for application in each
10 of these programmatic contexts so that metals assessments can be conducted with rigor,
11 precision, and accuracy.
12
13 1.5. Organization of Metals Framework
14 The framework includes sections on metals principles and conceptual models,
15 recommended methods for metals assessment, and metal-specific topics and methods. Section 2
16 begins with a discussion of metals principles and their importance in the assessment of inorganic
17 metals. A conceptual model is presented that highlights the areas where metal-specific
18 information is required to move through the risk assessment, criteria development, or
19 classification/ranking process. This discussion provides additional direction to where in the
20 document the guidance material is discussed for each issue identified. Also in Section 2, a
21 conceptual model on bioavailability is presented (McGeer et al., 2004) along with definitions •
22 developed by authors of the issue papers.
23 Section 3 provides recommendations to guide risk assessors in incorporating metal-
24 specific issues and application of tools and methods into the phases and levels of risk
25 assessment. Discussion is included about which tools are appropriate for screening-level
26 assessments and which information is most useful for detailed, in-depth analyses. Detailed
27 discussion of metal-specific tools and methods occurs in Section 4, where each subsection is
28 devoted to a particular metals issue. Specifically, Section 4 is organized as follows:
29
30 • Section 4.1. Environmental Chemistry
31 • Section 4.2. Human Exposure Pathway Analysis
32 • Section 4.3. Human Health Effects
33 • Section 4.4. Ecological Exposure Pathway Analysis
34 • Section 4.5. Characterization of Ecological Effects
35
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
1-9
-------
1 For each issue, a brief description of the concepts to be addressed is presented, and
2 reference is given to existing issue papers that cover it in greater depth. This is followed by a
3 presentation of currently available tools and methods for developing the needed information,
4 such as look-up tables, default values, appropriate models, and other sources. Each subsection
5 concludes with a brief discussion of the limitations of current tools and methods, any new
6 methodology currently under development, and a suggestion of where the process could be
7 improved in the future.
8 It is important to stress that Section 4 does not precisely follow the risk assessment
9 framework of exposure assessment and effects assessment but, rather, presents all the necessary
10 metal-specific attributes to consider when conducting a hazard and risk assessment. This is
11 because several of the issues are cross-cutting (e.g., environmental chemistry discussions in
12 Section 4.1) and may have application to both exposure and effects assessments. Furthermore,
13 because this framework addresses only the aspects of the risk or hazard assessment that are
14 specific to metals, it does not provide a comprehensive overview of the entire process.
15 The document concludes with a discussion about research under way, planned, and
16 needed to reduce uncertainty (Section 5). Although our understanding about metals issues is
17 broad based, specific methods and data are more readily available in some areas (e.g., freshwater
18 ecosystems) than in others (e.g., soils). This Section highlights the areas where active research is
19 expected to move the science forward within the next 5 years and identifies other aspects that
20 need further attention. The need for continued research and development of metals-specific risk
21 assessment methodology should in no way detract from the expectation of applying sound
22 science to our current way of doing business, as the framework provides substantial guidance on
23 process enhancements that can occur with currently available information.
24
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
1-10
-------
1 2. PROBLEM FORMULATION AND METALS PRINCIPLES
2
3 The risk assessment phase of identifying important issues and outlining the scope of both
4 human health and ecological risk assessments is commonly referred to as "problem formulation"
5 (U.S. EPA, 2003a, 2000a, 1998a). Metals have a number of characteristics that require special
6 consideration early and throughout the risk assessment process. This chapter states the major
7 principles underlying metals analyses and provides guidance on how to set up the conceptual
8 model and scope of the assessment to account for metals-specific differences in risk analysis.
9 Metals are naturally occurring constituents of the environment. Consequently, biota have
10 evolved and continue to evolve in their presence. Thus, naturally occurring levels of metals play
11 an important role in the biogeographic distributions of plants and animals and may, in fact, be
12 limiting factors in species distributions or landscape uses. Therefore, during the problem
/
13 formulation phase of an assessment of anthropogenically elevated metals, it is important to
14 clearly define the geospatial area to which the results will apply and to identify environmental
15 controlling factors (e.g., pH, organic matter, iron, aluminum) and the resulting naturally
16 occurring differences in biota composition and metal sensitivity (see Section 4.1, Environmental
17 Chemistry).
18 For metals, the type of assessment (i.e., screening or definitive) and the scale of the
19 assessment (i.e., site specific, regional, or national) will determine how information on metals
20 can be applied in the assessment. Site-specific assessments will involve only a single
21 geographical area of concern and, therefore, can incorporate locally relevant aspects of
22 environmental chemistry, natural background concentrations, and species sensm'vies. For
23 regional and national-scale assessments, more general assumptions about the form of the metal
24 in the environment, uptake and bioavailability parameters, and sensitive species or
25 subpopulations are useful, frequently producing results that are conservative in their assumptions
26 in order to be protective of sensitive species or locations. Regardless, the fundamental principles
27 that determine the form of metal in the environment and, consequently, the transport of metals
28 through environmental media to accumulate or cause toxic responses in biota should be
29 considered in all risk assessments.
30
31
11/24/2004 Peer Review Draft
DISCLAIMER: This infonnation is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
2-1
-------
1 2.1. PRINCIPLES OF METALS RISK ASSESSMENT
2 One of the purposes of this document is to present key principles that contain issues or
3 processes that differentiate inorganic metal compounds from other chemicals when assessing
4 their risk to human health and the environment. While we recognize that organic compounds,
5 for example, undergo bioaccumulation, there are unique properties, issues, and processes within
6 these principles that assessors should consider when evaluating metal compounds. Contributors
7 to the Metals Action Plan (MAP), members of the Science Advisory Board, and external
8 stakeholders, along with various contributors to and authors of the framework, have discussed
9 these metals principles for consideration in the
10 assessment of metals. The following discussion of topics
11 under each of these principles reflects this extensive
12 deliberation and describes unique aspects of inorganic
13 metal compounds that should be considered when risk
14 assessments are conducted. The principles focus on
15 unique properties of inorganic metal compounds, and this
Metals Principles
Environmental background concentrations
Essentiality
Environmental chemistry
Bioavailability
Bioaccumulation and bioconcentration
Acclimation, adaptation, and tolerance
Toxicity testing
Mixtures
16 chapter discusses why these principles are important for
17 risk assessments. These principles should be addressed and incorporated into metals risk
18 assessments to the extent practicable. They are visible throughout this document. In Chapter 3,
19 they are expanded upon with recommendations to guide assessors, and in Chapter 4, specific
20 topics are discussed in more detail.
21
22 2.1.1. Environmental Background Concentrations
23 Because metals are naturally present in the environment, it is important to consider the
24 background concentrations of metals when conducting risk assessments. How to incorporate
25 this unique aspect of metals in risk assessments is a common challenge. The following key
26 questions arise:
27
28 • How should the cumulative exposure and risk of background and anthropogenic or
29 "added metal" be considered in risk assessments?
30
31 • How do the natural background levels of metals influence the types of ecological
32 receptors that are naturally present and appropriate to consider in risk assessments?
33
34 Only the bioavailable fraction of background concentrations contributes to total metal exposure
35 and overall risk. Test organisms should be acclimated to background conditions, and
36 appropriately adapted organisms should be used for site-specific assessments.
37
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
2-2
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Background should be defined in a specific spatial and temporal aspect related to the
scope of the particular hazard or risk assessment Background metal concentrations can vary by
as much as five orders of magnitude, depending on soil type, geography, and other factors
(Chapman and Wang, 2000). Background may
exacerbate toxicological effects and accumulations of
metals from direct emissions or other regulated sources
Background
Background is defined as the amount
of metals occurring in soil, water, air, and
biota as a result of anthropogenic and
natural processes. Anthropogenic
contributions are limited to those that are
not influenced by current, direct releases
(i.e., emissions, discharges, or disposal)
from a source or site of concern. This
includes metals that may arise from
manmade substances (particularly
metalloids) or from natural substances
(metallic ores) present in the environment
as a result of human activity that is not
specifically related to the release in
question (U.S. EPA, 2003e).
or, conversely, it may result in adaptation of organisms to
higher metal concentrations and result in increased
tolerance to emissions. Furthermore, because metals
occur naturally, and some are essential macro- or
micronutrients, they are at least partially responsible for
how plants and animals are distributed within various
ecoregions. The distribution of plants and animals, local
species diversity, species survival,, and the vitality of
individuals can be profoundly affected by background
levels of metals in an area. Humans, on the other hand,
are distributed throughout the world, irrespective of naturally occurring levels of metals.
The contribution of the background level of a metal(s) to the cumulative exposure of
people and other organisms may be significant and so should be considered in any human health
assessment (see Section 4.2.2.1). Lifestyle choices expose people to metals in many different
contexts that warrant consideration when assessing the added risk caused by a particular source.
However, the added risk from dietary or other point sources should be considered in light of the
relative bioavailability of the background and additional sources. Background metals generally
are reduced in bioavailability as a result of aging in soils or sediments (see Section 4.1.6.4) or
transformation to less bioavailable salts.
2.1.2. Essentiality
Some metals are essential to maintaining proper organism health and may cause adverse
effects when present at deficient or excess amounts. The influence of metals essentiality on
exposure and effects of the metal(s) of concern should be addressed to the extent practicable in
the assessment.
As a practical matter, essentiality sets a lower bound on the range of metal exposures to
be considered with respect to the potential for toxic effects. The following discussion highlights
the issue of essentiality; further discussion is found in Sections 4,3.2 and 4.5.1. Seven metals are
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
2-3
-------
1
2
3
4
5
6
known to be nutritionally essential for humans, and four others have been shown to have
possible beneficial effects (Table 2-1); these are reviewed in detail in Goyer et al. (2004). Plants
and other animals also depend on certain metals (See Table 4-14), and thus the generalizations
about essentiality apply for these organisms as well. The list of metals with no known beneficial
human health effects is longer, and some examples are given in Table 2-1.
Table 2-1. Classification of selected metals based on characteristics of health effects
Nutritionally
essential metals
Chromium III
Cobalt
Copper
Iron
Manganese (animals but
not humans)
Molybdenum
Selenium
Zinc
Metals with possible
beneficial effects
Arsenic
Boron
Nickel
Silicon
Vanadium
Metals with
no known beneficial
effects
Aluminum
Antimony
Barium
Beryllium
Cadmium
Lead
Mercury
Silver
Strontium
Thallium
Tin
1 The response of humans and other organisms to exposure to these metals is
2 conceptualized as having three phases: the Deficiency zone, the Optimal or inactive zone, and '
3 the Toxicity or toxicological action zone (Figure 2-1).
4
5 2.1.3. Environmental Chemistry
6 The environmental chemistry of metals strongly influences their fate and effects on
7 human and ecological receptors.
8 Table 2-2 identifies factors governing the chemistry of metals in sediments, soils, and
9 waters. Metals do not degrade but, rather, transform and exist as multiple interconverting
10 species, the exact mixture of which depends on the environmental chemistry of the medium.
11 Because the behavior of metals differs, it is necessary to understand the chemistry of the
12 particular metal and the environment or medium of concern. Still, some generalizations can be
13 made about factors that control metal chemistry and environmental characteristics. These allow
14 risk assessors to develop preliminary estimates of metal exposure and effects and are discussed
15 in Section 4.1.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
2-4
-------
(a)
Deficiency
Low Critical
Gonconltation
: HIDDEN
! OR
: MILD
Limits of Homeostasis
Optimal
Tenacity
Upper Critical
Cbncfiittatisn
E?
o
Element concentration >
No effect ; Toxicity
'jfipt-f Critical
Bsment eonGantration >
Figure 2-1. Health effect curves for (a) essential elements and (b) nonessential
metals.
Source: Fairbrother and Kapustka, 1997.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
2-5
-------
Table 2-2. Factors that primarily control metal sorption to soils, aquifers, and
sediments
Soil solids
Soil solution
Solutes
Soil mineral composition
Specific surface areas of metal-sorbing solids
Surface site density or cation exchange
capacity of metal-sorbing solids
Aeration status
Microbial type, activity, and population
Organic matter content and character
Temperature
pH
Eh
Dissolved oxygen
Solute composition
Dissolved organic carbon
Ionic strength
Temperature
Chemical identity
Complexation chemistry
Solubility
Precipitation chemistry
Redox behavior
Vapor pressure
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
2.1.4. Bioavailability
The bioavailability of metals and, consequently, the associated risk vary widely
according to the physical, chemical, and biological conditions under which an organism is
exposed. To the extent that available data and methods allow, factors that influence the
bioavailability of a metal should be explicitly incorporated into assessments. In situations where
data or models are insufficient to address bioavailability rigorously, the assumptions made
regarding bioavailability should be clearly articulated in the assessment as should the
associated impact on results.
The bioaccessibility, bioavailability, and
bioaccumulation properties of inorganic metals
in soil, sediments, and aquatic systems are
interrelated and abiotic (e.g., organic carbon)
and biotic (e.g., uptake and metabolism).
Modifying factors determine the amount of an
inorganic metal that interacts at biological
surfaces (e.g., at the gill, gut, or root tip
epithelium) and that binds to and is absorbed
across these membranes. A major challenge is
to consistently and accurately measure
quantitative differences in bioavailability
between multiple forms of inorganic metals in
the environment.
The bioavailability issue paper authors
(McGeer et al., 2004) provided EPA with some
Bioavailability, Bioaccessibility, and
Bioaccumulation
Bioaccessibility of metals is the portion of total metal
in soil, sediment, water, or air that is available for
physical, chemical, and biological modifying influences
(e.g., fate, transport, bioaccumulation). It is termed the
environmentally available fraction and also known as
environmental availability.
Bioavailability of metals is the extent to which
bioaccessible metals adsorb onto or absorb into and
across biological membranes of organisms, expressed as
a fraction of the total amount of metal the organism is
proximately exposed to (at the sorption surface) during a
given time and under defined conditions.
Bioaccumulation of metals is the net accumulation of
a metal in the tissue of interest or the whole organism
that results from exposure from all environmental
sources, including air, water, solid phases (i.e., soil,
sediment), and diet, and that represents a steady-state
balance of losses from tissue and the body.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
2-6
-------
r
1 practical, standard, and defensible recommendations on concepts, terms, and definitions that can
2 serve as a paradigm for studying metals and their bioavailability. A conceptual framework along
3 with further discussion of metals bioavailability and bioaccumulation is presented in Section 4; 1
4 and Figure 2-2.
BioaccessiUe Fraction (BF)*:
Percent soluble metal ion
concentration relative to total
metal concentration (measured in
solution near biorrerrbrane) ^
Relative Bioavailability
(RBA)>>: Percent adsorbed or
absorbed compared to
reference material (measure of
nBmbrane dynarrics)
Absolute bioavailability
(ABA)0: Percent of rretal mass
absorbed Internally corrpared to
external exposure (measures
systemic uptake/accurrulation.
Bioaccessibility \
ivailability
Environmental availability \
,..\ ......§l9^9Mmy!?flon gf metal
Cationte
competition
^r
I
| Soluble fraction | | Paniculate fraction]
1
J
Total Metal concentration |
Membrane
uptake
\
Benign /
accumulation
/
J. Internal
Transport &
Distribution
\
Toxicoloical
Accumulation
Physiological
«^ membrane
Detoxification
& Storage
Essentiality
Excretion
_ Predation ^
Foraging
Bioaccumulated Metal
1
Figure 2-2. Conceptual diagram for evaluating bioavailability processes and
bioaccessibility for metals in soil, sediment, or aquatic systems.
°BF is most often measured using in vitro methods (e.g., artificial stomach), but should be validated by in
vivo methods.
bRBA is most often estimated as the relative absorption factor, compared to a reference metal salt (usually
calculated on the basis of dose and often used for human risk, but it can be based on concentrations).
"ABA is more' difficult to measure and used less in human risk; it is often used in ecological risk when
estimating bioaccumulation or trophic transfer.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
2-7
-------
1 2.1.4.1. Bioaccessibility or Environmental Availability
2 The portion of total metal in soil, sediment, water, or air that is available for physical,
3 chemical, and biological modifying influences (e.g., fate, transport, bioaccurnulation) is termed
4 the environmentally available fraction. Environmentally available metal is not sequestered in an
5 environmental matrix, and it represents the total pool of metal in a system that is potentially
6 bioavailable to (able to contact or enter into) an organism. The bioaccessible fraction of metal is
7 the portion (fraction or percentage) of environmentally available metal (e.g., <250 urn diameter
8 for vertebrates) that actually interacts at the organism's contact surface and is potentially
9 available for absorption or adsorption (if bioactive upon contact) by the organism (see Figure
10 2-2).
11 Environmental availability refers to the ability of a metal to interact with other
12 environmental matrices and undergo fate and transport processes. Environmental availability is
13 specific to the existing environmental conditions and is a dynamic property, changing with
14 environmental conditions. As an example of environmental availability, the divalent cation of
15 copper (Cu2+) is available for interaction with the gills of a sediment-dwelling invertebrate,
16 binding to dissolved organic matter, and advective transport, whereas copper in the form of a
17 sulfide in sediments is not. Resuspension of sediments with copper sulfide may introduce
18 oxygen and result in the release of divalent copper into the water column, making it
19 environmentally available.
20
21 2.1.4.2. BioavaUability
22 The concept of metal bioavailability includes metal species that are bioaccessible and are
23 absorbed or adsorbed (if bioactive upon contact) by an organism, with the potential for
24 distribution, metabolism, elimination, and bioaccumulation. Metal bioavailability is specific to
25 the metal salt and particulate size, the receptor and its specific pathophysiological characteristics,
26 the route of entry, duration and frequency of exposure, dose, and the exposure matrix. To date,
27 for most metals, the treatment of bioavailability for human health assessments is to assume that
28 the bioavailability of the metal exposure from the site is the same as the bioavailability derived
29 from the toxicity study that has been used to derive the toxicity value (Reference Dose or Cancer
30 Slope Factor).
31
32
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
2-8
-------
1 2.1.4.3. National Research Council Report
2 The National Academy of Science (NAS), National Research Council (NRC) report on
3 bioavailability of contaminants in soils and sediments provides a broad overview of chemical
4 availability issues in environmental media and within biota (NAS/NRC, 2002). As illustrated in
5 Table 1-1 of the NRC document, many variations of terms and definitions are used for the
6 concept of "bioavailability." The NRC report reviews the history and nuances of the various
7 terms and meanings involving "bioavailability processes." The metals framework responds to
8 the NRC report's recommendations on making bioavailability processes visible in risk
9 assessments and improving the scientific basis supporting bioavailability in risk assessments. It
10 sets forth principles and tools that are responsive to the NRC recommendations, although the
11 NRC report lists additional tools, with their strengths and limitations, that also may be applicable
12 to evaluating bioavailability. Although the NRC report states a preference for mechanistic
13 approaches, decisions also may be made using assumption and degrees of uncertainty,
14 particularly for national-scale assessments and ranking.
15
16 2.1.5. Bioaccumulation and Bioconcentration
17 Organisms bioaccumulate metals through multiple mechanisms of uptake, distribution,
18 metabolism, and elimination. The highly complicated and specific nature of metals
19 bioaccumulation substantially hinders the ability to accurately predict bioaccumulation and
20 extrapolate results across species and exposure conditions, particularly when simplified models
21 are used (e.g., Bioaccumulation Factor, Bioconcentration Factor).
22 Because plants and animals have evolved in the presence of metals, some of which are
23 required for proper physiological functioning, they have developed a variety of physiological
24 and anatomical means to regulate the amount of metals in their tissues. For some metals, this
25 includes storage in various compartments (e.g., lead in bone or cadmium in kidney). Such
26 bioaccumulation of metals may cause no effects or may eventually result in a toxic response.
27 Should the organism be eaten by another (e.g., an herbivore eating a plant that has stored metal
28 in its foliage or a predator eating a mollusc that has accumulated metal granules), the stored
29 metal may (or may not) result in toxicity to the consumer, depending on the form in which it was
30 stored. Therefore, the mere presence of a metal in an animal or plant cannot always be used to
31 infer toxicity to either the organism itself or to its consumers. If the concentration of metal in an
32 organism is greater than the environmental media (soil, water, or sediment), then the metal is
33 said to bioconcentrate. If the organism's metal burden is from both environmental media and
34 food intake, then the metal is said to bioaccumulate. Biomagnification occurs in the food web if
35 each trophic level has a higher amount of metal than the one below it. This occurs readily for
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
, 2-9
-------
1 persistent organic substances that are highly lipophilic, but generally it is not the case for metals.
2 Organisms at all trophic levels have developed mechanisms to regulate internal metal
3 concentrations and generally exhibit toxic effects and die before tissue levels become very high.
4 Exceptions are organometallic compounds such as methyl mercury or organoselenium and the
5 specialized metal hyperaccumulating plants that use metal storage as a means of detoxifying
6 their environment or discouraging feeding by consumer organisms.
7
8 2.1.6. Acclimation, Adaptation, and Tolerance
9 Metals naturally occur at a range of environmental concentrations and are influenced by
10 local biogeochemical controls on metal cycling. Within limits, organisms have developed
11 mechanisms for coping with excess metals exposure (e.g., acclimation, adaptation).
12 Organisms have developed various mechanisms
13 to cope with variable background metal concentrations
14 through either active or passive uptake and elimination
15 processes. Additionally, organisms can acclimate to
16 suboptimal metal levels by changing various
17 physiological functions, and populations can undergo
18 genetic change (adaptation) and develop increased
19 tolerance to different levels (Rusk et al., 2004; Wallace
20 and Srb, 1961). For example, the fact that plants of
21 diverse taxonomic relationships can grow on soils high
22 in metals provides evidence of adaptation for metal
23 tolerance.
24 This ability for organisms to tolerate various amounts of naturally occurring metals
25 makes it difficult to generalize about effects levels that are applicable and consistent to all
26 organisms in all habitats. Furthermore, this capacity for change makes it important to acclimate
27 organisms to test conditions when setting up bioassays for toxicity tests. Conversely, results of
28 tests conducted with organisms reared in media with low natural metal levels may not be
29 representative of effects to organisms that normally experience high metal concentrations (or
30 vice versa). This raises difficult questions about general applicability of test data and relevance
31 for site-specific assessments. Concerns about adaptation or acclimation have less relevance for
32 humans, as there are only a few examples of development of metal tolerance among specific
33 populations.
34
Tolerance, Acclimation, and
Adaptation
Tolerance is the ability of an organism
to maintain homeostasis under a variety of
environmental conditions, such as variable
metal concentrations.
Acclimation is how an individual
develops tolerance during its lifetime, and it
may be gained or lost. Acclimation is also
called phenotypic plasticity.
Adaptation is a genetic change over
multiple generations as a response to natural
selection. Traits are not lost during single
lifetimes. Adaptation is also known as
genotypic plasticity.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
2-10
-------
1 2.1.7. Toxicity Testing
2 Owing to limitations in available data and test methods, application of laboratory-
3 derived toxicity data often requires extrapolation of results across test species, metal
4 compounds, and exposure conditions that affect bioavailability. Toxicity data should be
5 expressed in a manner comparable to environmental exposure estimates, thus accounting for
6 bioavailability, tolerance (acclimation and/or adaptation), and species-response effects. Toxic
1 thresholds for essential elements should be set at levels higher than required daily intake.
8 Homeostatic mechanisms regulate the toxic response of an organism to metals. There are
9 a variety of ways, however, in which homeostatic control mechanisms can be overwhelmed or
10 circumvented, resulting in a toxic effect of the metal. Because soluble metal salts are often used
11 in toxicity testing, test experimental designs may need to be adapted to assess more common
12 environmental forms of metals, competition between essential and nonessential metals, and other
13 factors.
14
15 2.1.8. Mixtures . , .
16 Metals frequently occur as mixtures owing to their natural abundance in the environment
17 and the dietary essentiality of some metals for normal physiological functioning. Metals may
18 interact either synergistically, additively, or antagonistically in various ways, depending on the
19 combinations of metals and their relative amounts.
20 The presence of multiple metals can lead to competition among these metals for the
21 complexation capacity of the water, resulting in a decrease in complexation capacity relative to
22 what would be available for any single metal alone. This has direct implications to the
23 evaluation of metal availability and the potential for adverse effects. Such interactions are most
24 important when considering low-effect levels for the metal of interest, increasing in importance
25 as the concentrations of competing metals increases. Another problem with multiple metals is
26 that toxic interactions could exacerbate effects on the target organism. This could be in the form
27 of a single effect being exacerbated, as would be the case when the two metals have the same
28 mode of action (e.g., copper and silver affecting sodium regulation, or zinc and cadmium
29 affecting calcium regulation), or it could result in an organism being affected in different ways at
30 the same time when the modes of action differ. Metal interaction might also lead to a decrease in
31 the rate of uptake by one at the expense of the other, not only when trace metals such as
32 cadmium, strontium, or zinc interact with a hardness cation such as calcium, but when metals
33 such as lead and copper interact with each other as well.
34 Metals are normally found in the environment as mixtures with other metals as well as
35 organic compounds. Two key questions should be asked when assessing metals mixtures. First:
36 To what extent does each metal contribute to any observed joint effect? (Recognize that when
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
2-11
-------
1 the relative compositions of metals in the mixture change, this change the answer.) Second:
2 Are the effects significantly greater or lesser than the sum of the individual component effects?
3 Some rnetals can exert a protective or sparing effect when combined with others, thereby
4 mitigating risk (antagonistic) while others can enhance the toxic effect (synergistic). For
5 example, zinc may mitigate mercury toxicity, and copper may have a protective effect against
6 cadmium poisoning, while copper deficiency can enhance the effects of lead.
7 ' .
8 2.2. METALS CONCEPTUAL MODEL
9 The relationships between the sources, exposure, and effects of metals to human and
10 ecological receptors are complex and specific to the site, environmental condition, and receptor
11 organism. Because metals are naturally occurring substances, transition functions between
12 environmental loadings, media concentrations, exposed receptors, and the final organismal or
13 ecosystem responses are affected by natural processes to a much greater extent than occurs with
14 xenobiotic organic contaminants. These transition functions should be specifically identified in
15 the conceptual model for all metals assessments.
16 A conceptual model depicted in Figure 2-3 shows the interrelationship between the
17 metals or metal compounds of interest and the assessment process. It is a representation of the
18 actual and potential, direct and indirect relationships between stressors in the environment and
19 exposed humans (or particular subpopulations) or ecological'entities. The model depicts
20 possible pathways from sources of metals to receptors and includes environmental or biological
21 processes that may influence the predominant route of exposure or the physical/chemical
22 properties of the metal compounds.
23 The goals and scope of an assessment, in addition to the availability of data, methods,
24 and resources, are among the most important factors that determine the extent to which key
25 metal principles should be incorporated into an assessment. Generally, assessment endpoints are
26 selected during the problem formulation phase of all risk assessments based on their relevance to
27 management goals, societal values and laws, known adverse effects of metals, and endpoints of
28 importance to stakeholders. Risk assessors will incorporate metal principles to a lesser extent in
29 screening level assessments than in definitive risk assessments. Site-specific assessments can
30 account for more metal-specific processes (particularly, environmental chemistry) than can
31 national-level assessments that require generalization across multiple ecoregions. Therefore, it is
32 recommended that, when appropriate, regional- or national-level risk assessments be subdivided
33 into metal-related ecoregions, known as "metalloregions" (McLaughlin and Smolders, 2001),
34 such that protection levels, mitigation goals, and ranking results will be appropriate for the suite
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
2-12
-------
1 of species naturally present within each type of controlling environment. This is directly
2 analogous to the use of ecoregions when establishing water quality criteria (Griffith et al., 1999).
3 The problem formulation phase of the assessment should clearly identify whether a regional
4 approach is being used and, if so, how the metalloregions are defined in terms of species
5 composition and environmental controlling factors.
6 This concept of regional-based ecological assessments is significantly less important in
7 human health assessments. In these assessments, the environmental controlling factors (pH,
8 water hardness, etc.) may be important determinants in exposure calculations for dietary or
9 drinking water exposures. However, humans have not adapted to particular areas of metal
10 enrichment or impoverishment but, rather, choose to live in all environments. Therefore, the
11 differences in human sensitivity that should be considered are not geospatially correlated.
12 Rather, consideration should be given to the identification of potentially sensitive
13 subpopulations, such as the very young or the elderly, those with genetic predispositions to metal
14 sensitivity (e.g., Wilson's disease), or other similar groups (see Section 4.3, Human Health
15 Effects). Again, the problem formulation phase should clearly state whether the risk results will
16 be applied on a population-wide basis, such that protection is afforded to the most sensitive
17 individuals, or whether these groups are given additional scrutiny and separate risk analyses,
18 such mat results will be applicable only to the general population.
19 Areas in the conceptual model that stand out as metal-specific issues are identified in
20 Figure 2-3 as the transitions between environmental loadings, media concentrations, exposure
21 receptors, and the final organismal or ecosystem risk. Because metals are naturally occurring
22 substances with which organisms have evolved, it is particularly important to incorporate into
23 the risk assessment the natural processes that affect metal mobility, speciation, sequestration, and
24 toxicity. These may differ in details or approach, depending on the environment of concern
25 (water, land, air), the final receptor organisms (humans, animals, plants), and whether the
26 management goal is health of individuals or maintenance of populations and communities of
27 organisms. However, the same basic concepts always arise, regardless of the assessment
28 context.
11/24/2004 Peer Review D raft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
2-13 •
-------
1
(D
CO
4
1
0)
zr
O
CO
a
o
O
g
'fc
§
O
^
?l
BE
Is
fe w
•5 JS
Si I
fl !
CO* ><
cfc S
SS i
« SO TJ
8& S"
8-1 ^
g^'
8(0
ja
fis I| i
111 11 I
3 fi JS * o P
^r
=
£
of
I"
! » 3 a> c
is hi
i *'
.c. tii .
CO
•2
wa
•a
,e
^
e
E
•a
a
-------
1 The conceptual model identifies the following issues, indicates where within the risk
2 assessment process they occur, and helps direct the remainder of the assessment.
3
4 • Environmental chemistry (Ml). The partitioning of the metals or metal compounds
5 of concern into the various environmental media from the loading source is
6 dependent on the physical properties of the initial form of the material and the
7 particular chemistry of the receiving environment. Models are useful to estimate
8 speciation, transition kinetics, and potential resuspension of the material within the
9 context of natural background levels of the metal and other inorganic substances.
10 These can be very detailed for site-specific assessments, or they can provide a
11 potential range of processes that might occur over large regional scales for
12 assessments of a more generic nature (e.g., criteria development or ranking schemes).
13 The degree of influence of various environmental attributes on the final distribution
14 of the metals of concern into the various media also can be identified through the
15 application of appropriate models.
16
17 • Exposure models (M2). Estimating uptake of metals from environmental media into
18 biota follows many of the same processes as for organic substances, such as
19 understanding trophic relationships, dietary preferences, and movement patterns.
20 However, metal-specific issues arise owing to the variable solubility of metal
21 complexes, essentiality of some metals for organismal functions, and naturally
22 evolved processes for uptake, sequestration, or exclusion of these materials.
23
24 • Accumulation and bioaccumulation/physiologically based toxicokinetic (PBTK)
25 and toxicodynamic models (M3). Although many organic substances require
26 metabolic activation to become toxic or, conversely, to be detoxified and excreted,
27 metals do not. Metals may form a complex with proteins or other carrier molecules
28 for distribution to target organs or for sequestration and excretion. They typically do
29 not bioaccumulate or biomagnify within the whole organism, although they may do
30 so in particular tissues (e.g., lead in bone or cadmium in kidneys). For example, in
31 aquatic systems, the amount of metal taken up in the aquatic environment is
32 proportional to the external concentration (and a function of the internal
33 concentration), so it is not a constant. This is a particularly important distinction
34 between metals and organic substances and is a central aspect to the conceptual
35 approach for assessing risks of metals. It is equally important to understand how
36 different groups of organisms react to metal loading (as accumulators, excluders, or
37, sequesters) to accurately predict potential for immediate or delayed toxicity of metals
38 in the environment. Interactions among metals, particularly for the essential
39 elements, may significantly affect the toxicodynamics of the metal(s) of interest,
40 especially when exposure occurs via complex mixtures of substances. Finally, the
41 near-term experience of organisms with metals (acclimation) or long-term species
42 history (adaptation) can significantly affect how metabolic pathways are adjusted to
43 accommodate higher- or lower-than-normal metals loading.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
2-15
-------
1
2 • Residue-based toxicity models (M4). If risk to the organism(s) of concern is to be
3 based on an estimate of internal dose, then information about the relationship of
4 whole-body (or target organ) residue levels to toxic responses should be understood,
5 either from empirical data or PBTK/toxicodynamic models. Because of the processes
6 discussed in the previous paragraph, this can be particularly challenging for metals.
7 Metal speciation in the exposure matrix can significantly influence this relationship
8 because uptake and organ distribution kinetics are likely to differ.
9
10 • Accumulation and bioaccumulation/food web model (MS). This node of a
11 conceptual model applies to ecological risk assessments and, to a lesser extent,
12 human health assessments. Movement of inorganic metals and metal compounds
13 through the food web (or up the dietary pathway for humans) is complicated by
14 factors of bioavailability., essentiality., background concentrations, and natural
15 adaptive capacity of organisms.
16
17 • Indirect exposure model (M6). The exposure of an organism of concern is
18 dependent on its location within the trophic structure of the community and its dietary
19 preferences. Although this node of the conceptual model differs very little from risk
20 assessment approaches for organic substances, some metal-specific generalities about
21 the relative importance of exposure pathways can be applied to focus (and simplify)
22 the process.
23
24 • Exposure-based toxicity model (M7). Calculation of appropriate external dose (oral
25 intake, gill binding, etc.) for comparison with toxicity thresholds depends on
26 information about relative bioavailability (RBA), speciation of the metal or metal
27 salt, dietary preferences and rates, natural background concentrations, essentiality,
28 and metal interactions. Toxicity threshold considerations should be based on
29 comparable information., such as appropriate metal species in exposure media,
30 similarly acclimated or adapted organisms, similar exposure routes, and appropriate
31 combinations of essential metals.
32
33 • Media-based toxicity model (M8). This risk assessment model compares
34 environmental concentrations with organism response functions without calculating a
35 body burden or internal dose. It is used more frequently for aquatic and soil-dwelling
36 organisms, less frequently for wildlife, and very infrequently for human health
37 assessments. Consideration of RBA, trophic transfer rates, dietary preferences,
38 natural background concentrations, and organism adaptations is important for a
39 metals assessment.
40
41 • Population, habitat, ecosystem models (M9). Ecological risk assessments often ask
42 questions related to population growth, habitat change, or ecosystem functions in
43 addition to questions related to risks to individual organisms. Most of the models and
44 approaches are similar for both metal and organic substances. However, metals and
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
2-16
-------
1 other inorganic substances are among the fundamental determinants and delimiters of
2 ecoregions (in conjunction with climate, elevation, and day length associated with
3 latitude). Therefore, a knowledge of natural background and adaptation of organisms
4 to differing metal levels is essential in developing appropriate risk factors for
5 naturally occurring species.
6
7 In summary, the conceptual model lays out a series of working hypotheses about how the
8 metal(s) of concern might move through the environment to cause adverse effects in humans or
9 ecological systems. These hypotheses are examined through data analysis, models, or other
10 predictive tools to determine the probability and magnitude of occurrence of unwanted effects.
11 The approaches used to accomplish this are discussed in general within various Agency risk
12 assessment guidance documents.
13
14 2.3. NEXT STEPS
15 Chapter 3 of this framework provides the risk assessor with a tool box in the form of key
16 recommendations and important considerations for undertaking the risk analysis for human
17 health, aquatic, and terrestrial receptors potentially exposed to metals or metal compounds. It
18 includes recommendations for consideration of metals fate and transport, exposure, and effects.
19 The fundamental metals principles, outlined earlier in Section 2.1, that should be considered
20 throughout metals risk assessment are integrated into these, recommendations as appropriate.
21 Chapter 4 of the framework expands on the supporting components of the recommendations and
22 provides the risk assessor with a more indepth discussion of the strengths, limitations, and state
23 of the science of the tools and methods available for metals risk assessment. Section 5 of the
24 framework discusses metals research needs.
25
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
2-17
-------
1 3. METALS RISK ASSESSMENT RECOMMENDATIONS
2
3 The metals risk assessment recommendations are a formulation of the primary concepts
4 and factors that should be considered when conducting assessments of the ecological or human
5 health risks associated with metals exposures. The fundamental metals principles, outlined in
6 Section 2.1, that should be considered throughout metals risk assessment are integrated into
7 these recommendations as appropriate. The recommendations are not intended to provide a
8 prescriptive step-by-step guide on conducting metals risk assessment. Rather, they are intended
9 to promote the consistent application of the various tools and methods currently available to risk
10 assessors so that assessments take into consideration the unique properties of metals. Moreover,
11 these recommendations are presented with the intent that they be used in parallel with currently
12 available Agency guidance for human health and ecological risk assessment.
13 Section 4 provides the risk assessor with an expanded discussion of the underlying issues
14 and methods that form the basis for the recommendations in Section 3. In addition, the metals
15 recommendations rely on the final metals issue papers, which have been developed by scientists
16 commissioned under contract by EPA to identify current information, tools, and methods in
17 metals science. The final metals issue papers are available on the EPA Web site at
18 http://cfpub.epa.gov/ncea/raf/recordisplay.cfm?deid=86119. Section 5 reviews research needs
19 based on the limitations in the currently available tools and methods; it also reviews gaps in our
20 understanding of metals behavior and effects on humans and the environment.
21
22 3.1. HUMAN HEALTH RISK ASSESSMENT RECOMMENDATIONS
23 Assessing the risks of metals to human health is similar to conducting assessments of
24 organic substances. However, risk assessors should pay particular attention to the metal-specific
25 principles discussed in Section 2. These include the influence of environmental chemistry on
26 speciation, bioavailability, natural and ambient background levels of metals in the environment,
27 and the ubiquitous presence of metal mixtures. Additionally, certain properties associated with
28 toxicokinetics and toxicodynamics, carcinogenesis and non-cancer, and the sensitivity of
29 particular subgroups, are unique to metals and should be incorporated into the risk assessment
30 process. Data from standard toxicity tests should be interpreted appropriately in light of these
31 metal-specific attributes. Tools are available for addressing most of these metal-specific
32 attributes to one degree or another during a human health risk assessment
33
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
3-1
-------
I 3.1.1. Fate and Transport
I Transport or movement of metals through the physical environment (soil and sediment,' '
3 surface and ground water, and air) is a function of the source characteristics (e.g., discharge
4 point), the nature of the metal species or compound (e.g., chemically reactive or recalcitrant;
5 dissolved, complexed, or solid/precipitate), and the physical/chemical characteristics of the
6 receiving environment. Generally, the fate and transport of metals are not greatly affected by
7 macrobiota (e.g., via tracking), but the influence of local microbial processes can be significant
8 (including mobilization in the root zone, and methylation) (see Section 4.1.9) Therefore, such
9 analyses are equally relevant to human health and ecological risk assessments. Consequently,
10 recommendations pertaining to the fate and transport of metals are provided in Section 3.2. for
11 the aquatic environment (e.g., water, sediment, associated biota), and in Section 3.3. for the
12 terrestrial environment (e.g., air, soil, associated biota).
13
14 3.1.2. Exposure Assessment ' .
15 Assessing human exposures involves evaluating how people can be exposed to metals
16 and in what amounts. Specific steps include (1) identifying the forms (speciation) and
17 concentrations of metals in the media to which people are exposed (e.g., soil, water, air, or biota,
18 accounting for setting-specific biogeochemical conditions); (2) describing how people contact
19 these media, by what specific routes they are exposed (e.g., incidental ingestion of inhalation);
^ ' 20 (3) determining the appropriate exposure metric (e.g., oral intake, air concentration, or blood
21 concentration), and quantifying that metric (e.g., estimating the amount ingested, or determining
22 the blood concentration); and (4) characterizing the uncertainty and natural variability associated
23 with these estimates, where possible.
24
25 3.1.2.1. Background
26 Metals, being part of the periodic table of the elements, constitute a portion of the natural
27 background to which all humans are exposed. Metals are an important constituent of the diet
28 and the Earth's crust. For example, 95% of the general population's exposure to copper is
29 through the diet, and concentrations of natural levels of organic arsenicals in seafood can be
30 sufficiently high to cause the urinary arsenic concentration to resemble that of an occupational
31 exposure. Risk assessments addressing anthropogenically introduced metals should separate
32 such introductions from natural levels, or levels associated with other contributions beyond the
33 scope of the assessment, and account for the possibility of cumulative exposure. The former is
o
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
3-2
-------
1 Estimation of intake of metals in drinking water requires information about
2 concentrations of metals in the finished water and the amount of water consumed. Metal (]
3 concentrations in drinking waters are measured at the distribution point for municipal water
4 delivery systems. The contribution of metals from pipes (either from the distribution system to
5 the home or within the home) is rarely assessed. Water delivered from private wells or ambient
6 surface waters may contain higher levels of organic carbon or other ligands to which metals can
7 bind, thereby requiring an adjustment to account for differential bioavailability of the dissolved
8 metals. These factors can be incorporated into site-specific assessments, but local data will need
9 to be collected on a case-by-case basis. See Section 4.2.2.5 for further discussion.
10 Recommendations: . , .
" , '
12 • Regional differences in bioavailability of metals due to variation in water
13 characteristics (e.g., hardness), the contribution of household distribution systems
14 to total metal load at the tap, and lack of information from households on private
15 wells should be included as uncertainties in national-level risk assessments.
16
17 • It is recommended that site-specific assessments use measured metal
18 concentrations at the tap.
19 .
20 . . • Although people can be exposed to metals dissolved in ambient surface water
21 during swimming, other recreational activities, or various occupational activities,
22 dermal absorption can be considered a negligible exposure pathway.
23 . .
1
24 3.1.2.5. Integrated Exposure Approaches
25 The many pathways that might be associated with metals exposure represent an important
26 consideration for the risk assessor. For lead, EPA has developed the Integrated Exposure Uptake
27 Biokinetic (IEUBK) model to address these pathways. The model simulates the transfer of lead
28 into the blood, with consideration of absorption and elimination as a result of inhalation and oral
29 exposures. Similar models for other metals have not yet been developed. While similar models
30 for other metals have not yet been developed, there are a variety of exposure models that could
31 potentially be applied in metals risk assessment (e.g.; SHEDS, DEPM, TRIM, RESHRAD).
32 These models, and their applications and limitations, are discussed in Section 4.2.4.1.
33 ' Recommendation: v '
34 '••'•••
35 • The IEUBK model for lead in children is recommended for use in all site-specific
36 assessments. It is available on line at
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information,
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
3-5
o
-------
1 http://www.epa.gov/superfund/programs/lead/ieubk.htm. The IEUBK model
2 may be used to account for lead accumulation resulting from differential rates of
3 uptake and elimination. This model is not applicable to other metals.
4
5 3.1.2.6. Bioavailability
6 Bioavailability of metals is an important consideration in human health assessments,
7 particularly when assessing risk of metals in soil. Metals may form organic complexes in water,
8 bind to soil substrates or be incorporated into soil matrices, or form complexes with food
9 materials, and the bioavailability of the metal in the study on which the dose-response
10 • assessment is based (e.g., animal feed, water, or com oil) may differ from that of the metal in the
11 matrix being assessed (e.g., soil). Different approaches have been used to address
12 bioavailability, depending on the data available. These approaches have been advanced
13 particularly with regard to soil ingestion. EPA's OSWER has developed a relative
14 bioavailability approach to adjust the RfD. Further discussions on bioavailability issues are
15 presented in Sections 2 and 4.2.4.2., and in McGeer et al. (2004), NRC (2002), and U.S. EPA
16 (2004).
17 Recommendations:
18
19 To address relative bioavailability, the preferred approach is to use information
20 directly relevant to the conditions being assessed, e.g., data from animal
21 toxicology studies that use the metal form encountered in that environment.
22
23 • Where this information is not available (e.g., in screening-level assessments), the
24 recommended approach is to assume that the bioavailability of the metal in the
25 medium being assessed is the same as that of the metal in the study on which the
26 dose-response (e.g., RfD, RfC, or cancer slope factor) was based.
27
28 • For higher level, definitive risk assessments, a medium-specific default absorption
29 factor for the metal may need to be used, and in such cases the factor should be as
30 representative as possible of the conditions being assessed.
31
32 • For lead, the juvenile swine model presently is the preferred animal model to use
33 for development of relative bioavailability factors (U.S. EPA, 2004a).
34
35 • When validated, in vitro methods for determining lead relative bioavailability in
36 soils may be used in place of animal studies (U.S. EPA, 2004a).
37
38
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
3-6
-------
1 3.1.3. Effects Analysis
2 3.1.3.1. Physiologically Based Pharmacokinetic (PBPK) and Physiologically Based
3 Pharmacodynamic (PBPD) Modeling
4 Combined use of PBPK and PBPD models provides understanding of the complex
5 relationships between exposure and target organ effects. These models are valuable risk
6 assessment tools for purposes of interspecies, high dose/low dose, rout to route, and exposure
7 scenario extrapolation. PBPK models that include a fetal compartment are particularly valuable
8 for human health risk assessment of metals, as transplacental exposures cannot be directly
9 measured from environmental measurements. An additional specific recommendation regarding
10 the IEUBK model is provided in Section 3.1.2.5.
11 Recommendations:
12
13 • When using PBPK models or other dosimetric adjustments, absorption/distribution
14 and kinetic factors should be considered explicitly. The models require special
15 considerations for cellular uptake, interaction with nutritionally essential and
16 nonessential metals, protein-binding behavior and function, incorporation into bone,
17 metabolism, and excretion (see Table 4-11 in Section 4.2.6).
18
19 • Application of a PBPK model to risk assessment should satisfy the following key
20 criteria: (1) identifying toxic or active form(s) of the metal, (2) selecting the
21 appropriate dose metric, and (3) identifying the appropriate target organ or cells on
22 the basis of the health effect of greatest concern.
23
24 • These available models for lead (see section 4.2.6.1) can be used to estimate fetal
25 exposure but are limited in that they rely on assumptions of a steady state between
26 maternal and fetal blood lead concentrations, which is violated if the mother is no
27 longer exposed to lead during pregnancy. This assumption should be stated in the
28 risk characterization.
29
30 • Similar models for other metals are not available at this time, so transplacental
31 transfer cannot be estimated for other metals.
32
33 3.1.3.2. Essentiality
34 Some trace elements, such as cobalt, copper, iron, and zinc (Table 4-12 in Section 4.3.2),
35 are necessary for biological functions and the normal development of humans, other animals, and
36 plants. These metals are required for organism health at one range of concentrations and can be
37 toxic at higher quantities. For essential elements that exhibit biphasic dose-response curves,
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. - It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
3-7
-------
1 adverse effects resulting from deficiency should be considered as well as those that result from
2 excessive exposure.
3 Recommendations:
4
5 • In setting reference values (Reference Concentrations [RfCs]/RfDs) the
6 Recommended Daily Allowance (RDA) should be taken into consideration.
7
8 3.1.3.3. Toxicity Testing
9 At least five transition metals or metalloids—arsenic (through drinking water exposure),
10 cadmium, chromium VI, beryllium, and nickel (from pulmonary or, in some cases, dermal
11 exposures) — are accepted as human carcinogens in one form or another or in particular routes
12 of exposure (NTP, 2002). Inorganic lead compounds are considered probable human
13 carcinogens by EPA's IRIS program, and I ARC has concluded there is limited evidence of
14 carcinogenicity to humans. Target organ sites for metals as carcinogens are summarized by
15 Waalkes (1995). Many noncarcinogenic effects can also be caused by exposures to metals. See
16 Section 4.3.5 for further discussion.
17 Recommendations:
18
19 • It is recommended that metals risk assessments follow the same approach for
20 carcinogenicity assessment as used for all other substances (U.S. EPA, 2003b), with
21 modifications as indicated by unique metal-specific data or methodology that may
22 exist.
23
24 • In parallel, the same general approach applied to assess non-cancer endpoints for
25 other chemicals should also be followed for metals, except where modifications are
26 warranted.
27
28 For example, an extensive model has been developed to evaluate lead, but a completely
29 parallel model does not exist for assessing a wide range of other metals and other chemicals. For
30 arsenic, unlike many other chemicals, the standard toxicity test data for rodents are not the
31 primary basis for assessing human toxicity.
32
33 3.1.3.4. Metals Mixtures
34 Metals are normally found in the environment as mixtures, and risks may be mitigated or
35 enhanced by any associated interactions. There are generally three classes of interactions
36 between metals: (1) between nutritionally essential metals; (2) between non-essential metals;
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
3-8
-------
1 and (3) between essential and non-essential metals. Additionally, the relative dose and mixture
2 composition, as well as mode of action of each metal in the mixture, is especially important in
3 determining overall health consequences. One form of interaction, termed "molecular" or
4 "ionic" mimicry, is an important consideration in evaluating the health effects of metals.
5 Selenium, for example, may play a protective role with regard to ingested arsenic. See Section
6 4.3.6 for further discussion.
7 Recommendations:
8
9 • For metal mixtures, as for other mixtures, the assessor is referred to EPA's guidelines
10 and supplemental guidance for the assessment of mixtures (US EPA, 1986a; US EPA,
11 2000b).
12
13 • Unless other information is available, the default approach is to assume dose
14 additivity for individual metals that produce the same effects by similar mode of
15 action.
16
17 • In the case of metals with known differences in critical effects, separate effect
18 assessments are encouraged for each metal. Interactive effects (e.g., synergism,
19 antagonism) should be included if information as available or acknowledged as a
20 source of uncertainly.
21
22 • While additional research is needed to explore the issues of metal mimicry and its
23 effect on assessment, site assessments should consider the available information on
24 metal mimicry related to the metal(s) of interest, including that referenced in this
25 framework.
26
27 3.1.3.5. Sensitive Subpopulations and Life Stages
28 Metals have a relatively robust literature on susceptible populations. For example, the
29 efficiency of intestinal uptake of zinc declines in the elderly, absorption of cadmium may be
30 greater in premenopausal women with depleted iron stores, and various genetic polymorphisms
31 have been associated with the effects of metals, including copper, iron, lead, and nickel. See
32 Section 4.3.7 for further discussion.
33 Recommendations:
34
35 • For metals, as with other agents, Agency risk assessments should consider
36 subpopulations with differing sensitivities that may arise as a result of differential
37, exposure or susceptibility.
38
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
3-9
-------
1 3.2. METALS RISK ASSESSMENT RECOMMENDATIONS FOR AQUATIC
2 ENVIRONMENTS
3
4 This section presents recommendations for assessing exposure and effects of metals in
5 aquatic systems. Many of the suggested methods for fate and transport apply to human as well
6 as ecological risk assessments. Recommendations given are specific to aquatic biota residing in
7 the water column and sediments, unless otherwise noted. Particular methods and models
8 described in the recommendations are discussed in more detail in Section 4.4, the metal issue
9 papers, and referenced citations.
10
11 3.2.1. Fate and Transport
12 Approaches for evaluating the fate and transport of metals in aquatic environments
13 include measurements and models. Selection of methods will depend on the nature and scale of
14 the problem. For example, for applications where metals have already been released to the
15 environment (e.g., at a hazardous waste site), measurements of the distribution of the metals
16 among environmental compartments is often an important starting point for the assessment. For
17 applications where decisions need to be made on wasteload allocations., such as for
18 determination of Total Maximum Daily Loads (TMDLs), models play a central role; these are
19 typically supported by measurements. For national assessments, generic models that capture a
20 range of conditions may be most appropriate. A suite of measurement and modeling tools exists.
21 Selection of the most appropriate combinations should be commensurate with the problem at
22 hand. In many cases, relatively simple approaches will suffice, and in others more sophisticated
23 methods are warranted. A complicated model is not necessarily a "better" model to use than a
24 simple one; the appropriateness for use of a specific model needs to be decided on a case-by-case
25 basis.
26 Measurement programs that support evaluations of fate and transport are addressed in
27 other Agency guidance and are not addressed here; however, these should be considered an
28 appropriate part of assessments. Measurement approaches for evaluating the bioavailability of
29 metals are discus sed later in this subsection. Recommendations related to the use of fate and
30 transport models for the assessment of metals in aquatic environments are provided below.
31 An important parameter in modeling the transport and fate of metals in aquatic systems is
32 the partition coefficient. The partition coefficient sets the distribution of the metal between the
33 dissolved and sorbed phases; this, in rum, has important implications to the magnitude of
34 paniculate and diffusive transfers between the water column and sediment, and in assessing
11/24/2004 Peer Review Draft
DISCLAIMER: Tliis information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
3-10
-------
1 metal bioavailability. However, assuming equilibrium conditions is a potential
2 oversimplification in some situations (e.g., in the vicinity of a point source discharge to a
3 receiving water or during a pulse exposure), and this could be a significant source of uncertainty
4 in the exposure assessment. See Section 4.4.1 for further discussion.
5 Recommendations:
6
7 • The reviews by U.S. EPA (1997) and Paquin et al. (2003), include up-to-date
8 information with regard to the availability of models appropriate for use in evaluating
9 fate and transport of metals in aquatic environments. These reviews also include
10 descriptions of example applications of many of the models that are discussed. These
11 reviews and Section 4.2.2.1 provide a good starting point for selecting among
12 available models.
13
14 • Relatively simple steady-state and analytical solution models may be appropriate for
15 use in the screening level. Use of conservative assumptions at this level of analysis
16 may provide insight about whether or not it will be necessary to complete a more
17 detailed, definitive risk assessment.
18
19 • Relatively complex and sophisticated time-variable models are appropriate for use in
20 higher level, definitive assessments. These models, in particular, should always be
21 used by an analyst who is experienced in the use of models and familiar with the
22 structure of the model being employed. Predictions of fate and transport of metals in
23 aquatic systems may be accomplished by using integrated models that include
24 hydrodynamic, sediment transport, and chemical transport algorithms, or by using
25 stand-alone hydrodynamic and/or sediment transport models that should then
26 interface with a chemical fate model. The advantage of the former approach is that
27 integration of the hydrodynamic, sediment, and chemical transport results takes place
28 in a seamless manner with limited need for intervention by the analyst. This is in
29 contrast to use of the stand-alone models, where the output of one model needs to be
30 formatted in such a way that it is amenable to use by the subsequent models that are
31 to be applied. An advantage of the latter approach is that it can in some instances
32 reduce the need to rely on lengthy model run times.
33
34 • Most of the available transport models do not currently include chemical speciation
35 subroutines. In such cases, chemical equilibrium models such as MINTEQ serve as
36 useful alternatives for characterizing the forms of the metal that are present.
37
38 • Water quality analyses often require probabilistic results, as the Water Quality
39 Criteria (WQC) specify not-to-exceed concentrations for a once-in-3-year return
40 period. Steady-state models are not directly amenable to evaluation of a return period
41 for WQC exceedences. When using these models, a Monte Carlo analysis can be
42 conducted to generate a large number of model inputs and subsequent solutions,
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under-applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
3-11
-------
1 which can then be analyzed statistically to characterize the probability of an
2 exceedence of the effect level of interest. Time-variable models that generate long-
3 term time series results (e.g., 20 years) can also be statistically analyzed to evaluate
4 the frequency of exceedences.
5
6 • The risk assessor should consider the potential for situations where nonequilibrium
7 conditions exist, as this may have important implications to the partitioning of a metal
8 between dissolved and particulate phases and to the characterization of metal
9 speciation and bioavailability that is provided by a chemical equilibrium model (see
10 Section 4.1.4).
11
12 • An important parameter in modeling the transport and fate of metals in aquatic
13 systems is the partition coefficient. The partition coefficient sets the distribution of
14 the metal between the dissolved and sorbed phases; this, in turn, has important
15 implications to the magnitude of particulate and diffusive transfers between the water
16 column and sediment, and in assessing metal bioavailability. However, assuming
17 equilibrium conditions is a potential oversimplification in some situations (e.g., in the
18 vicinity of a point source discharge to a receiving water or during a pulse exposure),
19 and this could be a source of uncertainty in the exposure assessment. See Section
20 4,1.4 for further discussion.
21
22 • Special care should be taken when modeling metals (e.g., chromium) and metalloids
23 (e.g., mercury and arsenic) that can readily change oxidation state or undergo
24 transformation. Such changes affect physical and biological properties. Although
25 . many of the same transport models can be used, input parameters will require
26 modification.
27
28 3.2.2. Water Column Exposure, Bioavailability, and Effects
29 Potential exposure routes for aquatic species include inhalation/respiration, dermal
30 absorption, and dietary (from either food or incidental sediment ingestion). However, owing to
31 the diversity of aquatic organisms, the extent to which a metal is taken up by any one of these
32 ^ exposure routes is difficult to define for all relevant routes. With those limitations in mind, most
33 of the focus on exposure and effects has been on the binding of metals to the gill surface for
34 short-duration exposures and the resultant toxicity.
35 Approaches for evaluating exposure and effects include measurements and chemistry-
36 based exposure models. Laboratory and in situ toxicity tests and field assessments of biota have
37 all been used to evaluate metals. These will continue to have an important place in developing
38 chemical-based methods and in assessing conditions at particular water bodies. These
39 approaches have been described in other Agency guidance and are not repeated here.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
3-12
-------
1 Accounting for bioavailability should be an essential part of metal assessments; however,
2 a host of factors can influence the amount of metal that interacts and is taken up across
3 biological surfaces, and water quality can have a dramatic influence on toxicity. Depending on
4 the level of assessment, different methods and approaches can be used to incorporate
5 bioavailability.
6 " Ambient Water Quality Criteria (AWQC) are developed to support the Clean Water Act,
7 and since the 1980s aquatic life criteria for several cationic metals have been expressed as a
8 function of water hardness to address the relationship of certain cations on toxicity. However,
9 water hardness adjustments do not account for other important ions and ligands that can alter
10 metals bioavailability and toxicity. The Water Effect Ratio (WER) was developed in the 1990s
11 to address this issue directly. This is an empirical approach to make site-specific bioavailability
12 adjustments to criteria (U.S. EPA, 1994). This approach relies on comparing toxicity
13 measurements made in site water with those made in laboratory water to derive a WER. The
14 WER is then used to adjust the national criterion to reflect site-specific bioavailability.
15 Recent developments in understanding the physiology and toxicology of metals have
16 enabled further advances in incorporating bioavailability into assessments. Different forms of
17 the dissolved metal have differing bioavailabilities; free metal ions generally are the most
18 bioreactive, and complex forms generally are much less so. The relationship between speciation
19 and bioavailability has been developed as the Free Ion Activity Model (FIAM) (Campbell,
20 1995), and geochemical modeling software such as CHESS; MINEQL, WHAM, and MINTEQ
21 are powerful tools in understanding speciation. The development of speciation profiles for a
22 dissolved metal concentration via modeling and/or direct measurement will help to provide
23 information at a particular site of interest and can be used on a comparative basis between sites
24 to reduce the variability associated with different exposure conditions.
25 Although the link between bioavailability of metals and factors influencing speciation
26 (such as pH, temperature, and organic and inorganic anionic complexation) are of prime
27 importance, other abiotic factors, particularly cations, influence metal bioaccumulation and
28 toxicity. Dissolved cations such as Na+, K+, Ca2+, and Mg2+ can competitively inhibit metal
29 uptake and the recent development of integrated toxicity prediction models. Biotic Ligand
30 Models (BLMs) have successfully combined abiotic speciation, cationic competition, and
31 bioaccumulation at the presumed site of toxic action (as reviewed by Paquin et al., 2002).
32 Applicability of the BLM is being extended to chronic toxicity through, for example, the
33 use of the Ion Balance Model (IBM) of Paquin et al. (2002a, b, c). Although the approach,
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
3-13
-------
1 which was initially applied to silver, may ultimately provide a way to predict effects due to
2 metals over varying exposure durations, further development and testing are required.
3 Overall, the BLM approach has wide application in terms of understanding
4 bioavailability in relationship to toxicity because it incorporates speciation in the exposure
5 medium, bioaccumulation, and toxic impacts in a robust approach that has been possible to apply
6 in a variety of contexts to account for differences in bioavailability. For example, the BLM has
7 recently been incorporated as part of the revision of the WQC for copper, it has been used in the
8 context of risk assessment, and it is being applied as an alternative to the application of WER
9 approaches for setting site-specific discharge and cleanup objectives. When considering the
10 application of this approach, as with all models, care should be taken to understand and
11 explicitly account for the assumptions and potential sources of uncertainty. It should be
12 recognized that the BLMs are being developed only for a subset of metals (e.g., copper, nickel,
13 cadmium, silver, lead, and zinc) and are based on experimental data from a limited number of
14 test species. Additionally, most BLMs predict acute toxicity, although a few are being
15 developed that predict chronic toxicity (e.g., the IBM). Considerable research efforts are
16 ongoing, and it is likely that our understanding of metal bioavailability and the method for
17 integrating this knowledge into prediction models will improve quickly in the coming years. See
18 Section 4.5.7 for further discussion.
19 Recommendations:
20
21 • Methodologies for evaluating the bioavailability and toxicity of metals in water
22 include hardness adjustments, site-specific WERs, and use of the FIAM and BLM.
23 Selection of an appropriate approach will depend on the metal of interest and the
24 availability of data. The FIAM and BLM have been shown to offer dramatic
25 improvements over traditional approaches, and these models are preferred for metals
26 for which adequate supporting data have been developed.
27
28 • Hardness should be used as a normalizing function for metal toxicity only when
29 information on speciation is lacking.
30
31 • The BLM offers a mechanistic approach for relating the bioavailability of metals to
32 toxicity. It can be used to develop or revise water quality criteria (e.g., as in the case
33 of copper), in risk assessments, and as an alternative to the WER approaches for
34 setting site-specific discharge objectives.
35 '
36 • Quantitative Ion Character Activity Relationships (QICARs) are available for select
37 species and may be used to extrapolate availability and/or toxicity for a metal for
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
3-14
-------
1 which data are limited. However, the QICAR approach may require validation prior
2 to each application.
3
4 3.2.3. Background
5 Background concentrations of metals may contribute to and exacerbate toxicological
6 effects and accumulations of metal from various exposure sources or, conversely, it may result in
^
7 adaptation of organisms to higher metal concentrations and result in increased tolerance to
8 exposure. Thus, consideration of background metal concentration is an important part of risk
9 assessment.
10 Concentrations of metals in waters of the U.S. vary tremendously. Thus, use of a single
11 number to represent all areas within the U.S. is discouraged. See Sections 4.5.4.1 and 4.5.4.2 for
12 further discussion.
13 Regarding the discussion on the 'added risk approach", this approach has been used in
14 Europe, but has recently been reported to be unreliable by the European Commission Health and
15 Consumer Protection Directorate General, Scientific Committee on Toxicity, Ecotoxicity, and
16 the Environment (CSTEE) (equivalent to EPA SAB) (EC Health and Consumer
17 Protection/CSTEE, 2004). They report: "The CSTEE is of the opinion that current knowledge
18 on the geographic distribution of metal background concentration in aquatic systems is
19 insufficient to correctly implement the added risk approach... and,., not accounting for
20 bioavailability in both the MPA (maximum permissible addition, i.e., water quality criteria as an
21 amount added to background) and Cb fraction (background fraction) results in the incorrect
22 assessment of the risks of metals and thus prevents the establishment of science-based EQS
23 (environmental quality standard). The CSTEE suggests that an accurate assessment of the risks
24 (or EQS) posed by metals should be done by establishing - on a site-specific-, watershed/basin-
25 or regional basis - both the bioavailable total fraction in the environmental compartment/medium
26 (Ecbioavaitab,e - background) and the bioavailable total no effect concentrations (PHEC^^^)."
27 Recommendations:
28
29 • While work is underway to define average (and ranges) of background concentrations
30 for ecoregions, the recommended default is to use state averages where possible and
31 to always define the range that might be encountered within the spatial scale being
32 considered. See section 4.5.4.1 for further discussion.
33
34 • For site specific assessments, a physical and/or temporal boundary needs to be
35 defined then background should be described, estimated, or measured. (See section
36 4.5.4.1 for more information.)
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
3-15
-------
1 . . • • . .
2 • Although consensus on a methodology has not been reached at this point, it is.
3 recognized that the bioavailability of the background, as well as the anthropogenic
4 contribution, should be assessed in terms of both exposure and impact.
5
6 3.2.4. Bioaccumulation
7 • Characterization of metal bioaccumulation is an important component of many Agency
8 assessments that range in complexity (i.e., from screening-level to definitive assessments) and
9 overall scope (i.e., from to site-specific assessments to national ranking and characterization).
10 Assessments may consider bioaccumulation either in terms of the potential for the metal to
11 directly affect the organism and/or as the basis for estimating exposure to consumers via a
12 trophic pathway. Understanding this component is relevant for ecological risk assessments, and
13 also for human health assessments when the food chain pathway is involved.
14 The prevalence and importance of bioaccumulation in Agency assessments has brought
15 increasing attention on the scientific validity and uncertainty of the methods used to characterize
16 and quantify metal bioaccumulation. To reduce uncertainty in metals assessments, robust
17 connections should be established between the bioaccessible/bioavailable form(s) of metals in
18 various exposure media, their accumulation, metabolism and distribution in tissues, and the
19 form(s) of metals that exert their toxicity directly to the organism or indirectly to its consumers.
20 However, for many metal-organism combinations, data are lacking on the mechanism(s) of
21 absorption, distribution, metabolism, and excretion, as well as mode of action. Improved
22 understanding of these issues would improve linkages between exposure and effects or
23 accumulation and effects.
24 It has, however, been established that, within certain limits, organisms either regulate
25 their uptake of metal, store metals in forms not toxic to the organism (but potentially toxic to
26 their consumers), or are efficient metal excretors. Which one of these strategies is employed
27 may differ by metal and organism, particularly for those metals that are essential micronutrients.
28 For example, for aquatic organisms that regulate uptake or excretion of a nonessential metal,
29 tissue levels of metals are most frequently inversely related to water concentrations. Where the
30 organism is effectively regulating internal levels of an essential metal, the amount in the
31 organism can remain constant over a certain range of increasing environmental concentration.
32 Under both scenarios, toxicity will occur only when some threshold is reached where the
33 organism's regulatory mechanisms become overwhelmed. BAFs/BCFs should be derived using
34 mathematical relationships that represent the concentration in the organism or tissue as a
35 function of the concentration in the exposure medium/media (for BCF/BAF) for each set of
11/24/2004 Peer Review Draft
DISCLAIMER: Tills information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy. •
3-16
-------
1 vary significantly based on multiple factors, such as the form(s) of metal in tissues, the digestive
2 physiology of the predator, and the tissues in which the metal is stored (e.g., exoskeleton vs. soft
3 tissue). Development of analytical tools for quantifying the bioavailable fraction of accumulated
4 metals to consumers (e.g., analysis of tissue fractions such as cytosolic metals) is currently an
5 emerging area of research (e.g., Wallace and Luoma, 2003; Wallace et al., 2003). See Section
6 4.5.8 for further discussion. •
7 Recommendations:
8 .
9 • Bioaccumulation and trophic transfer of metals does occur and can be an important
10 source of exposure. In particular, consumers with high rates of metal uptake relative
11 to elimination rates and high metal assimilation efficiencies are generally of greatest
12 concern in aquatic food webs. However, biomagnification of inorganic forms of
13 metals across three or more trophic levels in aquatic food webs is rare. Therefore,
14 assessment of food web biomagnification alone is generally not an important aspect
15 of classifying hazards or risks of inorgani c metal compounds. .
16
17 • Although the prediction of toxicity due to dietary exposure to inorganic metals is
18 complicated by wide variation in the bioavailability and toxicity of accumulated
19 metals, it is a factor that should be considered in metals assessments. Direct
20 . approaches to accomplish this include quantifying the bioavailable fraction of
21 accumulated metals to consumers (e.g., analysis of tissue fractions such as cytosolic
22 metals). Bioassay methods, such as feeding field-collected contaminated prey to •
23 predators under controlled conditions; offer another way to assess trophic transfer of
24 metals, although such methods have not been widely standardized. These methods
25 may be particularly useful for site-specific assessments, but they require knowledge
26 of predator-prey relationships to be most applicable to field settings. Kinetic models
27 - also may be useful for predicting metal bioaccumulation in aquatic food,webs
28 (including trophic transfer); they show the most promise when applied and calibrated
29 on a site-specific basis using experimentally derived kinetic parameters (e.g., uptake,
30 " assimilation, elimination, and growth dilution) that reflect typical environmental
31 exposures. Efforts should continue to refine, evaluate, and apply models when
32 warranted.
33
34
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy. _
349 *,
-------
/~"N 1 3.2.6. Sediment Exposure and Effects
^ 2 Toxicity testing and benthic community assessments will continue to be important for
3 evaluating exposure and effects of metals in sediments. Chemical-based approaches specific to
4 evaluating metals in sediments have received increased attention. These include the
5 simultaneously extracted metals-acid volatile solids (SEM-AVS) approach and use of interstitial
6 pore water measurements. These approaches are focused primarily on estimating the
7 lexicological potency (bioavailability) of metals in sediments to sediment-dwelling invertebrates.
8 However, some uncertainties associated with estimating exposures to benthic invertebrates
9 should be recognized. These include the creation of microenyironments by sediment organisms
10 (e.g., irrigated burrows) and the influence of exposure through the diet sediment or metal-
11 contaminated prey. For these reasons, the exposure pathway involving direct exposure to pore
12 water is recognized as a simplification. Nevertheless, laboratory and field support exist for its
13 application. See Section 4.5.10 for further discussion.
14 Recommendations:
15
16 * The SEM-AVS approach is the best available tool for screening analysis based solely
17 on sediment chemistry. It is currently applicable to mixtures of copper, cadmium,
18 zinc, lead, nickel, and silver. .
19 •
O20 • It is recommended mat the use of SEM-AVS should be focused on identifying
21 sediments where direct toxicity to benthos is not expected (i.e., SEM-AVS < 0); it is
22 less powerful for predicting when toxicity will occur because it does not account for
23 all factors that influence toxicity. A modification of the method is available to
24 account for the presence of organic carbon within the sediment matrix.
25 '
26 • Consideration of uncertainties in the SEM-AVS approach include whether it is fully
27 applicable to organisms living in microenvironments (e.g., irrigated burrows) or
28 whether it adequately accounts for exposure via sediment ingestion. Some
29 bioaccumulation of metal in benthic organisms may still occur when SEM-AVS < 0.
30 • '
31 • The equalibrium partitioning (EqP) interstitial water approach may be'Combined with
32 the SEM-AVS approach and direct measures of toxicity as part of multiple line-of-
33 , evidence approaches.
34 - -
35 • The EqP interstitial water approach is best applied to identify nontoxic sediments;
36 sediments that exceed this guideline may or may not show toxicity.
37
38 • The EqP interstitial water approach has been evaluated primarily for copper,
39 cadmium, zinc, lead, nickel, and silver. Although the rationale may apply to other
r>
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
3-20
-------
1 speciation, and sorption processes as well as microbial activity. See Section 4.4.1 for further
2 discussion on soil transport models, and Section 4.4.1 for additional information on partition
o
3 distribution coefficients
4 Recommendations:
5
6 »A wide variety of analytical and chemical techniques may be used to characterize
7 metal speciation. For particle-bound metals, direct measurement tools include x-ray
8 absorption spectroscopy (XAS), x-ray diffraction (XRD), particle-induced x-ray
9 emission (PIXE and nPIXE), electron probe microanalysis-scanning electron
10 microscopy (EPMA-SEM), secondary ion mass spectrometry (SIMS), x-ray
11 photoelectron spectroscopy (XPS), sequential, extractions, and single-chemical
12 extractions. See Section 4.1.8, Table 4-9, for further information about the
13 applicability and limitations of each approach.
14
15 , • Computer models may be used to predict speciation of metals in soil solutions. These
16 include the Windermere Humus Aqueous Model (WHAM) (Tipping, 1998,1994) and
17 ' Non-Ideal Competitive Absorption (NICA) model (Gooddy et al, 1995). See
18 Lumsdon and Evans (1995) for a good review of metal speciation models.
19
20 • Models using partition distribution coefficients (KJ have significant inaccuracies for
21 metals, and the application of single partition coefficient values for individual metals
22 should be limited to site-specific assessments or to regional- and national-scale
23 studies where bounds of potential Kj values, or reasonably representative single
24 values are adequate.
25
26 • The MINTEQ model can be used to generate generic partition coefficients that may
27 be applied to regional or national mobility evaluations (epa.gov/ceampubl/mmedia/
28 minteq/index.htm or lwr.kth.se/english/OurSoftware/Vminteq/).
29
30 • Linearity of log Kj with pH may be assumed as a default approach for nationalrscale
31 , assessments. The value of log Kj for metal cation adsorption can be assumed to
32 increase linearly with pH, whereas the value of log Kj generally decreases with pH
33 for anion adsorption (Langmuir, 1997; Tessier, 1992). Partition coefficients tabulated
34 as a function of pH are available and a useful way to proceed; non-pH-dependent
35 values are available for lead (900), mercuric chloride (58,000), and elemental
36 mercury (1,000) (U.S. EPA, 1999a).
37
38 • For higher level, site-specific assessments, metal adsorption may be estimated fairly
39 accurately using literature information to estimate the sorption properties of metal
40 oxides, clays, and OM, which are then used in diffuse layer sorption models. If
41 greater accuracy or site specificity is required, it may be necessary to measure metal
42 adsorption in laboratory experiments.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
3-23
-------
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
• For site-specific assessments, Kj values developed for the site or for a similar
location are generally preferred over generic or default values.
• If fluid flow in porous soil is isotropic and adsorption is fast, reversible, and linear,
the Diffuse Layer (DL) model (also called the Generalized Two-Layer Model, or
GTLM) in MINTEQA2 can be used to predict adsorption and precipitation behavior
as a function of pH.
• PHREEQC, which has the DL metal adsorption model, can be used with the 3-
Dimensional Flow, Heat and Solute Transport model (HST3D, a ground water flow
and transport model), and the Chemical Transport (CHMTRNS) model also can be
used to model metal transport through porous media.
3.3.1.3. Transformation In Soils
Methylation and demethylation of organic mercury compounds in soils are mediated by
the same types of abiotic and microbial processes that occur in aquatic systems. Because soils
are primarily oxygenated systems, particularly in the root zone, conditions favorable to sulfide
formation and bacterial methylation occur infrequently. With the exception of peat bogs and
similar anoxic, highly saturated soils, methylation generally occurs only at very low rates in
soils. Plants also can transform metals and metalloids taken up from the soil. The most notable
example is selenium. Soluble inorganic oxanions of selenium are readily taken up by plants and
converted to organoselenium compounds, such as selenomethionine, selenocysteine, dimethyl
selenide, and dimethyl diselenide.
The dissolution and transformation of a metal compound in soil is influenced by chemical
and physical properties of the compound and of the soil. Environmental parameters, such as
temperature and humidity, have a strong influence on the rate of transformation. When metal
salts are added to soil, the form of the salt dictates the rate and amount of soluble metal that will
form in the pore water. For example, insoluble forms of metals (e.g., vanadium pentoxide
[V2O5]) will transform to soluble free ion (V) at a slower rate than will soluble metal salts (e.g.,
Na^O,,). However, it should be kept in mind that the rate of formation of the free ion is not
proportional to the dissolution rate of the salt, as aging reactions that take place at the same time
as transformation and dissolution will remove the free ion from the pore water. See Section
4.1.9 for further information.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
3-24
-------
1
2
3
4
5
6
7
8
' 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
.32
33
34-
35
36
Recommendations:
• In areas with well-oxygenated soils where methyl ation rates are very low, the
mercury in soils and plants can generally be assumed to be an inorganic form.
• In lieu of site specific information,, selenium in oxygenated soils should be assumed
to be in the inorganic form.
• Selenium in plants and soil invertebrates should be considered as an organic
compound for food chain analysis.
• Volatilization of all metalloids (mercury, selenium, arsenic) and organometallics
should be considered when conducting detailed site-specific fate and transport
models.
• Relative rates of dissolution and aging should be considered simultaneously to
accurately predict pore water concentrations.
3.3.2. Exposure Assessment
Terrestrial wildlife, plants, and invertebrates accumulate metals from direct contact with
soil or sediment, from ingestion of contaminated food (plants or other animals), and from
incidental soil or sediment ingestion. Pathways of exposure include movement from soils
through the food web and, to a lesser extent, air deposition either into soils or directly onto
terrestrial receptors (e.g., plants). Because of significant differences in exposure patterns, it is
convenient to discuss methods by receptor group (invertebrates, plants, wildlife) rather than by
pathways or environmental compartments,, with the exception of the contribution of natural
background to total exposure, which is applicable to all organisms.
3.3.2.1. Background
EPA has provided detailed guidance on estimating background concentrations for site-
specific assessments (U.S. EPA, 2002b, c, 2000c). Statewide average background soil
concentrations are available in the U.S. EPA ecological soil screening levels (EcoSSLs)
document (U.S. EPA, 2003c). Additional information on concentration of metals in soils at
smaller spatial resolutions is provided in Shacklette and Boerngen (1984). Some metals (e.g.,
iron, copper, zinc ) are included in the State Soil Geographic Database (STATSGO) available at
www.nrcs.usda.gov/technial/techtools/ stat_browser.html. These data can be grouped at
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
3-25
-------
1 whatever spatial s cale is required, but they are not screened for whether they represent true
2 background concentrations. See Section 4.5.4 for further information.
3 Recommendations:
4
5 • Use of a single number to represent background concentration for metals in all areas
6 . within the United States is of very little or no value owing to high variability across
7 such a large geographic area.
8
9 • Averages (and ranges) of background concentrations for various ecoregions should be
10 defined (Hargrove and Hoffman, 1999; Bailey, 1998; Omernick, 1986).
11
12 • If ecoregion-specific information is not available, a recommended default when
13 conducting national-scale assessments is to use state averages.
14
15 • The range of background concentrations that might be encountered within the spatial
16 scale being considered always should be defined.
17
18 * For a site-specific release of a highly bioavailable form of a metal, the background
19 .concentration may be a negligible contribution to ecological exposure and may be
20 ignored in assessing ecological risk for that metal.
21
22 • For areas of contamination where added metals have aged significantly, reduced
23 bioavailability should be considered. Thus, it becomes important to estimate
24 exposure in terms of the bioavailable fraction.
25
26 • For site-specific risk assessments, it may not be necessary to consider background
27 during the initial screen; bulk soil concentrations may be compared directly with
28 derived toxicity thresholds for soil organisms.
29
30 • For metals that do not pass the initial screen (based on bulk soil measurements),
31 adjustments for site specific bioavailability may be performed that will address the
32 significantly lower bioavailability of contributions from background.
33
34 • Natural background levels should be taken into consideration during any remediation
35 decisions because reducing soil metals below naturally occurring values will alter the
36 . plant and soil invertebrate community composition (potentially as dramatically as did
37 the anthropogenically elevated soil levels).
38
39 3.3.2.2. Soil Invertebrates and Plants
40 The assessment of exposure and effects of metals on terrestrial plants and soil
41 invertebrates should involve evaluating the bioavailability of the metals to the organism, the
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It: has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
3-26
-------
1 accumulation of metals in tissues, and the toxicity of the metals. The accumulation of metals in
2 plants or soil invertebrates may be used as an indication of toxicity but also is a consideration
3 when evaluating exposure to animals that feed on them. See Section 4.5.4 for further
4 information.
5 Recommendations:
6
1 • Bulk soil concentrations collected in the top 0-12 cm of soil can be used as an initial
8 estimate of exposure for soil organisms for all types of assessments.
9
10 • It is recommended that bulk soil concentrations be adjusted to account for
11 bioavailability factors in higher level, definitive assessments; bioassays using site
12 soils are encouraged for areas that do not meet initial screening level values.
13
14 • In higher level, definitive assessments, the organic matter on top of the soil (the
15 "duff) should be analyzed separately to provide further detail on exposure to
16 detritivores (such as Collembold) and deeper soil-dwelling organisms (e.g., various
17 species of earthworms).
18
19 • Normalizing methods may also be useful for addressing soil variability in soil toxicity
20 tests. Two normalizing metrics that have been used are percent organic matter and
21 cation exchange capacity (CEC). Notably, it should be remembered that CEC is a
22 function, at least in part, of soil pH. Therefore, normalization can be done only
23 among soils of similar pH ranges.
24
25 • Most accurate estimates of exposures are generally achieved through measuring or
26 modeling concentration of metals in soil pore water.
27'
28 • Use of metalloregions for national-level assessments is encouraged to account for
29 natural background levels and consequent adaptation of soil organisms.
30
31 • In characterizing risks to soil invertebrates and plants, it is important to recognize that
32 soil concentrations below natural background may be harmful to native biota.
33
34 3.3.2.2.1. Soil Invertebrates
35 Ecological and anatomical differences among soil invertebrates influence exposure to and
36 bioaccumulation of metals present in the soils. The soil ecosystem includes a complex food web
37 of soil invertebrates (both hard- and soft-bodied invertebrates) that feed on each other, decaying
38 plant material, and bacteria or fungi. For risk assessment purposes, however, exposure is
39 typically described as a function of soil concentration rather than as a detailed analysis of
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
3-27
-------
1 movement of metals through the food web. Although this is a reasonable approximation for soft-
2 bodied invertebrates (e.g., earthworms) whose exposure is primarily through soil pore water,
3 there is more uncertainty in correlating soil concentrations with effects in hard-bodied
4 invertebrates. These animals are exposed primarily through ingestion of food and incidental
5 amounts of soil. Direct measurements, when possible, provide a means of ascertaining
6 bioaccumulation for these hard-bodied species and can be used to judge the relationships
7 between those residues and the bioavailable concentrations in soils and the tissue levels in soft-
8 bodied animals such as earthworms. See Section 4.4.2.4.1 for further information.
9 Recommendations:
10
11 « For site-specific assessments, direct toxiciry testing of the soil of concern often is the
12 best method for assessing bioavailability and toxicity-to-soil biota (Fairbrother et al..
13 2002).
14
15 • Critical body residues (CBRs) provide an approach for judging the toxicity of metals
16 to soil invertebrates based on tissue measurements. However, only a few CBRs have
17 been developed for metals in soil invertebrates:. cadmium and zinc in the springtail
18 and cadmium in earthworms. See Section 4.5.9.2.
19
20 • For site-specific assessments, the concepts of pollution-induced community tolerance
21 (PICT) may be useful for assessing effects on soil invertebrate communities. See
22 Section 4.5.2.
23
24 3.3.2.2.2. Plants
25 Plants access metals through the pore water and have both active and passive
26 mechanisms for taking up or excluding metals, depending on internal concentrations and whether
27 the metal is an essential micronutrient or whether it is mistaken for an essential micronutrient.
28 Plants can be exposed to metals via aerial deposition onto leaf surfaces, trapping metals in hairs
29 or on rough articular surfaces that may provide an exposure route for herbivores. See Section
30 4.4.2.4.2 for further information.
31 Recommendations:
32
33 • The "soil-plant barrier" concept—developed for evaluating biosolids—influences the
34 potential for bioaccumulation as well as toxicity. General assumptions that may be
35 used include the following: strongly acidic soils (less than a pH of 4.5) increase plant
36 uptake of zinc, cadmium, nickel, manganese, and cobalt and increase the potential for
37 phytotoxicity from copper, zinc, and nickel; in alkaline soils, the high pH (greater
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
3-28
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
than 8) increases uptake of molybdenum and selenium; and lead and chromium are
not absorbed by plants to any significant extent at any pH.
• The influence of soil chemistry on metal speciation should be considered when
judging the potential for accumulation and/or toxicity to plants. Reduced soils can
form sulfide, and sulfide forms low-solubility compounds for most metals, including
lead, zinc, cadmium, copper, and nickel. For essential elements (e.g., zinc, copper,
nickel), low-solubility species can result in deficiency syndromes. Upon oxidation of
the soil, sulfide is quickly oxidized, and the metals are returned to more normal
equilibrium reactions of aerobic soils.
• Aerial deposition of metals onto leaf surfaces can be assumed inconsequential for
plant exposure, with the exception of volatilized forms of metals, in particular
mercury and lead.
• General categories of uptake based on bioavailability of metals to plants are shown in
Section 4.
3.3.2.3. Wildlife
With some exceptions, risks of metals to wildlife species will typically involve
estimating dietary exposures and relating these to Toxicity Reference Values (TRVs). For some
vertebrates (e.g., especially amphibians), dermal exposures will be important. There also may be
circumstances where inhalation exposures are important. Better understanding of when to
consider dermal and inhalation pathways is needed. The recommendations given below
emphasize the dietary route of exposure because this appears to be the most important pathway
for most wildlife species and is better understood. This pathway includes food, water, and
incidental soil or sediment ingestion. The relative importance of the dietary versus incidental
soil ingestion pathways is dictated by (1) the degree to which food items are able to
bioaccumulate metals from the soils, (2) the degree of incidental soil ingestion as a fraction of
overall food intake, and (3) the relative bioavailability of the metal in the soil as compared to the
metal in food items. See Sections 4.5.9 and 4.5.12 for further information.
Recommendations:
• Incidental ingestion of soil can be assumed to be an important route for exposure to
wildlife when (1) the BAF from soil to food (e.g., to plants or soil invertebrates) is
less than 1 and (2) the fraction of soil in the diet is greater than 5%.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.'
3-29
-------
1 • The same variables that restrict uptake by plants or other soil organisms can be
2 assumed to reduce bioavailability to wildlife that ingest soil directly; an exception
3 may be the lower pH that occurs in the digestive systems of many animals.
4
5 • Measured concentrations of metals in soil, surface water, and food items can be used
6 in food, chain models to estimate total exposure, or concentrations in food can be
7 modeled using literature or site-specific bioaccumulation models. The
8 recommendations given earlier for estimating bioaccumulation of metals in plants and
9 invertebrates are applicable to estimating food concentrations for wildlife exposure
10 models.
11
12 • Although bioaccumulation and trophic transfer of metals does occur,
13 biomagnification (i.e., increases in concentration through the food web) is rare and
14 may be assumed to be unimportant. Exceptions are the organometallic compounds,
15 such as methyl mercury, that do exhibit biomagnification. It follows that with the
16 exception of organometallic compounds, biomagnification is not an important
17 consideration when evaluating the hazards of metals or when ranking metals based on
18 hazard. -
19
20 • CBRs reduce uncertainties because they account for site-specific bioavailability and
21 multiple exposure pathways and should be used when available. However, very little
22 information is available for metal CBRs in terrestrial wildlife; exceptions are methyl
23 mercury, lead, selenium, and cadmium (Beyer et al., 1996). Tissue residues of other
24 metals can be used to indicate that exposure is occurring but will not allow a
25 determination of risk to be made.
26
27 3.3.2.4. Food Chain Modeling
28 Food chain modeling is used to estimate the exposure of wildlife to metals based on
29 ingestion of soil, food, and water, and may also be useful in human health risk assessments. The
30 basic format of the model is the same as that for organic substances. Detailed explanations are
31 available in several related documents, such as U.S. EPA (2002h), (2003c), and (2004d). See
32 also Sections .4.4.3 and 4,5.12 for additional information.
33 • EcoSSLs guidance -
34 http://www.epa.gov/ecotox/ecossl/pdf/ecossl_guidance_chapters.pdf
35 • Ecological Committee on FIFRA Risk Assessment Methods (ECOFRAM) -
36 http://www.epa.gov/oppefedl/ecorisk/rra_chap_three.htm
37 • OAQPS Fate, Transport and Ecological Exposure Model (TRIM.FaTE) -
38 http://www.epa.gov/ttn/fera/data/trim/tsdv2-sect7b-sept.pdf
39
11/24/2004 Peer Review Draft
DISCLAIMER: Tlus information is distributed solely for the purpose of peer review under applicable information
quality guidelines, tt has not been formally disseminated by the EPA and should not be construed to represent any
Agency determinati on or policy.
3-30
-------
1 Recommendations:
.2
3 • Measured or modeled concentrations of metals in soil, surface water, and food items
4 may be used in food chain models. Available food chain models vary in their
5 complexity (e.g., ranging from the use of soil-biota relationships and trophic transfer
6 factors to bioenergetic models that simulate transfers of metals via inhalation and
•7 ingestion.)
8
9 • Information on diet, foraging area, similar topics is found in U.S. EPA (1993).
10
11 *In using wildlife exposure estimates to assess wildlife risk in screening assessments,
12 relative (not absolute) bioavailabiliry generally should be assumed to be either 1
13 (default value) or an appropriate site-specific estimate. However, very little
14 information is available on dietary bioavailability for most wildlife species,
15 particularly owing to differences in digestive physiology and anatomy across the
16 board and the diverse range of mammalian and avian species. General guidelines are
17 provided for some metals (e.g., lead) in NRC (1980), and values derived for human
18 risk assessments (e.g., for lead) in U.S. EPA (2004) may be useful for animal species
19 that have digestive systems similar to those of humans (generally omnivores); such
20 information would be inappropriate for herbivorous animals. If data for one species
21 (e.g., humans, cows, chickens) are used for another species, the uncertainties
22 associated with animal to animal extrapolations for absorption of metals should be
23 described.
24
25 • Food chain modeling should be used in national criteria-setting or hazard-ranking
26 exercises. Screening level approaches and default values are detailed in the U.S. EPA
27 EcoSSL documentation (U.S. EPA, 2003c).
28
29 • For site-specific risk assessments, it is recommended to begin with using default
30 parameters and then successively adding more site-relevant data as particular species-
31 exposure route combinations are not screened out; for methodology, see Fairbrother
32 (2003).
33
34 3.3.2.5. Bioaccumulation
35 Characterization of metal bioaccumulation is an important component of many Agency
36 assessments that range in complexity (i.e., from screening-level to definitive assessments) and
37 overall scope (i.e., from to site-specific assessments to national ranking and characterization).
38 Assessments may consider bioaccumulation either in terms of the potential for the metal to
39 directly affect the organism and/or as the basis for estimating exposure to consumers via a
40 trophic pathway. Understanding this component is relevant for ecological risk assessments, and
41 also for human health assessments when the food chain pathway is involved.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
3-31
-------
1 The prevalence and importance of bioaccumulation in Agency assessments has brought
2 increasing attention on the scientific validity and uncertainty of the methods used to characterize
3 and quantify metal bioaccumulation. To reduce uncertainty in metals assessments., robust
4 connections should be established between the bioaccessible/bioavailable form(s) of metals in
5 various exposure media, their accumulation, metabolism and distribution in tissues, and the
6 form(s) of metals that exert their toxicity directly to the organism or indirectly to its consumers.
7 However, for many metal-organism combinations, data are lacking on the mechanism(s) of
8 absorption, distribution, metabolism, and excretion, as well as mode of action. Improved
9 understanding of these issues would improve linkages between exposure and effects or
10 accumulation and effects.
11 It has, however, been established that, within certain limits, organisms either regulate
12 their uptake of metal, store metals in forms not toxic to the organism (but potentially toxic to
13 their consumers), or are efficient metal excretors. Which one of these strategies is employed
14 may differ by metal and organism, particularly for those metals that are essential micronutrients.
15 Where the organism is effectively regulating internal levels of an essential metal, the amount in
16 the organism can remain constant over a certain range of increasing environmental
17 concentration. Under this scenario, toxicity will occur only when some threshold is reached
18 where the organism's regulatory mechanisms become overwhelmed. BAFs should be derived
19 using mathematical relationships that represent the concentration in the organism or tissue as a
20 function of the concentration in the exposure media (for B AF) for each set of exposure
21 conditions. It is important to emphasize that these relationships should always be based on the
22 environmentally available or bioaccessible fraction of the metal in those media (e.g., soil), to
23 normalize the relationships for site-specific differences in environmental chemistry and
24 associated metal speciation! See Section 4.4.3 for further information.
25 Furthermore, with many metal-organism combinations, the bioaccumulation process
26 cannot be quantified with sufficient rigor to enable unambiguous predictions of metal residues in
27 tissues that are lexicologically meaningful for evaluating impacts to soil organisms. Generally,
28 the linkage between bioaccumulation as measured by whole body concentration and the potential
29 for lexicological impact (i.e., the hazard), is lacking. In the case of direct toxic impacts, the
30 rates of metal accumulation often are more meaningful than tissue residues. Thus, the latest
31 scientific data or.i bioaccumulation do not currently support the use of BAF data as generic
32 threshold criteria for the acute or chronic hazard potential of metals. See Section 4.5.6.3 for
33 further information.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
3-32
-------
1 Recommendations:
2
3 • Measurements of metals in plant tissues, when possible, provide the most direct
4 method for assessing bioaccumulation. When the data are to be used to evaluate
5 exposure to higher trophic levels, measurements should be made of the appropriate
6 edible tissues.
7
8 • Because BAFs of metals are influenced by organism background concentrations and
9 metabolism or storage mechanisms, they should be expressed as a function of soil
10 concentration.
11
12 • . The BAF derivation should also account for the concentrations and forms of the metal
13 in the exposure media, to reflect speciation, bioaccessibility, and bioavailability as
14 indicated. These BAF values should be limited to site-specific applications.
15
16 • The latest scientific data on bioaccumulation do not currently support the use of BAF
17 values when applied as generic threshold criteria for the hazard potential of inorganic
18 metals (e.g., for classification as a "PBT" chemical).
19
20 • For national-level assessments, equations from Sample et al. (1999,1998a, b) for
21 uptake of metals from soil for soil invertebrates and for vermivorous wildlife (e.g.,
22 songbirds, voles, and shrews) and from Efroymson et al. (2001) for plant uptake may
23 be used. However, the soil invertebrate models are not specific to soil type and
24 therefore do not account for bioavailability factors such as pH, clay content, or cation
25 exchange. Furthermore, they do not adequately predict chromium or nickel uptake.
26
27 • Soil parameter values from the 5th to 95th percentile of the area of concern should be
28 used to bound the possible uptake rates (alternatively, Monte Carlo approaches can be
29 used, assuming appropriate distributions for each parameter value).
30
31 • If using models for screening-level assessments, site-specific assessments may use
32 measured soil parameter values. For detailed assessments, site-specific assessments
33 may use either measured values of metals in soil organisms or site-specific bioassays
34 to determine uptake rates (see Fairbrother, 2003).
35
36 • The highest accumulation of metals in plants occurs in the roots, and, except for
37 hyperaccumulator species, most plant trophic transfer rates can be assumed to be <1.
38
39 • Plants are quite sensitive to some metals and may die before achieving levels high
40 enough to be toxic to animals.
41
42
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy,
3-33
-------
1 j 3.3.3. Toxicity Assessment
2 Because metals are naturally occurring substances, toxicity assessment for ecological
3 receptors exposed to metals requires an understanding of both the natural mechanisms for
4 tolerance for (or, in the case of micronutrients, the use of) metals and the toxicological responses
5 that occur when exposure exceeds the capacity of the organism to regulate its body burdens.
6 Interactions between metals in either their uptake or toxicity (such as cadmium/calcium/zinc,
7 mercury/selenium, copper/molybdenum) also should be considered in toxicity assessments. Risk
8 assessments for metals are further complicated by the need to express the dose-response (or
9 concentration-response) functions of bioavailable units that are functionally equivalent to
10 measures of exposure. Issues of essentiality, appropriate toxicity tests, and acclimation or
11 adaptation to continued exposures also should be considered.
12
13 3.3.3.1. Adaptation and Acclimation
14 Organisms have developed various mechanisms to cope with variable background metal
15 concentrations, particularly for the metals that are essential elements. The genetic makeup of an
16 organism defines its ability to cope with variable environmental conditions. The shifting of
17 tolerance to a metal within the genetically defined limit of an organism is known as acclimation
18 and generally involves physiological changes that may not always be passed on to offspring.
19 Genetic adaptation results from increased survival of tolerant genotypes and subsequent changes
20 in gene frequencies. Laboratory experiments conducted with Fj generations obtained from
21 metal-contaminated habitats provide the strongest evidence to support a genetic basis of
22 tolerance (Klerks and Levinton, 1993), and new methods in toxicogenomics (e.g., microarrays)
23 are providing additional insights. The ability of organisms to adapt or acclimate to metals in
24 their environment should be considered when assessing chronic risks or determining the relative
25 ranking hazards of metals. See Section 4.5.2 for further information.
26 Recommendations:
27
28 • For national-scale assessments, the country should be subdivided into metalloregions,
29 such that protection levels, mitigation goals, and ranking results will be appropriate
30 for the suite of species naturally present within each type of controlling environment.
31
32 • For site-specific assessments, the concepts of PICT can be applied.
33
34 • All reports of toxicity effects used in setting TRVs or benchmarks, whether for
35 generally applicable national assessments or for site-specific assessments, should be
11/24/2004 Peer Review Draft
DISCLAIMER: Tltis information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
3-34
-------
1 critically examined for information about how test organisms were raised and housed.
2
3 • For national-scale assessments, organisms should be acclimated to standard soils or
4 (preferentially) to the various metalloregion soils.
5
6 • For site-specific assessments, acclimation should take place in soils with similar
7 characteristics (e.g., percent clay, pH, OM, CEC) as the site of concern.
8
9 3.3.3.2. Essentiality
10 Essentiality is the requirement for metals in normal organism metabolic function. It is
11 one of the primary factors that differentiates risk assessment for metals and metal compounds
12 from that of synthetic organic chemicals and, therefore, requires specific consideration in risk
13 assessments. Some trace elements, such as cobalt, copper, iron, manganese, selenium,
14 molybdenum, and zinc, are necessary for the normal development of plants and animals. Other
15 metals, such as arsenic, cadmium, lead, and mercury, have no known functions. Others may
16 have beneficial metabolic effects but have not been shown to be essential. For essential
17 elements, TRVs should not be below the optimal concentration range (or safe intake range). If
18 set too low (i.e., in the range where deficiency can occur), the determination of risk will be
19 erroneous and organisms actually may be harmed owing to lack of essential nutrients. See
20 Section 4.5.1 for further information.
21 Recommendations;
22
23 • Metals that plants and wildlife require as micronutrients for normal organism
24 metabolic function are identified in Table 4-14 in Section 4.5.1.
25
26 • Derived toxicity threshold values for essential elements should be used in
27 screening-level risk assessments for both national and site-specific applications, if
28 they are no more than 10-fold lower than the nutritional requirements. Otherwise, the
29 required levels should be used as a threshold for allowable exposures. See Section
30 4.5.1.
31
32 • Higher level, definitive assessments may require additional bioassays to characterize
33 the biphasic dose-response curve and determine both required and excessive
34 threshold levels.
35
36 • For wildlife, the literature on dietary requirements of essential elements for livestock
37 can be consulted. The National Research Council has published useful summaries
38 (NRC, 1994,1980), and a recent publication updates this information (McDowell,
39 2003).
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
3-35
-------
1
2 • Minimum concentrations required for plant growth are summarized in Epstein (1972,
3 1965) and Kabata-Pendias and Pendias (2001).
4
5 33.3.3. Metals Mixtures
6 Mixtures of metals (including metalloids and other contaminants) are commonly
7 encountered in the natural environment as a result of anthropogenic inputs and the natural
8 co-occurrence and enrichment of many metals. Interpretation of available information on the
9 toxic effects of metal mixtures is complicated by differing measures and definitions of the
10 bidavailable fraction of metals, whether it is the fraction that is available for uptake from the
11 environment or at the site of toxic action. Two key questions should be addressed in risk
12 assessments of metal mixtures: To what extent does each metal contribute to any observed
13 effect? Are the effects significantly greater than or less than the sum of the individual
14 component effects? See Section 4.5.3 for further information.
15 Recommendations:
16
17 • Metals mat have the same mode of action initially can be assumed to be additive in
18 effect
19
20 • For metals with assumed additivity, either Concentration Addition or Effect Addition
21 .models can be used (see Section 4.5.3 for descriptions of these models).
22
23 • Effect Addition models, especially if based on body or tissue concentrations, are
24 more accurate than Concentration Addition models but require reliable dose-response
25 and bioaccumulation curves for all single metals and then careful testing of the
26 models.
27
28 • National criteria for mixtures may not be possible because the combined effects
29 depend on relative amounts of each metal and their relative bioavailability.
30
31 3.3.3.4. Toxicity Testing
32 Variability among soil toxicity test results is due in part to the influence of soil properties
33 on bioavailablily of metals. Additionally, incorporation of sparingly soluble substances, such as
34 many environmental forms of metals, into the soil matrix is difficult, and acclimation/adaptation
35 of test organisms can further complicate test results. Use of soluble metal salts with the addition
36 of organism to the test matrix immediately after mixing is not representative of most
37 environmental situations, where aging and other physical/chemical processes affect metal
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
3-36
-------
1 speciation and uptake. However, a large body of literature on the toxicity of metals to soil
2 organisms and wildlife already has been developed. The risk assessor should use these data,
3 taking into account the test-to-test variability in "soil chemistry parameters and differences
4 between laboratory test matrices and real-world conditions. See Section 4.5.11 for further
5 information.
6 Recommendations:
7
8 • Existing toxicity data for soil organisms (plants and invertebrates) should be adjusted
9 on the basis of OM, pH, and (if available) CEC to account for bioavailablity
10 differences before use in any type of risk assessment or hazard ranking.
11
12 • Field data may be compared with laboratory toxicity response information by
13 measuring metals in soil pore water from field assessments and comparing such data
14 to spiked laboratory soils. This will increase the accuracy of the comparison by
15 increasing the similarity of the bioavailability of the measured fraction.
16
17 • The effects of metals on plants can be evaluated by using either toxicity tests or
18 published data. Toxicity tests performed by adding bioavailable metal salts to soils
19 may provide different results than seen in field conditions with long-term metal
20 exposures. The differences in bioavailability could be corrected by either allowing
21 the added metals in the bioassays to age or by expressing metals in test soils in the
22 environment as the bioavailable fractions (for protocols, see Fairbrother et al., 2002).
23 Alternatively, measurement of metals in pore water from either field or laboratory
24 studies will increase the comparability of results. The guidance provided for
25 development of EcoSSL values (U.S. EPA, 2003c) is useful for judging the
26 applicability of literature studies to plant toxicity threshold determinations.
27
28 • Sequential extractions of soil samples also may be used to estimate the bioavailable
29 fraction. These treat soil samples with a succession of reagents intended to
30 specifically dissolve different and less available phases of metals. Many of these
31 techniques are a variation on the classical method of Tessier et al. (1979), in which
32 metal associated with exchangeable, carbonate-bound, iron-manganese bound,
33 organically bound, and residual species are determined. Although these methods are
34 preferable to total metal extraction, relating the sequential fraction extracted to metal
35 bioavailability is open to debate. For the application and limitations of extraction
36 techniques, assessors should review Beckett (1989), Kheboian and Bauer (1987), and
37 Foerstner (1987).
38
39 • Toxicity of metals to wildlife species can be expressed either as dose-response
40 relationships or as TRVs; the latter have received the most application in risk
41 assessments. Toxicity information is usually obtained from the literature because
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
3-37
-------
1 case-specific toxicity studies are rarely done. The best sources of information on
2 metal toxicity thresholds are NAS/NRC (1994,1980), McDowell (2003), and the
3 documentation supporting development of EcoSSL values (U.S. EPA, 2003c). The
4 EcoSSL document also includes a general approach for screening studies for
5 acceptability for use in derivation of TRVs for risk assessments. Uncertainty factors
6 can be carefully applied if there is concern for extrapolation of data to species in a
7 different taxonomic category (e.g., genus, family, class). General summaries for
8 some metals are available in Beyer et al. (1996) and Fairbrother et al. (1996).
9
10 • Site-specific toxicity data can be developed for wildlife when the uncertainties
11 associated with use of published data are large and the implications of the decisions
12 warrant better information. These typically involve feeding studies with
13 contaminated foods or soils contaminated with metals such as lead.
14
15 • Cross-species extrapolations should be conducted with some knowledge of animal
16 physiology and specific responses to metals. Digestive physiology is the most
17 important distinction because most metal exposures in wildlife are by the dietary
18 route. However, other specific organismal responses should be understood as well.
19 See Section 4.5.12.
20
21 • Use of critical tissue residues as an alternative approach to toxicity endpoints is
22 conceptually sound but may requires significant research to establish critical tissue
23 levels. Exceptions include some essential elements in plants, where
24 deficisncy/sufficiency concentrations in foliage have been developed, and tissue
25 levels of lead, selenium, and cadmium in wildlife (Beyer et al., 1996). See Sections
26 4.5.6.3 and 4.5.12.
27
28 3.3.3.5. Extrapolation of Effects
29 Extrapolating the results of toxicity tests to untested species is necessary because of the
30 paucity of data on the toxicity of metals to these receptors. However, extrapolation of results
31 should be approached with caution owing to the large amount of uncertainty that can be
32 introduced into the risk assessment process. Toxicological responses among species vary
33 because of many physiological factors that influence the toxicokinetics (absorption, distribution,
34 and elimination) and toxicodynamics (relative potency) of metals after exposure has occurred.
35 These include (but are not limited to) differences in gut physiology, renal excretion rates, and
36 egg production. Methods for extrapolating metal effects data among species are not unique to
37 metals risk assessment. However, for metals, some species are able to regulate or store metals in
38 their tissues without experiencing toxic effects (i.e., biota-specific detoxification), which makes
39 extrapolations between species especially problematic. The essentiality of other metals adds
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
3-38
-------
1 additional complexity to interspecific toxicity extrapolations. See Sections 4.5.11 and 4.5.12 for
2 further information.
3 Recommendations:
4
5 • Methods for extrapolating toxicity thresholds among species (e.g., species sensitivity
6 distributions (SSDs) are the same for metals as those used for organic substances and
7 are subject to the same assumptions of physiological similarities and degree of
8 required margin of safety (for a review of methods, see Kapustka et al., 2004).
9
10 • All extrapolations should account for different requirements for essential elements
11 and the factors that modify metal toxicity (e.g., acclimation, essentiality, mixtures).
12
13 • SSDs should be used for extrapolating effects data among species (Van Straalen,
14 2001; Posthuma, 2001).
15
16 • A review of cross-species extrapolation methods can be found in U.S. EPA (2003).
17
18 • Across-metal extrapolations (even within the same species) should be avoided unless
19 mechanisms of action are known to be similar.
20
21 • For soil invertebrates, extrapolating toxic response to one metal to potential effects of
22 a nontested metal can be done using QICARS (Kapustka et al., 2004; Lewis et al.,
23 2000).
24
25 • For plants and wildlife, there is sufficient information about modes of action of most
26 common metals to make informed judgments about relative toxicity, so empirical
27 extrapolation models such as QICARS would be applicable only for initial screens of
28 minor elements.
29
30 • Caution should be exercised when using published or measured toxicity data to
31 evaluate risks to plants in particular regions. This is because naturally occurring
32 levels of metals play an important role in biogeographic distributions of plants and
33 animals and may, in fact, be limiting factors in species distributions or use of
34 landscapes. Thus, it becomes very important to define the geospatial location of the
35 area to which the assessment results will apply.
36
37 • For assessments conducted for regional or national assessments, criteria development,
38 or ranking purposes, it should be acknowledged that results will be based on
39 organisms and soil types that result in greatest bioavailability and sensitivity. Care
40 should be taken, however, that the organism-environment combinations assessed are,
41 - in fact, compatible with real-world conditions.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
3-39
-------
1 4. METAL-SPECIFIC TOPICS AND METHODS
2
3 This chapter discusses metal-specific topics and methods to be used in the assessment of
4 risk to humans and ecological entities from exposures to inorganic metals. It applies information
5 and text from the metals issue papers and reflects contributions by EPA scientists and external
6 experts. The final metals issue papers are available on the EPA Web site at
7 http://cfpub.epa.gov/ncea/rafi'recordisplay.cfm?deid=86119.
8 Key topics and tools in this section are presented in subsections on environmental
9 chemistry, exposure pathway analysis, human health effects, and ecological effects. The
10 applications and limitations of the various models and methods for conducting metals
11 assessments are presented to inform the reader. Topics and tools related to bioavailability and
12 bioaccumulation are discussed throughout Chapter 4 because they have far reaching impact that
13 crosses many aspects of metals assessment.
14
15 4.1. ENVIRONMENTAL CHEMISTRY
16 4.1.1. Introduction and Terminology
17 A general review of factors pertaining to the chemistry of metals in sediments, soils,
18 waters, and the atmosphere is presented in this chapter in the context of risk assessment.
19 Because the behavior of metals defies simple generalities, it is necessary to understand the
20 chemistry of the particular metal and the environment of concern. However, we can generalize
21 factors that control metal chemistry and environmental characteristics where this generalization
22 allows us to progress with estimates of metal fate and effects.
23 Metal speciation determines the behavior and toxicity of metals in the environment.
24 Speciation refers to the occurrence of a metal in a variety of chemical forms. These forms may
25 include free metal ions, metal complexes dissolved in solution and sorbed on solid surfaces, and
26 metal species that have been coprecipitated in major metal solids or that occur in their own
27 solids. The speciation of a metal affects not only its toxicity but also its volatilization,
28 photolysis, sorption, atmospheric deposition, acid/base equilibria, polymerization, complexation,
29 electron-transfer reactions, solubility and precipitation equilibria, microbial transformations, and
30 diffusivity (Bodek et al., 1988).
31 The following sections address the application of hard and soft acid and base (HSAB)
32 concepts to metal behavior, including the formation of metal complexes, and the importance of
33 pH and oxidation-reduction reactions to metal mobility and toxicity. The chapter then examines
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. Jit has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-1
-------
1 the occurrence and interactions of the metals of concern in natural media, including in surface
2 and ground waters, soils and aquatic sediments, and the atmosphere. Important topics
3 considered in these sections are metal sorption behavior, aging in soils, metal dissolution and
4 transformation and transfer to plants, and methods of determining metal speciation in soils and
5 sediments.
6
7 4.1.2. Hard and Soft Acids and Bases: The Stability of Complexes
8 Complexes are formed between metals (acids) and
9 ligands (bases), both in solution and at the surfaces
10 minerals and organisms. The toxic reaction of organisms
11 to metals can be directly related to the nature of the metal
12 complexes formed in solution and at the surface of the
13 organism.
14 A useful concept that helps to explain the strength
15 of metal complexing and metal toxicity is that of hard and
16 soft acids and bases (HSAB), which was introduced by
17 Pearson (1973). In this concept, metal cations are Lewis
18 acids and ligands are Lewis bases, with the metal cation
19 and Hgand in a complex acting as electron acceptor and
20 donor, respectively. "Soft" implies that the species'
21 electron cloud is deformable or polarizable and the electrons are mobile and easily moved. Soft
22 species prefer to participate in covalent bonding. Hard species are comparatively rigid and
23 nondeformable, have low polarizability, hold their electrons firmly, and prefer to participate in
24 ionic bonds in complex formation (Langmuir, 1997). Hard acids form strong, chiefly ionic
25 bonds with hard bases, whereas soft acids and soft bases form strong, chiefly covalent bonds
26 when they form complexes. In contrast, the bonds formed
27 between hard-soft or soft-hard acids and bases are weak,
28 such that their complexes tend to be rare. Table 4-1
29 summarizes hard and soft acid and base relationships for
30 the metals of concern. The first text box summarizes the
31 applicability of hard and soft concepts to the formation of
32 metal complexes; the second text box defines ligands.
Hard and Soft Acids and Bases
Hard acids and hard bases.
Complexes formed between divalent hard
acid cations and monovalent or divalent
hard bases are ionic and relatively weak
and are often termed "ion pairs." ,
Complexes formed between Be2+ or
trivalent hard acids and hard bases tend to
be ionic and relatively strong.
Soft adds and soft bases. Strong,
relatively covalent bonds are formed in
complexes between soft and borderline
soft acid cations and soft bases. Ligand
binding sites on the external or internal
surfaces of organisms are often of soft
base character and thus bond strongly
with soft and borderline soft acid cations.
Ligands
Ligands are anions or molecules that
form complexes with metal ions.
Depending on whether a ligand shares
one, two, three or more electron pairs
with metals, it is called a mono-, bi-, tri-
or multidentate ligand.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-2
-------
1
2
3
4
5
6
7
8
1
2
3
4
5
6
7
S
9
Hard metals (hard acids), which are the least toxic, preferentially bind with hard bases
that contain oxygen, forming weaker bonds with soft nitrogen and sulphur species. The strength
of binding between hard metals and hard ligands is usually a function of pH. Many of the hard
metals are macronutrients. Soft metals (acids) bind preferentially with soft S and N ligands,
forming weaker bonds with hard base species such as hydroxide and sulfate. Soft and
borderline metals, and Mn2+, which is hard, form bonds of decreasing strength with soft ligands
such as
Table 4-1. Hard and soft acids (metal cations) and bases (ligands)
Hard acids
Borderline acids (between hard and soft)
Soft acids
Hard bases
Borderline bases (between hard and soft)
Soft bases
A13+, Ba2+, Be2*, Co3*", Cr3*, Fe3+, Mn2+, Sr2*, U4+,
UO/+, V02t
Co2+, Cu2+, Fe2+,Ni2+, Pb2*, Zn2*
Ag*. Cd2+, Cu*, Hg2*, Hg+, CH3Hg+, T13+, Tl+
F, H20, oxyanions: OH', SO/', CO/', HCO,',
C2042-, Cr042-, Mo042' H^O/-3, H^AsO/*
SeOA H2VO4-, NH3, RNH2, N2H4> ROH, RO',
RA CHjCOCr, etc.
C1-, Br, NO2', SO32-, HnAsO3"-3, C6H5NH2,
C5H5N,N3-,N2
I", HS-, Ss; CN-, SCN-, Se2-, S2(V", -SH, -SCH3,
-NH2) R-, C2H4, C6H6, RNC, CO, R3P, (RO)3P,
R,As. R7S. RSH. RS-
Source: Modified after Huheey et al. (1993) and Langmuir (1997). "R" refers to an organic molecule.
sulfide, generally in the following order: Pb2+> Cu2+> Cd2+> Co2+ • Fe2+ > Ni2+ > Zn2+ > Mn2+.
The tendency of metals to bind to soft ligands or to organic substrates (which are usually soft) is
greatest for soft and borderline metals (soft acids), followed by the hard metals (hard acids),
typically in the order Pb2+> Cu2+>Cd2 K> Zn2+ > Ca2+ > Mg2+ » Na+ (Pickering, 1986).
The tendency of metals to form solid phases, such as sulfides in sediments, is also related
to their HSAB qualities. For example, extremely insoluble metal sulfides are formed in anoxic
sediments by soft acid metal cations, such as Hg2+ (log K^ = -57.25) or Ag+ Qog K^ = -49.7),
whereas borderline hard and hard metal cations such as Mn2+ (log K^ = -19.25) or Fe2+ (log ^
= -22.39) form slightly more soluble, although still highly insoluble, metal sulfides.1 These
'Solubility products for all sulfides except Ag2S are from Di Toro et al. (1990). The
product for silver sulfide is from Stumm and Morgan (1970).
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-3
-------
1 differences in solubilities are fundamental to the method of acid-volatile sulfide (AVS)
2 normalization of sediment-associated metals (see Section 4.1.5, on Sediment Chemistry).
3 Finally, according to the Biotic Ligand Model (BLM) (see Sections 3.1.4.2 and 3.4), effects of
4 metals are related to or correlated with metal interactions with biological ligands, which are
5 generally soft base species.
6
7 4.1.3. Aquatic Chemistry
$ 4.1.3.1. Speciation and Complexes
9 Metal species dissolved in water may occur as free ions, or aquo-ions, or as complexes.
10 Free metal cations are generally surrounded by coordinating water molecules and so have been
11 termed aquocations, although by convention the water molecules are ignored when writing
12 chemical reactions involving metal cations.
13 The total analytical concentration of a given metal in water is the sum of the
14 concentrations of its free ion and its complexes and any metal associated with suspended solids,
15 whether organic or mineral. For example, the total molal concentration of lead, SPb, in a natural
16 water might equal:
17
18 SPb = mPb2+ + mPbOrT + mPbC03° + mPbHC
-------
1 where the terms Ypb and YSCW are the activity coefficients of the ions. The product of the ion
2 activity coefficient and the molal concentration of each species equals the activity of the ion.
3 This equation shows that the activity of free lead ion controls the solubility of lead sulfate. For a
4 given total lead concentration (see the previous equation), the more of the lead that is
5 complexed, the lower will be the concentration of free lead ion. This means that as the extent of
6 lead complexing increases, the total lead concentration must also increase to reach saturation
7 with lead sulfate. In other words, metal complexing increases total metal solubility.
8 Metal complexing also has a direct influence on metal adsorption to organic matter or
9 mineral surfaces. For example, metal carbonate, sulfate, and fluoride complexes are usually
10 poorly adsorbed, whereas metal hydroxide complexes are strongly adsorbed (Langmuir, 1997).
11 In summary, metal complexing generally increases the solubility and mobility of metals in
12 surface and ground waters.
13 For many metals, the free metal ion is thought to be the primary metal species that causes
14 toxicity to aquatic organisms. This is consistent with the free ion activity model (FIAM), which
15 assumes that the free or aquo-ion is the most biologically active form of the dissolved metal.
16 Accordingly, the key parameters that can modify the degree of toxicity are those that affect
17 speciation, such as pH and the amount of inorganic and organic ligands (e.g., dissolved organic
18 carbon, DOC) that can form metal complexes and so provide alternative binding sites for the
19 metal ion. Metal toxicity is also affected by other dissolved ions (e.g., Na, Ca) that compete with
20 metals for binding sites on the gills of fish or on respiratory surfaces of other aquatic organisms.
21
22 4.1.3.2. Importance of pH and Redox Conditions
23 The pH is probably the single most important variable that influences the behavior of
24 metals in the environment. Thus, metal complexes with sulfate, fluoride, chloride, and
25 phosphate are most stable and important below pH 7, whereas metal,carbonate and hydroxide
26 complexes become increasingly more important above pH 6-8.
27 Also, as discussed in Section 4.1.4, hydrogen ion competes with metal cations for
28 adsorption sites, so that adsorption of metal cations by hydrous ferric oxide (HFO), for example,
29 is low in acid systems but increases with increasing pH. In contrast, oxyanions of As, Mo, Se,
30 and Cr tend to be desorbed from HFO with increasing pH because of competition between the
31 oxyanions and OH" ion for sorption sites. Furthermore, the solubility of most metal-containing
32 minerals is greatest under acid conditions and decreases with increasing pH.
33 Figure 4-1 shows the locus of measured values of oxidation potential (Eh) and pH in
34 aquatic systems. The principal controls on Eh are atmospheric oxygen and organic matter. High
35 Eh (oxidizing or aerobic) conditions are maintained in the atmosphere and in most surface waters
11/24/2004 Peer Review Draft
DISCLAIMER: Tliis information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-5
-------
1 and shallow soils in contact with atmospheric oxygen. The lowest Eh values and reducing or
2 anaerobic conditions are found in water-logged soils and sediments that contain organic matter,
3 and in ground waters that contain a few milligrams per liter or more of dissolved organic carbon
4 (DOC). Intermediate Eh conditions are found in waters and sediments that are only partially
5 oxidized because of their relative isolation from the atmosphere. Measured Eh values may not
6 agree with Eh values computed from the concentrations of redox-sensitive species. The
7 difference between measured and computed Eh values is discussed at length by Stumm and
8 Morgan (1996) and Langmuir (1997).
9
+ 1.0
4-0.8
+0.6
+0.4
+0.2
8 0.0
iS
-0.2
-0.4
-0.6
-0.8
-1.0
•,
I
I
6 8
pH
JO 12 14
Figure 4-1. Approximate positions of some natural environments in terms of Eh
and pH. The dashed line represents the limits of measurements in natural environments,
as reported by Baas-Becking et al. (1960).
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-6
-------
Microorganisms play an essential role in defining the redox conditions in aquatic
systems. The effect of the common sequence of microbially mediated redox reactions on Eh
when conditions in a water or sediment become more reducing (depleted in oxygen) or oxidizing
is shown in Figure 4-2. These reactions importantly affect or are affected by the redox behavior
of the major elements C, S, N, Fe, and Mn and also affect the mobility of most of the metals of
concern, which have multiple redox states (Sb, As, Cr, Co, Cu, Mn, Hg, Mo, Ni, Se, and Ag).
Thus, Cr(VI) (oxidized) in chromate ion is highly mobile in aqueous environments compared to
Cr(III) (reduced). Cr(VI) is considered a known human inhalation carcinogen, whereas Cr(III) is
generally considered to have low human toxicity. Among the microbially mediated reactions in
Figure 4-2, the one that most affects possible metal toxicity is sulfate reduction, which requires
the presence of organic matter. Sulfate reduction produces hydrogen sulfide, which reacts with
-0,5
+0.5
+1.0
eM (VOID
-10
+5
+10
+15 +20 p£
F
\
sductior
s
• < Ot • Reduction ]
| i
< DeniMficatton |
I 1
i
1 Sulfide *SO4»'^
s
OxWat. of Fe(ll£>
| NH44*NCV }
OxidalofMn(tl)>
I
[ Nj, *N03-
I
?
|Oj - Foimatton>
-10
+5 +10 +15
+20
Figure 4-2. Sequence of microbially mediated oxidation-reduction reactions.
11/24/2004 Peer Review Draft
DISCLAIMER: Tliis information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-7
-------
1 metals such as Sb, As, Cd, Co, Cu, Pb, Hg, Mo, Ni, and Ag and can cause their almost
2 quantitative precipitation as insoluble metal sulfides. This immobilizes these metals and so
3 makes them unavailable to plants and animals.
4 Much of the preceding discussion of the importance of metal speciation, hard and soft
5 acids and bases, oxidation potential, and metal sulfides, is summarized in Table 4-2. The table
6 shows, for example, that the hard acid metals Al, Ba, Be, and Sr have but one oxidation state and
7 do not form insoluble sulfides; thus, their solubilities and mobilities are not directly affected by
8 redox conditions, although they are strongly affected by changes in pH. Also indicated'are
9 forms of the metals that occur in soils and waters, with As, Mo, Se, and sometimes Cr and V
10 occurring as oxyanions, and most of the other metals occurring usually as metal cations. Eh-pH
11 diagrams for the redox-sensitive metals of concern and of major elements such as Fe, S, and C
12 are given in Langmuir et al. (2004). These show the detailed occurrence of the metal species
13 under oxidizing and reducing conditions. In simplified terms, the Eh-pH diagram and Table 4-2
14 show that Fe, Mn, and Tl are most mobile under reducing conditions, whereas S and the
15 remaining metals of concern are usually most mobile under oxidizing conditions.
16
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-8
-------
I
V
1
"8
£
I
•s
J
I
T-
O
OQ
"o
co
u
u
oa
09
03
u
o
u
I
1
s
I
o
a
si
as
p.
R
R
O
o
I
m
CQ
o
•-4
|
i
9
f?
CQ
o
CQ
"a
a
o.
co
i
+•
A
•?
c.
a
CO
Bari
oa
u
CO
Cadmi
OB
a
OQ
U
ffl
£
I
U
1
'
Selenium
Stronti
g
Va
-------
1 4.1.3.3. Effects Concentration and Bioavailability
2 The traditional method of predicting effects of metals on aquatic organisms is to estimate
3 an effects concentration for the water column and the sediment. For the water column, the
4 EPA's Water Quality Criteria (WQC) (U.S. EPA, 1996a; Stephen et al., 1985) or the Predicted
5 No-Effect Concentration (PNEC) derived following the European Union's (EU's) technical
6 guidelines (EU, 1996) are frequently used as default screening values. However, these criteria
7 make only limited corrections for bioavailability, taking into account only water hardness and
8 ignoring other modifying factors such as pH and competing ligands. The BLM was developed to
9 remedy this situation (Paquin et al., 2002a; U.S. EPA, 2000d; Di Toro et al., 2000). It
10 incorporates the WHAM speciation model and it also models the competitive metal binding at
11 the toxic site of action (the biotic ligand). BLMs are currently available for copper and silver
12 (Di Toro et al., 2000; Santore et al., 2000) and zinc (Santore et al., 2002) and are under
13 development for cadmium, nickel, and lead (see www.epagov/waterscience). The development
14 and use of the BLM is covered in detail in Section 4.5, Characterization of Ecological Effects.
15
16 4.1.4. Ground Water and Metals Mobility
17 Site-specific risk assessments for EPA programs often need to predict the rate of,-
18 movement of metals through soils and their subsequent movement and concentrations in ground
19 water. The primary processes governing the environmental fate and transport of metals in the
20 subsurface are advection, dispersion, matrix diffusion, and retardation (U.S. EPA, 1994c).
21 Advection and dispersion are functions of the system rather than of the contaminant. Matrix
22 diffusion, which is a function of the contaminant, is relatively unimportant and is omitted in most
23 model transport algorithms. Retardation depends on a number of factors (Langmuir, 1997; U.S.
24 EPA, 1994c) and may involve or be affected by the following:
25
26 • Sorption. The attachment of metal species to mineral surfaces or other surfaces.
27
28 • Speciation. The distribution of a given constituent among its possible chemical
29 forms, including metal complexes, which have differing tendencies to be adsorbed or
30 desorbed.
31
32 • Precipitation. The process by which dissolved species exceed the solubility limits of
33 their solids, so that some of the species precipitate from solution. When a metal
34 species reaches mineral saturation, addition of further amounts of the species to
35 solution are precipitated, not adsorbed.
36
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy. •
4-10
-------
1 • Colloid formation. The process of forming colloids and the association of metal
2 species with them. The metals may be sorbed or coprecipitated with colloidal-sized
3 particles.
4
5 • Biofixation. The binding of metals to solid materials due to the interactions of
6 microorganisms or plants.
7
8 • Natural organic matter (NOM) interactions.
9
10 « Other important processes, such as changes in pH, oxidation potential, salinity,
11 concentrations of competing ions, the nature of sorbent phases and their surface areas,
12 and surface site densities.
13
14 4.1.4.1. Metals Sorption
15 4.1.4.1.1. Introduction. The ability of a trace or minor metal such as a metal of concern to sorb
16 to a substrate is usually the determining factor in its mobility. Physical adsorption, which is
17 important for molecular organic compounds, is largely inapplicable to the sorption of the trace
18 metals, which are usually adsorbed as ionic species. In fact, trace metal adsorption is ofen
19 relatively independent of the concentrations of the major ions. Trace metal adsorption is
20 strongly pH-dependent and a function of ionic strength, metal complex formation, competitive
21 ion sorption, redox conditions, and the amounts and reactive surface areas of specific sorbing
22 solids. The most accurate and mechanistic approach to modeling and predicting trace metal
23 adsorption is surface complexation modeling, using a model such as the Diffuse Layer (DL)
24 model, which ideally can account for all of these variables (Langmuir, 1997; Stumm and
25 Morgan, 1996).
26 The degree of mobility of organic contaminants is often expressed by means of a single
27 partition coefficient (Kj) value that describes the distribution of a species between sorbed and
28 dissolved forms (U.S. EPA, 1995a). Mobility is then calculated from the partition coefficient.
29 Such an approach is applicable to metal adsorption only when the conditions listed above are
30 practically constant and are the same in the environment as in the laboratory where the Kj was
31 determined, which is rarely the case. In fact, when metal adsorption is described using partition
32 coefficients, the value of such coefficients typically should be varied by two or more orders of
33 magnitude to reproduce metal adsorption behavior at a specific site. Therefore, single partition
34 coefficients for metals are of little value except in broad regional surveys when a large
35 uncertainty in a ICj value may be acceptable.
36
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-11
-------
1 4.1 .4.1.2. Surface complexation models: Diffuse Layer model It has been observed that in
2 many soils and sediments low in organic matter OM), HFO is the most important metal sorbent
3 and the only sorbent that needs to be considered in predictions of metal sorption behavior
4 (Benjamin and Leckie, 1981). The DL model (also called the Generalized Two-Layer Model, or
5 GTLM) in MINTEQA2 has been extensively applied in aquatic environmental studies of metal
6 transport and attenuation. Loux et al. (1989) used the DL model and MINTEQA2 to predict the
7 adsorption and precipitation behavior of eight metals in an oxidized, sandy aquifer as a function
8 of pH. Assuming that HFO was the only sorbent, DL model adsorption adequately described
9 changes in the concentrations of Ni, Pb, and Zn in the sediment. Cadmium behavior was better
10 understood, assuming its precipitation in CdCO3 (otavite). Changes in Cu, Ba, Be, and Tl were
11 not simply explained. Copper may have been adsorbed by organic matter, which was not
12 considered in the modeling.
13 More recently, adsorption of metals by OM and Al oxyhydroxides as well as HFO has
14 been included in DL modeling with MINTEQA2 (Paulson and Balistieri, 1999). These authors
15 studied neutralization of acidic ground waters by ambient surface and ground waters using a
16 mixing model approach. Particulate organic matter (POM) and HFO were the chief metal
17 sorbents. In pristine systems, Cu is usually the chief metal associated with particulate organic
18 carbon (POC); in the Paulson and Balistieri study, Zn and Cd were mostly adsorbed by POM,
19 and Cu was mostly absorbed by HFO.
20 It may be possible to estimate metal adsorption with acceptable accuracy without having
21 to measure it, depending on the information available on a specific soil, surface water, or ground
22 water system. What is needed minimally is the weight and surface area of potentially sorbing
23 materials (e.g., metal oxides, clays, and organic matter) in a volume of soil or sediment, or in
24 suspension in a stream. Literature information can then be used to estimate the sorption
25 properties of these materials for use in a sorption model. For example, as noted above in the
26 discussion of the DL adsorption model, where HFO is the dominant sorbent and the amount
27 suspended in a stream is known, estimation of metal adsorption can be accurate to within
28 10-20%, as shown in Table 4-3 (Smith et al., 1998). As a general observation, other factors
29 being equal, it has be found that the surface charge density—and thus the metal adsorption
30 capacity—of most minerals is largely a function of their surface areas exposed to water (Pabalan
31 et al., 1998). Thus, the adsorption of metals by Al and Fe(III) oxyhydroxides in a system at a
32 given pH may be assumed to be the same if they have the same surface areas.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-12
-------
o
o
Table 4-3. Comparison of model predictions" and measured values of percent metals
associated with the suspended particulate fraction of mine drainage waters from select sites
Metal
As
Pb
Cu
Zn
Ni
Cd
Argo-3
(pH 5.6, HFO = 0.007 g/L)
Predicted
-
82
18
<]
<1
<].
Measured
-
<71"
27
Oto8
<1
<1
Rawley-3
(pH 6.2, HFO = 0.11 g/L)
Predicted
98
80
60
<1
<1
<1
Measured
<78"
<93"
63,
Oto9
1
6
Leadvilk drain
(pH 7.2, HFO = 0.001 g/L)
Predicted
_
86
-
2
-
<1
Measured
-
<71b
-
3
-
<1
"Model predictions made with the Diffuse Layer model and MINTEQA2.
"Dissolved concentration was below the detection limit; value was computed using the limit of detection for the
dissolved concentration.
Source: Smith etal., 1998.
Cederberg et al. (1985) and Yeh and Tripathi (1991) considered surface complexation
modeling of metal adsorption and metal transport in ground water. Parkhurst (2002) developed a
computer model called PHAST,2 which is a three-dimensional reactive transport model that
4 combines PHREEQC, which has the DL metal adsorption model, with HST3D, a ground water
5 flow and transport model.
6 Several recent studies have measured and modeled trace metal adsorption and metal
7 transport in streams using a surface complexation approach to adsorption. U.S. Geological
8 Survey researchers of the Toxic Substances Hydrology Program have published a number of
9 papers using the OTEQ and OTIS models. OTEQ is a one-dimensional model for studying the
10 fate and transport of metals in streams and rivers. The model couples the OTIS transient storage
11 model with MINTEQ, which includes DL model adsorption of metals by HFO (Ball et al., 1999;
12 Runkel et al., 1999). Runkel et al. (1999) considered in-stream metal transport, metal oxide
13 precipitation-dissolution, and pH-dependent sorption of copper and zinc.
14 If greater accuracy or site specificity is required, it may be necessary to measure metal
15 adsorption in laboratory experiments. Such measurements can be performed on pure minerals or
16 on whole (usually sieved) soils. The sorption results may be used to develop DL model
O
2 See http://wwwbrr.cr.usgs.gov/projects/GWC_coupled/.
11/24/2004 Peer Review Draft
DISCLAIMER: TWs information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-13
-------
1
2
3
4
5
6
7
g
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
values should be limited to regional studies and should be adjusted to account for regional
variations in soil properties.
Table 4-4. Partition coefficients as a function of pH for several important elements
of potential concern » .
Element
Arsenic(V)
Barium
Beryllium
Cadmium
Chromium(III)
Chromium(VI)
Nickel
Selenium(VI)
Silver
Thallium(I)
Zinc
pH 4.9
25
11
23
15
1,200
31
16
18
0.1 - ,
44
pH 6.8
29
41
790
73
1,800,000
19
65
5
8.3
71
62
pH8
31
52
100,000
4,300
4,300,000
14
1,900
2.2
110
96
Source: U.S. EPA(1998b).
The distribution coefficient approach may apply if fluid flow in the porous media (soil or
sediment) is isotropic and adsorption is fast, reversible, and linear (Freeze and Cherry, 1979).
These assumptions are often not valid for metal adsorption. Some transport models assume a
constant partition coefficient or linearity of the partition coefficient over all concentration
ranges. To the extent that sorption is not constant and follows a nonlinear isotherm (which is the
usual case for metals), these models will be inaccurate. The best that can be hoped for when
single partition coefficients are used to describe metal adsorption is that they represent bounding
values in a given application.
4.1.4.1.5. Use indecision making. U.S. EPA (1999a) discusses the advantages and
disadvantages of several methods for measuring partition coefficients, including laboratory batch
testing, in situ field batch testing, flow-through testing, and field modeling. In many national
assessments, EPA has used the MINTEQ model and its subsequent versions to generate generic
partition coefficients that may be applied to regional or national mobility evaluations
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy. • . ,
4-16
o
-------
1 (http://www. epa.gov/ceampubl/mmedia/ minteq/index.htm or http://www.lwr.kth.se/english/
2 OurSoftware/Vminteq/).
3 For screening assessments:
4
5 • For screening assessments, partition coefficients have been tabulated as a function of
6 pH by EPA (U.S. EPA, 1998b), and the Agency has also presented non-pH-
7 - dependent values for lead (900), mercuric chloride (58,000), and elemental mercury
8 (1,000) (U.S. EPA, 1999b).
9 ' - '
10 • In simple systems, the value of log Kj for metal cation adsorption usually increases
11 linearly with pH, whereas the value of log Kj generally decreases with pH for anion
12 adsorption.
13
14 For definitive assessments:
15
16 • It may be possible to estimate metal adsorption with some accuracy without having to
17 measure it, depending on the information available on a specific soil, surface water,
18 or ground water system. What is needed minimally is the amounts and surface areas
19 of the potentially sorbing materials (e.g., metal oxides, clays, and OM) in a soil or
20 sediment or in suspension in a stream, and the detailed chemical composition of the
21 water, especially its pH and metal concentration. Literature information can then be
22 used to estimate the sorption properties of these materials for use in the DL sorption
23 model, for example.
24
25 • If greater accuracy or site specificity is required, it may be necessary to measure
26 metal adsorption in laboratory experiments designed to parameterize the DL model
27 for application to a specific study area. The experiments could be batch tests that
28 attempt to reproduce the composition of waters and sorbing solids in the study area.
29
30 4.1.5. Sediment Chemistry
31 In addition to the challenges posed by metal chemistry, the sedimentary environment is
32 complex and often highly heterogeneous. Fortunately, we can generalize about the sedimentary
33 environment and the main controlling factors to progress toward a method for risk assessment.
34 This brief review summarizes information on the composition of sediments; processes that act on
35 sediments and their metal burden; and the chemistry of the sedimentary environment that
36 influences the fate, bioavailability, and effects of metals. It is important to consider these factors
37 in light of the aim of estimating potential biological effects of metals in sediments.
38 Sediment solids can hold up to a million times more metal than an equivalent volume of
39 water. The exact proportions of a chemical held by sediment relative to water is a function of a
11/24/2004 Peer Review Draft
DISCLAIMER: Tliis information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-17
-------
1 metal's chemistry as well as the chemistry of the sediment solid and the surrounding
2 environment, and this distribution is dynamic (Diamond and Mudroch, 1990). Because of their
3 large capacity to "hold" metals, sediments have been characterized as "sinks." Although this is
4 largely true, sediments are now appreciated to be
5 temporary sinks, wherein some of the metal can enter
6 ecological and human food webs through several
7 routes (e.g., Diamond, 1995), primarily through
8 accumulation in benthic organisms. These organisms
9 include those that fully or partially live in the
10 sediments (e.g., tubificids, chironomids, trichopteran
11 larvae) or those that feed from the sediment bed (e.g.,
12 suckers, carp). Some organisms obtain their chemical
13 dose from both pelagic and benthic routes (e.g., lake
14 whitefish, walleye), but because of high chemical
15 concentrations in sediments, the benthic route can be
16 the dominant route of uptake (Morrison et al., 2000).
17 For humans, the route of entry of metals from
18 the sediments is through water used for drinking,
19 bathing, and swimming. The availability of these metals is mediated by sediment-water
20 exchange processes that can result in the release or remobilization of chemicals from the
21 sediment bed. However, due to the ability of Hg to bioaccumulate in its monomethyl form, fish
22 consumption is the critical route of exposure to this metal for humans.3
23 . Many important chemical reactions involving the metals of concern occur in the fine-
24 grained materials that accumulate in the deep parts of water bodies. The controlling factors or
25 master variables that influence metal chemistry are redox potential and pH. A depth profile of
26 the sediments will reveal decreasing sediment porosity and concentrations of dissolved oxygen
27 because oxygen is consumed as organic matter decomposes. pH is often relatively constant or
28 may decrease with depth, but alkalinity may increase owing to mineralization of organic matter
29 (Stumm and Morgan, 1996). As dissolved oxygen is consumed, anaerobic microbes use other
30 electron acceptors in redox or oxidation-reduction reactions in the order of nitrate, ferric iron,
31 ammonium, sulfate, and bicarbonate to produce carbon dioxide, ammonia, sulfide, and methane
Sediments
Bed or bottom sediments are found at
the bottom of lakes, rivers, and estuaries.
Sediments have several sources that
influence their composition and chemistry.
The type and chemistry of sediments is
also determined by their location in the
water body as well as the characteristics of
the water body. At any given site, metals
can be associated with solid-phase
minerals, organic matter, colloids, and
pore water. The solid phase can vary from
sand (>63 urn), to silt (2-63 um), to clay
(<2 nm). Because clays have more active
binding sites than do the other grain sizes
and because of their high surface area-to-
volume ratio, fine-grained particles are of
greatest significance in terms of metal
binding.
3Marine biota can also be a significant route of exposure to arsenic in its organic forms, such as
methylarsonic acid; however, these arsenic species are significantly less toxic than the inorganic forms (Fowler,
1983).
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-18
-------
1 The sequence of microbially mediated redox reactions that produce these reduced species are
2 given in Figure 4-2.
3 The redox status of the sediments can be assessed by measuring the concentration of
4 dissolved oxygen or other redox-sensitive species, by calculating pE (a measure of electron
5 availability in solution), or by measuring Eh (millivolts) or the electromotive force of the pore
6 water solution. The zone of transition from oxic to anoxic conditions is the redoxycline, which
7 can migrate vertically, depending on the mixing of the overlying water column (e.g., Diamond
8 and Mudroch, 1990). For example., the redoxycline may be 5 to 10 cm below the sediment-water
9 interface in a well-oxygenated oligotrophic lake or river, but it may be above the sediment-water
10 interface in a thermally stratified eutrophic lake or river.
11
12 4.1.5.1. Metal Chemistry in Sediments
13 In this discussion, two pools of metals should be considered. The first pool consist of
14 metals that exist as aqueous (or dissolved) species bound to colloids or DOM and those bound to
15 sediment particles through an exchangeable binding process. This pool is often referred to as the
16 exchangeable or labile pool. The second pool consists of metals found within the mineral matrix
17 of the sediment solids. This pool is largely unavailable to biota, and its release will occur over
18 geologic time scales through diagenetic processes. Because the latter pool is largely unavailable,
19 we will consider only the exchangeable pool of metals. Note that the exchangeable pool will be
20 composed of naturally occurring metals that are released into solution due to weathering and
21 diagenetic processes as well as metals released into the environment due to anthropogenic
22 activities.
23 The exchangeable pool of metals is subject to speciation in the aqueous phase (e.g.,
24 within the pore water) and sorption to solid phases, where sorption is a general term that includes
25 adsorption (the accumulation of matter at the solid-water interface or a two-dimensional process)
26 and absorption (inclusion in a three-dimensional matrix) (Stumm and Morgan, 1996). Here,
27 speciation refers to the distribution of metal species in a particular sample or matrix or species
28 distribution (Templeton et al., 2000). In the aqueous phase, metal will react or bind with
29 dissolved ligands according to the pH, Eh, ionic strength, and abundance of ligands (see above
30 discussion on aquatic chemistry).
31 The concentration of metal in the dissolved phase is controlled by sorption to the solid
32 phase. Although sorption can occur rapidly, desorption or dissolution of metal from the solid
33 phase may be a two-phase process, where the second phase is rate limiting (e.g., Nyffeler, 1986;
34 Santschi et al., 1986). If we neglect the kinetic limitation of reactions, the distribution of metals
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-19
-------
1 among aqueous species and between the aqueous phase and the solid phase can be estimated.
2 Several speciation/complexation models are available to perform this calculation, such as
3 MINEQL+ (Schecher and McAvoy, 2001), the Windermere Humic Aqueous Model (WHAM)
4 (Lofts and Tipping, 1998), and MINTEQA2 (Allison et al., 1991). These models work well.
5 under oxic conditions, but estimates of metal binding are less reliable under anoxic conditions,
6 where metal concentrations are most often controlled by the solubility of metal sulfides.
7 Furthermore, in some circumstances equilibrium may not be achieved, particularly when the
8 redoxycline moves more quickly than the rate of metal reaction or when the reaction is governed
9 by microbial processes, as occurs with the methylation of mercury or arsenic.
10 In oxic sediment pore waters (above the redoxycline), metals will exist as aqueous
11 species, that is, as freely dissolved ions or metal complexes (e.g., phosphate, sulfate, or
12 carbonate complexes), and associated with colloids. Solid-phase reactions are controlled by iron
13 oxyhydroxides and manganese oxides that may exist as colloids, sediment particles or surface
14 coatings of particles, OM that may also exist as colloids or coat sediment particles, and clay
15 colloids and particles.
16 As Eh declines, the solid-phase manganese oxides are the first to be reduced and thereby
17 dissolve, which releases metals that have been sorbed or coprecipitated with these minerals.
18 Some of the metals released into tiie pore water may then be adsorbed by iron(III)
19 oxyhydroxides, which are the next to dissolve as the Eh continues to drop. Under reducing
20 conditions, particularly as sulfate is consumed and the sulfur is converted to sulfide, metal
21 concentrations in pore waters again drop as solid-phase metal-sulfides are formed (see discussion
22 below about the role of AVSs in regulating toxicity).
23 As a result of redox chemistry, metals can undergo seasonal redox-driven cycling
24 between the water column and sediments or within the sediments, depending on the position of
25 the redoxycline. The stages in the cycling are, first, the adsorption or coprecipitation of metals
26 with iron and manganese hydroxides under oxidizing conditions; then with the development of
27 moderately reducing conditions, the reduction and dissolution of the manganese and iron
28 oxyhydroxides, and consequent release of the associated metals into the water or pore water;
29 followed by their diffusion upward toward the zone of low metal concentrations under oxidizing
30 conditions. It is also possible for dissolved metals to diffuse downward toward the zone of low
31 metal concentration owing to their precipitation as sulfides. As a result of this vertical cycling,
32 me depth profile of metals in pore water may not match that of the solid phase (e.g., Carignan
33 and Tessier, 1985). Moreover, it is possible, but less usual, that the cycling can occur relatively
34 rapidly and involve a significant portion of the solid-phase metal. Under these conditions, the
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-20
-------
1 solid-phase sediment profile reflects this reworking rather than the historical record of metal
2 loadings (MacDonald et al., 2000).
3 pH controls metal speciation and binding by affecting the species distribution of
4 dissolved ligands (e.g., phosphate, sulfate, carbonate, humic substances) and the surface charge
5 of binding sites on DOM and solid phases such as iron oxyhydroxides. Generally, at low pH,
6 when surface sites are protonated, the sorption of cationic metals decreases, and, hence, metal
7 mobility increases. The converse occurs at high pH, which results in low metal solubility and
8 greater sorption. The patterns of dissolution and sorption are reversed for metalloids, such as
9 arsenic, that exist as anionic species.
10
11 4.1.5.2. Methods of Estimating Metal Distribution in Sediments
12 4.1.5.2.1. Application. A main objective in terms of assessing the hazard or risk posed by
13 metals is estimating metal in the dissolved phase that is potentially bioavailable. Accordingly,
14 several methods have been developed to estimate the distribution of metals among dissolved and
15 solid phases in sediments. These methods have been thoroughly reviewed by Mudroch et al.
16 (1999,1997). Although bioavailability is also a function of aqueous phase speciation (see
17 Section 4.1.3, Aquatic Chemistry), limited research has been conducted to estimate metal
18 speciation in pore waters. Generally, ecological risk assessments assume that the exposure of
19 benthic organisms to sediment-associated metal is proportional to the metal concentrations in
20 interstitial water, although some studies indicate that uptake from overlying water (Hare et al.,
21 2003; Roy and Hare, 1999) or ingested sediment may be a significant source of body burdens of
22 metals (see Section 3.2, Metals Risk Assessment Recommendations for Aquatic Environments.,
23 for more discussion of this topic).
24 Distribution of metals in sediment pore waters may be determined by field
25 measurements, experimental methods, and mathematical modeling, with the latter also requiring
26 some field measurements. Concentrations of metals in pore waters may be determined in the
27 field by use of pore water dialysis chambers or peepers arid by methods that separate the solid
28 phase from the pore water, although the latter have been shown to be less reliable (Mudroch et
29 al., 1997). Several extraction schemes have been developed to determine the distribution of
30 metal among operationally defined fractions (e.g., Forstner, 1995a; Tessier et al., 1979). It is
31 well known mat sequential extraction methods do not cleanly distinguish the occurrence and
32 speciation of different forms of metals in sediments and soils (Tye et al., 2003; Verloo, 1999).
33 Other experimental methods include leaching tests (e.g., Reuther, 1999). The results of any of
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. II has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-21
-------
1 these methods are concentrations of metals in pore water, which can be related to toxiciry
2 benchmarks.
3 Because of the need to develop Sediment Quality Guidelines (SQGs) for metals that
4 explicitly address toxicity and are based on readily measured parameters, several methods have
5 been developed. For oxic sediments, Tessier and co-workers (Tessier et al., 1993,1989,1984;
6 Tessier, 1992) compiled partition coefficients of metals that were derived from field studies of
7 freshwater sediments. The partition coefficients are dependent on pH (because Eh is held
8 constant) and are generally linear over a range of pore water pH values (see above discussion
9 under ground water chemistry for the theoretical basis for development of partition coefficients).
10 Speciation/complexation models also may be used to estimate fractions of dissolved and bound
11 metal species. These models rely on measurements of pH, dissolved oxygen, or Eh to establish
12 redox conditions. The models assume that solid-phase binding is governed by sorption to iron
13 and manganese oxides. Model estimates are less reliable when other solid-phase substrates are
14 dominant (e.g., clay minerals) and are a function of the availability and accuracy of the stability
15 constants for the metal-ligand reactions that are used in the calculations.
16 For anoxic sediments, the availability of sulfide controls metal distribution and solubility.
17 Operationally, AVSs—mainly iron monosulfide—have been considered as a measure of reactive
18 sulfides (Forstner, 1995a). Studies have demonstrated an inverse relationship between sediment
19 toxicity and AVSs for marine and freshwater sediments (Di Toro et al., 1992,1990; Ankley et
20 al., 1991) as well as between pore water concentrations and AVSs (Brumbaugh et al., 1994;
21 Casas and Crecelius, 1994). As a screening-level tool, the toxicity of anoxic sediments can be
22 assessed by determining the ratio of AVSs to simultaneously extracted metal (SEM). Low
23 sediment toxicity is indicated when AVSs are in excess (AVS > SEM), which implies sufficient
24 capacity of the AVS to bind essentially all free metal. This topic is further discussed in Section
25 4.5.
26 For estimating effects concentrations in the sediment, there are many different SQGs,
27 which vary in their derivation and the degree to which they incorporate bioavailability
28 considerations. Many of the published SQGs are based on empirical relationships between
29 biological effects and the total (dry weight) concentrations of sediment contaminants (e.g.,
30 McDonald et al., 1996; Long and Morgan, 1991). Although these empirically based guidelines
31 do show general relationships between the degree of sediment contamination, they do not
32 explicitly account for site-specific differences in bioavailability of contaminants. Although EPA
33 has not formally adopted any single SQG approach as an Agency standard, it has been active in
34 . developing the "equilibrium partitioning" (EqP) approach to SQG development The EqP
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-22
-------
1 approach considers effects of sediment chemistry on bioavailability by comparing the
2 concentrations of AVSs, SEM, and organic carbon (U.S. EPA, 2000c; Ankley et al.,1996,1993;
3 Di Toro et al., 1992,1990). This approach is reviewed in more detail in Section 4.5,
4 Characterization of Ecological Effects.
5
6 4.1.5.2.2. Limitations. Model estimates are less reliable when other solid-phase substrates are
7 dominant (e.g., clay minerals), and they are a function of the availability and accuracy of the
8 stability constants for the metal-ligand reactions that are used in the calculations.
9
10 4.1.6. Soil Chemistry
11 The cationic metals occur naturally in soils as oxides and hydroxides (Fe, Mn, Al); to a
12 lesser extent as carbonates, phosphates, and sulfates; and in reducing (usually wet or
13 waterlogged) soils as sulfides, which are highly insoluble. The soil parameters important in
14 affecting sorption and precipitation reactions and the extent of their influence—and thus
15 contaminant bioavailability—depend on the intrinsic properties of the contaminants. In the soil
16 environment, metals can exist as cations, anions, or neutral species. Their form significantly
17 affects their sorption, solubility, and mobility. For example, most soils are chiefly negatively
18 charged; thus, metal cations have a higher propensity to be sorbed by soil particles than do metal
19 anions (U.S. EPA, 2003e).
20 Cationic metals can react with inorganic soil constituents (e.g., carbonates, sulfates,
21 hydroxides, sulfides), forming aqueous complexes, which may be adsorbed or precipitated in
22 mineral form. Most complexation and precipitation reactions are pH dependent (U.S. EPA,
23 2003e).
24 Arsenic, chromium, selenium, and vanadium complex with oxygen and typically exist as
25 anionic species under most environmentally relevant scenarios (Bohn et al., 1985; Lindsay,
26 1979). The most common forms of arsenic are arsenate (arsenic(V)) and arsenite (arsenic(III)),
27 which are present in soil solution in the form of AsO43" and AsO33", respectively. The chemistry
28 of arsenic resembles that of phosphate (Barber, 1995; Bohn et al.51985). Chromium can exist as
29 chromate (chromium(VI) or CrO42~), which is usually considered more soluble, mobile, and
30 bioavailable than the sparingly soluble chromite (chromium(III)), which is normally present in
31 soil as the precipitate Cr(OH)3 (Barnhart, 1997; James et al., 1997). Similarly, selenium can be
32 present as selenates (SeO42') and selenites (Se032~). For vanadium, vanadate (VOf) is the most
33 common form (U.S. EPA, 2003e).
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-23
-------
1 Metals can exist in the pore water as aquo-ions or soluble complexes. The bonding of
2 metal species to soil particles can range from ionic to covalent. For most soils in the United
3 States, negatively charged sites are more plentiful; less than 5% of the total available charge on
4 the soil surface is positively charged. Metals existing as cationic species have a greater
5 propensity to associate with such soils. This makes them less bioavailable, but it also results in
6 greater loading of metals into the soil ecosystem. Anionic metals generally move into pore
7 water—and so are more bioavailable—but leach out of the system much more rapidly, In
8 summary, soil pH and availability of.charged sites on soil surfaces are the primary soil factors
9 controlling release of metals to pore water and, subsequently, bioavailability (U.S. EPA, 2003e).
10
11 4.1.6.1. Key Parameters Affecting Metal Bioavailability in Soils
12 From the preceding overview of how the metals and metal compounds interact with soil
13 constituents, it is clear that soil plays a very significant role in reducing the potential
14 bioavailability of metals in the environment. Given the types of contaminant-soil interactions
15 presented, the primary soil factors controlling the potential bioavailability of metals are soil pH,
16 the availability and character of sorption sites on soil surfaces, the content of Fe and Al
17 oxyhydroxides and soil organic matter, and least important, the soil clay mineral content The
18 following discussion briefly details the key soil parameters affecting the various contaminants
19 availability to the pore water.
20 Soil pH is often termed the master soil variable because it controls virtually all aspects of
21 contaminant and biological processes in soil. These processes include dissolution and
22 precipitation of metal solid phases, complexation and acid-base reactions of metal species, and
23 metal sorption as well as microbial activity. Increasing soil pH also results in an increase in the
24 number of negatively charged soil sites, with a concomitant decrease in the positively charged
25 sites. Therefore, increasing the soil pH increases the sorption and removal from pore water
26 (Bonn et al., 1985).
27
28
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-24
-------
1 4.1.6.2. Adsorption Behavior of the Metals of Concern
2' 4.1.6.2.1. Surface area and surface density. In porous media, the most important sorbent solids
3 for metals are oxyhydroxides of Fe and Mn. Their important surface properties are discussed in
4 Langmutr et al. (2003). For a given weight of sorbent, metal sorption capacity is proportional to
5 surface area and surface site density. The greatest surface site densities (positively or negatively
6 charged sites) are those of organic material and the oxyhydroxides. These phases are the
7 strongest and most important sorbents of trace metals. Except for kaolinite, the clays (0.02-2
8 mmol sites/g) have a surface charge that is largely independent of pH, whereas the surface
9 charge of organic matter and the oxyhydroxides is strongly pH dependent.
10
11 4.1.6.2.2. Importance ofpH. The pH at which a solid surface changes sign is referred to as the
12 zero point of charge (ZPC). The ZPC of organic matter and kaolinite are below the range of
13 usual pH and while their surface charge decreases with decreasing pH, their surfaces remain
14 negatively charged above pH 2-3. In contrast, Fe(in) and Al oxyhydroxides, such as goethite
15 (a-FeOOH), have ZPC values near pH 7-8 and so have a positive surface charge at low pH and a
16 negative surface charge at high pH. Thus, the negative surface charge of the oxyhydroxides, and
17 their sorptive capacity for metals, increases with increasing pH. Conversely, the positive surface
18 charge of the oxyhydroxides increases as the pH drops, making these phases more effective
19 sorbents for anions under low pH conditions.
20 These effects'are shown in Figure 4-3, which is a plot of percentage sorbed versus pH for
21 metal adsorption by ferrihydrite, or HFO. The curves are called sorption edges for each metal.
22 The diagram shows that the oxyanions are strongly adsorbed by HFO under acid conditions but
23 are desorbed (become mobile) with increasing pH. Based on Figure 4-3, combined with
24 adsorption data assembled by Dzombak and Morel (1990), the order of desorption from HFO
25 with increasing pH is selenate, antimonate, molybdate, chromate, vanadate, arsenate, and
26 phosphate. Selen ate desorbs between pH 3 and 8, whereas arsenate is strongly held at lower pH
27 values and desorbed between pH 9 and 11. Also based on Figure 4-3 and Dzombak and Morel
28 (1990), with increasing pH, HFO preferentially adsorbs metals in the order
29 H^Be^Ba^-Cr^Pb^Cu^Cd2* * Zn2+>Ni2+.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-25
-------
(a)
100
so
60
40
20
0
100
80
60
40
20
0
SeOj-
8
pH
10
12
(W
pH
Figure 4-3. Adsorption of various metal cations and oxyanions, each at 5 *
107 M, by ferrihydrite (EFe[III] = 103 M) as a function of pH at an ionic
strength of 0.1 mol/kg. There are 2 x 10"" M of reactive sites on the
oxyhydroxide. The dashed curves are calculated after Stumm (1992).
1 4.1.6.2.3. Organic matter (organic carbon) content. Organic matter includes plant and animal
2 remains in various stages of decomposition, cells and tissues of soil organisms, and substances
3 exuded from plant roots and soil microbes (Sumner, 2000). Organic matter is primarily
4 composed of carbon, oxygen, and minor amounts of nitrogen and phosphorus. On average,
5 approximately 58% of organic matter is organic carbon. Soils encompass a range in organic
6 matter, from <1% for a sandy soil to almost 100% for a peat soil, with most soils having organic
7 matter contents <10% (Bohn et al., 1985). Also, organic matter content is usually higher in
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the 'purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-26
-------
I surface soils or in the root zone and decreases with depth in the soil profile. The organic acid
2 functional groups typically present in organic matter have a high affinity to attract metal cations.
3
4 4.1.6.3. Aging of Metals in Soil
5 A distinction should be made between persistence of total metals in soil and persistence
6 of bioavailable forms of the metal. As metals age in soils, they decrease in bioavailability over
7 time. It has been well documented that metal chemistry in solutions freshly added or spiked into
8 soils vary from metal forms in field-contaminated soils. Typically, the metal contaminant pool
9 requires time to diffuse into micro- ornanopores and to be absorbed onto organic matter and soil
10 particles. These slow reactions are attributed to micropore diffusion, occlusion in solid phases
11 by (co)precipitation and (co)flocculation, and cavity entrapment. Although the slow reactions
12 play a key role in metal bioavailability, their rates, mechanisms, and controlling factors have not
13 been comprehensively elucidated. Evidence of aging processes is provided by studies of metal
14 extractabilily and lability (Young et al., 2003; Hamon et al., 1998). It has been frequently
15 observed that easily extractable pools revert with time (~1 year) to more strongly bound forms.
16 Isotopic dilution provides a useful way to quantify changes associated with progressive
17 attenuation of metals in soil. Aging reactions are almost over after about 1 year and are
18 reversible. At present, information regarding the aging reactions of different metals and
19 metalloids, and sorbing solids, is very limited, so it is not possible to generalize which metal(s)
20 age the fastest or with greater/less reversibility.
21
22 4.1.6.3.1. Steady-state calculations for metals in soils
23 4.1.6.3.1.1. Application. Aging reactions can be determined empirically by calculating
24 partitioning through measurement of the soil pore water concentration of a metal in well-
25 equilibrated soils. If KjS are calculated from adsorption isotherms, aging should be considered
26 separately. This is related to the high affinity of metals for soil solid phases. In fact, for most
27 metals, metal losses by leaching, erosion, or removal by a crop that is harvested are small when
28 compared with the total metal concentration. However, for some elements, such as Se, the half-
29 life in soil is significantly shorter. Critical factors that affect the mass balance of metals are the
30 anthropogenic and natural inputs and outputs via leaching to ground water, the removal through
31 surface erosion, and crop harvesting. Excluding erosion processes, the elimination half-life of
32 metals in soil (t,/2) can be predicted from a soil mass as follows:
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-27*
-------
0.69 x d x 10000
Mfl
y* TF +
R
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
where d is soil depth in meters,
y is annual crop yield (t ha"1 y"1),
TF is the ratio of the metal concentration in plant to that in soil,
R is the net drainage loss out of the soil depth considered (m3 ha"1 y'1),
Kj is the ratio of the weight of metal adsorbed to the weight of the sorbent,
divided by the metal concentration in a volume of soil solution (L Kg'1).
The time required to achieve 95% of steady state is about four half-lives. This is shown
in Table 4-5 for select metals and metalloids. Selenium approaches steady state after only 1
year. As a consequence, Se soil concentrations after 100 years and at steady state are identical.
In contrast, concentrations of Cu, Cd, Pb, and Cr(III) are still well below steady-state values after
100 years and, consequently, their concentrations in soil are very similar. The time necessary to
approach steady state is a function of the loading rate and Kd.
Table 4-5. Time to achieve 95% of steady-state metal concentration in soil and total
soil metal concentrations after 100 years and at steady state3
Metal
Se
Cu
Cd
Pb
CrIII
Loading rate
(ghaV)
100
100
100
100
100
Kd
(L kg V
0.3
480
690
19000
16700
T
(years)
1.3
1860
2670
73300
64400
Soil metal concentration
(mg added metal kg'1)
Steady state
0.01
16
23
633
556
After 100 yrs
0.01
2.4
2.4
2.6
2.6
'Assumes a soil depth of 25 cm and a net drainage loss of 3000 m3 ha'1 y"1; background was zero at the start of
loading.
k Mean Kd (ratio of total metal concentrations in soils to that in pore water) of 49 Dutch soils (De Groot et al., 1998).
Source: Adapted from Smolders et al. (2004).
4.1.6.3.1.2. Limitations. It should be noted that the time needed to approach steady state for all
the metals considered, except Se, is in the order of thousands of years, and it is difficult to
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-28
-------
1 envision that soil conditions would not change in this time frame. However, the concept is
2 useful because it provides a time frame within which to consider risk as a function of loading
3 capacity of the soils and the potential for continued exposure even after cessation of
4 anthropogenic inputs of metals to soils. Limitations on the application of Kj are discussed in
5 Section 4.1.4. .
6 '
7 4.1.6.3.2. Laboratory methods to simulate aging in soils
8 4.1.6.3.2.1. Application. The aging effect requires laboratory studies on soils to apply atime-
9 dependent weathering or aging treatment of spiked soils. Critical toxicity values generally are
10 based on toxicological tests performed during the period of relatively fast metal adsorption that
11 follows metal addition to soil. Such values would be lower than those derived from a similar
12 study conducted with soils a year or more after addition of the metal. McLaughlin et al. (2000)
13 proposed that toxicity thresholds be set using a sequential testing procedure. Tests would be
14 conducted within 2 to 7 days following incorporation of the test substance to generate an
15 estimate of acute hazard. Another set of soils would be tested 60 days after mixing, and a third
16 would be subject to a leaching process and also tested after the 60-day period. It has been
17 estimated in preliminary studies that toxicity is reduced up to 10-fold owing to aging of metals in
18 soils. However, further studies are warranted and standardized processes should be agreed upon
19 before metals aging can be properly accounted for in soil toxicity testing.
20 "
21 4.1.6.3.2.2. Limitations. The leaching and equilibrium times are limited for practical reasons
22 and are meant to simulate some degree of aging and dissolution as a result of weathering. If
23 aging occurs at faster rates than does dissolution, then toxicity will decrease with time.
24 Conversely, if dissolution occurs at a greater rate, then toxicity will increase.
25
26 4.1.6.4. Dissolution and Transformation of Metals
27 4.1.6.4.1. Application. The dissolution and transformation of a metal compound in soil is
28 related to a series of chemical and physical properties characteristic of the compound itself and
29 of the soil. Environmental parameters such as temperature and humidity have a strong influence
30 on the rate of transformation. When assessing the transformation of a compound in soil, it
31 should be remembered that aging reactions may take place at the same time as transformation
32 and dissolution. When metal salts are added to soil, the form of the salt dictates the rate and
33 amount of soluble metal that will form in the pore water. Insoluble forms of metals (e.g.,
34 vanadium pentoxide [V2O5]) will transform to soluble free ion (V) at a slower rate than will
11/24/2004 Peer Rev lew Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-29
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
soluble metal salts (e.g., Na3VO4). However, the rate of formation of the free ion is not
proportional to the dissolution rate of the salt because aging reactions will remove the free ion
from the pore water. The relative rates of dissolution and aging should be considered
simultaneously to accurately predict pore water concentrations.
4.1.6.4.1.1. Solution speciatlon (computer-based models). Computer-based models employ
either equilibrium constants or Gibbs free energy values to determine metal speciation from
solution chemistry conditions (concentration, pH, Eh, organic complexes, adsorption/desorption
sites, and temperature). Both approaches are subject to mass balance and equilibrium conditions
that should be defined. In recent years, as more accurate thermodynamic data have become
available, the models have undergone extensive development and can provide useful predictive
estimates of metal behavior. A good review of these models and their applications is provided
by Lumsdon and Evans (1995). Examples of computer-based speciation models include
MINTEQL, REDEQL2, ECOSAT, MINTEQA2, HYDRAQL, PHREEQC, and WATEQ4F.
Both MINTEQA2 (U.S. EPA, 1991b) and VMINTEQ (Gustafson, 2003) contain
subroutines that allow estimates of the importance of metal-organic complexing if the
concentration of DOC is known. Perhaps more useful in studies of metals in soil solution are
programs such as WHAM (Tipping, 1998,1994) and the Non-Ideal Competitive Absorption
(NIC A) model (Gooddy et al., 1995). Application of the chemical speciation model WHAM has
been discussed by Tye et al. (2003), who successfully predicted Zn2+ and Cd2* activities in soil
pore water by assuming the metals were adsorbed by soil humus according to a pH-dependent
Freundlich isotherm model. Competitive adsorption between Ca2+ and Zn2+ and Cd2+ could be
ignored because it did not improve model fits.
4.1.6.4.2. Limitations. In some instances, metal speciation is controlled by simple reactions.
However, in many cases (particularly in contaminated media), the state of equilibrium and the
reversibility of metal reactions are unknown. In addition, mathematical thermodynamic
equilibrium models suffer from other limitations, such as lack of reliable thermodynamic data for
relevant species, inadequacies in models to correct for high ionic strength, poorly known
reaction kinetics, and complex reactions and lack of models for co-precipitation/adsorption. The
first limitation is perhaps the most significant for contaminated media. As an example, none of
the models can predict the behavior of the common, anthropogenic lead phases in paint, solder,
or slag.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-30
-------
1 4.1.6.5. Soil Metal Transfer to Plants
2 The "soil-plant barrier" concept was introduced to communicate how metal addition, soil
3 chemistry, and plant chemistry affect risk to animals from metals mixed in soil (Chancy,
4 1980). Reactions and processes that take place at the soil-plant barrier are influenced by the
5 following factors: (1) soil solid phases (e.g., Fe, Al, and Mn oxyhydroxides and organic matter)
t
6 may have adsorptive surfaces that influence soil chemistry; (2) adsorption or precipitation of
7 metals in soils or In roots limits uptake-translocation of most elements to shoots; and (3) the
8 phytotoxicity of Zn, Cu, Ni, Mn, As, B, Al, and F, for example, limits residues of these elements
9 in plant shoots. More recently, inductively coupled plasma-mass spectrometry (ICP-MS) and
10 other very sensitive analytical methods have been used to examine soil solution and soil-plant
11 transfer of 60 elements as a function of soil pH (Tyler and Olsson, 2001 a, b). These studies
12 provide evidence that further supports the concept of the soil-plant barrier
13 For strongly adsorbed metal cations, the pattern of plant response to metals contained in
14 biosolids is strongly curvilinear (i.e., plant metal concentration approaches a plateau with
15 increasing soil metal concentration), rather than being linear with increasing concentration.
16 Several areas for potential errors in the research methodology should be avoided when making
17 these comparisons:
18
19 • First, comparison of application rates is valid only after the system has equilibrated
20 utilizing accepted methods.
21
22 • Second, soil pH levels should be equal across rates studied; co-variance of soil pH
23 should be used to correct for unequal soil pH (Bell et al., 1988).
24
25 • Third, the metal concentration in the source applied affects the slope of metal uptake:
26 higher metal concentration in the source means higher phytoavailability at equal
27 metal applications (Jing and Logan, 1992).
28
29 4.1.6.5.1. Application/limitations. Strongly acidic soils increase plant uptake of Zn, Cd, Ni,
30 Mn, and Co and increasl the potential for phytotoxicity from Cu, Zn, and Ni. In alkaline soils,
31 the high pH increases uptake of Mo and Se. Lead and Cr are not absorbed by plants to any
32 significant extent at any pH (Chaney and Ryan, 1993). However, each element should be
33 considered separately because of its unique chemistry. For example, arsenate is more strongly
34 adsorbed than is arsenite; when a soil is flooded to grow rice, soil microbes can reduce arsenate
35 to arsenite, and the higher concentration of dissolved arsenite can be phytotoxic to rice in more
36 highly contaminated soils. Most other elements have little potential for redox change with
37 change in the redox status of soils. Reduced soils can form sulfide, and sulfide forms low-
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-31
-------
1 solubility compounds with most of the metals of concern in soils, including Pb, Zn, Cd, Cu, and
2 Ni (see above discussion on sediment chemistry). For essential elements (e.g., Zn, Cu, Ni), low-
3 solubility species can result in deficiency syndromes. Upon oxidation of the soil, sulfide is
4 quickly oxidized, and the metals are returned to more normal equilibrium reactions of aerobic
5 soils.
6
7 4.1.7. Atmospheric Behavior/Chemistry
8 Most metals and metal compounds exist in the solid phase under ambient conditions and
9 thus occur almost exclusively in the particle phase of the atmosphere, where they are ubiquitous.
10 Anthropogenic sources include combustion from fossil fuels, and metal industries, as well as
11 industrial sources employing specific metal compounds in specific processes. Some airborne"
12 metals (e.g., Mn and Ni) may derive largely from crustal sources (U.S. EPA, 1996b).
13 Richardson (2002) included volcanic eruptions and emissions, entrainment of soil and dust,
14 entrainment of sea salt spray, and natural forest fires as significant metals emission sources.
15 For purposes of risk assessment, particle size is important. The aerodynamic size and
16 associated composition of particles determine their behavior in the mammalian respiratory
17 system. Furthermore, particle size is one of the most important parameters in determining the
18 atmospheric lifetime of particles, which may be a key consideration in assessing inhalation
19 exposures, as well as exposures related to exposure pathways involving deposition onto soil or
20 water (U.S. EPA, 1996b). Metals emitted by combustion processes (e.g., the burning of fossil
21 fuels or wastes) generally occur in small particles or the fine fraction, which is often
22 characterized by particles less than 2.5 microns in diameter (PM25). In contrast, the larger sized,
23 course mode particles result from mechanical disruption, such as crushing, grinding, evaporation
24 of sprays, or suspensions of dust from construction and agricultural operations. Accordingly,
25 metals in course mode particles (i.e., those larger than approximately 1-3 microns) are primarily
26 those of crustal origin, such as Si, Al, and Fe (U.S. EPA, 1996b). It is noted that the fine versus
27 coarse distinction simply differentiates two relatively distinct size distributions of particles, the
28 separation point of which occurs in the range of 1 to 3 um. The distinction does not refer
29 directly to particle sampling methods or size fractionations particular to risk assessment (U.S.
30 EPA, 1996b).
31 Fine and coarse particles typically exhibit different behavior in the atmosphere; fine
32 mode particles exhibit longer atmosphere lifetimes (i.e., days to weeks) than coarse particles and
33 tend to be more uniformly dispersed across a large geographic region (U.S. EPA, 1996b).
34 Relatively lower dry deposition velocities of fine particles contribute to their persistence and
35 uniformity throughout an air mass (U.S. EPA, 1997c). The larger coarse particles (i.e., greater
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
• Agency determination or policy.
4-32
-------
1 than 10 um) tend to rapidly fall out of the air and have atmospheric lifetimes on the order of
2 minutes to hours, depending on their size and other factors (U.S. EPA, 1996b).
3 In most cases, metals do not undergo transformation while in the particle phase; thus,
4 their removal from the atmosphere is regulated by the rate at which the particles themselves
5 participate in wet and dry deposition processes. For example, metals such as As, Be, Cd, Pb are
6 generally found in airborne compounds with a single predominate oxidation state (As0II),
7 Be(II), Cd(II), Pb(II)). S ome metals (e.g., the transition metals Cr, Mn, and Ni) present the
8 possibility of changing oxidation state in situ in the particle, although little is known of these
9 processes (U.S. EPA, 2003d). This is an important consideration for risk assessment as the
10 different oxidation states also differ in toxicity (such as for Cr).
11 For metals that can change oxidation states, much of the atmospheric chemistry takes
12 place in the aqueous phase, such as cloud droplets or water films on particles. Metal salts and
13 oxides that dissolve in water can undergo several reversible reactions, including hydration,
14 hydrolysis, polymerization, and reaction with other anions. The equilibrium between these
15 forms depends on the atmospheric conditions, the equilibrium and solubility constants, and the
16 concentrations of other chemicals. Transformations between oxidation states can occur either to
17 increase the oxidation state (such as oxidizing Cr(III) to Cr(VI)) or to reduce it. These oxidation
18 or reduction reactions can occur through reaction with other species, such as dissolved metals,
19 reduced sulfur species, and organic compounds (Seigneur and Constantinou, 1995). Although
20 models exist that can be used to estimate metal speciation in aerosols with liquid water, the
21 reactions are still highly uncertain.
22 Mercury is an exception among the commonly occurring metals; it exists primarily in the
23 vapor phase under ambient conditions but can also occur in particle and aqueous phases. At
24 least three species of mercury should be considered: elemental (Hg(0)) mercury, which is largely
25 present as a gas; divalent (Hg(II)) inorganic mercury compounds, which are more water soluble;
26 and particulate-phase mercury (Shroeder and Munthe, 1998; U.S. EPA, 1997b). The behavior of
27 mercury in the atmosphere depends strongly on the oxidation state. Elemental mercury is
28 . capable of being transported long distances, even globally; divalent mercury deposits within a
29 few hundred kilometers of sources; and paniculate mercury is deposited at intermediate
30 distances, depending on the particle size (Shroeder and Munthe, 1998). Elemental mercury that
31 is deposited can be reemitted from the surface, as can divalent and paniculate mercury after
32 biological or chemical reduction to the elemental form.
33 In the gas phase, elemental mercury can be oxidized to divalent mercury by O3, OH,
34 . H202 and molecular chlorine, although other halogen atoms might also be important (Shroeder
35 and Munthe, 1998). In the aqueous phase, elemental mercury can be oxidized by OH, O3, and
i 1/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by, the EPA and should not be construed to represent any
Agency determination or policy.
.4-33
-------
1 dissolved chlorine, and divalent mercury can be reduced by processes such as reaction with HO2
2 and S(FV). Both vapor-phase and aqueous atmospheric chemistry may involve multiple phases.
3 EPA has made a substantial effort to evaluate the atmospheric fate of mercury as a result
4 of the requirements of the Clean Air Act. U.S. EPA (1997b) contains a comprehensive
5 evaluation of mercury's atmospheric fate, but this is an area of ongoing research and
6 controversy. EPA continues to be active in investigating mercury behavior in the atmosphere
7 (e.g., Landis and Stevens, 2003; Jaffe et al, 2003; Bullock and Brehme, 2002; Bullock, 2000a, b;
8 U.S. EPA, 2001a).
9
10 4.1.7.1. Application/Limitations
11 Most sampling and analytical techniques published by EPA for metals in air are oriented
12 toward evaluation of particular-phase total metals rather than metal species (U.S. EPA, 1999a).
13 These methods involve collection of a particular size fraction of particles (e.g., PM25, PM,0,
14 TSP), with subsequent analysis by x-ray fluorescence, atomic absorption, inductively coupled
15 plasma, proton-induced x-ray emission, or neutron activation analysis gamma spectroscopy
16 techniques. The one notable exception is a method for mercury (Method IO-5) that speciates
17 vapor and particulate forms. To the extent that metals are sorbed to particulate phases, analysis
18 of individual metal species can, at least theoretically, be accomplished by the same techniques
19 used to analyze those species in other solid media.
20
21 4.1.8. Metals Speciation Techniques
22 A wide variety of analytical and chemical techniques have been used to characterize
23 metal speciation in various media (Isaure et al., 2002; Manceau et al., 2000,1996; Welter et al.,
24 1999; Szulczewski et al., 1997; Charlatchka et al., 1997; Lumsdon and Evans, 1995; Ma and
25 Uren, 1995; Hunt et al., 1992; Gupta and Chen, 1975). These techniques provide information on
26 speciation, particle size, and the source of the metal and also quantitatively determine the metal
27 level present. Of the techniques tested (physicochemical, extractive, and theoretical), the tools
28 that have been used most often to evaluate speciation include the following.
29
30 4.1.8.1. Particle-Bound Metal
31 For particle-bound metal, tools include x-ray absorption spectroscopy (XAS), x-ray
32 diffraction (XRD), particle-induced x-ray emission (PIXE and uPIXE), electron probe
33 microanalysis-scanning electron microscope (EPMA-SEM), secondary ion mass spectrometry
11/24/2004 Peer Review Draft
DISCLAIMER: This infonnation is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy. <.
4-34
-------
1 (SIMS), x-ray photoelectron spectroscopy (XPS), sequential extractions, and single-chemical
2 extractions.
3 Over the past decade, numerous advances in materials science have led to the
4 development of a wide range of analytical tools for determining metal concentrations, bonds, and
5 valences of individual particles on a scale that can be considered useful for the speciation of
6 environmentally important materials (soils, wastes, sediments, and dust). Although most of
7 these tools are scientifically sound and offer important information on the mechanistic
8 understanding of metal occurrence and behavior, only a few provide currently useful information
9 on metal bioavailability for use at a "site" level (see Table 4-9 at the end of this section).
10 However, other techniques still may be essential for conducting a detailed characterization of a
11 selected material to describe the chemical or kinetic factors controlling the release, transport,
12 and/or exposure of a metal.
13 An indirect approach to speciation, in contrast to the direct methods previously described,
14 includes functional or operational extraction techniques that have been used extensively (Tessier
15 and Campbell, 1988; Tessier et al., 1979; Gupta and Chen, 1975). These methods use either a
16 single or a sequential extraction procedure to release species associated with a particular metal
17 within a medium.
18
19 4.1.8.2. Single-Chemical Extractions
20 These methods generally are used to determine the bioavailable amount of metal in a
21 functional class (e.g., water soluble, exchangeable, organically bonded, Fe-Mn bound, or
22 insoluble). In a similar approach, sequential extractions treat a sample with a succession of
23 reagents that are intended to specifically dissolve different and less available phases. Many of
24 these techniques are a variation on the classical method of Tessier et al. (1979), in which metals
25 associated with exchangeable, carbonate-bound, Fe-Mn bound, organically bound, and residual
26 species are determined. A number of excellent reviews on the use and abuse of extraction
27 techniques are available (Beckett, 1989; Kheboian and Bauer, 1987; Forstner, 1987). These
28 techniques can be useful in a study of metal uptake by plants and soil invertebrates, where
29 transfer takes place predominantly from a water solution phase. However, these methods are not
30 "selective" for metal species, and these leachable fractions have never actually been correlated to
31 bioavailability.
32
33 4.1.8.3. Plants
34 When considering the bioavailability of a metal to plants from soil and sediments, it is
35 generally assumed that both the kinetic rate of supply and the speciation of the metal to either the
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
"4-35
-------
1 root or the shoot are the most important factors. In soils and sediments, there is generally a small
2 volume of water in contact with the chemical form of the metal. Although the proportion of
3 soluble metal in pore water is small when compared with the bulk soil/sediment metal
4 concentrations, it is this phase in pore water that is directly available to plants at the root tips.
5 Therefore, understanding pore water chemistry is critical; that is, measuring metal concentrations
6 as simple inorganic species, organic complexes, or colloid complexes is most important. Tools
7 currently used for metal speciation in plants include the following:
8
9 • In situ measurements using ion selective electrodes (Wehrli et al., 1994; Gundersen et
10 al., 1992; Archer et al., 1989).
11
12 • In situ collection techniques using diffusive equilibrium thin films and diffusive
13 gradient thin films followed by laboratory analyses (Zhang et al., 1995; Davison et
14 al., 1994,1991; Davidson and Zhang, 1994).
15
16 • Equilibrium models (WHAM/FREEQC/MINTEQA2). .
17
18 4.1.9. Organo-Metals/Metalloids Transformation Processes
19 Metals/metalloids can exist in the environment in several valence forms and as
20 organometallic compounds. Organometallic compounds (referred to in this section collectively
21 as "organometallics") are compounds that have a metal/metalloid-carbon bond. The bonds in
22 organometallic compounds are generally covalent and between soft metals and soft ligands.
23 Metal/metalloid transformation processes, such as metal methylation, occur through interactions
24 with other chemicals and biota in the environment. Cycling and distribution of organometallic
25 compounds between terrestrial, water, and atmospheric phases may be physically, chemically, or
26 biologically mediated. Examples of some commonly occurring environmentally stable
27 organometallic compounds are shown in Table 4-6.
28
29
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-36
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
W 20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Table 4-6. Some stable organometallic compounds
Metal/metalloid
Arsenic
Lead
Mercury
Selenium
Tin
Stable organometallic compound
Methylarsenic acid, dimethyl arsenic acid, trimethyl arsine, trimethylarsine oxide
Tetramethyl/ethyl lead, trimethyl/ethyl lead, dimethyl/ethyl lead
Methyl mercury, dimethyl mercury
Dimethyl selenide, dimethyl diselenide, seleno-amino acids
Triburyltin, bis(tributlytin) oxide
Environmental methyl-metal concentrations reflect the net methylation rather than simple
rates of methyl-metal synthesis. Metal methylation and demethylation rates in ecosystems are
influenced by the speciation and biochemical availability of the metal. Metals involved in
abiotic or biotic methylation/demethylation processes are presented in Table 4-7. With the
exception of arsenic and selenium, the metals listed in the table form stable complexes with
either methyl or ethyl groups. In addition to methyl/ethyl compounds, stable organometallic
compounds such as lipids and arsenic and amino acids and selenium are incorporated
biochemically.
Table 4-7. Metals/metalloids involved in methylation processes
Process
Environmentally stable organometallics
Abiotic chemical methylation
Abiotic demethylation
Biotic methylation
Biotic demethylation
Metals affected
Si, Ge, Sn, Pb, Hg, As, Sb, Se
Hg,Pb,Sn
Sn,Pb
As, Cd, Hg, Pb, Se, Sn, (others? Sb, Pt, )
As, Hg, Sn, Pb
Source: Bodek et al. (1988).
Organometallic environmental transformations may affect both the mobility and the
toxicity of these metals. The rates of transformation and the organometallic products are
dependent on environmental conditions and the population of microorganisms available. For
example, methylation/demethylation rates are dependent on the speciation of the metal, the
microbial community., the environmental variables (e.g., pH, temperature, reduction oxidation
potential, organic matter, dissolved oxygen, nutrient availability, salinity, complexing agents)
and the distribution of the metal between compartments (sediment, water, gaseous). The inter-
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-37
-------
1 relatedness of these processes has made research into unraveling the factors controlling net
2 methylation difficult and, to date, incomplete, However, some general trends can be predicted
3 with some certainty, anti these are discussed in this section.
4
5 4.1.9.1. Abiotic Transformations
6 Organometallic compounds that are composed of metals with electronegativities >1 .7 are
7 the most stable under environmental conditions. Carbon-metal bonds with more polar (metal
8 electronegatives <1.7) bonds will undergo hydrolysis (reaction with water). Abiotic chemical
9 methylation can occur by three mechanisms: transmethylation reactions between mercury and
10 tin/lead alkyls, humic/fulvic substances, and photochemical reactions.
11
12 4.1.9.2. Biotic Methylation Transformations
13 Biotic methylation occurs when organisms, primarily microorganisms, transfer alkyl
14 groups to bioavailable metals. In general, it is thought that anaerobic sulfate-reducing bacteria
15 are the principal methylators in freshwater and estuarine environments. However, methylation
16 rates are not always correlated with sulfate-reducing bacteria. Not all sulfate-reducing bacteria
17 are capable of methylating, and efficiency of methylation is dependent on the activity and
18 structure of the bacterial community. Other bacteria may be involved in methylation. Biotic
19 methylation occurs" predominantly in the sediment column; however, because the water column
20 by volume is much larger, water column methylation is important.
21 Maximum methylation rates typically occur at the redox boundary, which varies
22 seasonally and frequently coincides with the sediment-water interface (Ullrich et al, 2001).
23 Methylation rates decrease with increasing sediment depth, probably due to a decrease in biotic
24 habitat Microorganisms may also demethylate (or dealkylate) organometallic compounds.
25 Microbial-mediated transformations are frequently the most important environmental
26 organometallic processes. Generally, as the amount of organic material increases in a system the
27 microbial populations also increases. Examples of typical bacterial populations in natural waters
28 and sediments are shown below (Ullrich et al., 2001).
29 High temperatures and anaerobic conditions generally favor metal-methylation
30 formation, and demethylation processes are generally favored under low temperatures and/or
31 aerobic conditions. Studies on the effects of pH are not consistent. Interconnected parameters
32 include pH effects on the microbial communities and effects on the speciation distribution of the
33 metals/metalloids in the water and the sediment as well as adsorption rates. Organometallic
34 compounds appear to increase in the water column in low pH environments, but this may be due
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-38
-------
1 to release of methylated metals from the soil and subsequent depletion of organometallic
2 compounds in the soil. Therefore, pH effects on net methylation in a system are not fully
3 understood. In freshwater ecosystems, where sulfate concentrations are typically low, increase
4 in sulfate concentration increases methylation rates. However, in reducing environments,
5 increasing sulfide concentration decreases methylation rates.
6 The inhibitory effect of sulfide is probably not due to metal sulfide formation but, rather,
7 to the formation of less bioavailably charged metal-sulfur complexes. High organic matter may
8 increase abiotic methylation through humic/fulvic metal reactions; however, this mechanism is
9 poorly understood and confounded because biotic methylation rates may increase in
10 environments with high organic matter. In ecosystems with high DOC concentrations, DOC
11 may bind with metals/metalloids, rendering them unbioavailable and thereby reducing biotic
12 methylation rates.
13
14 4.1.9.3. Organometallic Transformation
15 Organometallic methylation and demethylation rates are influenced by both speciation
16 and bioavailability of the metal, the microbial community, and a large number of environmental
17 factors, many of which are interrelated. Sulfide and organic matter are important environmental
18 variables that significantly affect methylation; however, their effect on
19 methylation/demethylation is as yet poorly understood. Which variables dominate differs among
20 locations and between seasons, although it is clear that methylation is predominantly a
21 biologically mediated process. Methylation/demethylation rates are strongly influenced by the
22 metal/metalloid speciation and bioavailability. General trends in methlyation/demethylation
23 rates are outlined in Table 4-8.
24 ' '
11/24/2004 Peer Review Draft
DISCLAIMER: This! information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-39
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
29
30
31
32
33
34
Table 4-8. General trends of environmental factors affecting rates of
methylation/demethylation
Organometallic
transformations
Net methylation
Methlyation aq
Methylation sed
Demethylation
Temperature
High
t
t
t
4-
Low
4
4
7
t
pH
High
?
1?
t
i
Low
?
t?
4
t
scv
High
7
4
4
9
Organic
matter
7
4t
7
7
Redox
Oxic
I
I
?
t
Anoxic
t
t
7
4
Salinity
High
4
4
7
7
T indicates an increase in rate.
4 indicates a decrease in rate.
? indicates conflicting data or insufficient data to indicate a likely trend.
4.1.9.4. Atmospheric Transformations
4.1.9.4.1. Abiotic chemical methylation/demethylation transformations. Boiling points for
organometallics for a given metal/metalloid decrease with increasing alkyl substitution and with
shorter alkyl chains. For example,, the boiling points of organotin compounds decrease with
dimethyltin dichloride > trimethyltin > tetramethyltin. Fully methylated metals such as dimethyl
mercury may be transported great distances in the atmosphere owing to the combined low
boiling point and low water solubility. Methylation of Se, Hg, Pb, and As volatilizes these
compounds, contributing to their air concentrations.
Demethylation of organometallic compounds in the atmosphere occurs by primarily by
photolysis, such as:
(CH3)2Hg(g)
light
Hg° + 2CH3
Demethylation may also occur by reaction with ozone, hydroxyl radicals, nitrate radicals,
and sorption to paniculate matter. Organometallic compounds are also removed from the
atmosphere by wet/dry deposition.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-40
-------
1 4.1.9.5. Aquatic Transformations . • '
2 4.1.9.5.1. Abiotic chemical methylation/demethylation transformations. In stratified aquatic
3 systems, methyl-metal/metalloid formation occurs predominantly at the oxic/anoxic interface.
4 Organosiloxanes and other silicone-related substances have been considered as possible abiotic
5 methylating agents. Overall, abiotic methylation is probably of minor importance except in
6 ecosystems with high organic matter.
1
8 4.1.9.5.2. Biotic Methylation/Demethylation Transformations. Microbial methylation
9 processes play the major role in methylation of Hg, Sn, and As, with methylcobalamin the most
10 likely environmental methyl donor. Metal speciation is a prime factor regulating the methylation
11 potential in a system. Until recently, Hg2+ was considered the main mercury species methylated
12 by bacteria; however, current research indicates that uncharged Hg complexes are more likely
13 the principal species methylated (Ullrich et al, 2001). Arsenate can be reductively methylated
14 (via arsenite) under anoxic conditions to dimethylarsine. Selenium and selenite can be
15 methylated via microorganisms; demethylation of organoselenium via other biotic processes is
16 also known.
17
18 4.1.9.5.3. Environmental Factors affecting methylation. Seasonal variation of methylated
19 mercury is thought to be related to temperature, redox potential, and productivity. Seasonal
20 variation for organic arsenic (dimethylarsenic) has also been reported, with organoarsenic
21 species decreasing in late fall and winter.
22
23 4.1.9.6. Terrestrial Transformations
24 4,1.9.6.1. Abiotic chemical methylation/demethylation transformations. Methylation and
25 demethylation of organic mercury compounds in soils appear to be mediated by the same types
26 of abiotic and microbial processes that occur in aquatic systems. The frequency and magnitude
27 of soil moisture play an important role in availability and transformation processes. Because
28 soils are primarily oxygenated systems, particularly in the root zone, conditions favorable to
29 sulfide formation and bacterial methylation occur infrequently. With the exception of peat bogs
30 and similar anoxic, highly saturated soils, methylation generally occurs only at very low rates in
31 soils.
32
33 4.1.9.6.2. Biotic methylation/demethylation transformations. Plants have the capacity to
34 transform metals and metalloids that are taken up from the soil. The most notable example is
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-41
-------
1 selenium. Soluble inorganic oxyanions of selenium are readily taken up by plants and converted
2 to organoselenium compounds, such as selenomethionine, selenocysteine, dimethyl selenide, and
3 dimethyl diselenide. Demethylation/dealkylation of organoselenium (e.g., trimethylselenonium),
4 organomercury, organoarsenic, and organotin can occur in soil.
5
6 4.1.9.6.3. Environmental factors affecting methylation. Formation and degradation of
7 organometallic compounds in soils appears to be mediated by many of the same types of
8 microbial and environmental processes as in aquatic systems. Speciation of metals/metalloids
9 dominates the methylation and/or uptake. Methylation/demethylation rates are affected by soil
10 moisture: low moisture decreases biotic processes in soils. Soils high in iron and aluminum
11 oxide, silts, and clay minerals interfere with methylation of metals/metalloids owing to the
12 reduced bioavailability of metals/metalloids. Plants methylate selenium, predominantly to
13 selemethionine and some selenocysteine. Plant uptake and methylation of selenium or arsenic is
14 specific to plant species. Many soil organisms are capable of converting arsenate/arsenite to
15 volatile methylated arsines. Losses of 15-30% per year due to volatilization of arsenic in soil
16 have been reported (ATSDR, 2003). Organolead complexes, on the other hand, are thought to be
17 relatively stable in soils.
18
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-42
-------
I
a
~
e
o,
•2
w
o
tfl
vt
i
B9
I
i
9
»$oa
XniiqB|iBAV
aoijniosff}j
A"HAT}isu«>s juaraaig
japtUEip aidojosi
Ajpijpads }naui3|3
arniEpnnqe sapadg
uoijisodaiDS sspadg
Suipuoq sapadg
a)Bj$ 30ue|BA sapsdg
azis apijjttd sapadg
A4l|iq[B| sapads
A
o
|2
W
1-H
1
iS
£
T
m
2"
1
1
1
*
£
1
£
g
8
ts
3.
7
1/1
d
g,
a,
8
r-H
S
S
<1
M
0>
>H
cn
#
«
S
1
1
I
8
rr
i
0
t-H
S
t-H
1
D
d
w«
•s
"8
>-
-
2
1
2
£3
w
<«
r~l
3.
8
S
*
^
§
•IH
1
5
en
£
w
#
1
'i2
E
i
««
m
i
r-J
•&
o<
1
D
rS
«
^
%
£
w
#
1
2
«
1
«
•*
^
T
a
o
r-^
£
S
*s
^
V3
U
>-
5
£
£
I
u
H
E
ts.
;al and reduction.
"Identifies crystalline compounds and stoichiometric compositions only.
'Valence detennined by charge balance of complete analyses.
0 Technique has limitations based on particle-counting statistics.
dLimited when combined with ICP/MS/LA
Technique requires that each element be tuned and standardized, requiring unreasonable time limi
techniques designed and tested only on simple systems. Multiple species require lengthy analyti<
to
T
rr
-------
1 4.2. HUMAN HEALTH EXPOSURE PATHWAY ANALYSIS
2 4.2.1. Introduction
3 This section and Section 4.3 discuss issues of importance when performing human health
4 risk assessments for metals and metal compounds. The information provided will complement
5 other general Agency guidance on the risk assessment process (e.g., Carcinogen Risk
6 Assessment (U.S. EPA, 2003b), Exposure Assessment (U.S. EPA, 1992b), Developmental
7 Toxicity (U.S. EPA, 1991b) and focuses on the unique and specific characteristics of metals and
8 metal compounds that might be applied in metals risk assessments for human health. This
9 section provides some of the scientific basis that underlies metal-specific characteristics of
10 . human health effects assessment, but it is not intended to be comprehensive. Appropriate and
11 sufficient reference material provided in this framework document will guide readers to
12 additional details on any of the topics addressed.
13
14 4.2.2. Human Exposure
15 Assessment of human exposures to any chemical agent includes (1) identifying how
16 people come into contact with metals in the environment; (2) determining the concentrations of
17 specific forms (speciation) of the metal in specific media (e.g., soil, water, air, and biota); (3)
18 identifying the pertinent exposure metric (via consideration of dose-response assessment); (4)
19 estimating the exposure metric (e.g., oral intake, inhalation exposure concentration, blood
20 concentration), which may involve quantifying relationships between exposure concentrations
21 and intakes; and (5) identifying sources of uncertainty and natural variability and, where
22 possible, quantifying these in estimates of exposure. Although these components are common to
23 exposures to human and nonhuman receptors, and to metal as -well as nonmetal toxicants, some
24 specific aspects of human metal exposure assessment are discussed below.
25
26 4.2.2.1. Environmental Background Concentrations
27 For assessments performed to assess impacts associated with particular human activities
28 (e.g., hazardous waste sites, environmental releases), the term "ambient background" generally
29 refers to all other sources of the metals of interest. The contribution of the background to human
30 metal exposure may be significant For example, metals are natural components of the
31 environment and are repeatedly cycled throughout the biosphere (this component of the ambient
32 background is referred to as "natural background"). Metals are also present as background from
33 persistent anthropogenic activities. During the early 1970s, for instance, industrial sources
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-44
-------
1 accounted for more than 90% of the airborne lead, the deposition of which supplied a large
2 fraction of the lead in many ecosystems (Nriagu and Pacyna, 1988). Various strategies are
3 available to estimate background metal concentrations in environmental media (soil, water, and
4 sediment) (U.S. EPA, 1995a, b; 1989b), and a number of documents dealing with the
5 background concentrations of metals in soils have been published by the U.S. EPA (2002b, c, d;
6 2001b,c).
7 Dietary pathways account for the major impact of background sources of metals. A
8 number of sources provide information on dietary exposure background for metals (Capar and
9 Cunningham 2000; Schoof et al., 1999a, b; Thomas et al., 1999; Bolger et al., 1996; Dabeka and
10 McKenzie, 1995; Gunderson 1995; Tsuda et al., 1995; Dabeka et al., 1993). The prominence of
11 diet in exposure assessment is discussed by Thomas et al. (1999). The United Kingdom also has
12 extensive archive;; of metal content of beverages and infant foods
13 (hjtp://archive.food.gov.uk/maff/archive/food). Additional human dietary exposures may occur
14 secondary to commercial processing of foods (e.g., use of preservatives, emulsifiers, taste
15 enhancers, and packaging products that contain metals). Products such as liquid diets for weight
16 loss, infant formula, and supplements for geriatric patients contain metals.
17 Exposure to metals in the diet can be high enough to approach or even exceed
18 occupational or o ther well-known exposures to the metal (e.g., arsenic in drinking water). Thus,
19 biomarkers (e.g., urinary arsenic, blood mercury) should take diet into account when estimating
20 exposure. For example, in the 24 hours following ingestion of a seafood meal, the urinary
21 arsenic concentration can often rise to 1 mg/L (WHO/1PCS, 1981). In contrast, persons living in
22 Taiwan in an area with endemic arsenic contamination of the water supply of 50-300 ng/L had
23 urinary arsenic concentrations of 140 jig/L (WHO/IPCS, 2001). Workers at a copper smelter
24 where there was considerable arsenic exposure were found to have urinary arsenic
25 concentrations in the range of 200-600 ug/L (WHO/IPCS, 2001). In a study of 380 American
26 dentists, Brady et al. (1980) reported a mean concentration of 8.5 jig per litre, 7.4% of the
27 participants having blood mercury levels greater than 15 ug/L (WHO/IPCS, 1991). In contrast,
28 median cord blood mercury concentrations in a cohort with high fish consumption in the Faroe
29 Islands was 24 u.g/L (NRC, 2000). Mercury concentrations in the blood of people with long-
30 term exposure to methyl mercury from fish were reported to be as high as several hundred ng/L
31 (WHO/IPCS, 1990a).
32 Lifestyles expose people to metals in many different contexts and contribute to ambient
33 background and total exposures. These exposures occur in the workplace, in nutritional
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-45
-------
1 supplements, in Pharmaceuticals taken by the oral and injection routes, in medical diagnostic
2 procedures, in cosmetics, in recreational drugs (e.g., smoking), in folk medicines, and in paints
3 and pigments. For example, cadmium from tobacco smoking can significantly affect a person's
4 total exposure.
5
6 4.2.2.1.1. Application. Depending on the scope, purpose, and approach employed for the
7 exposure assessment, background may be considered in different ways. For example, in the case
8 of predicting exposure estimates from information on environmental releases of interest,
9 background estimates may be included as well, and depending on the methodology, the risk
10 estimates may be developed for the aggregate exposure as well as for the different sources of
11 interest, including background. In this approach, differences in bioavailability among the
12 various metal sources should be considered and incorporated into the calculation of total
13 exposure. Alternatively, biomarkers may be used for exposure assessment, which already
14 account for bioavailable fractions. Therefore, while it may not be feasible to partition out
15 contributions specific to the pollutant source or exposure pathway of interest, biomarkers will
16 provide an estimate of aggregate exposure.
17
18 4.2.2.1.2. Limitations. The approach taken for background may depend on the dose-response
19 relationship for the health endpoint of interest. In the case of cancer assessments, where the
20 Agency has derived a probabilistic relationship and the risk management decision framework is
21 particular to the incremental risk of the pollutant source or human activity of interest, the basic
22 background evaluation can simply focus on providing the foundation for that incremental
23 context. Yet for many other health endpoints, the Agency's traditional approach for toxics
24 involves the Reference Dose (RfD) or Reference Concentration (RfC) approach, which may be
25 • conceptually described as an uncertainly bound on a No-Observed-Adverse-Effect Level
26 (NOAEL). The NOAEL by itself provides no information about the slope of the dose-response
27 relationship or, therefore, the incremental change in risk expected for an incremental increase in
28 exposure above the background level. Thus, in those cases, it will be more important for the
29 exposure assessment to consider estimates of total, rather than incremental, exposures.
30
31 4.2.2.2. Air Pathways of Exposure
32 Air pathway inhalation is the major pathway for human intake of metals in which air
33 serves as the primary medium of contact. Indirect pathways in which air serves as an antecedent
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for .the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-46
-------
1 medium also include the following: deposition of metals to surface dusts and subsequent intake
2 from ingestion or inhalation; deposition to surface water and sediment and intake from ingestion;
3 and uptake of deposited metals into or onto aquatic and/or terrestrial biota, entrance into the
. 4 human food chain, and intake from ingestion: Both systemic bioavailability and local actions on
5 nasal mucosa, airways, and lung tissue should be taken into account when considering human
6 health effects incurred via the air pathway.
7
8 4.2.2.2.1. Applications. Most airborne metals, with a few important exceptions (e.g., mercury
9 and arsine) occur in paniculate form, which necessitates certain considerations for inhalation
10 exposure assessment. For example, particle size affecting respirability (i.e., how much of the
11 pollutant enters the respiratory system). Additionally, inhalation dosimetry for particles involves
12 some distinctly different processes than for gases (i.e., deposition, clearance, dissolution, etc.),
13 which are also influenced by particle size (U.S. EPA, 1997c, as revised in 2004b). Particle size
14 is thus an important factor in assessing metals exposure, with the focus generally being on
15 particles less than or.equal to 10 microns (n) in diameter (PM10). Larger particles do not
16 generally penetrate far into the respiratory tract and can be cleared to the ingestion route. They
17 may play a larger role in irritant and other effects on eyes and nasal passages, and if deposited in
18 the uppermost reaches of the respiratory tract may be transferred to the digestion tract. Thus, for
19 exposure assessments involving measurements (e.g., using area or personal samples), the size of
20 particles sampled is an important consideration.
21 For metals for which the Agency has developed RfCs and ITJRs, the exposure estimate
22 should be for the form of metal used in the dose-response assessment that established the
23 reference values (e.g., exposure concentrations, usually with focus on the respirable fraction)
24 (U.S. EPA, 2004c). For more information on the consideration of particle size in the dose-
25 response assessment for RfCs and lURs, refer to U.S. EPA, 1990. For metals for which the
26 Agency has developed alternative dose-response metrics (e.g., blood lead concentration),
27 respirability, deposition, and clearance as well as absorption into the circulatory system may
28 need to be addressed as part of the exposure assessment.
29
30 4.2.2.2.2. Limitations. In developing inhalation exposure estimates, attention should be given to
31 the form of the metal pertinent to the dose-response assessment (e.g., RfC, IUR). Simply
32 measuring the total amount of a metal without regard to speciation may introduce uncertainties
33 into inhalation exposure estimates, as it can with all exposure routes. Metal speciation affects a
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-47
-------
1 range of processes that change how the metal is deposited in the respiratory tract and
2 subsequently distributed throughout the body and, consequently, its potential toxicity (Bailey
3 and Roy, 1994; Oberdoerster, 1992). For example, in assessing the risk of inhaled chromium,
4 the assessor should consider speciation (e.g., Cr+3 vs. Cr+6), as the dose-response assessment
5 includes that specification. The bipavailability of metals via inhalation can be much higher than
6 that of other routes of intake. This may result in relatively high internal doses, even when
7 intakes are similar to those from other routes. An example is the large contribution made by
8 cigarette smoking to the body burden of cadmium (e.g., Friis et al., 1998; Ellis et al.31979).
9 Variations in airway structure and respiratory conditions (e.g., as with age) may alter the
10 deposition pattern of inhaled particles and contribute to variations in bioavailability (James,
11 1994; Xu and Yu, 1986; Phalen et al., 1985).
12
13 4.2.2.3. Dust and Soil Pathways of Exposure
14 Surface dusts and soil are particularly important media of human contact with metals.
15 Both serve as long-lasting repositories for airborne metal particles, and soils also contain metals
16 due to direct contamination from runoff or mixing with solid wastes. Humans are exposed to
17 metals in surface dust and soil primarily through incidental ingestion or inhalation of suspended
18 dust particles. Dermal contact with metals in soil represents another potential route of exposure,
19 but the relatively low lipid solubility of most metals generally limits absorption through the skin
20 (Paustenbach, 2000; Hostynek et al., 1998). Few studies have actually attempted to quantify the
21 extent or kinetics of dermal penetration of metals deposited on the skin, and the applicability of
22 these studies to metal species and complexes that occur in surface dust or soil is highly
23 uncertain. Therefore, this exposure route is of lesser importance for most metal assessments than
24 the ingestion and inhalation exposure routes.
25 Infants and children are particularly vulnerable to exposures to metals through the surface
26 dust pathway because (1) their crawling and play activity put them in close proximity to surface
27 dust, and (2) they often mouth their hands (e.g., finger sucking) and objects in their environment.
28 This causes intakes of surface dust that are generally greater than those normally found in adults
29 (e.g., Barnes, 1990). The amount of soil ingested by children can be expected to vary with
30 numerous factors, including age, activity patterns, and accessibility to soil and dust. Data are
31 limited with regard to distinguishing between the quantity of dust ingested and the quantity of
32 soil ingested. This parameter is important in connecting measured soil metal concentrations with
33 surface dust ingestion that occurs in the indoor and outdoor environments (U.S. EPA, 1994b).
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-48
-------
1 Exposure assessment methods for direct soil ingestion is described in the Risk Assessment
2 Guidance for Superfund (RAGS) (U. S. EPA, 1989b;
3 http://www.epa.gov/superfund/programs/risk/ragsa/index.htm). Additional guidance with
4 respect to children (e.g., amount of soil a child may ingest) may be found in the Child-Specific
5 Exposure Factors Handbook (U. S. EPA, 2002e).
6 Few studi es of soil ingestion in adults have been conducted; however, the estimates
7 support the general assumption that average daily soil ingestion rates of adults who do not
8 participate in activities in which intensive exposure to surface dust and soil occur (e.g.,
9 occupational gardening, construction work) are lower than those of children (Calabrese et al.,
10 1990; Hawley, 1985). Because concentrations of the metal contaminants in soil can be expected
11 to vary with depth, exposure assessments should consider soil metal concentrations at the depth
12 appropriate to the metal(s) of concern and human behaviors and activities. The size of metal-
13 bearing particles also varies with depth. For example, higher concentrations of lead and smaller
14 particles are found near the soil surface (Duggan and Inskip, 1985; Duggan et al., 1985;
15 Fergusson and Ryan, 1984). For review, see Chaney et al. (1988). Because bioavailability of
16 lead and other metals increases with decreasing particle size (U.S. EPA, 2004a; Barltrop and
17 Meek, 1979), both particle size and depth of contamination become very important
18 considerations in metal exposure assessments.
19
20 4.2.2.4. Dietary Pathway
21 Food can be a major contributor to human metal exposures from contaminant point
22 sources as well as containing trace amounts of naturally occur metals. Failure to accurately
23 account for the dietary contribution can result in significant errors in exposure and risk estimates
24 for metals (Choudhury et al., 2001). Human dietary exposures to metals may occur through
25 various pathways. Livestock grazing in metal-contaminated areas can take up metals from soils
26 or surface water. Metals also can migrate into surface water and sediments and can be taken up
27 by aquatic organisms that are consumed by humans. Human food crops also can take up metals
28 from soils and surface water or become contaminated with metals through deposition of airborne
29 particles. Metals can enter food during harvesting or processing of produce and livestock, during
30 food storage from metals in food containers, and during preparation of foods for meals.
31
32 4.2.2.4.1. Application. Estimation of intakes of metals in food requires information or estimates
33 on the levels of trie metal in food and the amount of food consumed. Although large-scale
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-49
-------
1 surveys of the metal contents of foods and food consumption patterns have been conducted (e.g.,
2 Egan et al., 2002; Ryan et al., 2001; U.S. FDA, 2001; O'Rourke et al., 1999; Thomas et al.,
3 1999; U.S. DHHS, 1996), these surveys have several limitations for applications to risk
4 assessment at a contaminated site, Analysis is often conducted with "market basket" samples of
5 packaged processed foods. With few exceptions, such applications have not been empirically
6 evaluated against biomarkers of exposure (Clayton et al., 2002,1999; Choudhury et al., 2001).
7 For some risk assessments, multimedia fate and transport models may be used to predict
8 concentrations in locally grown or raised foods or wildlife consumed by the population of
9 interest.
10
11 4.2.2.4.2. Limitations. Because some metals (e.g., cadmium and lead) could have long
12 residence time in the body, reconstructing historic dietary exposures can be a challenge. Food
13 consumption surveys generally are limited to short-term consumption (e.g., 1 to 3 days) and do
14 not capture intra-individual variability that would affect long-term averages. Furthermore, food
15 consumption patterns can be expected to change over time; thus, patterns discerned at any given
16 time may not accurately represent historical exposures. An additional challenge is the
17 integration of data from separate metal residue surveys and food consumption surveys (e.g.,
18 Tomerlin et al., 1997). This leads to considerable uncertainty in estimates of metal exposure
19 through the dietary route.
20 Furthermore, estimates of dietary intakes of metals in food based on national or regional
21 data cannot be expected to accurately reflect intakes of metals in locally harvested foods owing
22 to differences in natural background levels of metals. This can be a particularly important
23 limitation when the receptors of concern are subsistence fishermen or hunters. These surveys
24 also may not accurately reflect the amount of metals that enter the food pathway during local
25 food preparation or storage (e.g., in the home).
26
27 4.2.2.5. Water and Sediment Pathway
28 Part of the human population obtains drinking water from untreated sources, such as
29 wells. However, people also consume water specifically treated for human consumption.
30 Treatment of ambient water for human consumption removes DOC and suspended organic
31 sediments that can form complexes with metals (AWWA, 1999). Thus, the exposure context of
32 metals in human drinking water will be very different from that of ambient water. That is,
33 inorganic forms of metals in treated drinking water will consist of the more bioavailable, water-
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy,
4-50
-------
1 soluble species. Treatment also removes bacteria that can participate in organification reactions
2 of toxicological significance to humans (e.g., methylation of inorganic mercuric mercury).
3
4 4.2.2.5.1. Applications. Estimation of the intake of metals in drinking water requires
5 information about or estimates of concentrations of metals in the water and the amount of water
6 consumed. Data on the metal content of tapwater can be obtained from EPA's Office of
7 Drinking Water (U.S. EPA, 2000d), EPA's Exposure Factors Handbook contains exposure
8 information on daily drinking water ingestion and incidental ingestion of water during swimming
9 and showering (U.S. EPA, 1997h). People also can be exposed to metals dissolved in ambient
10 surface water or in association with suspended sediments; the latter can serve as a long-term
11 repository for waterbome metal particles. Such exposures occur during swimming or other
12 recreational activities as a result of incidental water ingestion or during occupational activities in
13 which the sediments are disturbed or resuspended in the water column. Metal bioavailability in
14 ambient surface water can be expected to be much more diverse than in treated drinking water or
15 in ground water because of the presence of organic carbon, inorganics, and suspended organic
16 material that can serve as ligands or reactants for metals. Speciation and concentration will also
17 vary with pH of the surface water. Therefore, estimation of intake of metals from surface water
18 will require appropriate adjustment for relative bioavailability.
19
20 4.2.2.5.2. Limitations. Metals can enter treated drinking water at various stages of water
•21 treatment, distribution, or delivery. Generally, water metal concentrations are measured at the
22 distribution point for municipal water delivery systems. Distribution systems within homes
23 (pipes, storage containers, etc.) and, in the case of lead, glassware, can contribute metals to the
24 water (Graziano et al, 1996). The contribution of metals from pipes (either from the distribution
25 system to the home or within the home) is rarely assessed. This can be highly variable, both
26 within the system and temporally; water that remains in pipes overnight frequently has a higher
27 metal load than water used during the day. Furthermore, the concentration of organic material
28 and other ligands in the water may vary across drinking water sources and can affect the
29 bioavailability of the metal. There is also wide variability in the form of metals in ambient
30 surface water. These factors can be incorporated into site-specific assessments, but local data
31 should be collected on a case-by-case basis. Inhalation of inorganic metal contaminants in water
32 can result from aerosolization, or in special cases, from volatilization (e.g., Hg°). The
33 aerosolization pathway can be a maj or source of intake of inorganics under certain conditions.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-51
-------
1 As an example of aerosol exposure, sea spray can be a significant contributor to iodide intakes in
2 populations that live near the seashore (Whitehead, 1984). A pathway more typical of human
3 exposure to contaminated water is showering, in which areosolization can occur at the water tap.
4 Although models have been developed to predict human inhalation exposures to volatile
5 organics from showering (e.g., Moya et al., 1999; McKone, 1987) (see Guo, 2002, for review),
6 comparable models do not exist for aerosolized metals (Wilkes, 1998), and the magnitude of
7 exposure from showering is unknown.
8
9 4.2.3. Routes of Entry
10 The most frequently encountered routes of entry of metals into humans are ingestion and
11 inhalation. The dermal route is of less concern for most metals but can be important for skin
12 toxicity of some metals (e.g., nickel, chromium). Other routes can be important in specific
13 circumstances, for example, explosions of metal-bearing materials can result in intra- or trans-
14 dermal exposures (Robinson et al., 1983). It is also important to remember that metals can
15 produce toxicity at the point of entry. While metal absorption from skin may be minimal, dermal
16 irritation and sensitization can occur without absorption. Similarly, human health effects such as
17 lung disease and lung cancer can occur from inhalation exposures that do not result in substantial
18 systemic uptake. Routes of entry do not necessarily correspond to the expected exposure
19 pathway. For example, uptake of airborne metals associated with larger particles can actually be
20 attributed to ingestion of surface dust rather than absorption from the respiratory tract. The
21 significance of this process relative to other pathways will depend on the exposure scenario
22 generally being less significant than diet or soil ingestion pathways (see sections 4.2.1 and
23 4.2.2.2). Thus, uptake in the lung can be the source of exposure by two routes.
24 Although dermal contact with metals occurs through soil, air, and water pathways, the
25 relatively low Hpid solubility of most metals limits absorption through the skin (Paustenbach,
26 2000; Hosrynek et al., 1998). An exception is Hg°, which is dermally bioavailable (Hursh et al.,
27 1989). In general, empirical information on dermal absorption of metals should be consulted
28 when available (Stauber et al., 1994; Hostynek et al., 1993; Wester et al., 1992; Hursh et al.,
29 1989; Ilyinetal., 1975).
30
31 4.2.4. Integrated Exposure Approaches
32 Approaches to integrating exposure across pathways and physiological routes of uptake
33 include modeling, relative bioavailability estimates, and biomarker assessment.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-52
-------
1 42.4.1. Modeling
1 Few specific exposure models have been developed for metals with the exception of lead.
3 The Integrated Exposure Uptake Biokinetic (IEUBK) model for lead in children (White et at,
4 1998; U.S. EPA, 1996c, 1994b) was specifically developed for translating exposure
5 measurements into risk estimates at sites contaminated with lead. The IEUBK model and
6 background documentation are available on line at
7 http://www.epa.gov/superfund/programs/lead/ieubk.htm. The IEUBK model assumes that 10
8 ng/dL is a no effect level, while current science indicates there may be no safe level of lead
9 exposure. The model is not readily generalized to other metals because blood lead levels are
10 used as the dose metric, and dose-response assessment for most other metals use estimates of
11 oral or inhalation intake rates.
12 A stochastic human exposure model for lead that is linked to a lead pharmacokinetics
13 model (O'Flaherty et al., 1995) has also been developed (Beck et al., 2001). Less complex
14 models linking adult exposures and blood lead concentrations are available as well (Carlisle,
15 2000; Stem, 1996,1994; U.S. EPA, 1996c; Bowers et al., 1994; Carlisle and Wade, 1992).
16 These models have not been reviewed by the Agency authors. An exposure model for arsenic
17 has also been reported (Cohen et al., 1998).
18 Other general exposure models used in risk assessment can be potentially applied to
19 metals, including population-based models. The EPA's Stochastic Human Exposure and Dose
20 Simulation (SHEDS) mode! is a probabilistic, physiologically based model that simulates
21 aggregate human exposures and doses (i.e., via inhalation, dietary, dermal, and nondietary
22 routes) for population cohorts and multimedia, multipathway chemicals of interest (Zartarian et
23 al,, 2000). EPA recently has applied the SHEDS model to estimate arsenic exposure of children
24 from chromated copper arsenate (CCA)-treated wood (Dang et al., 2003). EPA has also
25 developed a Dietary Exposure Potential Model (DEPM) that links national food consumption
26 and chemical residue data to allow estimates of average dietary intakes of metals and other food
27 contaminants (Tomerlin et al., 1997). EPA's Total Risk Assessment Methodology (TRIM) is
28 also being developed for multipathway risk assessment for air pollutants, including metals (See
29 http://www.epagov/ttn/fera/trimjgen.html). A generic exposure model, RESRAD, developed
30 by the U.S. Department of Energy for risk assessment of radionuclides (U.S. DOE, 2001;
31 LePoire et al., 2000), includes an extensive human exposure module applicable to other metal
32 contaminants.
33
11/24/2004 Peer Review Draft
DISCLAIMER: Tltis infonnation is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-53
-------
1 4.2.4.2. Relative Bioavailabililty
2 4.2.4.2.1. Discussion. Evaluating bioavailability is
3 important, because a given dose of a metal in an
4 environmental medium may be absorbed to a
Evaluating metals bioavailability can
significantly reduce uncertainty in human
health risk assessments and more
accurately characterize potential risks to
5 different extent than the same dose administered in exposed populations.
6 the study used to derive a toxicity value (e.g., oral
7 RfD or cancer slope factor (CSF)). In addition, it is usually assumed that the bioavailability of
8 all metal species is the same, regardless of exposure media Studies clearly show that the
9 bioavailability of metals does vary by environmental medium and the species present.
10 Therefore, bioavailability information can significantly reduce uncertainty in risk assessments
11 and more accurately characterize potential risks to exposed populations.
12 The EPA Office of Solid Waste and Emergency Response (OSWER) has used
13 bioavailability information in making quantitative adjustments, hi 1989, OSWER published a
14 risk assessment guidance for use in human health risk assessments at Superfund sites and has
15 updated this guidance periodically (U.S. EPA, 1989b). This guidance recognizes that the
16 toxicity of an ingested chemical depends on the degree to which it is absorbed from the
17 gastrointestinal (GI) tract into the body. Thus, adjustments to bioavailability assumptions were
18 developed to account for differences in absorption efficiencies between the medium of exposure
19 and the medium from which the toxicity value was derived. Further, because RJDs and CSFs are
20 generally expressed in terms of administered-dose rather than absorbed dose, it also discusses the
21 need to adjust for differences in the expression of dose between the exposure and toxicity value
22 (e.g., absorbed vs. administered dose). The Agency guidance recommends that the relative
23 bioavailability adjustment (RBA) of a chemical should be assumed to be equal in food, water,
24 and soil in the absence of data to the contrary.
25 Estimating bioavailability of metals is particularly difficult because it is dependent on
26 many variables, including the physical and chemical form of the metal, the physical and
27 chemical characteristics of the association between the metal and soil particles, particle size of
28 metal species, and the metal source, hi addition, metal species continuously undergo reactions in
29 soil, referred to as "aging or weathering," that affect bioavailability.
30
31 4.2.4.2.2. Current Practice. The Agency currently addresses bioavailability through the use of
32 default values and, in some cases, through the development of site-specific and medium-specific
33 values. To date, the most common treatment of bioavailability for human health assessments is
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-54
-------
1 to assume that the bioavailability of the metal exposure from the site is the same as the
2 bioavailability of ilie source used to derive the toxicity value (RfD or CSF). The RfD and CSF
3 are typically developed from laboratory toxicity tests using highly bioavailable forms and are
4 usually based on administered rather than absorbed doses. This is true for all chemicals, but it is
5 of special importance for ingested metals because metals can exist in a variety of chemical and
6 physical forms, and not all forms of a given metal are equally well absorbed. For example, a
7 metal in contaminated soil may be absorbed to a lesser extent than when ingested in drinking
8 water or food.
9 It is important to recognize that a default RBA value of 1 (100%) is not necessarily
10 conservative (i.e., more protective of human health). The bioavailability of the metal in the
11 exposure medium of concern at the site may actually be greater than in the exposure medium
12 used in the critical toxicity study that formed the basis of the RfD or CSF. If this is the case,
13 assuming RBA of.I for the medium of concern would result in an underestimate of risk at the
14 site. The Agency recognizes that some cases may exist where sufficient data are available to
15 support development of medium-specific default absorption factors for a particular chemical.
16 The purpose of these medium-specific and chemical-specific default values is to increase the
17 accuracy of exposure and risk calculations even when site-specific studies are not available.
18 Lead is an example of a chemical for which the Agency has established medium-specific
19 default absorption factors for both children and adult populations. The IEUBK model for lead in
20 children predicts PbB concentrations for a hypothetical child or population of children (birth to
21 84 months of age) resulting from exposure to environmental sources of lead, including soil, dust,
22 air, drinking water, and diet (U.S. EPA, 1994b; White et al., 1998). An assumption in the model
23 is that the ABA of lead in soil and dust, at low intake rates, is 0.3 (30%) and the ABA of soluble
24 lead in water and food is 0.5 (50%). This corresponds to an ABA of 0.6 (60%) for lead in soil
25 (or dust) compared to lead in water or food. The model also allows for the input of site-specific
26 values.
27 The Agency has developed the Adult Lead Methodology (ALM) for assessing lead risks
28 in adult populations (U.S. EPA, 1996c). An assumption in the ALM is that the ABA of lead in
29 soil is 0.12 (12%). This value is based on assumptions that the ABA of soluble lead in water is
30 0.2 (20%) and that the RBA of lead in soil, compared to soluble lead, is 0.6 (60%).
31 The Agency has also derived RfDs that are specific for an exposure medium, based on
32 consideration of bioavailability or other factors that might suggest unique dose-response
33 relationships in that medium. For example, separate RfDs for cadmium in food and drinking
11/24/2004 Peer Review Draft
DISCLAIMER: Tliis information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-55
-------
1 water have been derived through the rationale that the bioavailability of cadmium in water is
2 greater than.that of cadmium in food by a factor of 2 (i. e., 5% vs. 2.5%, respectively [U. S. EPA,
3 2003fJ). Similarly, the Agency recommends that a modifying factor of 3 be applied to the
4 chronic oral RfD for manganese when the RfD is used to assess risks from drinking water or soil
5 to account, in part, for potential differences in bioavailability of manganese in water and soil
6 compared to that in food (U.S. EPA, 2003g). Therefore, use of default values should not
7 substitute for site-specific assessments of bioavailability, where such assessments are deemed
8 feasible and valuable for improving the characterization of risk at the site.
9 As described in the NRC report (NAS, 2002), a
While a variety of tools are
available to attempt measurement of
bioavailability, the use of whole-
animal approaches are the most
feasible.
10 variety of tools are available to attempt to measure
11 bioavailability. The approaches include biomarkers of
12 exposure (e.g., ALA activity from lead exposure), cell
13 culture studies, isolated GI tract tissue, whole-animal
14 approaches, and clinical studies. Of these options, the <
15 use of whole animals is most feasible (Weis and Lavelle, 1991); clinical studies offer desirable
16 advantages but present many obstacles (Maddaloni et al., 1998). The following discussion
17 focuses on the oral route of exposure. ,
18
19 4.2.4.2.3. Animal Models. Historically, a variety of experimental animal models (in vivo) have
20 been used to evaluate bioavailability, including rats, rabbits, monkeys, guinea pigs, and swine.
21 Within an animal model, evaluation of bioavailability has included measuring the amount of
22 metal in blood, body tissues, or excreta (e.g., feces and urine). The appropriate study design is
23 dependent on the pharmacokinetics of the metal in the animal model and differences between the
24 selected animal species and humans, hi other words, it is important to consider how soluble
25 forms of the metal are absorbed, how the metal is excreted, and whether there are any tissues
26 where the metal might accumulate. The most common methods for measuring bioavailability in
27 vivo are blodd, urine, fecal and tissue measures.
28 The principal advantage of whole-animal oral chemical absorption studies is that they
29 measure bioavailability in its most clinically relevant form, that is, from the Gl tract and into the
30 systemic circulation. This integrates all of the relevant biological components related to
31 systemic absorption, including pre-systemic elimination, if present By using the animals as
32 surrogates for humans, these studies avoid the experimental and ethical problems associated with
33 the use of human subjects. Currently, certain in vivo bioavailability studies conducted with an
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-56
-------
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
W 19
20
21
22
23
24
25
26
'27
28
29
30
31
appropriate species are considered the gold standard for developing bioavailability information
suitable for use in quantitative human health risk assessments, and they are often used to validate
olher bioavailability tools. For example, the young swine model for lead bioavailability has been
used to validate in vitro extraction tests.
The juvenile swine model is
presently being used as the
preferred animal model by EPA
for making site-specific
bioavailability adjustments for
humans exposed to lead in soil.
Site-specific bioavailability
adjustments based on successful
application of the model have
been accomplished at several
sites across the US.
4.2.4.2.3.1. Animal Model Applications. Animal
models have served as the basis for making site-specific
bioavailability adjustments at several Superfund sites.
Scientists from EPA Region 8 sponsored the
development of whole-body in vivo bioavailability
studies in juvenile swine as a model of young children
who were exposed to lead in soil contaminated with
various forms of mine wastes (Lavelle et al., 1991; Weis
et al., 1992,1993a). The results of these efforts were
subjected to outside peer review and found to be valid
and acceptable for use in adjusting the RBA for lead in
human health risk assessments. As a result, the juvenile swine model is presently being used as
the preferred animal model for lead (U.S. EPA, 2004a). Site-specific bioavailability adjustments
based on results from the juvenile swine model have been accomplished at several sites across
the country, including the Murray Smelter in Colorado; Palmerton, PA; Jasper County, MO;
Smuggler Mountain, CO; and the Kennecott site in Salt Lake City, UT.
Interim draft guidance has also been developed by EPA Region 10 for making
bioavailability adjustments with arsenic-contaminated soil (U.S. EPA, 2000a).
Recommendations are based on literature data on arsenic bioavailability and the results of a '
Region 10 animal study in which immature swine were dosed with arsenic-contaminated soil
derived from the Ruston/North Tacoma Superfund site, which was a former smelter site (U.S.
EPA, 1996d). This interim guidance recommends default values of RBA for arsenic in soil
ranging from 60 to 100%, depending on the'source of contamination (e.g., mineral processing,
fossil fuel combustion, pesticides/wood treatment processes). As with lead, the juvenile swine is
the recommended animal model for supporting departures from the default RBA assumptions.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-57
-------
Animal models are generally
complex, time consuming, and more
resource intensive than other tools,
which may limit their feasibility to large
sites where it is difficult to adequately
characterize variability in bioavailability
across the site.
1 4.2.4.2.3.2. Animal Model Limitations. Currently,
2 bioavailability studies conducted with an appropriate
3 animal species are considered the most reliable for
4 making quantitative bioavailability adjustments in
5 human health risk assessments. They are also used as
6 a means to validate other tools, such as
7 physiologically based extraction tests. However,
8 animal models are generally complex, time consuming-, and much more expensive than other
9 tools. As a result, use of animal models is usually limited to large sites where it is difficult to
10 adequately characterize the variability in bioavailability across the site. An investigator must
11 ensure mat the study design and animal model selected are appropriate for the metal being tested.
12 The best measure of bioavailability (blood vs. tissue vs. feces vs. urine) is dependent on the
13 pharmacokinetics of a particular compound. As discussed above, each measure has its own
14 inherent limitations. Further discussion of the limitations of animal models are discussed
15 elsewhere (NAS/NRC, 2002; NFESC, 2000a, b).
16
17 4.2.4.2.4. In vitro Methods. Recently, significant effort has been expended on developing in
18 vitro methods for assessing the RB A of metals, due to their ease of use and potential cost savings
19 when compared against more traditional in vivo methods using laboratory animals. Several
20 researchers have investigated in vitro models that attempt to simulate the conditions in the GI
21 tract (Drexler et al., 2004; Medlin, 1997; Rodriquez et al., 1999; Ruby et al., 1993, 1996,1999)
22 and which are often referred to physiologically based extraction tests. These methods are based
23 on the concept that the rate and/or extent of metals solubilization in GI fluid is likely to be an ,
24 important determinant of metals bioavailability in vivo. These assays provide a measure of
25 bioaccessibility or the amount solubilized in the GI fluid and available for potential absorption
26 (Ruby etal., 1993).
27 Model development has focused on the simulation of complex physiological and
28 biological functions within the GI tract, including considerations of pH, solids:fluid ratios,
29 motility/transit, and solution chemistry. The most common approach has been the two-solution
30 method, which addresses pH changes in the GI tract by providing an exposure to both the low
31 pH of the stomach (1.3 to 3) and the higher pH of the small intestine (5.5 to 7). Solution pH is
32 usually maintained either by titrations with drop-wise addition of acid or base while solutions are
33 continuously monitored or by using buffers.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-58
-------
1
2
3
4
5
6
7
8
9
10
11
12
13.
14
15
16
17
IS
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
The solids:fluid ratio used for in vitro models has ranged from 1:10 to approximately
1:150 (g/mL). None of these ratios reflects the 2:1 ratio observed in adults (3000 g daily food
intake vs. the 1500 mL stomach volume) (Washington et al., 2001). It is recommended that this
ratio be dictated by practical considerations. Therefore, a sample mass that can be accurately
weighed and is representative should be provided, along with a volume that can help maintain
good particle-to-solution contact and minimize any unusual kinetics.
Motility and transit time within the GI tract are difficult to model with standardization.
Both processes vary greatly and can be affected by diet and daily cycles. Historically,
investigators have used either diffusers, stirrers, or rotation devices to mimic these factors. All
methods are adequate; however, the diffuser system is difficult to control and clean. Also,
rotation mechanisms are not favorable to techniques that require constant pH monitoring.
Variations observed in extraction fluid chemistry are by far the greatest source of method
deviations. Although most methods have the gastric solution dominated by HC1, other acids,
proteins, and peptides have been added, with extraction times of about 1 hour. Intestinal
solutions have their pH adjusted by addition of sodium bicarbonate and/or other biological salts
and are extracted for 3 to 5 hours. Most systems were maintained at a temperature of 37 °C, and
some methods have used argon to maintain anaerobic conditions, even though the GI tract is
aerobic in humans.
4.2.4.2;4.1. Applications. Physiologically based extraction tests have been conducted for a
variety of metals, including lead, arsenic, cadmium, nickel, and mercury. Model results are
currently being used as a screening tool only until adequate model validation has occurred. To
date, EPA has not endorsed the use of in vitro techniques for making site-specific bioavailability
adjustments.
4.2.4.2.4.2. Limitations. In vitro methods are clearly significantly cheaper and less time
consuming than standard in vivo studies. However, these extraction techniques are based on the
premise that solubility or bioaccessibility is the primary factor controlling bioavailability, which
is not necessarily the case. In addition, these tests cannot reflect the complex physiological or
pharmacokinetic aspects of human absorption.
One of the key limitations of this approach is adeptly defined in NAS/NRC (2002):
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-59
-------
1 Regulatory acceptance of the tools used to generate bioavailability information in
2 risk assessment is expected to be influenced by several factors, including the
3 relevance of the tools to the site conditions and the extent of tool validation.
4 Validation variously refers to the performance of a tool or approach in terms of
5 reproducibility, reliability, and multi-lab calibration. An appropriate body of
6 experimental work to validate a tool would (1) clarify where and when a tool
7 yields a definitive response; (2) clarify that the tool can be linked to a biological
8 response of a similar magnitude, and that the linkage stands up across a range of
9 conditions in the type of environment that is being managed; (3) test the
10 prediction of bioavailability using different types of experiments and field studies;
11 (4) clarify which types of biological responses are best predicted by the approach;
12 and (5) include critiques of the best applications and the limits of the tool,
13 especially compared to alternatives. A tool that is well accepted and validated
14 should be given greater weight than one that is new or experimental.
15
16 OSWER is currently developing a bioavailability document on metals to advise risk
17 assessors and managers on whether to collect site-specific information on the bioavailability of
18 metals in soil and how to evaluate bioavailability data for use in human health risk assessments.
19 The document will outline a decision framework that explains how to use bioavailability data
20 consistently as part of a human health risk assessment. The decision framework will consist of a
21 two-tiered approach. The first tier presents general guidelines for determining whether
22 bioavailability is worth considering at a particular site. The second tier involves an ordered
23 process for the actual collection and analysis of bioavailability data.
24
25 4.2.4.3. Biomarkers
26 Integration of exposures across media, route, and time of exposure can be reflected in
27 biomarkers of exposure. The World Health Organization (WHOAPCS, 1993) defines a
28 biomarker of exposure as "an exogenous substance or its metabolite or the product of an
29 interaction between a xenobiotic agent and some target molecule or cell.that is measured in a
30 compartment within an organism." Ideally, there should also be a well-established relationship
31 between biomarker of exposure and outcome, in that the biomarker not only provides
32 information about exposure levels but also can be predictive of an effect For example, urinary
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-60
-------
1 cadmium is directly correlated to the concentration of cadmium in the renal cortex, which is one
2 site for toxicant action of this metal.
3
4 4.2.4,3.1. Application. In the case of metals, urinary cadmium and blood lead are examples of
5 exogenous substances used as biomarkers of exposure. The measurement of metals in biological
6 fluids has been used as the primary means of quantifying biomarkers of exposure for metals by
7 occupational health organizations such as the American Congress of Governmental Industrial
8 Hygienists. An interaction between a metal and a target molecule, such as the adduction of
9 chromium(VI) with DNA and protein, is used to a more limited extent Some biomarkers of
10 exposure such as the DNA adducts of chromium(VI) fall into the area of transition in the
11 continuum from exposure to effect.
12 For many of the metals of interest, biomarkers of exposure and effect are used as basic
13 tools for population or molecular epidemiology studies of effects of exposure to humans of
14 various metals. The Centers for Disease Control and Prevention (CDC) conducts an extensive
15 biomonitoring program of human blood and urine that includes lead, mercury, cobalt, uranium,
16 antimony, barium, beryllium, cesium, molybdenum, platinum, thallium, and tungsten (CDC,
17 2003). The data are summarized in age, gender, and ethnicity categories.
18
19 4.2.4,3.2. Limitations. A biomarker of exposure is a measure of cumulative exposure to a metal
20 and also of metal actually existent in the body, as occurs with chronic exposure for metals.
21 However, such an approach may not be appropriate for metals that are not extensively
22 accumulated in tissues, and it does not differentiate between metal present in a tissue in a
23 sequestered or inactive form and metal engaged in toxic or pathological processes.
24 There are environmental (water, air, soil, dust), occupational, medicinal, and dietary
25 sources of metal exposure. For this reason, use of biomarkers increases the need for
26 comprehensive, multi-pathway assessments of exposure. Reference or background levels of
27 ' biomarkers of exposure are essential for any assessment. Failure to consider background dietary
28 sources of metals may result in a misinterpretation of the exposure. For example, arsenobetaine
29 is a nontoxic organic form of arsenic found naturally in shrimp and other seafood. The analysis
30 of total unspeciated urinary arsenic of individuals who consume seafood, without recognition of
31 their diet history, will lead to an overestimation of exposure to potentially toxic (inorganic)
32 arsenic species—some assessments of arsenic exposure have assumed that 10% of total
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-61
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
elemental arsenic in seafood and 100% of arsenic in all other foods is in a toxic, inorganic form
(NAS/NRC, 1999).
The correct frequency and timing of sampling of biological fluids and tissues, as well as
the correct interpretation of the results, depend on knowing the elimination half-life of the metal.
The half-life of lead in plasma, blood, soft tissues, and bone ranges from hours to months to
years (Sakai, 2000). A detection of lead in plasma above background levels would be indicative
of an acute exposure, whereas a detection in bone would be indicative of chronic exposure.
Thus, sampling plasma every other day or week, or analyzing bone, would not be the best way to
determine whether an acute exposure to lead occurred.
The validity of a biomarker is supported by three kinds of relevance: analytical,
toxicokinetic, and biological (Schulte and Talaska, 1995; Grandjean et al., 1994; WHO/IPCS,
1993). Key analytical issues include specificity, sensitivity, standardization of methodologies (to
reduce intra- and interlaboratory variability), speciation, quality assurance, and the availability of
reference samples. Analytical methods for the detection of metals include ICP-MS, hydride
generation atomic absorption, and fluorescence spectrometry. When coupled with
high-performance liquid chromatography, these methods are enhanced because of the ability to
detect speciated parent metal and metabolites. Although these methods can be very reliable for
the analysis of metals in biological fluids, using them for tissue analysis is more difficult.
Digestion and extraction make it difficult to fully speciate the metal and produce interfering
matrix factors. Reference standards for tissues are seldom available. X-ray fluorescence
spectrometry, used to detect lead in bone (Ambrose et al., 2000), and neutron activation analysis,
used for manganese in liver (Arnold et al., 1999), are highly powerful noninvasive in vivo
techniques; however, the sensitivity is extremely limited. A disadvantage of any in vivo method
is that the metal species in the environmental exposure cannot be estimated correctly.
4.2.5. Toxicokinetics and Toxicodynamics
Several specific properties of metals that affect their absorption, distribution, metabolism,
and elimination can be identified. These properties of metals can influence their physiological
handling as compared to organics (Table 4-10) and should be taken into account in
pharmacokinetic analyses related to human health risk assessment. Properties that affect
toxicodynamics should also be taken into account in these analyses.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-62
-------
Table 4-10. Metal (versus organic) compound properties affecting
absorption, distribution, metabolism, and elimination
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Organic compounds
Metabolism is generally extensive and often
species-specific.
Persistence in body fat is common
because of lipid solubility (not capacity-limited).
Predominantly eliminated by excretion in urine
and exhaled air after biotransformation from
lipophilic forms to hydrophilic forms.
Tissue uptake is most commonly a blood
flow-limited process, with linear portioning into
tissues.
Interactions with other structurally similar
compounds may occur, especially during
metabolism.
Metals
Metabolism is usually limited to oxidation state
transitions and alkylation/dealkylation reactions.
Often sequestered, bound to specific plasma or
tissue proteins (intrinsically capacity -limited) or
bone.
Predominantly eliminated in urine and bile.
Metal compounds are hydrophilic.
Metals and their complexes are often ionized,
with tissue uptake (membrane transport) having
greater potential to be diffusion-limited or to use
specialized transport processes.
Interactions among metals and between metals
and organics are numerous and occur commonly
during the processes of absorption, excretion, and
sequestration.
The evaluation of toxicodynamics addresses the sequence of biochemical events at the
cellular and molecular levels that begin when lexicologically active form of the metal interacts
with the target (e.g., from molecule to a protein, enzyme, or other cellular molecule, that leads
to a toxic physiological response. Toxicodynamics involves the biological processes that
underlie the severity of an effect as well as its reversibility, recovery, and adaptive response.
These evaluations are applied for single metals as well as for metal mixtures, recognizing that
one metal can induce multiple effects (from critical to secondary and more, as the exposure
levels increase) and common effects can be exerted across many metals. To assure that effects
are appropriately combined, it is important to reflect the underlying mechanism or mode of
action; the potential influence of toxicokinetic processes should be considered (e.g., changes in
gastrointestinal absorption and liver and kidney retention can change liver and red blood cell
functions).
Two properties of metals, their hydrophilic nature and their characteristic protein
binding, are discussed in more detail below, followed by a discussion of available
physiologically based pharmacokinetic (PBPK) models.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-63
-------
1 4.2.5.1. Hydrophttic Properties of Metals
2 Solubility in aqueous media is one of the major factors influencing absorption of metals
3 and metal compounds. The water solubility of a metal compound depends on its chemical
4 species, on the pH of its medium (H+ ions), and on the presence of other chemical species in the
5 medium (see Section 4.1). Nitrates, acetates, and all chlorides of most metals except silver,
6 mercury, and lead are soluble. Sulfates of most metals are also soluble, except for barium and
7 lead. On the other hand, most hydroxides, carbonates, oxalates, phosphates, and sulfides are
8 poorly soluble. Another factor influencing absorption of poorly soluble compounds is particle
9 size: fine particles are usually more soluble. Metallic lead in body tissues (as may occur
10 following gunshot wounds) is probably absorbed after being oxidized to soluble salt. Metallic
11 mercury is corrosive when embedded in body tissues, but metallic mercury swallowed into the
12 gastrointestinal tract is not soluble (Goyer and Clarkson, 2001).
13 The relative hydrophilic nature of metals versus organic compounds influences
14 absorption at different sites. Absorption of metals in the gastrointestinal tract is hindered by the
15 lipid nature of intestinal cell membranes but is favored by solubility in the hydrophilic contents
16 of the gastrointestinal tract (preabsorption). In the lungs, the absorption of aerosols of
17 paniculate forms of metals and metal compounds and of lipophilic organic compounds may not
18 be as dependent on the lipophilic or hydrophilic nature of the substance, depending more on
19 particle size and on whether the substance is presented as a vapor or a gas (e.g., elemental
20 mercury). Human skin is not very permeable and provides a good barrier against absorption of
21 metals and metal compounds as well as highly lipophilic organic compounds. Elemental
22 mercury and dimethyl mercury (Siegler et al., 1999) are notable exceptions. When dermal
23 absorption does occur, the mechanism may differ between organic substances and metals. Polar
24 substances, like metal compounds, appear to diffuse through the outer surface of protein
25 filaments of the stratum corneum, which is hydrated, whereas lipophilic nonpolar organic
26 molecules diffuse through the lipid matrix between the protein filaments (Rozman and Klaassen,
27 2001).
28 Hydrophilic metal ions do not readily diffuse into richly lipophilic tissues, such as the
29 brain, liver, and neutral fat stores, where they would be difficult excrete. Although
30 biotransformation of metals occurs, the products typically maintain a hydrophilic profile.
31 Entrance of metals or inorganic metal compounds into lipid-rich tissues like the brain depends on
32 hydrophilic pathways. Retention in tissues of metals or metal compounds is generally related to
33 formation of inorganic complexes or metal protein complexes (e.g., lead in bone and cadmium in
34 tissues bound to the low-molecular-weight protein metallothionein).
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-64
-------
1 Although the low lipid solubility of metal ions could limit their accessibility to tissues
2 and cells, recent rapid progress in identifying metal transporters (Foulkes, 2000) suggests that
3 generalizations are not appropriate, and each metal should be assessed in terms of its ability to
4 access transporters and of the presence of transporters in potential target organs. Furthermore,
5 complex lipids can offer high-affinity binding sites for metal ions that promote their distribution
6 to lipid compartments, and some metals, such as thallium, have a demonstrated affinity for
7 adipose compartments.
8 Generally, the primary factor that influences the uptake and distribution of organic
9 substances within the body is the substances' lipophilicity. Passage through cellular membranes
10 and partitioning to organs occurs primarily through passive diffusion through the lipid portions
11 of cellular membranes. Metabolism (or lack thereof) can, of course, influence the distribution
12 and excretion of organic chemicals primarily because of the accompanying changes in chemical
13 structure and, therewith, lipophilicity. The uptake, distribution, metabolism and excretion of an
14 organic chemical can often be predicted by consideration of the substance's chemical structure
15 and lipophilicity (i.e., octanol/water partition coefficient).
16 In contrast to organic substances, the uptake and distribution of metals in inorganic forms
17 is generally influenced by atomic size and charge, and the availability of active cellular
18 processes that naturally exist for the uptake and distribution of nutritional metals (e.g., calcium,
19 sodium, potassium, magnesium) that can also transport other metals into or across cell
20 membranes. The assumption of linear partitioning in tissues commonly applied to organics, and
21 which is largely lipophilicity-based, is therefore not appropriate for the capacity limited
22 processes that generally control the uptake and disposition of metals. Rather, the uptake,
23 distribution, metabolism, sequestration, and mechanisms of action of metals is generally
24 considered in the context of their kinetic behavior.
25 Metals can have residence times in the body as long as months or years, particularly
26 when they are bone-seeking elements or are associated with tissue storage proteins. Metals, such
27 as lead, strontium, and uranium, may be incorporated into bone, stored there, and many years
28 later may be released from the bone tissue into the systemic circulation to sites of toxic action,
29 by mechanisms closely linked to bone metabolism. These mechanisms include incorporation
30 and loss with bone formation and resorption and, depending on the ionic radius and charge of the
31 metal, can include migration within the bone complex as well as out of the bone (O'Flaherty,
32 1998). While long-term deposition of a metal (e.g., lead) in a tissue (e.g., bone) may not
33 necessarily result in toxicity to that tissue, subsequent release of that metal from that tissue
34 enables transportation of the metal to its site of toxic action elsewhere in the body. It is
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-65
-------
1 important for the risk assessor to consider the potential toxicological consequences of storage or
2 deposition of metals in body compartments'as part of the overall human health assessment.
3 Modeling the kinetics of metals with extended tissue residence times presents challenges
4 . in that anatomical and physiologic functions, along with changes in exposure, need to be
5 characterized as a function of age and clinical status. For example, bone turnover is rapid in
6 children and relatively quiescent in adults, but is increased in adults during conditions such as
7 menopause, pregnancy, and lactation, to name a few. Ninety to ninety five percent of the adult
8 human body burden of lead is contained in the bone. However, lead is not permanently fixed in
9 bone tissue, but returns to the blood as bone is resorbed and by return from the bone surface
10 (O'Flaherty, 1998). Gulson et al (1995) observed that between 45 and 75 percent of blood lead
11 in a group of adult women had originated from the bone. Physiologically based kinetic models
12 have been developed or are in the process of development for arsenic, lead, chromium, and
13 mercury (O'Flaherty, 1998).
14
15 4.2.5.2. Metal-Binding Proteins
16 Metals react with many different proteins in the body that may modify kinetics. Many
17 metals bind with albumin for purposes of transport in the circulatory system and across cell
18 membranes and within cells. However, research is identifying a growing number of proteins that
19 play specific roles in transport, cellular uptake, and intracellular storage of metals (Goyer and
20 Clarkson, 2001), including the following:
21
22 • Transferrin. Transferrin is a glycoprotein that binds most of the ferric ion in plasma
23 and has a role in,transporting iron across cell membranes. This protein also
24 transports aluminum and manganese.
25
26 • Ceruloplasmin. Ceruloplasmin is a copper-containing glycoprotein oxidase in
27 plasma that converts ferrous to ferric iron, which then binds to transferrin.
28
29 • Membrane carrier proteins. A number of recently discovered carrier proteins
30 transport metals across cell membranes. Many of these carrier proteins are
31 multispecific (e.g., divalent metal transporter 1 and 2, metal transporter protein 1),
32 Some metals are transported as complexes with endogenous ligands; no transport
33 systems are intended for the ligand itself, accepting substrates that vary considerably
34 but are recognized by the attached metal ion (Dawson and Ballatori, 1995).
35
36 • Metallothioneins. The metallothioneins are a group of low-molecular weight (MW)
37 proteins (MW about 6,000 daltons), rich in sulfhydryl groups that serve as ligands for
11/24/2004 Peer Review Draft
DISCLAIMER: This infoimation is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-66
-------
1 several essential and nonessential metals. In vitro studies have found that the highest
2 affinity is for silver, then in descending order mercury, copper, bismuth, cadmium,
3 lead, and zinc (Kagi and Kogima, 1987). However, studies of in vivo
4 metallothioneins from various sources included zinc, copper, and cadmium.
5 Metallothioneins have multiple binding sites that have different affinities for metals.
6 Also, the types of metals bound to metallothioneins differ depending on the species,
7 the organ, and previous exposures to metals, but most of them contain at least two
8 different types of metals. For example, metallothioneins isolated from adult or fetal
9 human livers contain mainly zinc and copper, and those from human kidneys contain
10 cadmium, copper, and zinc (Cherian and Goyer, 1995), In most cases, the
11 metallothioneins are inducible and perform a number of functions, including serving
12 as a storage protein for zinc and copper in the liver, kidney, brain, and possibly skin
13 and having an important protective role in cadmium toxicity (Goyer and Clarkson,
14 2001). Although metallothioneins have an affinity for lead in vitro, in vivo binding to
15 lead has not been demonstrated. Also, mercury may induce synthesis of
16 metallothionein in vivo, but binding is only temporary regardless of the demonstrated
17 in vitro affinity.
18
19 • Ferritin. Ferritin is primarily a storage protein for iron in reticuloendothelial cells of
20 the liver, spleen, and bone. It plays an important role in turnover of iron. It has also
21 been suggested that ferritin may serve as a general, metal agonist because it binds a
22 number of metals, including cadmium, zinc, beryllium, and aluminum.
23
24 • Lead-binding protein(s). Lead binds with a number of lead-binding proteins, but
25 their identity or function is not as well defined as that of other metal-specific proteins.
26 The most studied lead-binding protein is the denatured lead-protein complex
27 identified as the intracellular inclusion body occurring in cells, particularly in the
28 liver and kidney in persons with high-level lead exposure. It has been suggested that
29 lead-binding proteins may have a protective effect for lead (Goyer and Clarkson,
30 2001).
31
32 4.2.6. Pharmacokinen'c/Pharmacodynamic Modeling of Metals
33 Physiologically based pharmacokinetic/physiologically based pharmacodynamic
34 (PBPK/PBPD) modeling of metals entails the mathematical description and modeling of a
35 substance's behavior in the body (e.g., absorption, distribution, metabolism, excretion, and toxic
36 effects). PBPK and PBPD models are valuable risk assessment tools for interspecies, high-
37 dose/low-dose, route-to-route, and exposure scenario extrapolation (Krishnan and Andersen,
38 1994). PBPK models consist of multiple compartments representing tissues or tissue groups that
39 are linked by blood flow. PBPD models describe the relationship between target tissue dose and
40 health endpoints or target tissue effects. PBPK models that include a fetal compartment are
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-67
-------
1 particularly valuable for human metal risk assessment. In utero exposure can be important for
2 establishing the body burden of certain metals before birth that result from transplacental
3 transfer.
4 PBPK models have historically been developed and used for risk assessment of volatile
5 organic compounds (VOCs) (e.g., methylene chloride) (Andersen et al., 1987) but have also been
6 applied to many metals (White et al., 1998; Clarke, 1995). Metals differ in their kinetic behavior
7 from VOCs in a number of ways, as discussed by O'Flaherry (1998). Guidelines exist for model
8 evaluation, and these should be used as a framework to determine whether a particular model is
9 appropriate for use in risk assessment (Clark et al., 2003; Andersen et al., 1995).
10
11 4.2.6.1. Application
12 Combined use of PBPK and PBPD models provides understanding of the complex
13 relationships between exposure and target organ effects. PBPK models are often capable of
14 predicting aggregate exposures for comparison to exposure models.
15 For metals that have long retention times in tissues, the maternal tissues can serve as a
16 reservoir for exposures during fetal development. This can be a particularly important exposure
17 pathway for metals that accumulate in the inorganic matrix of bone (e.g., lead, strontium,
18 uranium) because mobilization of bone minerals to develop the fetal skeleton can result in a
19 transfer of maternal bone stores of metals to the fetus (e.g., Gulson et al., 1999a, b; Tolstykh et
20 al., 1998). Transplacental exposures cannot be directly measured from environmental
21 measurements but require the use of PBPK models. A few models of transplacental transfer of
22 lead in humans have been developed; models for other metals are not available for use in risk
23 assessment. The lead models reported to date are limited in that they rely on assumptions of a
24 steady state between maternal and fetal blood lead concentrations (U.S. EPA, 1996c, 1994b;
25 Leggett, 1993; O'Flaherty, 1993). This assumption will be violated if the mother is no longer
26 exposed to lead and if the fetal exposure is due to remobilization of lead from the mother's
27 bones. Furthermore, if the mother experiences lead exposures for the first time during
28 pregnancy, the lead will be partitioned among the various body compartments (bone, hair, blood,
29 fetus) in a dynamic manner.
30 - When using PBPK models or other dosimetric adjustments in the risk assessment process
31 for metals, one should explicitly consider the absorption/distribution and kinetic factors
32 discussed above. There are special considerations for cellular uptake, interaction with
33 nutritionally essential and nonessential metal, protein-binding behavior and function,
34 incorporation into bone, metabolism, and excretion as outlined in Table 4-11.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-68
-------
Table 4-11. Kinetic factors to consider when evaluating the use of PBPK
models or other dosimetric adjustments in the risk assessment process for
metals in humans
Cellular uptake
Interaction with nutritionally
essential and nonessential
metal
Protein-binding behavior and
function
Incorporation into bone
Metabolism
Excretion
Carrier-mediated uptake (e.g., phosphate or sulfate transporters)
Facilitated transport in the form of organic complexes
Competition for binding sites on membrane transport proteins
Interactions at enzyme active sites?
Systemic level interactions altering absorption
•• Capacity limited to binding to specific proteins
*• Inducibilily of binding proteins
> (Zn, Cu, Cd, As, Ni, Hg to metallothionein)
* Protein binding as sequestration mechanism
»• Pb-binding protein in inclusion bodies
* Lead sequestered in bone
* Relative contribution to overall elimination compared to excretory
mechanisms
> Relative contribution of urinary and biliary excretion
» Capacity limitation (saturation kinetics)
1 Pharmacokinetic models for use in human metal risk assessment incluide three models
2 for lead. The O'Flaherty Model is a PBPK model for children and adults. It includes the
3 movement of lead from exposure media (i.e., intake via ingestion or inhalation) to the lungs and
4 gastrointestinal tract; subsequent exchanges between blood plasma, liver, kidney, and richly and
5 poorly perfused tissues; and excretion from liver and/or kidney (O'Flaherty, 1995). The IEUBK
6 model was developed by EPA for predicting lead levels in children (U.S. EPA, 1994b). The
7 Leggett model allows simulation of lifetime exposures and can be used to predict blood lead
8 concentrations in both children and adults (Leggett, 1993). EPA has a research program for the
9 development of an all ages lead (biokinetic) model and a cadmium biokinetic model based, at
10 least initially, on the Kjellstrom and Nordberg model (Kjellstrom and Nordberg, 1978).
11
12 4.2.6.2. Limitations
13 Many of the processes controlling the disposition of metals are intrinsically capacity -
14 limited and highly metal-specific. This makes it necessary to understand physiology well
15 enough to model these processes and methods to estimate binding constants. Another
16 overarching theme is that metal-metal interactions of multiple types (e.g., competition,
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-69
-------
r
1 antagonism, and synergism as well as essential-nonessential metal interactions) commonly occur
2 at multiple points during the processes of absorption, distribution, metabolism, and excretion.
3 Another distinctive characteristic of metals is that common sequestration mechanisms, such as
4 incorporation into bone and binding to storage proteins, can result in extended residence times.
5 In addition to kinetic factors, constructive use of PBPK and PBPD models in the risk
6 assessment process also requires some consensus concerning mode(s) of action and the form of
7 the chemical responsible for the effect of greatest toxicological concern to select an appropriate
8 dose metric. The issue of which endpoints are matched with what form or species of the metal
9 will influence the functional form of the model and hence the dose metric selection. A critical
10 consideration will be to match the toxic endpoint with the active form of the metal in cases
11 where sufficient data exist to suggest that there are one or more active forms of the metal or
12 metalloid (e.g., arsenic).
13
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-70
-------
1 4.3. HUMAN HEALTH EFFECTS
2 4.3.1. Introduction
3 Issues of importance when conducting the Effects Assessment (dose-response
4 determination) for human health risk assessments for metals and metal compounds are discussed
5 in this section. The discussion provides some of the scientific basis that underlies metal-specific
6 characteristics of human health effects assessment, but it is not intended to be a comprehensive
7 review. Further detailed discussions on the points raised here can be found in Goyer et al.
8 (2004). Suggested approaches for including metal-specific information in risk assessments are
9 intended to complement other general Agency guidance (e.g., Carcinogen Risk Assessment,
10 Exposure Assessment (U.S. EPA, 2003), Developmental Toxicity (U.S. EPA, 1991).
11
12 4.3.2. Essentiality Versus Toxicity
13 Seven elements are designated as nutritionally essential for humans by the National
14 Academy of Sciences (NAS) (Table 4-12, first column). The categorization as an "essential
15 nutrient" includes the identification of Recommended Dietary Allowances (RDAs) or Adequate
16 Intakes (AIs) by the Food and Nutrition Board (FNB) of the NAS. The FNB has also examined
17 the possible beneficial effects of other elements (Table 4-12, second column), which have not yet
18 been categorized as essential (or not) (NAS/IOM, 2002). The extent to which these elements can
19 currently be considered beneficial to humans and animals varies and is discussed in Goyer et al.
20 (2004) and in NAS/IOM (2001, 2002).
21
Table 4-12. Metals classified by their known essentiality
Nutritionally essential metals
Cobalt
Chromium(III)
Copper
Iron
Manganese (animals but not
humans)
Molybdenum
Selenium
Zinc
Metals with possible
beneficial effects"
Arsenic
Boron
Nickel
Silicon
Vanadium
Metals with no known
beneficial effects
Aluminum
Antimony
Barium
Beryllium
Cadmium
Lead
Mercury
Silver
Strontium
Thallium
Tin
a Possible beneficial effects in some physiological processes and some species as reviewed by the Food and
Nutrition Board (FNB) (NAS/IOM, 2000). Beneficial effects have been posited for other metals but have not
been officially reviewed by the FNB.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-71
-------
Table 4-12 (third column) also lists some metals of interest to the EPA; the FNB has not
yet reviewed their status as essential/beneficial nutrients. This category includes the metals lead,
mercury, and cadmium. Dose-response curves for essential elements are shown in Figure 4-4.
EAR
RDA
ObmvMlUvd of Intake
Figure 4-4. Dose-response curves for essential elements. The response of
organisms to metals that have both toxic and nutritional properties is
conceptualized as having three phases: an area where there is a risk of
inadequacy, a neutral area, and an area where there is a risk of adverse effects.
Source: FNB (2000)
Note: This figure simply illustrates basic concepts and does not reflect an exact scale. (That is,
although it might appear that the RDA is set at a level associated with a risk of inadequacy above
0.02-0.03, that is not the intent.
Dietary Reference Intakes
Recommended Dietary Allowance (RDA)—The average daily dietary nutrient intake level
sufficient to meet the nutrient requirement of nearly all (97 to 98 percent) healthy individuals in a
particular life stage and gender group.
Adequate Intake (AI)—The recommended average daily intake level based on observed or"
experimentally determined approximations or estimates of nutrient intake by a group (or groups)
of apparently healthy people that are assumed to be adequate; used when an RDA cannot be
determined.
Tolerable Upper Intake Level (UL)—The highest average daily nutrient intake level that is likely
to pose no risk of adverse health effects to almost all individuals in the general population. As
intake increases above the UL, the potential risk of adverse effects may increase.
Estimated Average Requirement (EAR)—The average daily nutrient intake level estimated to
meet the requirement of half the healthy individuals in a particular life stage and gender group.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-72
-------
1 Considerable concern for human health has been focused on the lexicological aspects of
2 essential metals. Because of this concern, the FNB of NAS has provided Tolerable Upper Level
3 Intake Values and RDAs for these metals where data permit and has provided guidance for
4 assessing risk from dietary exposures to these elements (NAS/IOM, 2001). The World Health
5 Organization (WHO/IPCS, 2002) also has provided guidance on methods for assessing risks
6 from excessive exposure to nutritionally essential metals.
7
8 4.3.3. RDAs and RfDs/RfCs
9 The RDA and Reference Dose (RfD)/Reference Concentration (RfC) differ significantly.
10 The RDA is indexed to a clinical effect of deficiency; and it is used to guide people to take in
11 enough of that material to maintain their health. Conversely, RfDs/RfCs are used to assess the
12 potential for toxic effects from elevated exposures (note that RfDs/RfCs are based on many
13 endpoints, which have differing levels of relevance to whole organism effects). In any case, the
14 RDA is designed to minimize adverse effects associated with nutritional deficiency, so is
15 applicable only to dietary exposures (notably food and beverages). Note that these values are not
16 necessarily synonymous with optimal health (for example, it has been suggested that zinc intakes
17 above the RDA might have certain beneficial/protective effects, including for the prostate in •
18 men, for limiting menopausal calcium loss in women, and assisting copper metabolism). RDAs
19 have been identified for different age groups and genders, with RDAs listed for 16 different
20 age-sex and 6 age-pregnancy combinations (NAS/IOM 2003). These RDAs are based on
21 distributions from empirical data or general assumptions that cover 97 to 98% of the population
22 in the given age/gender category, and for. women, additional pregnancy/lactational categories.
23 The RDA is given as mass (milligram, or mg) per day.
24 In contrast, the RfC is used to.assess the potential for a noncancer effect. It represents
25 the amount of a given metal or other material that humans can take in every day without
26 appreciable risk of any harmful effect during a lifetime, including for sensitive subpopulations.
27 The RfD is chemical- and route-specific (oral). The RfC represents a similar measure, but for
28 the air concentration to which someone can be exposed (via inhalation). Sensitive subgroups are
29 commonly addressed through the use of uncertainty factors (which also account for variability)
30 incorporated into these values. Dietary exposures to commercial foods (not locally grown) are
31 typically categorized as "background" and are not often included in environmental risk
32 assessments. The The RfD is given as a mass (mg) per kilogram (kg) body weight that can be
33 safely ingested orally every day, and the RfC is given per m3 of air (i.e., amount safely inhaled).
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-73
-------
1 Both RDAs and RfDs/RfCs have been developed for several metals; for example, values
2 exist for chromium, copper, manganese, molybdenum, selenium, and zinc), and ULs are
3 available for boron and vanadium, and RfDs (but not RDAs) are available for certain compounds
4 of those metals. This highlights the basic concept underlying all of toxicology that the dose
5 makes the poison (from Renaissance physician Paracelsus). That is, dose is the key factor in
6 determining whether the effect of a given metal will be beneficial or adverse, considering other
7 exposure conditions.
8
9 4.3.4. Biology Relevant to Toxic and Essential Properties of Metals
10 The bioaccessibility and bioavailability of metals in locally grown food will depend on
11 the species of the metal in the environment and after ingestion; this exposure may be affected by
12 other chemicals such as metal-chelating agents that restrict uptake (e.g., phytates in plant foods)
13 or facilitate uptake (e.g., ascorbic acid). Metals in food can be found in organic forms or
14 complexes not found in soil, sediment, water, or air. For these reasons, the bioavailabilityof a
15 metal within food may differ considerably from its bioaccessibility in an environmental media;
16 to illustrate, different oral RfDs have been developed for cadmium in food and cadmium in
17 water.
18 The term "molecular or ionic mimicry" has been applied to situations in which a metal
19 forms a complex with an endogenous ligand, and the resulting compound mimics the behavior of
20 a normal substrate, disrupting normal function. Such interactions play an important role in the
21 health assessment for specific metals. One well-studied example is the replacement of zinc by
22 lead in heme-synthesis, which inhibits the function of heme-synthesizing enzymes (Goyer and
23 Clarkson, 2001).
24 Single nutrient deficiencies, such as iron deficiency anemia or iodine deficiency goiter,
25 are well characterized. Generally, intake of essential elements is sufficient in affluent
26 populations to avoid such diseases. However, mild multiple nutrient deficiencies may occur that
27 may be indistinguishable from the pathophysiological effects of primary deficiency of a single
28 metal. In conducting environmental exposure analyses, it is important to consider whether any
29 nutritional deficiencies exist in the population being assessed, or if intake of all trace elements
30 needed to maintain health is met, so results do not confuse adverse health effects due to
31 nutritional deficiency with those resulting from toxic responses to excessive levels.
32 . The uptake of essential metals from dietary food and water and their subsequent
33 distribution within the body are regulated by homeostatic mechanisms that allow enhanced
34 absorption or excretion at low dietary intakes and diminished absorption/excretion with high
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-74
-------
f
1 dietary intakes. Furthermore, dietary factors can reduce the uptake of essential metals. For
2 example, high amounts of phytates in the diet complex with endogenous and exogenous zinc,
3 thus preventing reabsorption and increasing zinc elimination. Specialized carrier proteins mat
4 may originally have developed as a transfer mechanism for the essential elements also are used
5 to sequester or eliminate nonessential metals (see above, Section 4.3.3.3).
6 However, homeostatic control mechanisms can be overwhelmed (e.g., by Ac a high dose
7 level) or otherwise circumvented to produce a toxic effect. (Note that actions of metals on skin
8 or pulmonary membranes are not driven by homeostatic controls.) For example, at high doses
9 chromium salts can produce severe effects in the skin or respiratory tract mucosa, and inhalation
10 of zinc oxide fumes can give rise to inflammatory response (per cytokine excretion) that can
11 produce a local tissue effect (unrelated to systemic zinc exposure levels)..
12 Toxic effects of metals also result from interactions that block the availability or activity
13 of essential metals. For example, lead can block the utilization of iron in heme synthesis by
14 inhibiting the enzyme ferrochelatase, and there is evidence that cadmium can block the entry of
15 zinc into the fetus, thereby causing a variety of developmental defects in the newborn.
16 Cadmium, lead, and mercury, in combinations or by themselves, may reduce the availability of
17 zinc, copper, and selenium when these essential elements are present in marginal amounts in the
18 diet. In addition, a competitive interaction between one or more essential metals could lead to
19 toxic effects, e.g., copper toxicity may be enhanced by reduced levels of molybdenum or vice
20 versa
21
22 4.3.4.1. Limitations
23 The RDAs have been developed for essential metals (as well as other materials, including
24 vitamins) and are designed to provide adequate nutritional intake for 97 to 98% of the
25 population. Should there be a narrow window between the required and toxic amounts of an
26 element (e.g., as for selenium), then it is possible that sufficient amounts for those who need the
27 - most may be above toxic levels for the majority of the population. This potentiality should be
28 evaluated on a case-by-case basis using what is known about potential sensitivities to that
29 element. As a note, RDAs have been established for single chemicals, not mixtures, while metal
30 RfDs to date are also for single chemicals. Under this situation, the component-based approach
31 to assess mixtures is commonly applied (see section 4.3.6).
32
33
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-75
-------
1 4.3.5. Toxicity
2 Metals are associated with numerous health effects that are reviewed in detail in reports
3 from EPA (IRIS reports), the Agency for Toxic Substances and Disease Registry (ATSDR)
4 Toxicological Profiles, reports from the World Health Organization's International Programme
5 for Chemical Safety, and toxicology textbooks. At least five transition metals—arsenic,
6 cadmium, chromium(VI), beryllium, and nickel—are accepted as human carcinogens in one
7 form or another or in particular routes of exposure (NTP, 2002), and inorganic lead compounds
8 are considered probable human carcinogens by EPA's IRIS program, while IARC (2004) has
9 concluded there is limited evidence of carcinogenicity to humans (See:
10 http://www.epa.gov/iris/subst/0277.htnrfcarc and
11 http://monographs.iarc.fr/htdocs/announcements/vol87.htm). Other effects of metals are also
12 well documented, including effects on the neurological, cardiovascular, hematological,
13 gastrointestinal, musculoskeletal, immunological, and epidermal systems. For example,
14 following oral exposure, beryllium can cause intestinal lesions and copper can cause intestinal
15 irritation, while nickel can decrease kidney weight and cadmium can cause proteinuria, and both
\
16 trivalent chromium and nickel can decrease liver and spleen weights. Many metals, including
17 those that are toxic (including carcinogenic), follow the metabolic pathways of similar essential
18 metals, the result of similar binding preferences among various metals (Clarkson, 1986).
19 Carcinogenic metals typically do not require bioactivation, at least not in the sense that an
20 organic molecule undergoes enzymatic modification that produces a reactive chemical species
21 (Waalkes, 1995). Enzymatic modification is generally not a mechanism available for detoxify
22 metals. The body may use other detoxification mechanisms, such as long-term storage (e.g.,
23 cadmium) and biliary and/or urinary excretion.
24
25 4.3.5.1. Application
26 Many factors related to pharmacokinetics and susceptibility act as determinants of
27 toxicity following exposure to a metal. Short-term exposures may produce target organ effects
28 very different from those produced by a similar dose over a longer period of time. Short-term,
29 high-level exposure by ingestion may give rise to well-recognized acute toxicity syndromes,
30 usually involving the gastrointestinal tract initially and possibly secondarily involving the renal,
31 cardiovascular, nervous, or hematopoetic systems. Survivors of acute high-dose arsenic
32 ingestion usually experience multiple organ effects, sometimes with long-term sequelae. Long-
33 term, low-dose exposures from ingestion of metals in food and water generally cause an
34 accumulation in target organs over time. Such exposures can involve any organ system but do
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-76
t
-------
t
1 not usually produce overt gastrointestinal symptoms. For example, low-level, long-term
2 exposure to cadmium in food—sometimes combined with inhalation exposure from cigarette
3 smoking—will cause cadmium to accumulate in target organs (e.g., kidney) but will not produce
4 any obvious clinical effects until "excess" capacity is diminished to a point where the normal
5 function is lost (e.g., onset of renal disease and/or osteoporosis later in life).
6 Nickel and nickel compounds and chromium and chromium compounds are well-
7 established contact allergens. Other metals that have been cited as contact allergens include
8 copper (WHO/IPCS, 1998), cobalt salts (AIHA, 2003), organomercurials (AIHA, 2003),
9 beryllium (WHO/IPCS, 1990b), palladium (Kimber and Basketter, 1996), and gold (Kimber and
10 Basketter, 1996). Toxic interactions with the immune system that result in exaggerated
11 responses are known as hypersensitivity or allergic reactions. Allergic contact dermatitis (or
12 delayed hypersensitivity) is one such example and occurs as a result of allergy to a substance
13 (antigen) through cell-mediated immunity. In sensitized persons, such reactions can be provoked
14 by minute amounts of the allergen. There are two main phases in cell-mediated immunity, the
15 sensitization phase (in which the person becomes allergic to the antigen) and the elicitation
16 phase. Sensitization usually takes at least 10 days. When sensitization has been achieved and
17 the individual is then re-exposed, a reaction is obvious after a characteristic delay of 12-48 hours
18 (hence the term "delayed") (AIHA, 2003). Although there is some connection between skin and
19 respiratory sensitization, it does not follow exact rules, and the dermal mode is a much more
20 common reaction to metals.
21
22 4.3.6. Metal Mixtures
23 Metal mixtures present interesting challenges for the risk assessor. As described below,
24 certain metals can mimic other metals, and depending on the dose and composition of other the
25 metal mixture can result in either a toxic or protective effect. Because metals are naturally
26 occurring substances, they more frequently occur as mixtures than as single, toxic-level
27 exposures. Thus, it is important to account for multiple, simultaneous metal exposures to
28 realistically assess health risks.
29
30
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-77
-------
1 4.3.6.1. Molecular or Ionic Mimicry
2 A large body of literature provides examples of molecular or ionic mimicry of metals.
3 This phenomenon is central to aspects of uptake and biokinetics for metals within the body. The
4 term molecular or ionic mimicry has been applied to situations in which a metal forms a complex
5 with an endogenous ligand and the resulting compound mimics the behavior of a normal
6 substrate, disrupting normal function. A number of reviews discuss this phenomenon, giving
7 examples of the mechanism of toxicity for specific metals (Ballatori, 2002; Clarkson, 1993).
8 Most of these examples involve replacement of an essential metal with a nonessential metal, and
9 molecular or ionic mimicry may be viewed as a form of metal-metal interaction. For example,
10 the protective effects of zinc against copper toxicity are most likely due to diminished
11 gastrointestinal uptake of copper. Lead replaces zinc in heme synthesis by inhibiting the
12 function of heme-synthesizing enzymes (Goyer and Clarkson, 2001). The substitution of
13 calcium by lead results in toxicity of several vital enzyme systems in the central nervous system.
14 This toxicity impaired the development and function of enzymes involved in the production and
15 transport of neurotransmitters (NAS/NRC, 1993). The uptake of lead from the gastrointestinal
16 tract likely occurs via both passive diffusion processes and via active transport mechanisms used
17 in the uptake of essential minerals such as calcium. Calcium deficiency increases the uptake of
18 lead into the body, presumably as a result of lead uptake via calcium active transport processes.
19 Calcium supplementation then diminish lead uptake via both competitive binding to uptake.
20 proteins and down-modulation of active transport activity. Lead is actively taken up into the
21 body and sequestered into bone because of ionic mimicry for calcium. Similarly, cadmium
22 uptake may in large part be related to ionic mimicry of zinc. Divalent inorganic mercury forms
23 linear bonds that form a complex that structurally mimics oxidized glutathione. Arsenate -
24 complexes with phosphate in the sodium-dependent transport system in renal cells, and the
25 arsenate replace the phosphate in mitochondria, impairing synthesis of ATP and energy
26 metabolism.
27 As a converse of the enhanced toxicity that can result from nutritional deficiency, the
28 effects of moderate doses of naturally occurring metals that are not required for nutrition can be
29 reduced (or antagonized) by essential metals found in foods. Diet, therefore, can be a major
30 factor in determining whether potential adverse health effects of additional metal exposures are
31 moderated or enhanced. For example, humans can be exposed to mercury by consuming fish
32 that have absorbed mercury from contaminated water, whereas selenium present in the same
33 water can act as a natural antagonist for mercury toxicity, and vice versa (although the protection
34 is inferred through formation of non-reactive mercury-selenide complexes, rather than as result
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-78
t
-------
t
1 of ionic mimicry); cadmium in soil can be ingested via fruits and vegetables grown on that soil,
2 while zinc in nuts also eaten as part of that diet can limit (or antagonize) cadmium toxicity.
3 Relative intakes of zinc, sulfur, or iron play a significant role in modulating copper deficiency or
4 toxicity. Suttle and Mills (1966) showed that dietary levels of copper at 425 mg/kg caused
5 severe toxicosis in pigs. However, all signs of toxicity were prevented by simultaneously
6 supplementing the diet with 150 mg/kg zinc and 150 mg/kg iron.
7
8 43.6.2. Studies of Metal Mixtures
9 Few controlled studies exist on the toxicologic interactions of metals relevant to levels
10 found in the environment. ATSDR has compiled and evaluated interaction studies for various
11 combination of chemicals, including two sets consisting only of metals: (1) arsenic, cadmium,
12 chromium, and lead; and (2) copper, lead, manganese, and zinc. Additional interaction profiles
13 that include two or more metals are: (3) cesium, cobalt, strontium, trichloroethylene (TCE), and
14 poly chlorinated biphenyls (PCBs), and (4) arsenic, strontium-90, TCE, hydrazine, and jet fuels
15 (available online at http://www.atsdr.cdc.gov/iphome.html). Few studies quantified the
16 magnitude of the interaction, whether using the authors' definitions of toxicologic interaction or
17 EPA's definitions based on dose and response addition. (This same limitation also applies to
18 non-metal mixtures.) The summaries below indicate some of the qualitative conclusions
19 available regarding potential toxic interactions.
20 A study of a mixture of cadmium, lead, and zinc study in rats found slightly more marked
21 adverse hematological effects with the ternary mixture exposure than with the cadmium-lead,
22 cadmium-zinc, or lead-zinc mixtures (Thawley et al., 1977); inconsistencies in dietary levels of
23 calcium and vitamin D in this study, however, may have made comparisons problematic. A
24 well-controlled rat study has reported significant synergism between cadmium and lead
25 regarding testicular atrophy (Saxena et al., 1989). That study also demonstrated protective
26 effects of high dietary levels of zinc, which effectively reduced the testicular effects of the
27 cadmium-lead mixture to control levels. No studies have been located that would allow
28 extrapolation of those high exposure results to more common, lower environmental levels.
29 Fowler and Mahaffey (1978) investigated a relatively wide range of endpoints in studies that
30 covered each metal singly and all possible binary and ternary mixtures. Body weight gain was
31 depressed equally by the ternary mixture and the cadmium-lead mixture, and to a lesser extent by
32 the arsenic-lead and cadmium-lead mixtures, whereas food utilization was depressed more by the
33 ternary and arsenic-cadmium mixtures than by the other binary mixtures. In general, the
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-79
-------
1 biological parameters studied in this report indicated changes of smaller magnitude and
2 inconsistency in direction for binary mixtures compared with ternary mixtures.
3 For some endpoints, the data are not sufficiently robust to show even the direction of
4 interaction (i.e., whether the joint action will be dose additive or greater or less than additive).
5 Many animal studies use commercial diets or semi-purified diets that may have higher or lower
6 levels of essential metals than human diets. Much higher doses of the metals appear to be
7 required to elicit effects when commercial diets are used than when semi-purified diets are used.
8 At the other extreme, effects are seen at very low doses when deficient diets are used.
9 Comparisons among studies are therefore problematic, particularly when the diets are not
10 specified. Experimental efforts to identify and quantify interaction mechanisms among metals
11 are still needed.
12
13 43.6.3. Human Health Assessment
14 Approaches for assessing joint toxicity following exposure to mixtures have been
15 developed by EPA (U.S. EPA, 2000b, which updates U.S. EPA, 1986a). The selection of risk
16 assessment approach begins with the assessment of data quality. Available data are classified
17 into three categories: data on the mixture of concern, data on a mixture of a lexicologically
18 similar nature, and data on individual effects of each component chemical in the mixture. The
19 first approach—assessment of the data based on the mixture of concern—is the preferred
20 approach when the data allow and is essential when considering a complex mixture (e.g.,
21 containing hundreds of component chemicals). Such data include epidemiologic studies on the
22 complex mixture or in vitro data on the complex mixture. The typical example of a complex
23 mixture is coke oven emissions; few metal mixtures relevant to environmental exposures would
24 be expected to fall within this category. The second approach is an assessment of data on a
25 mixture of a lexicologically similar nature to that of the mixture of concern, and a similar
26 limitation applies to the usefulness of this approach. The third approach is based on the toxic or
27 carcinogenic properties of the components in the mixture. In this approach, information on
28 toxicologic interactions of components is incorporated into the assessment. This last approach
29 will likely be most useful for assessing metals in the environment, because this situation is most
30 common (i.e., something is known about the components of the mixture but nothing is known
31 about the mixture itself, or a similar mixture).
32 When no information on interactions exists (i.e., to indicate higher-than-additive or
33 lower-than-additive results),*a dose-additive or risk-additive model is applied as a default The
34 decision on which of these to use depends primarily on how toxicologically similar the metals in
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-80
-------
1 the mixture are. For metal mixtures, a dose-additive approach would be suggested if each metal
2 could be thought of as a concentration or dilution of every other metal in the mixture. Three
3 approaches have been used when the components are considered to be dose additive: the Hazard
4 Index, the Relative Potency Factors, and the Toxicity Equivalence Factors (a special case of the
5 Relative Potency Factors). The choice of approach will depend on the quality of the data. If the
6 metals were assumed to behave independently of each other, then response addition would be the
7 approach applied. This approach is often used for cancer assessments at Superfund sites and
8 other environmental contamination situations; note that response addition is not valid for high
9 exposure concentrations. For mixtures containing metals known to interact, then an interaction-
10 based Hazard Index can be estimated (U.S. EPA, 2000b; Hertzberg et al., 1999).
11 Dietary information is an important part of evaluating the potential for mixture toxicity
12 and interactions. As described above, some quantitative data exist for some metal combinations,
13 and where sufficient data exist, they can be used to predict the pattern of interactions for various
14 proportions of the mixture components or to quantitatively modify the risk assessment.
15 Although at this time few such data are available, some studies do exist, and in some cases the
16 results can offer at least qualitative insights. For example, in-vitro studies showed that
17 chromosome mutagenicity resulting from co-exposure to arsenic and antimony was less-than-
18 additive. That is, less cell damage was observed than would have been expected from an
19 additive effect from the two metals (Gerbel, 1998). However, it may be difficult to validate
20 laboratory data in the absence of comparable field (epidemiological) data (McCarty et al., 2004).
21 Exposures to other types of stressors also must be considered. A recent progress report on a
22 mixtures study of metals and polycyclic aromatic hydrocarbons (PAHs) indicated that
23 environmental metals decreased levels of the enzyme needed to catalyze PAH bioactivation (a
24 precondition of carcinogenesis), affecting induction at both transcriptional and post-translational
25 levels (Kaminsky et al., 2003). (Note that the carcinogenicity of the metals themselves has not
26 yet been assessed.)
27 A number of in vivo studies of multiple metals and metal compounds were published
28 before 1980. The draft interaction profiles developed by the Agency for Toxic Substances and
29 Disease Registry and EPA include evaluations of these metals and more recent publications for a
30 selected set of metals.
31 The EPA provisional interaction profile for arsenic, cadmium, lead, and zinc includes
32 information relevant to chronic exposures as well as insights from acute toxicity studies. Some
33 acute studies considered a variety of endpoints - mortality, testicular necrosis, and
34 hepatotoxicity (Hochadel and Waalkes, 1997) - while others focused only on mortality (Yanez et
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
v. . Agency determination or policy.
4-81
-------
1 al., 1991; Diaz-Baniga et al., 1990; Yoshikawa and Ohta, 1982). The studies captured in these
2 draft profiles often include measurements of changes in chemical distributions in target tissues,
3 which can be difficult to interpret in terms of potential health risks. Nevertheless, some specific
4 joint toxicity insights can be gained. An example of how this information can be summarized for
5 at least qualitative consideration in a health risk assessments is given in Table 4-13.
6 Current studies of toxicological interactions are limited, often relying on in vitro assays
7 of bacteria and cell lines. These studies have tended to focus on mechanisms at the cellular and
8 molecular levels, although their implications for health risk assessment on the tissue or whole
9 organism level are not yet clear. For example, Pounds et al. (2004) identified responses in a
10 recent study of interaction toxicily among simple chemical mixtures of cadmium, mercury,
11 methylmercury, and trimethyltin, using cultured murine renal cortical cells that are targets for
12 metal toxicity. Meanwhile, a model for synergistic metal activation leading to oxidative damage
13 of DNA (genetic material, deoxyribonucleic acid) has been developed by Sugden et al. (2004).
14 This model used in vitro assays to assess the oxidative activation of chromate by arsenite.
15 Although toxicokinetics are a common focus of recent work, interesting new studies are also
16 breaking ground in the area of toxicodynamics, including those under way at TNO in the
17 Netherlands (by J. Groten and collaborators). Future studies are expected to produce key
18 insights mat will improve the ability to assess human health implications of exposures to
19 multiple metals.
20 4.3.7. Variations in Susceptibility
21 Many factors can contribute to human variation in susceptibility to metals. In addition to
22 diet/nutritional status discussed above, these include other lifestyle factors such as smoking or
23 alcohol consumption. Smoking can damage the lung, which can affect the ability to withstand
24 the insult caused by simultaneously or subsequently inhaling metals, particularly those that act
25 directly on the lung such as beryllium, cadmium, chromium, and nickel. Similarly, alcohol
26 damage to the liver can be exacerbated by metals that also cause liver pathology. Other factors
27 influencing susceptibility include age, gender, concurrent damage or disease, and genetic
28 polymorphisms. The discussion below focuses on susceptibility issues considered key for
29 metals; such issues should be at least qualitatively considered in identifying and evaluating
30 potentially susceptible subpopulations for metals risk assessments. Examples for each of these
31 additional factors are offered below.
32
33 4.3.7.1. Age
34 Differences in the pharmacokinetic behavior of metals exist at different life stages,
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-82
-------
1 particularly for the nutritionally essential metals (WHO, 1996a). During the immediate postnatal
2 period, absorption of essential metals is poorly regulated (e.g., chromium, iron, zinc) until
3 homeostatic regulatory mechanisms become established with increasing gut maturity. Much of
4 what is known about gastrointestinal absorption during infancy is derived from animal studies;
5 few studies have been conducted on humans. On the other hand, numerous studies have been
6 conducted on the effects of lead and the developing nervous system in humans (IPCS, 1995;
7 NAS/NRC, 1993). It is suspected that the human placenta is resistant to transport of cadmium
8 (Goyer, 1995). It has also been shown that neonate experimental animals have a higher
9 absorption of both lead and cadmium (Kostial et al, 1978). The efficiency of intestinal uptake of
10 some trace metals, particularly zinc, declines in the elderly, but differences between mature
11 adults for other metals of interest to EPA were not identified in a study conducted several years
12 ago (WHO, 1996c). For comparison, the RDA for copper for children up to 8 years old is just
13 under half that identified for teenagers to adults (NAS, 2001).
14 The sensitivity to skin irritants is considered to generally decrease with age, so children
15 could be more sensitive to metal irritants. Meanwhile, the general loss of renal function with age
16 means older adults will be less able to withstand the harmful effects of metals that affect the
17 kidney, (e.g., cadmium; U.S. EPA, 1999). Note that the RDAs specifically tabulated for older
18 adults (FIU 2004) remains similar to those identified by NAS for this population group within
19 the overall tables published in 2001 (FIU 2004). Thus, when assessing metal risks across
20 different population groups, it will be important to review the current status of metal-specific
21 information with age implications for the metal(s) in that study.
22
23 4.3.7.2. Gender
24 Pregnancy and lactation increase demand for some essential metals, particularly copper,
25 zinc, and iron (NAS/IOM, 2003; Picciano, 1996). Because of physiological changes that include
26 higher iron (and calcium) requirements, hormonal changes, and susceptibility to respiratory
27 disease, it has been suggested mat pregnant women could be predisposed to the toxic effects of
28 beryllium, lead, and manganese (U.S. EPA, 1999f). As a note, the RDAs for copper, iron,
29 molybdenum, and selenium are higher for pregnant and lactating women than for other women
30 or men; for selenium, the RDA for other women is the same as that for men (NAS/NRC, 2000);
31 in contrast, the RDA for zinc is the same for pregnant women 19 to 50 years old as it is for men
32 aged 19 to 70 and older. For additional comparison, the RDA for chromium for pregnant women
33 is roughly the same as that for men, which is higher than for non-pregnant women and lower
34 than for lactating women (NAS/NRC, 2001).
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-83
-------
1 References to women as being highly susceptible to metal toxicity usually refer to effects
2 on the fetus during pregnancy (e.g., of lead and mercury), but basic gender differences
3 independent of pregnancy also may account for differences in toxicokinetics between women
4 and men. Women have only about two-thirds the fat-free body mass of men—so their protein
5 and energy requirements are lower—while having a larger percentage of body fat. The
6 male/female ratio for urinary creatinine excretion (an index of body muscle mass) is 1.5. Men
7 are generally larger than women. Skeletal size as well as body calcium are a function of height,
8 and calcium can also be lost as women age. These differences have an impact on body content
9 of minerals (WHO/IPCS, 2002). Women also have significant loss of iron during menstruation,
10 and it has been shown that absorption and toxicity of cadmium are greater in women, related to
11 decrease in iron stores (Berglund et al., 1994).
12
13 43.7.3 Concurrent Damage or Disease
14 People with higher-than-average biological sensitivity to environmental stressors include
15 allergies and those with pre-existing medical conditions (e.g., with compromised immune
16 systems as a result of a disease or treatment for it, such as chemotherapy). For example,
17 chemotherapy can damage the kidney over time, and at certain levels other medicines (such as
18 acetaminophen) can damage the liver. This may increase metal sensitivity for those metals that
19 cause liver dysfunction.
20 Skin abrasions or other irritations also can alter exposures to and subsequent effects of
21 metals (although dermal absorption is not a primary route of metals exposure for intact skin).
22 For example, both nickel and chromium can cause allergic contact dermatitis, so their combined
23 presence could result in joint toxicity with a potential for interaction; broken skin could
24 potentially increase the absorption of other metals (and other toxic agents) across the exchange
25 boundary. Beyond these well-established contact allergens, other metals have also been cited as
26 possible contact allergens, including copper (WHO/IPCS, 1998), cobalt salts (AIHA, 2003),
27 beryllium (WHO/IPCS, 1990b)3 palladium (Kirhber and Basketter, 1996), and gold (Kimber and
28 Basketter, 1996).
29 With regard to effects associated with high-dose metals exposures, elevated arsenic
30 intakes - from drinking water with levels 17 times above the U.S. drinking water standard - have
31 been linked to blackfoot disease, a vascular complication that represents a severe form of
32 arteriosclerosis (EPA 2004). like many such diseases, the incidence of blackfoot increases with
33 age and thus is also linked to the age factor. Some studies suggest that higher doses of copper
34 and manganese might be linked with Parkinsonism (which could be reversible) (Feldman, 1992;
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-84
-------
1 Gorell et al., 1999). Thus, these studies offer insights for specific metals risk assessments where
2 exposure conditions, including levels and durations, were sufficiently similar.
3
4 4.3.7.4. Genetic Polymorphisms
5 Individuals vary considerably in the nature and severity of their response, and potential
6 repair and recovery, from exposure to metals and metal compounds. Such differences may be
7 due to genetic polymorphisms that can alter the transport and metabolism of a metal. The most
8 apparent of these genetic polymorphisms affecting metabolism and toxicity of metals are
9 disorders in homeostatic mechanisms for nutritionally essential metals.
10 Two disorders affect copper metabolism: Wilson disease and Menkes disease. Wilson
11 disease is an autosomal recessive abnormality (prevalence of 1 in 30,000) that is believed to be
12 due to impaired biliary excretion of copper, resulting in accumulation in and damage of various
13 tissues, particularly the liver, brain, kidney, and cornea; hemolytic anemia can also result.
14 Menkes disease is an X-linked recessive disorder of copper metabolism (prevalence of 1 in
15 200,000) that resembles copper deficiency regardless of level of copper intake (WHO/IPCS,
16 2002).
17 Hemochromatosis is a common inherited disorder related to iron homeostasis. This
18 disorder is characterized by excessive iron absorption, elevated plasma iron concentration, and
19 altered distribution of iron stores (altered iron kinetics). One long-term effect is liver cirrhosis,
20 with increased risk of liver cancer (NAS/IOM, 2003).
21 ATSDR (2003) reported that a relationship between human leukocyte antigens (HLA)1
22 and nickel sensitivity was observed in patients who had a contact allergy and positive results in a
23 patch test for nickel. The nickel-sensitive group had a significant elevation in HLA-DRw6
24 antigen compared with normal controls. The relative risk for patients with DRw6 to develop a
25 sensitivity to nickel was approximately 11-fold.
26 A genetic polymorphism for a heme-metabolizing enzyme affecting lead metabolism was
27 identified in 1973 (Granick et al., 1973), but the molecular characteristics and potential clinical
28 implications did not receive attention until about ten years ago (Smith et al., 1995). Fleming et
29 al. (1998) found that the relationship of bone lead to the cumulative blood index for workers
1 The major histocompatibility complex is a group of genes on chromosome 6 that code
for the antigens that determine tissue and blood compatibility. In humans, histocompatibility
antigens are called human leukocyte antigens because they were originally discovered in large
numbers on lymphocytes. There are thousands of combinations of HLA antigens.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-85
-------
1 with occupational exposure was greater in those workers with the ALAD1 allele, suggesting that
2 the ALAD2 genotype decreased transfer of lead from blood to bone. This effect was only
3 demonstrated in workers with higher blood lead levels than the general population.
4 It is suspected that genetic polymorphisms also exist for arsenic metabolism (NAS/NRC,
5 2001), but these have not yet been defined. Other genetic polymorphisms that may affect the
6 metabolism of chemicals continue to be described. The human MT gene locus is very complex;
7 four individual isoforms and more than ten subforms of the MT1 gene exist, including functional
8 and nonfunctional genes. Because of the protective role that MT plays in metal metabolism, an
9 understanding of the biological role of significant variations in this MT gene locus (both within
10 and among individuals) can offer insights into differential sensitivity or susceptibility and will
11 help characterize variability across a population.
12 Although all MT proteins are associated in some way with a protective role, each of the
13 multiple isoforms appears to have a unique function, as reflected by their unique tissue
14 distributions, the specific conditions under which each is expressed, and unique properties of the
15 metal binding clusters (Bogumail et al.s 1998). The multiplicity of the human MT gene loci
16 makes them potentially important for understanding genetic variability. Emerging technologies
17 are enabling new research in the areas of genomics, proteomics, and metabolomics that will
18 eventually improve human health risk assessments for metals and other stressors, including the
19 role that MT isoforms play in variability in response.
20 Other creative approaches have coupled human tissue studies with animal toxicology
21 studies. For example, studying lung slices from humans (from transplant or tumor resection)
22 and from rodents exposed to the same metals offers promise for improving our understanding of
23 potential effects from inhalation exposures. Comparison of study results between rats and mice
24 under similar exposure regimes and linking them to similar human exposures will provide
25 insights into whether human responses are better predicted by rats or by mice.
26
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-86
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25.
26
27
28
29
30
31
32
4.4. ECOLOGICAL EXPOSURE PATHWAY ANALYSIS
Risk is a function of both hazard (i.e., the toxicity of a substance) and level of exposure.
The route of exposure, as well as other details such as timing and duration, is also important
because exposures vary in their effectiveness in delivering a dose across a biological boundary.
(Ryan, 1998; Ott, 1985). No risk exists unless an effective exposure to a receptor occurs.
Exposure routes include'inhalation, ingestion, and
dermal or, for plants, root uptake and leaf exposures.
Pathways describe the specifics of any exposure and
include transport of the contaminant in the environment as
well as exposure route for organisms of concern (e.g.,
dietary ingestion of a soil contaminant that has been taken
up by plants). Phase association and chemical speciation
influence metal movement through pathways and the
Metal Exposure Assessment
Metal exposure assessment includes
consideration of the exposure routes and
pathways specific to metals, the phase
associations and chemical forms of the
metals, and the expression of exposure and
target doses in a manner consistent with
defining hazard thresholds for particular
organisms.
availability of a metal for delivery of the dose to receptors of concern.
Major metal sources to waters and lands include diverse manufacturing., mining,
combustion, and pesticide activities. Major atmospheric sources are oil and coal combustion,
mining and smelting, steel and iron manufacturing, waste incineration, phosphate fertilizers,
cement production, and wood combustion (Haygarth and Jones, 1992). Metals from these
atmospheric sources can find their way into soils, sediments, and water. Other major sources to
aquatic and terrestrial systems include chlor-alkali, acid, pigment, electronics, and copper sulfate
production.
Some exposure assessments do not involve anthropogenic releases of metals to the
environment. Rather, they focus on changes in exposure to ambient metals that result from other
aspects of human activities. For example, acidification of freshwater changes aluminum phase
association and speciation, resulting in an increased dose of naturally occurring metals to aquatic
biota (e.g., Campbell et al., 1992). Intensive irrigation mobilizes selenium that is naturally
present in relatively high concentrations in western soils, and consequent evaporative
concentration in wetlands, impoundments, and other low-lying areas in arid regions of the
United States can lead to toxic exposures (e.g., Wu et al., 1995).
Methods for assessing transport of metals through various media (water, soil, air) are
described in this section. Exposure to aquatic organisms through water and dietary routes
(looking again at models to account for site-specific differences in bioavailability) is also
r'
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-87
-------
1 discussed. The section concludes with discussions about quantifying exposure to metals for
2 terrestrial organisms.
3
4 4.4.1. Aquatic and Terrestrial Transport Pathways for Metals
5 When an exposure assessment is completed for a site, the available data are usually
6 limited in their coverage with respect to the characterization of exposure levels over both time
7 and space. Depending on the situation, it may be advantageous to use a transport and fate model
8 (i.e., a computational model) to fill in the data gaps, such that an improved characterization of
9 exposure levels is available for usel by the risk assessor. Models are also useful in situations
10 where it is desirable to have an estimate of future exposure levels that are expected to result from
11 the implementation of remediation measures. These results can be used to quantitatively
12 evaluate the effectiveness of alternative remediation scenarios that are being considered. The
13 models can also be used to refine the design of the viable alternatives so that an optimal
14 remediation strategy can be developed.
15 Although numerous models are available for use, most are based on the same
16 fundamental principles. That is, metals are ubiquitous in the environment, as they are found in
1? the aquatic, terrestrial, and atmospheric compartments. Within each compartment, they are
18 present in association with water (freely dissolved metal or as organic and inorganic metal
19 complexes), particles (sorbed, precipitated, or incorporated within a mineral phase), and air. The
20 evaluation of metal transport therefore requires evaluation of the distribution of the metal among
21 these phases, within each compartment, as well as the movement (i.e., the transport) of each of
22 these within and among the various compartments. It is important to simulate the movement of
23 water and particles explicitly because this provides a way to evaluate differences in the degree to
24 which various chemicals/metals may be transported in association with particles (i.e., via settling
25 and resuspension) or in association with the dissolved phase (diffusive flux of dissolved metal).
26 The analyst represents the environmental setting of interest as a series of discrete, interconnected
27 volumes. Mass balance equations for air, water, solids, and metal are then formulated for each
28 volume to obtain a system of mass balance equations that may then be solved for the
29 concentrations of interest over both time and space. Note that it is not necessary to include all of
30 the compartments in every model. For example, models for a site that is impacted by a smelter
31 might call for use of a model of an atmospheric compartment (e, g., to simulate transport of a
32 release from a stack) and a terrestrial compartment (to simulate fate of atmospheric inputs to the
33 soil). Alternatively, for an aquatic setting dominated by previously contaminated sediments, it
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-88
-------
1 may be sufficient to consider water and sediment transport alone and to neglect the atmospheric
2 and terrestrial compartments.
3 Although the underlying principles of most models are similar, the features included vary
4 widely from one model to the next. The output from a relatively simple model may be adequate
5 for decision-making purposes in some instances. Some relatively simple models are limited in
6 their applicability to steady-state analyses and spatially uniform conditions, but if this will
7 provide a reasonable and/or conservative representation of conditions at a site, it may provide the
8 analyst with a useful and cost-effective modeling alternative. In other cases, the analysis may
9 require the completion of time-variable simulations to properly represent conditions that vary
10 over time, such as daily or seasonal variations in flow and upstream boundary concentrations,
11 point source loads, and pulse exposures. The analyst should select is an appropriate model
12 because not all models will be applicable to every situation. Although in principle the more
13 sophisticated models provide the risk assessor with the capability to complete a more detailed
14 and mechanistically based analysis than will a simple model, successful application of these
15 models will require greater resources (data, time, and funding) than will the use of a simpler
16 model, and that the analyst also must possess a relatively high level of modeling expertise. It is
17 for this reason that the more sophisticated models are usually reserved for use in higher level,
18 definitive assessments.
19 Many of the models available for use in evaluating the transport and fate of metals were
20 originally developed for application to neutral organic chemicals. As a result, these models
21 frequently include a variety of reactions that are not necessarily germane to an analysis of metal
22 transport and fate (e!g., biodegradation, photo-oxidation, and volatilization). Although these
23 models still may be of use in an exposure assessment for metals (the nonapplicable processes
24 often may be bypassed), a more significant problem is that they-often fail to represent some
25 important metal-specific processes. For example, the evaluation of metal speciation and metal
26 partitioning between dissolved and particulate phases will be represented only in a very simple
27 manner in such models. This limitation may be overcome, at least in part, by performing the
28 requisite metal-specific analyses with a stand-alone chemical equilibrium model, but this
29 approach will place an added burden on the analyst to integrate the results of the two models in a
30 technically defensible manner.
31 An overview of some of the aquatic fate and transport models available for use is
32 presented in the following subsection. Although models that include some metal-specific
33 capabilities will be noted, no single model that is currently available for use includes all of the
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-89
-------
1 metal-specific features that would be desirable for use. More detailed discussions of these and
2 other transport and fate models, as well as a number of chemical equilibrium models, may be
3 found in Paquin et al. (2003) for aquatic systems and Allen (2002) for terrestrial systems.
4 Additionally, work is in progress to develop updated models that will offer improved metal-
5 specific capabilities (e.g., see the discussion of the Unit World model in Section 5). As a result,
6 metal fate and transport models should be viewed as an evolving technology, with new models
7 expected to become available in the not too distant future.
8
9 4.4.1.1. Aquatic Transport Models
10 Modeling of metal transport and fate within aquatic systems involves the representation
11 of hydrodynamic transport to simulate movement of water, particulate transport to simulate the
12 movement of particles, and chemical transfers and kinetics to simulate exchange of metal
13 between dissolved and particulate phases and between the water column and benthic sediment
14 (Figure 4-5). The analyst has the option of using independent hydrodynamic transport, sediment
15 transport and chemical fate models, or an integrated model mat incorporates all of these
16 processes. Although the models described below are mainly in the latter category, their use does
17 not preclude the use of a stand-alone model (e. g., a hydrodynamic or sediment transport model)
18 as an aid in the development of inputs to the integrated model.
19 Modeling the movement of metals through an aquatic system begins with a
20 characterization of the movement of water through the system. The time scale for the
21 hydrodynamic analysis should be represented in a way that will satisfy the needs of the sediment
22 transport and chemical fate analyses that are also being performed. For example, low-flow
23 conditions associated with minimum dilution may be judged to be the most critical conditions in
24 a setting involving a point source discharge, while peak flow conditions may need to be
25 simulated in a setting where resuspension of contaminated sediments is the primary concern. A
26 steady-state model might be appropriate for use in the former case, and a time-variable model
27 would likely be needed in the latter case. Thus, the details of the specific problem setting will
28 necessarily have an influence on both model selection and how the model will be used.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-90
-------
Cwfmiwc:::;:; :.::: ;^D*
Figure 4-5. A generalized model framework for chemical fate and transport in an
aquatic system.
Source: Paquin et al.s 2003.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-91
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
The reader also should be aware of s everal reviews of fate and transport models that have
been completed since the early 1980s. In earlier reviews, Delos et al. (1984) reviewed models
for wastewater load allocations, and Mills et al. (1985) described screening-level analyses for
water quality assessments of conventional pollutants. The Agency subsequently prepared an
updated summary of the features included in the water quality models that were available for use
in exposure assessments (U.S. EPA, 1987). Schnoor et al. (1987) simultaneously published their
review and included detailed descriptions of fate and transport models and their required input
parameters. Later, EPA published a review describing the use of modeling tools for the
development of total Maximum Daily Loads (TMDLs) in watersheds (U.S. EPA, 1997a). More
recently, Paquin et al, (2003) completed a review of exposure, bioaccumulation, and toxicity
models for aquatic systems, with a focus on their applicability to metals (exclusive of
organometallics). Because of the advances that have been made since the mid-1990s with regard
to the development of fate and transport models, including some recent efforts to couple these
models with metal speciation models and more sophisticated stand-alone hydrodynamic and
sediment transport models, the latter reviews by U.S. EPA (1997a) and Paquin et al. (2003) tend
to include the most up-to-date information with regard to the availability of models that are
appropriate for use. These reviews also include example applications of many of the models
discussed.
Integrated Model
The fate and transport of metals in
aquatic systems is most reliably
predicted using integrated models,
rather than stand-alone hydrodynamic or
sediment transport models.
4.4.1.1.1. Applications. As noted above, fate and transport
analyses may be performed by using an integrated
hydrodynamic, sediment, and chemical transport model or
by employing what tend to be relatively sophisticated stand-
alone versions of these three submodels. The advantage of
the former approach is that integration of the hydrodynamic,
sediment, and chemical transport results takes place in a seamless manner with limited need for
an analyst's intervention. This is in contrast to the use of stand-alone models, where the output
of one model should be formatted in a way that ensures it is amenable to use with the other
models that are to be applied. A distinct advantage of the latter approach is that it has the
potential to reduce the time needed to complete a model run, an important consideration for a
multiyear simulation of a large and complex problem setting. For example, it may not be
necessary to repeat the simulation of hydrodynamic and sediment transport if the model input
being modified affects only chemical transport (e.g., partition coefficient). A similar line of
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-92
-------
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
reasoning would apply to the use of an integrated transport model that includes metal speciation
versus one that uses a stand-alone chemical equilibrium model to satisfy this need.
The fate and transport models that may be considered
for use in the analysis of an aquatic setting are listed in Table
4-13 and discussed below. The models range in complexity
Analytical Solution Models
Analytical solution models of
aquatic system fate and transport are
the simplest models with the lowest
computational requirements. They are
solved analytically to provide
concentrations over metal of time or
space.
from simple mass-balance calculations that can be performed
on a hand calculator, to one-dimensional steady-state
models, to multidimensional time-variable models. The
models are listed in groups that correspond to the
mathematical solution technique used to solve the governing mass-balance equations (analytical
solution, steady-state numerical solution, and time-variable numerical solution), a categorization
that is approximately in accordance with their ease of use as well. Paquin et al. (2003) describe
example applications of many of these models to illustrate how they have been previously
applied. Although publications of successful modeling applications serve as useful illustrations
of how models may be used, it is important to bear in mind that "past performance should not be
viewed as a promise of future returns." That is, the successful application of any model will
depend on many factors other than the model framework itself, including the complexity of the
problem setting, the data that available for use, and the experience of the analyst. Another
important point to keep in mind is that a complicated model is not necessarily a "better" model
to use man a simple one. Models should be selected on a case-by-case basis.
The simplest modeling analysis to consider involves simple screening-level calculations.
While clearly applicable to a screening level of analysis, such calculations can frequently
provide a quantitative perspective on the severity of a problem, one that will be useful in
defining the nature of subsequent modeling analyses that may be required.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-93
-------
Table 4-13. List of fate and transport models
Abbreviation
Name
References
Analytical Solution Models:
WQAM
DJOC-m
RIVRISK
SLSA
USES
QWASI
TRANSPEC
Water Quality Assessment Methodology
Donald J. O'Connor (1 988), a triad of papers
RIVerRlSK
Simplified Lake and Stream Analysis
Uniform System for Evaluation of Substances
Quantitative Water Air Sediment Interaction
Transport and Speciation
Mills etal. (1985,1 982a,b)
O'Connor (1988)
Grieb (1995); EPRI (1996)
DiToro etal. (1981)
, HydroQual(1982)
RIVM et al. (1994)!, as described in Johnson and
Luttik(1995)
Mackay (1991); Mackay et al. (1983)
Bhavsar etal. (2004)
Steady-State Numerical Solution Models:
CTAP
PAWTOXIC
SMPTOX3
(Version 2)
MEXAMS
Chemical Transport and Analysis Program
PAWtuxentTOXICs
Simplified Method Program - Variable -
Complexity Stream Toxics Model
Metals EXposure Analysis Modeling
System Includes EXAMS and MINTE
HydroQual (1982, 1981)
Wright (1987)
LTI (1992)
Dilks etal. (1995,1994)
Felmy etal. (1984)
Bums etal. (1982)
Time-Variable Numerical Solution Models:
EXAMSII
RIVEQLII
WASTOX
RCATOX
(AESOP)
WASPS
META4
DELFT3D
MIKE21
HSPF
RECOVERY
EFDC
EXposure Analysis Modeling System - 11
RIVEr Quality II
Water Quality Analysis Simulation of
TOXics
Row-Column AESOP for TOXics
(Advanced Ecological Systems Operating
Program)
Water Quality Analysis Simulation
Program, Version 5
Distributed w/ DYNHYD5, Dynamic
Hydrodynamics 5
Metal Exposure and Transformation
Assessment Model
Delft 3D Model
MIKE21
Hydrologic Simulation Program -
FORTRAN
RECOVERY
1
Environmental Fluid Dynamics Code
Burns and Cline (1985); Bums et al. (1982)
Chapman (1982)
Connolly and Winfield (1984)
HydroQual (2003)
Ambrose etal. (1993)
Ambrose et al. (1993)
Medine(1995)
Martin and Medine (1998)
Delft Hydraulics (1998)
Danish Hydraulic Institute
Bicknell etal. (1993)
Boyer etal. (1994)
Ruiz et al. (2000)
Hamrick (2002)
Source: Paquin et al. (2003).
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-94
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Water Quality Assessment Methodology (WQAM) is an excellent resource for use in
these types of analyses. WQAM describes a variety of simple screening-level procedures that
may be used in water quality assessments. It provides a useful overview of the types of
simplified analyses that may be performed and is a generic framework that does not generally
attempt to address metal-specific modeling needs. An analyst having limited water quality
modeling experience would probably find this to be a useful reference.
A series of three papers by O'Connor (1988) also provide a helpful introduction to the
basic concepts and use of fate and transport models. These publications include analytical
solutions and example applications for a variety of conditions, including spatially varying
suspended solids concentrations. Although these solutions are not currently available in the form
of a computer program, they could be readily adapted for this purpose.
Analytical solution models tend to be inherently simple models, a requirement of the
solution technique used. As a result, they tend to have modest development and computational
requirements (an advantage), but, at the same time, they limit the analyst with regard to the level
of detail that can be used to represent the problem setting with the model. For example, these
models normally allow the user to represent the system as a one-dimensional, uniform cross-
section stream or a completely mixed water body, with constant inputs over time, and participate
and dissolved exchange of chemicals between the water column and sediment Their simple
structure makes these models well suited for use in screening level analyses. Models in this
category include the River Risk (RTVRISK) model, Simplified Lake and Stream analysis (SLSA)
model, and Uniform System for the Evaluation of Substances (USES) model. Each of these
models includes a water column and a single-layer sediment compartment. RIVRISK is a
steady-state model that may be applied to a one- or two-dimensional stream setting. It is one of
the few models in Table 4-13 that includes a bioaccumulation subroutine. However, it is a
proprietary model of the Electric Power Research Institute and is available only to registered
users. The SLSA program (HydroQual, 1982; Di Toro et al., 1981) may be used to represent a
one-dimensional stream or a completely mixed lake. Its utility in time-variable mode was
demonstrated by its success in simulating the long-term (about 5 years) recovery of a quarry that
had been dosed with instantaneous releases of the organic chemicals DDE and lindane (Di Toro
and Paquin, 1984). Finally, the USES model was developed in Europe to model dissolution and
movement of antifoulant paints on boats (Johnson and Luttik, 1995). It provides the analyst with
a simplified way to estimate concentrations of organic chemicals and metals in a completely
mixed water body, with successful model applications having been reported for metals (Johnson
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-95
-------
1 and Luttik, 1995). USES has recently been replaced by a considerably more complex version,
2 the European Union System for the Evaluation of Substances (BUSES) model (Berding et al.,
3 1999). .
4 Another analytical solution model that has been refined to include metal-specific features
5 is the Quantitative Water Air Sediment Interaction (QWASI) model (Mackay, 1991; Mackay et
6 al., 1983). QWASI, developed as a fugacity model for application to lakes and rivers, is used to
7 simulate the exchanges of chemicals with measured vapor pressure between air, water, and
8 sediments (the Unit World Model for organic substances). A numerical solution version was
9 subsequently developed, one that considers multiple species, as occurs with metals (Diamond et
10 al., 1992). More recently, it has been coupled with MINEQL to allow for interconversion of
11 species and consideration of metals in dissolved, colloidal, and particle phases in water. It is
12 known as the TRANSport and SPECiation (TRANSPEC) model (Bhavsar et al., 2004) and is not
13 yet applicable to sediments.
14 The remainder of the models listed in Table 4-13 and described below are numerical
15 solution models. The first four are applicable to steady-state conditions and the remainder to
16 time-variable conditions. The steady-state models are appropriate for use in a screening level or
17 definitive level of analysis. Their main advantage in comparison to analytical solution models is
18 their capability to represent more complex system geometry and transport regimes (at the
19 expense of an increased level of effort required to set up the model). The Chemical Transport
20 and Analysis Program (CTAP) is essentially a numerical solution version of SLSA, with the
21 added capability of being able to represent two- and three-dimensional systems and multiple-
22 particle size classes. Pawtuxent Toxics (PAWTOXIC) is a relatively simple one-dimensional
23 model for estimating particle deposition or resuspension in river and streams, assuming EqP
24 conditions. PAWTOXIC does not include exchange of dissolved metals between water and
25 sediments, and it can be used only to simulate net settling or resuspension.
26 The Simplified Method-Program Variable-Complexity Stream Toxics (SMPTOX) is of
27 particular interest for metals because it represents one of the first models to include the
28 simulation of AVS and SEM (see Section 3.4) for copper, cadmium, nickel, lead, and zinc. It
29 does not consider the potential for oxidation of metal sulfides. SMPTOX, another model that has
30 evolved from the SLSA framework, is applicable for use in simulating a one-dimensional river
31 or stream and it also includes both particle settling and resuspension and a diffusive flux of
32 dissolved metals between water column and sediments. As a Windows-based model, it is
33 relatively simple to use.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-96
-------
1 The last steady-state model to be considered, Metals Exposure Analysis Modeling
2 Systems (MEXAMS), links EXAMS (a widely used fate and transport model developed by
3 EPA) with MINTEQ, a chemical equilibrium model (Felmy et al., 1984) (see Section 4.1,
4 Environmental Chemistry, for further discussion of MINTEQ). MEXAMS has not yet been
5 coupled with EXAMSII (a time-variable model), so it can handle only steady-state conditions.
6 EXAMS can be applied to a one-, two-, or three-dimensional system and includes an interactive
7 bed layer. By transferring total metal concentrations to MINTEQ in an iterative process,
8 EXAMS is able to calculate fate and transport of metal species. It requires the user to specify a
9 bulk exchange coefficient to represent water-sediment exchange.
10 The remaining models listed in Table 4-13 include the capability to perform time-
11 variable simulations. As a general rule, such models also can be used to simulate steady-state
12 conditions, such as determining simple waste-load allocations during critical low-flow
13 conditions. Input requirements will be relatively high for time-variable models, as the user may
14 need to specify time variable inputs (e.g., upstream and tributary flows and concentrations, point
15 source loads). The Exposure Analysis Modeling System II (EXAMSII, Version 2.97) is a
16 relatively well-known model that was originally developed for steady-state applications.
17 EXAMSII also represents bulk exchange of water and solids between the water column and
18 sediment, rather than separate terms for particulate and diffusive fluxes of metal. It has the
19 advantage of being very flexible with regard to execution of both simplified steady-state
20 assumptions and more realistic (albeit more complex) time-variable analyses. It has also been
21 widely used for many years and is well documented.
22 River Quality II (RIVEQLII) (Chapman, 1982) was developed specifically for inorganic
23 substances. It incorporates relatively sophisticated chemical equilibrium calculations through
24 linking with MINEQL (Westall et al., 1976) (see Section 4.1.5.1 for further discussion of
25 MINEQL). RTVEQLII is applicable to one-dimensional rivers and streams. It assumes
26 equilibrium conditions and incorporates particle setting to the bed, sorption reactions, and
27 precipitation and dissolution of chemical to and from the sediment bed.
28 The next four models listed in Table 4-13, WASTOX, RCATOX, WASP5, and META4,
29 evolved from an earlier model, the Water Quality Analysis Simulation Program (WASP)
30 (DiToro et al., 1981). The Water Quality Analysis Simulation of Toxics (WASTOX) (Connolly
31 and Winfield, 1984) incorporates particulate and dissolved transport processes that are very
32 similar to SLS A and CTAP. Row-Column AESOP for Toxics (RCATOX) (HydroQual Inc.,
33 2003) is a more recent implementation of WASTOX, one that has the potential to take advantage
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-97
-------
1 of parallel processing capabilities. Since RCATOX is still under development, it has not yet
2 been released for general use.
3 The Water Quality Analysis Simulation Program, Version 5 (WASPS) (Ambrose et al.,
4 1993) is the version of WASP that is currently. supported by EPA. Essentially, WASPS
5 incorporates many of the EXAMS capabilities, plus refined sediment transport capabilities.
6 WASPS is distributed with two different subroutines: EUTRO, which is used in eutrophication
7 problems, and TOXI5 for simulating fate and transport. The WASPS package also includes an
8 associated food chain model. Notably, WASPS and RCATOX are designed to interface with
9 state-of-the-art hydrodynamic and/or sediment transport models, a feature not shared by any of
10 the other available fate and transport models. This capability would be of use in estuarine and
11 coastal systems having complex hydrodynamic conditions. WASPS interfaces with the PC-
12 compatible hydrodynamic model DYNYD5 (Ambrose et al., 1993). WASPS and RCATOX can
13 both interface with the Estuary, Coastal, Ocean Model (ECOM), a family of hydrodynamic
14 models (Blumberg and Mellor, 1987), and with ECOMSED (HyroQual, 1998), a sediment
15 transport model.
16 The Metal Exposure and Transformation Assessment Model (META4) model is
17 essentially WASPS combined with MINEQL. It is used for developing TMDL waste-load
18 allocations and evaluating remedial actions and TMDLs (Martin and Medine, 1998). It is
19 applicable to a variety of receiving waters, including ponds, streams, rivers, lakes, and estuaries,
20 and can be run as one-, two-, or three-dimensional systems. META4 addresses some of the
21 shortcomings in WASP4 to more accurately describe metal dynamics by, for example, the recent
22 addition of subroutines to represent the interactions of dissolved metals with iron oxyhydroxides
23 in the water column and sediment under variable pH and the ability of the model to predict future
24 concentrations of the major cationic metals under variable regimes of water chemistry. It also
25 handles numerous point and nonpoint loads and sequential deposition or scouring of sediment
26 bed layers.
27 The Delft 3D model (DELFT3D), developed by Delft Hydraulics Lab in The
28 Netherlands, is another model that is quite sophisticated with regard to its capabilities. It is a
29 flexible integrated model that includes the following modules: hydrodynamics, water quality
30 (including sediment transport), chemistry, and wave generation. Although the model software is
31 proprietary, the model itself is commercially available.
32 MIKE21, developed by the Danish Hydraulic Institute (DHI), is another commercially
33 available fate and transport model that is applicable to rivers. This three-dimensional model
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-98
-------
1 includes both particulate and dissolved transport processes. One of the modules of a recent
2 release may be used to simulate the fate, transport, and bioaccumulation of metals.
3 The Hydrologic Simulation Program-FORTRAN (HSPF) (Bicknell et al., 1993) is the
4 most recent version of a family of models originally developed in the 1960s to simulate
5 agricultural runoff. It is unique among the models listed in Table 4-13 because that it includes
6 hydrology and nonpoint-source terrestrial runoff modules, in addition to in-channel fate and
7 transport capabilities. HSPF is frequently used to simulate complex watersheds and provides
8 information for continuous input simulations or for storm events. It models well-mixed
9 reservoirs as well as branched river systems and can include up to three sediment types. It has
10 been widely used throughout the United States, but its high level of sophistication makes it most
11 appropriate for use by experienced modelers.
12 RECOVERY, a model supported by the U.S. Army Corps of Engineers, has been
13 developed with features that make it amenable to evaluating sediment capping alternatives (Ruiz
14 et al., 2000; Boyer et al., 1994). The model is structured as a well-mixed surface-water layer
15 overlying a vertically stratified but horizontally well-mixed sediment bed. The sediment is
16 defined by three zones (surface, deep contaminated, and deep clean), which can be further
17 subdivided on the basis of porosities, contaminant concentrations, and other factors.
18 RECOVERY is particularly useful for evaluating sediment-capping scenarios and sites with old
19 contamination (where clean sediment has layered over the contaminated bed).
20 One other model that is nearing release for use as a fate and transport model is the
21 Environmental Fluid Dynamics Code (EFDC) (Hamrick, 2002). EFDC, originally developed as
22 an advanced level hydrodynamic model, was subsequently modified to include sediment
23 transport capabilities. It is a state-of-the-art model that incorporates a variety of options for
24 representing sediment transport. Most recently, it was further refined for use with chemicals,
25 with an initial application to metals in which two-phase linear partitioning was used (Ji et al.,
26 2002); a second application was the ongoing analysis of poly chlorinated biphenyls (PCBs) on
27 the Housatonic River in Connecticut. The chemical fate subroutine does not currently evaluate
28 metal speciation. EFDC is scheduled for release in the near future.
29 ' Water quality analyses often require probabilistic results, as the WQC in the United
30 States are expressed in probabilistic terms, with an allowable l-in-3-year exceedance frequency.
31 Steady-state models cannot evaluate a return period for exceedences. When using these models,
32 a Monte Carlo analysis can be conducted to generate a large number of model inputs and
33 subsequent solutions that can then be analyzed probabilistically to obtain a characterization of
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-99
-------
1 the probability of exceeding a particlar concentration. Time-variable models that generate long-
2 term simulations (e.g., 20 years) can also be statistically analyzed to evaluate the exceedence
3 frequencies. Several statistical models have been developed specifically for conducting these
4 types of analyses. The Probabilistic Dilution Model (PDM) (Di Toro, 1984) is one such model.
5 Program Monte (HydroQual, 1997) is a Monte Carlo program that generates a time series of
6 daily concentrations. The Dynamic Toxics (DYNTOX) model (LTI, 1994) is an EPA model that
7 uses three statistical methods to predict the frequency of water quality standard violations.
8 RIVRISK, one of the steady-state analytical solution models described above, also contains a
9 built-in Monte Carlo simulator.
10 Transport and fate models will vary considerably in their ease of use, with the simpler
11 one-dimensional and analytical solution models placing relatively modest demands on the user.
12 Although it may be a simple matter to set up and run this type of model, their proper use requires
13 the user to exercise a considerable amount of judgment. More complex models become
14 increasingly difficult to use and interpret, and errors associated with setting up the model are not
15 always easy to detect. Hence, it is important that the relatively sophisticated modeling analyses
16 be performed by an experienced analyst, one who is familiar with the details of the processes
17 included in the model being used. Owing to the complexity of setting up and parameterizing the
18 more complex models listed in Table 4-13, particularly for analysis of a complex problem
19 setting, their use will often be limited to relatively refined definitive assessments.
20 Further detailed discussions about features, limitations, and example applications of the
21 models discussed in this section are provided by Paquin et al. (2003). This review also identifies
22 additional models designed for use in specialized regions of water bodies, such as mixing zones
23 or plumes.. The appendix includes information about where to obtain many of models discussed
24
25 4.4.1.1.2. Limitations. The partition coefficient, which controls the distribution of metal
26 between the dissolved and particulate phases, is considered to be a key model parameter in
27 chemical/metal fate and transport evaluations (see Section 4.1.4). It is important for several
28 reasons. First, the distribution of metal between the dissolved and particulate phases has a direct
29 bearing on the magnitude of particulate fluxes of metal that occur in association with the settling
30 and resuspension of sorbed metal. Second, it also controls the magnitude of diffusive fluxes of
31 metal between the sediment interstitial water and the overlying water column, as this flux is
32 proportional to the concentration gradient of total dissolved metal (free ionic metal + metal-DOC
33 and metal-inorganic ligand complexes) between these compartments. Partitioning reactions also
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-100
-------
1 affect the metal bioavailability and the route of exposure to a metal (food vs. water). Given its
2 importance, it is necessary to recognize that the assumption of equilibrium partitioning is a
3 potential oversimplification in some situations (e.g., near a point source discharge to a receiving
4 water, or immediately following a pulse exposure of such as an overflow from a storm sewer
5 system). Under these conditions, the assumption of equilibrium between the dissolved and
6 paniculate phases may be less appropriate than is otherwise the case and additional uncertainly
7 will necessarily be associated with the model results. If this situation exists, consideration
8 should be given to conducting special studies (e.g., measurement of free metal concentrations
9 over time in a water sample) to test the validity of the equilibrium assumption.
10 For neutral organic chemicals., the magnitude of the partition coefficient is often assumed
11 to be proportional to the dissolved organic carbon content of the particles. Use of linear partition
12 coefficients is likely to be an oversimplification in the case of metals, where a variety of sorption
13 phases may be important Several models are available that provide a relatively detailed
14 representation of these metal-particle interactions (see Paquin et al., 2003, for a review).
15 Many models that were initially developed for organic substances assume first-order
16 decay processes. These are of little importance for metals, but they sometimes are used in
17 screening-level assessments to represent removal of sorbed material from the water column.
18 This approach provides a relatively simplistic representation of the underlying processes and
19 should be used with caution. Given the capacity of most computing environments today, the
20 potential exists to use relatively sophisticated metal-specific sorption and water-sediment flux
21 models, even for simple assessments. However, until such subroutines become more widely
22 available, their use will be problematic, especially for an inexperienced user. Modification of
23 existing models may be an option, but this should be attempted only by an individual who is
24 experienced in model development and programming.
25 Modeling of mercury and metalloids, such as arsenic and selenium, is complicated by
26 transformation processes that change the form of the metal. Methylation of mercury and arsenic
27 or binding of selenium to amino acids (e.g., selenomethioine) changes both their physical and
28 biological properties (see Section 4.1. for a discussion of transformation processes). Similarly,
29 metals that readily change oxidation state (e.g., chromium) also require additional considerations
30 in fate modeling. Many of the same transport models can be used, but input parameters will
31 require modification. Such models are beyond the scope of this framework, which is focused on
32 inorganic forms of metals. See, for example, the Mercury Cycling model by Hudson et al.
33 (1994) for.further guidance on fate and transport models for mercury.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-101 '
-------
1 All modeling exercises are limited by the validity of the model framework, the accuracy,
2 of input parameters, and the experience of the analyst (Dzombak and Ali, 1993). All exposure
3 assessments should include an explicit description of model assumptions and associated
4 uncertainties.
5
6 4.4.1.2. Terrestrial
1 Movement of metals through soils is dependent on the chemical properties controlling
8 speciation, the presence of ligands that control complexation of metals within pore water (and
9 ground water) and adsorption onto mineral surfaces, and the rate of water flux through the soil.
10 Metals are lost from the soil primarily by leaching into ground water, although in particular
11 instances uptake by plants can represent a significant loss. Section 4.1.4 reviews the processes
12 and models that predict movement of chemicals through soils or partitioning onto mineral
13 surfaces (i.e., partition coefficients). These serve the same role as the fate and transport models
14 discussed above for aquatic systems and therefore are not repeated here.
15
16 4.4.2. Routes of Exposure to Aquatic and Terrestrial Species
17 4.4.2.1. Aquatic Species
\ 8 Potential exposure routes for aquatic species include inhalation/respiration, dermal
19 absorption, and diet (either food or incidental sediment ingestion). The extent to which a metal
20 is taken up by any one of these exposure routes is difficult to define for all relevant routes.
21 Inhalation/respiration, which for aquatic species means general exchange across respiratory
22 surfaces, can involve diverse gill and lung types. Respiratory surfaces include fish gills, various
23 molluscan and arthropod gill types, the pseudo-lungs of pulmonate gastropods, cells surrounding
24 the sponge spongocoel, mammalian and avian lungs, and plant leaves. Absorption includes
25 movement across the skin; the walls of such diverse structures as spongocoels, the cnidarian
26 gastrov'ascular cavity, and the echinoderm water-vascular system; the filtration-based feeding
27 structures of many Crustacea, insects, and polychaetes; diverse phytoplankton cell membranes;
28 and plant roots and leaves. Ingestion pathways can involve direct consumption or consumption
29 after exchange through a trophic web. The diversity of potential receptors makes definition of
30 exposure pathways more difficult for aquatic receptors than for humans.
31 The respiration/inhalation route is a particular challenge in aquatic exposure assessments
32 because of the differing types of respiratory organs, the dynamic nature of the respiratory
33 process in water, and the intimate contact between a receptor and metals dissolved in waters. •
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-102
-------
1 Further complicating the issue, some respiratoiy organs can also be involved in locomotion,
2 excretion, ion regulation, and food capture, sorting, and ingestion. The absorption route can
3 involve uptake across a phytoplankton cell membrane, amphibian skin, arthropod exoskeleton,
4 the general integument of an infaunal clam or annelid, or the egg membrane of a bird or
5 amphibian. Absorption can occur from overlying or interstitial waters; these sources have very
6 distinct chemistries that influence metal bioavailability.
7 Similarly, the ingestion route is difficult to define for aquatic receptors because of the
8 diversity of feeding modes and food sources, such as sediments, suspended solids, microflora,
9 animal tissues, and plant tissues. The combining of respiratory, locomotive, or feeding
10 structures complicates description of the ingestion pathway for some species. The presence of
11 life stages that feed differently also confounds exposure assessment.
12 Despite the complexities associated with quantifying exposure of aquatic animals to
13 metals from multiple routes of uptake, the relative importance of the different uptake pathways
14 has received considerable attention in recent years (Wang 2002; Hook and Fisher, 2001; Fisher
f
15 et al., 1996; Bjerregaard et al., 1985). Bioenergetic-based kinetic models used to describe the
16 accumulation of contaminants in aquatic animals have been developed relatively recently and
17 have been successfully applied to a variety of organic and inorganic contaminants. These
18 models provide a broad framework for addressing controls on contaminant bioaccumulation for
19 diverse organisms and can be used for studying contaminant bioavailability and determining the
20 relative importance of different routes of contaminant accumulation, including that of metals
21 (Landrum et al., 1992; Wang et al., 1996). The models are flexible enough to incorporate
22 environmental variability in contaminant sources, contaminant concentrations, food availability,
23 and organism growth rates in their predictions of organism contaminant levels.
24 Applications of one-compartment biokinetic models using laboratory-based
25 measurements of key model parameters (assimilation efficiency, uptake rates from water and
26 food, elimination rates) have been extended to field situations for populations of marine mussels
27 (Wang et al., 1996; Fisher et al., 1996); Ag, Cd, Co, and Se in clams (Griscom et al., 2002;
28 Luoma et al., 1992); Po in copepods (Stewart and Fisher, 2003; Fisher et al., 2000); Se in fish
29 (Baines et al., 2002); and freshwater mussels (Roditi et al., 2000). Site-specific model
30 predictions for metal concentrations in animal tissues are strikingly close to independent field
31 measurements for diverse water bodies, suggesting that it is possible to account for the major
32 processes governing contaminant concentrations in aquatic animals and that the laboratory-
33 derived kinetic parameters are applicable to natural conditions.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-103
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Temporal Aspects of Exposure
Temporal aspects of exposure are
important considerations in assessing
aquatic risks to metals. Organisms may
respond vary different to episodic
exposure than they respond to the
constant exposures incorporated into
classic toxicity bioassays. Sound risk
analyses clearly articulate all.
assumptions about exposure durations.
Temporal aspects of exposure should be considered in
assessments of any toxicant. Rapid speciation and phase
changes associated with changes in pH/Eh make temporal
issues particularly germane to metals. Fluctuating or pulsed
exposures occur in situations such as rapid changes in pH/Eh
associated with photosynthesis and respiration, hypolimnetic
discharge from stratified reservoirs, biocide (e.g., copper
sulfate) spraying, ingestion of prey items with seasonally
high metal concentrations, surface waters receiving wastewater treatment plant effluent, urban
storm water, snowmelt, and acid precipitation runoff. Transient metal concentrations may be
orders of magnitude higher than background concentrations but may last for only a few hours.
These episodic exposure scenarios have been poorly characterized for metals (Hoang et al.,
2005). Any risk assessment for metals should clearly state all assumptions about duration of
exposure.
Water pathway and respiratory route. For acute exposure of most water column
organisms, binding of metals to the gill (respiratory) surface is the primary route of exposure.
Binding of metals to the gill surface is the primary route of
exposure for most water column organisms, at least for short-
duration exposures. Chemical kinetics play an important role in
Primary Route of Exposure
Binding of metals to the gill
surface is the primary route of
exposure for most water column
organisms for short-duration
exposures.
this context, as gradients in pH and ionic composition exist at the
gill surface microlayer due to respiration, excretion, and ion
regulation. Dissolved aluminum toxicity to freshwater fish is a
good illustration of this point (Playle and Wood, 1990). The deposition of aluminum on gills is
determined largely by the rapid shift from dissolved ionic aluminum to an A1(OH)3 precipitate on
the gill surface due to the more alkaline state of gill microlayer water. How much aluminum
speciation and phase association changes occur is a function of the initial pH of the bulk water.
Exposure assessment should address metal speciation kinetics in such situations (see Section 4.1,
Environmental Chemistry) or else inaccuracies will emerge in subsequent effects assessment.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-104
-------
Free Ion Activity Model (FIAM)
The FIAM states that, in general,
metal ion availability and effect are
correlated with the free ion concentration
or activity in the water. It also
recognizes the role of competing ions in
reducing metal bioavailability ( e.g.,
calcium reducing zinc bioavailability).
The FIAM is a powerful tool as long as it
is applied with enough understanding to
anticipate or recognize exceptions.
1 4A.2.1A. Application. Predicting exposure at the
2 respiratory surface should include analysis of chemical
3 speciation, chemical kinetics, and binding with biological
4 ligands on the gill (or equivalent respiratory surface). The
5 FIAM states that, in general, metal ion availability and
6 effect are correlated with free ion concentration or activity
7 in the water (Brown and Markich, 2000; Campbell, 1995).
8 Consequently, knowledge of free ion concentration or
9 activity is crucial to fully describing exposure. The BLM
10 further suggests that the bioactivity of a metal is a result of its interaction with biological ligands
11 (i.e., biological macromolecules on the surface of the respiratory organ) (see Section 4.5,
12 Characterization of Ecological Effects), It follows that a dissolved metal ion's bioactivity is a
13 function of its complexation with dissolved ligands (which determines how much free ion will be
14 available for binding with biological ligands) and the affinity and stability of the metal
15 complexes with the biological ligands. Further complexation of metals takes place within the
16 cells of the respiratory organ, with the remaining free metal available for binding to transport
17 macromolecules for delivery to the organism's circulatory system; this results in the true
18 delivered dose.
19 Equilibrium speciation can be used to estimate the aquatic free ion concentration and its
20 resulting activity. In some cases, the free ion can be measured directly during an exposure
21 assessment. Combining insights from the FIAM approach with those from the BLM and HSAB
22 (see Section 4.1, Environmental Chemistry) theories allows general prediction of metal activity
23 on biological surfaces as different as fish gills (Janes and Playle, 1995; Reid and McDonald,
24 1991; Pagenkopf, 1983), green algae (Parent and Campbell, 1994; Crist et al., 1988), and
25 bacteria (Azenha et al., 1995). Entry across the integument could also be addressed with
26 conceptual tools of the FIAM and BLM (Krantzberg and Stokes, 1988), but this application is
27 not yet sufficiently developed for immediate use.
28 In the absence of sufficient information about speciation, exposure concentrations for
29 dissolved metals such as cadmium, lead, and zinc can be normalized on the basis of water
30 hardness. Most practitioners fit linear models (log of toxicity endpoint = log a + b [log of
31 hardness] and back-transform them to produce a normalizing function:
32
33 Toxic endpoint = 10a (hardness)1*
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-105
-------
1 This power function contains an easily corrected back-transformation bias that should be
2 removed from predicted hardness-adjusted exposure concentrations (Newman, 1991). The use
3 of hardness as a normalizing function should be done only when information on speciation is
4 lacking, as there is greater uncertainty using this method.
5
6 4.4.2.1.2. Limitations. The FIAM is a powerful tool as long as it is applied with enough
7 understanding to anticipate or recognize exceptions. As examples of important exceptions,
8 charged uranium complexes are toxic, in addition to the free ion (Markich et al., 2000), and the
9 neutral mercury complex HgCl2° is bioavailable due to its high lipophilicity (Simkiss, 1996).
10 Small organic ligands bind metals forming nonionic complexes that are also exceptions to the
11 FIAM. Copper bound to ascorbate was bioavailable to the green alga, Scendesmus sp.
12 (Campbell, 1995). Silver bound to glutathione or cysteine was bioavailable to Ceriodaphnia
13 dubia (Bielmyer et al., 2002). In addition, some nonionic metal-inorganic ligand compexes,
14 such as NiCO3, may also be bioavailable (Hoang et al., 2004).
15 The BLM has been parameterized for copper, nickel, silver, and zinc in fish, algae, and
16 invertebrates (Daphnia, ceriodaphnid). Some initial work has been completed for other species
17 and metals, but rigorous parameterization/validation is limited to the organisms and metals listed
18 above. Recent efforts in Europe have focused on extending BLM principles to describe chronic
19 toxicity; the results should become available in the near future. At present, the BLM approach
20 assumes that delivered dose is equivalent to the amount of metal bound to the gill (or other
21 respiratory organ), which may overestimate exposure. Models needed to predict within-cell
22 complexation of metals into available and nonavailable pools have not yet been developed.
23
24 4.4.2.2. Food Chain Pathway and Dietary Exposure **
25 Defining the particulars of metal exposure by ingestion is complicated by the diversity of
26 feeding modes, digestive systems, and physiology of candidate receptors. Thatdietbome
27 exposure to metals can result in accumulation in aquatic
28 organisms is well established, although the rate and
29 magnitude vary among organisms. What is less well
30 understood is how best to express dietary exposure in a
31 way mat can be linked to potentially toxic effects. For
32 nonionic organic chemicals, evidence is strong that
33 whole-body burdens of chemical (normalized to lipid
Currently, no standard approaches exist
to assess dietbome exposures of metals to
aquatic organisms in water quality
assessments, and consequently, the most
significant limitation involves a lack of
broad understanding of the mechanisms
underlying dietbome metal toxicity.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer-review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-106
-------
1 content) are robust metrics of toxicological dose, and these relationships appear to be
2 independent of whether exposure was via water or diet. In fact, the greatest strength of the body
3 residue approach for organic chemicals is that it effectively integrates different exposure
4 pathways into a single expression of dose and toxicological potency.
5 For metals (aside from organo-selenium and methyl mercury), the situation is far more
6 complex, and whole-body residue does not appear to be a robust indicator of dose when
7 compared across a range of exposure scenarios and/or organisms. The reason that whole-body
8 residue/effect relationships are not as effective for assessing metals probably stems from the fact
9 that although distribution of nonpolar organic chemicals in organisms is influenced largely by
10 passive partitioning, the uptake, distribution, and disposition of metals is governed by a number
11 of active biochemical processes. For example, some organisms take up metal and sequester it
12 into "storage" compartments in chemical forms that have little toxicological potency, whereas
13 other organisms actively excrete excess metals. Even for a particular organism, uptake and
14 disposition of metal may vary between waterbome and dietborne exposure (e.g., Kamunde et al.,
15 2002; Szebedinszky et al.s 2001).
16 Although these issues confound the development
17 of simple dose/effect metrics based on whole-body metal
18 residues, one should presume that residue/effect
19 relationships could be established if there were better
20 understanding of more specific concentration/response
21 relationships for the site (or tissue) of toxic action. For
22 example, some studies have suggested that the metal
23 concentration in the cellular cytosol (as opposed to that bound to cell walls or sequestered in
24 nonbioavailable metal granules) may provide a better expression of internal metal dose
25 associated with toxic effects (Wallace and Luoma, 2003; Wallace et al., 2003; Wallace and
26 Lopez, 1996). Others have suggested that whole-body residue residue-effect relationships are
27 confounded because the factor that determines the effects is not whole-body concentration per
28 se, but the rate of metal uptake in relation to metabolic capacity for detoxification and storage,
29 and, therefore, effects are governed by factors that influence the rate of uptake. When uptake is
30 elevated, the concentration of metabolically active metal at the site(s) of action increases (e.g.,
31 the spillover hypothesis) and effects ensue (Rainbow, 2002).
32 For these reasons, assessment of dietborne metal exposure should be considered in two
33 contexts: (1) dietborne exposure leading to accumulation and exposure to higher trophic levels
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-107
Assessment of dietborne metal
exposure is best considered in two
contexts: (1) dietborne exposure leading
to accumulation and exposure to higher
levels in the food chain (e.g., humans,
wildlife) and (2) dietborne exposure
leading to direct effects on exposed
organisms.
-------
1 in the food chain (e.g., humans, wildlife) and (2) dietborne exposure leading to direct effects on
2 exposed organisms. For the former, the primary challenges are to predict the aggregate uptake of
3 metal from both waterbome and dietborne exposure and to express that accumulation in terms
4 that are appropriate to assessing risk to the consuming organisms. Issues that come into play in
5 this scenario include whether forms of metal that are sequestered and detoxified by the aquatic
6 organism are bioavailable to organisms that then consume those aquatic organisms. The
7 available data to date suggest that metals can be sequestered/detoxified via different
8 mechanisms/forms in an organism, with metals sequestered as inorganic granules having greatly
9 reduced to no bioavailability to consumer organisms, while metals detoxified via
10 metal!qthioneins have relatively higher bioavailability to consumer organisms (Wallace et al.,
11 2003; Wallace et al., 1998; Mason and Jenkins, 1995; Nott and Nicolaidu, 1994). However,
12 owing to different digestive physiologies and other factors, changes in the bioavailability of
13 metals sequestered/detoxified by these different mechanisms or forms may occur (Wang, 2002).
14 With respect to the second scenario, the literature is mixed on the degree to which this is an
15 important pathway for inducing toxicity to aquatic organisms and how best to quantify exposure
16 in that context.
17
18 4.4.2.2.1. Application. There currently are no standard approaches to assess dietborne
19 exposures of metals to aquatic organisms in water quality assessments (Schlekat et al., 2001).
20 For the organometallic compounds of organo-selenium and methyl mercury, dietborne exposure
21 has been clearly shown to be a primary route of both uptake and toxic effects and should be
22 considered in assessments of these metal compounds. Beyond those two compounds, the picture
23 is much less clear. Toxicity to aquatic organisms from dietborne exposure to metals has been
24 demonstrated where exposure is sufficiently high, although in some cases these concentrations
25 are extreme (e.g., 10,000 ng/g Cu) (Handy et al., 1993). If the dietborne exposure necessary to
26 elicit effects is exceptionally high, it is not clear that this pathway will drive ecological risk, as
27 the environmental concentrations necessary to produce these exposures may be so extreme that
28 ecological risk will occur via other pathways (e.g., direct toxicity of waterbome metal).
29 In other studies, however, effects from dietborne exposure have been demonstrated at
30 relatively low exposure concentrations, such as in zooplankton studies (Hook and Fisher, 2002;
31 2001a, b). This raises much greater concern for metals assessment because it raises the potential
32 for lexicologically significant exposures occurring in cases where risk via a waterbome pathway
33 is low. That potential notwithstanding, other studies evaluating dietborne exposure with the
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-108
-------
1 same organisms and metals but somewhat different methods have reached different conclusions
2 regarding the significance of dietborne exposure (De Schamphelaere and Janssen, 2004; Metzler,
3 2003). Dietbome exposure to metals is an area of active research, and it is likely that new
4 insights will be forthcoming to rationalize what now appear to be conflicting data into a more
5 comprehensive understanding of dietborne metal effects. Until that time, the absence of a
6 standard approach will require that decisions regarding the handling of dietborne metal exposure
7 in aquatic assessments be made on a case-by-case basis. Some bioaccumulation models for
8 metals include consideration of dietborne exposure, although few, if any, link this accumulation
9 directly to effects.
10 In the context of screening for exposure and potential risks to consumers, the use of
11 whole-body inorganic metal concentrations in prey species may have some utility despite the
12 uncertainties associated with trophic transfer and bioavailability of dietary metals (i.e., in cases
13 where whole-body residues are below dietary toxic thresholds). For more definitive
14 assessments, further research is needed on quantifying the bioavailabilty and effects of inorganic
15 dietary metals (with the exception of certain organometallics where dietary toxicity has been
16 well established).
17 In cases where trophic relationships are not well understood, stable isotope techniques
18 can aid in defining the trophic status of species of concern and can be used to delineate the food
19 web (i.e., who is feeding on whom). The 15N concentration increases relative to 14N
20 concentration with each trophic exchange, and statistical models can link trophic status to metal
21 concentration (e.g., Cabana et al, 1994). Additionally the potential for using stable (and radio)-
22 isotopes of a metals can serve as a tools to understand dietary versus waterborne uptake,
23 particularly in laboratory studies. Other uncertainties include how behavioral changes induced
24 by dietborne metal exposure may affect the survival of the exposed organism (Irving et al.,
25 2003).
26
27 4.4.2.2.2. Limitations. As indicated above, the primary limitation in assessing dietborne metal
28 exposure for aquatic organisms is the absence of a broad understanding of the mechanisms
29 underlying dietborne metal toxicity and the consequent lack of a standard assessment approach.
30 Establishing rigorous residue/effect relationships mat integrate waterborne and dietborne metal
31 exposure is a critical consideration. When assessing dietborne exposure through direct
32 measurement from field samples, there are methodological issues to be resolved, such as whether .
33 to depurate the digestive system of prey organisms. Metal contained in food or sediment within
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-109
-------
1 the digestive system will increase measured body burdens, but it may not have the same
2 biological availability as metal residing within organismal tissues (Chapman et al., 2003).
3 Fractionating body burdens of metals (e.g., cytosolic metal vs. metal granules) has been
4 suggested as a better means of identifying the most readily accumulated fraction of dietborne
5 metal (Seebaugh and Wallace, 2004; Wallace et al., 2003; Fisher and Reinfelder, 1995;
6 Reinfelder and Fisher, 1994), although the interpretation of this information in the context of risk
7 assessment has not been rigorously developed.
8
9 4.4.2.3. Sediment Exposure
10 Most situations involving dissolved metals can be addressed adequately by assuming
11 near-equilibrium conditions between water and sediment concentrations (Hoffmann, 1981), as
12 long as one is sufficiently aware that some conditions can lead to important nonequilibrium
13 dynamics (Paquin et al., 2003). Bioavailability, sediment
14 transport, and chemical speciation all affect exposure of
15 benthic organisms to metals. Additional issues of trophic
16 transfer and routes of exposure (gills or equivalent vs.
17 dietary) should be considered in the same manner as
18 discussed above for water column organisms.
19 The most widely used approach for assessing metal
20 exposure in sediment is based on EqP theory, with sulfides as the primary partitioning phase (Di
21 Toro et al., 1990; 1991). Other approaches include sequential extractions with different media
22 (water, weak acids, strong acids) or normalization to total iron or organic carbon. Each has its
23 strengths and limitations. Detailed discussions are provided in Paquin et al. (2003) and Newman
24 et al. (2004); additional information is provided in Section 4.5.9.
25
26 4.4.2.3.1. Application. Several approaches have been taken for estimating exposures to
27 sediment-associated metals that account for bioavailability differences among various sediments.
28 The EqP approach assumes that chemical activity in the sediment, as indexed by chemical
29 concentration in the interstitial water, is proportional to the chemical's bioavailability to
30 sediment-dwelling organisms. In anoxic sediments, sulfides provide the primary binding phase
31 for many cationic metals. These metal sulfides are highly insoluble and are thought to have very
32 low toxicity. Thus, in sediments where there is more sulfide than metal, most metal should be
33 present as sulfides and therefore relatively nontoxic. The amount of reactive sulfide is quantified
Sediment Exposure Estimates
Not all benthic organisms are
exposed to sediments. Some feed and
respire in the overlying water column.
Sediment exposure estimates apply
only to those organisms that extract
nutrients or oxygen from sediments
and pore water.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality, guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-110
-------
1 by measuring the amount of sulfide liberated when sediment is extracted with 1 N HC1. This
2 procedurally defined quantity is known as acid-volatile sulfide or (AVS). The amount of
3 reactive metal is determined from the same extraction by measuring the metal concentration in
4 the acid extractant This quantity is known as simultaneously-extracted metal or SEM. The
5 potential bioavailability of metal is determined by comparing the relative molar concentrations
6 of the two, When SEM-AVS < 0, sufficient sulfide exists to bind all SEM and metal toxicity is
7 not expected. When SEM-AVS > 0, metal is present beyond the binding capacity of sulfide, and
8 toxicity may occur if there is sufficient excess metal but not sufficient other binding phases to
9 bind the metal. Use of this SEM-AVS as exposure estimates that are correlated with toxicity of
10 metals in sediment has been explored closely for Class B or borderline Class B metals (Berry et
11 al., 1996; Hansen et al., 1996; Ankley et al., 1996; Ankley et al., 1991; Carlson et al., 1991; Di
12 Toro et al., 1990). See Section 4.5.9 for further discussion.
13 Although the correspondence of SEM-AVS to
14 toxicity was found to be strong in these studies, some
15 question the applicability of the approach to all benthic
16 organisms because it is based on the chemistry of bulk
17 anoxic sediment, and many organisms live in oxygenated
18 burrows. In addition, several studies have shown some
19 degree of metal accumulation in organisms exposed to
20 sediments where sulfide is in excess and metals are thought to be nonbioavailable (or at least
21 nontoxic). A better understanding of the mechanisms of metal accumulation from sediment and
22 their relationship to toxic effects is needed to help interpret these issues. Until such information
23 becomes available, the SEM-AVS model can be used in exposure estimations as long as its
24 shortcomings are acknowledged appropriately.
25 Other tools for determining the exposure
26 concentration of sediment-bound metals include
27 metal concentrations in chemical (Fan and Wang,
28 2001; Babukutty and Chacko, 1995; Tessier et al.,
29 1984), acid (Langston, 1980; Luoma and Bryan,
30 1978), or biomimetic (Weston and Maruya, 2002;
31 Mayer et al., 2001; Chen and Mayer, 1998) extracts.
32 However, no consensus yet exists on their best use
33 for different types of metals or metalloids, Several
Sem-AVS
When the molar concentrations of
acid-volatile sulfide (A VS) in sediment
exceed the amount of simultaneously
extracted metal (SEM), the metals are
expected to associate with the solid phase
and not be bioavailable.
Exposure Assessments for
Benthic Organisms
The following information would enhance
exposure assessments for benthic organisms: (1)
improved computational or analytical methods for
analyzing distribution of metal among components
of the sediments, (2) improved computational
methods for assessing the influences of metal form
in sediments on sediment-water metal exchange,
and (3) a better understanding of the processes
controlling bioaccumulation of metals from
solution and food by metazoan species directly
exposed to the sediments.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-111
-------
1 other methods have been proposed. Based on the premise that iron oxides in oxic sediments.
2 lower metal bioavailability, iron in a 1 N HC1 sediment extract has been used to normalize metal
3 exposure concentrations (Luoma and Bryan, 1978). Increasing concentrations of organic carbon
4 can decrease metal bioavailability (Crecelus et al, 1982), so normalization of sediment metal
5 concentrations to organic carbon content has been useful in other cases. The more readily
6 extracted metals from sequential chemical extraction schemes tend to be the most bioavailable
7 (Young and Harvey, 1991; Tessier et al., 1984) and can be used as exposure metal
8 concentrations.
9
10 4.4.2.3.2. Limitations. Exposure assessment for benthic receptors could be enhanced if a clearer
11 consensus were reached about the utility of each method for different classes of metals,
12 biological species, and sediment types. Specifically, the following information is needed to
13 improve exposure analysis: (1) improved computational or analytical methods for analyzing
14 distribution of metal among components of the sediments, (2) improved computational methods
15 for assessing the influences of metal form in sediments on sediment-water metal exchange, and
16 (3) a better understanding of the processes controlling bioaccumulation of metals from solution
17 and food by metazoan species directly exposed to the sediments (Luoma, 1989).
18 EPA (U.S. EPA, 2002f) published a report on the application of solid phase AVS
19 equilibrium partitioning sediment benchmarks (ESBs) and interstitial water ESBs as No-Effect
20 guidelines to predict sediments that are acceptable for the protection of benthic organisms.
21 Details of the SEM-AVS method are discussed in Section 4.5.9 and the metals issue papers. The
22 method has been applied to predict toxicity of metals in sediment for Class B or borderline Class
23 B metals. An ESB based on the difference between the concentration of SEM and AVS is
24 appropriate for protecting benthic organisms from the direct effects of sediment-associated
25 metals, and not for estimating metal bioaccumulation. Chapman (2003) discusses limitations
26 concerning the AVS-SEM approach, including the degree to which the EqP approach adequately
27 represents exposures for organisms living in microenvironments in the sediment and/or who are
28 exposed via ingestion of sediment. Site-specific measurements, where possible, are useful for
29 confirming that generalized approaches like SEM-AVS appropriate for specific assessment
30 scenarios. Furthermore, one should recognize that bioaccumulation of metal may still occur
31 even when sulfides are in excess, and the potential consequences of this accumulation should be
32 considered.
33
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-112
-------
Pathway of Exposure for
Terrestrial Organisms
Pathways of exposure far
terrestrial organisms to metals
include movement from soils through
the food web, and to a lesser extent,
air deposition either into soils or
directly onto terrestrial receptors (e.g.,
plants).
1 4.4.2.4. Terrestrial Species
2 Terrestrial wildlife, plants, and invertebrates
3 accumulate metals from direct contact with soil or sediment,
4 from ingestion of contaminated food (plants or other animals),
5 and from incidental soil or sediment ingestion. A conceptual
6 model for direct and indirect exposure of terrestrial receptors
7 to metals in soil is presented in Figure 4-6. Pathways of
8 exposure include movement from soils through the food web,
9 and to a lesser extent, air deposition either into soils or directly
10 onto terrestrial receptors (e.g., plants). Because of significant differences in exposure patterns, it
11 is more convenient to discuss methods by receptor group (invertebrates, plants, wildlife) rather
12 than by pathways or environmental compartments.
13
14 4.4.2.4.1. Soil invertebrates. The soil ecosystem includes a complex food web of soil
15 invertebrates (both hard- and soft-bodied) that feed on each other, decaying plant material, and
16 bacteria or fungi. For risk assessment purposes, however, exposure is described as a function of
17 soil concentration rather than a detailed analysis of movement of metals through the food web.
18 This is a reasonable approximation for soft-bodied invertebrates (e.g., earthworms) whose
19 exposure is primarily through soil pore water (from both dermal absorption and soil ingestion)
20 (Allen, 2002). There is more uncertainty in correlating soil concentrations with effects in hard-
21 bodied invertebrates because they are primarily exposed through ingestion of food and incidental
22 amounts of soil. (Sample and Arena!, 2001). Regardless, for all types of soil invertebrates,
23 exposure estimates should account for differences in bioavailability among soil types, which
24 include differential partitioning to soil particles and pore water, metal speciation, and aging.
25 Each of these processes is discussed in detail in Section 4.1, Environmental Chemistry.
11/24/2004 Peer Review' Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-113
-------
Receptors
Soil Zones
Surficiat
(-8-8")
Surface
(-fl'-S1)
Sub-surf3C8
<•»*)
V*
VjJ
s\
< \
^
^
f*nntart
Ingestion
Inhalation Dust
or Vapors (during
burrowing)
>
/
' »
^
\
y
/\
/< ^
™_>
[ riarns
Soil Invertebrates
(Mammals Birds)
. Herptiles
m. MorhtwtrAc
• Insectivores
* Predators
<
1^
II
Figure 4-6. Conceptual model for direct and indirect exposure of ecological
receptors to metals in soil zones.
Source: Menzie and Little (2000).
1 4.4.2.4.1.1. Application. Currently, soil invertebrate exposure is calculated on the basis of total
2 metal concentration in bulk soils collected in the top 0-12 cm of soil (U.S. EPA, 1989c). In
3 detailed, higher level assessments, the organic matter on top of the soil (the "duff) may be
4 analyzed separately to provide further detail on exposure to detritivores (such as Collembold)
5 and deeper-soil-dwelling organisms (e.g., various species of earthworms). However, such
6 measures of exposure are limited, as they do not account for differences among soils in
7 bioavailability factors.
8 Cation exchange capacity (CEC) recently has been shown to be an important factor
9 modifying zinc bioavailability in soils for both invertebrates and plants, and presumably it will
10 be important for other canonic metals as well. However, CEC is strongly dependent on the type
11 and amount of organic material and oxyhydroxides present in the soil and is strongly pH
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-114
-------
1 dependent. Surface charge on organic material and oxyhydroxides increases with increasing pH,
2 thereby increasing their sorptive capacity for metals (thus decreasing metal bioavailability).
3 Conversely, positive surface charges increase as the pH drops, increasing sorption of anions
4 (e.g., arsenic or selenium) under low pH conditions and decreasing sorption of cation ionic
5 metals. Clays, on the other hand (except for kaolinite), have a surface charge that is largely
6 independent of pH. Therefore, normalization of toxicity data to CEC can be done only within
7 specific soil types and pH ranges, which frequently are not specified either in laboratory
8 bioassays or many field studies. Furthermore, it is important to note that most published values
9 of CEC are measured at pH 7.
10 Soil chemical models are being developed to predict how aging will modify bulk soil
11 concentrations when soils are amended with soluble salts. Aging reduces the bioavailable
12 fraction of metals over time (see Section 4.1,6.3 for a discussion of aging in soils). Preliminary
13 studies suggest that consideration of aging may result in estimates of the bioavailable fraction as
14 low as 0.1 x bulk soil concentrations. Until the data become available for metals of concern,
15 toxicity values derived from soluble-salt amended soils (which have not simulated aging) cannot
16 be reliably corrected to approximate aged metals in field situations. Appropriate adjustments
17 should be included in toxicity test protocols to simulate aging (McLaughlin et al., 2002), except
18 when assessing acute (short-term) risks of spills.
19
20 4.4.2.4.1.2. Limitations, Data on CEC for field soils are often available, but similar information
21 from laboratory studies of the toxicity of metal-spiked soil currently in not Furthermore, the
22 dependency of CEC on soil type (amount and type of organic matter, type of clay, and pH) also
-23 complicates the comparison among studies. Therefore, although exposure concentrations can be
24 adjusted across field locations of similar soil type and pH, it is more difficult to make
25 appropriate comparisons of field exposures with laboratory-generated concentration-response •
26 functions. Expressing exposure on the basis of pore water concentration is the goal, to reduce
27 the variability in toxicity among sites; however, there are currently significant limitations to
28 collecting and interpreting metal-related data from soil pore waters. Such information is not
29 available in the published literature and, therefore, should be estimated using EqP theory (as with
30 sediment pore water analyses). Published soil binding coefficients (IQs) can be used, although
31 these values are inherently uncertain as well (published value depends on derivation method, soil
32 type, etc.; see Section 4.1.4 for a discussion on the limitations of
33
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-115
-------
Metal Accumulation
in Plants
The highest accumulation
of metals in plants generally
occurs in the roots.
1 4.4.2.4.2. Plants. Plants access metals through the pore water,
2 although mycorrhyzae, protons, and phytosiderophores released by
3 the root can significantly influence the microenvironment and
4 change uptake rates of metals (George et al., 1994; Sharma et al.,
5 1994; Laurie and Manthey, 1994; Arnold and Kapustka, 1993).
6 Furthermore, plants have both active and passive mechanisms for taking up or excluding metals,
7 depending on internal concentrations and whether or not the metal is an essential micronutrient,
8 or whether it is mistaken for an essential micronutrient. Plants can be exposed to metals via
9 aerial deposition onto leaf surfaces, trapping metals in hairs or rough cuticular surfaces. This
10 might provide an exposure route for herbivores; it may also provide an exposure route for plants,
11 as there are ion channels through the cuticle that are able to transport ionic metals from the leaf
12 surface to other locations in the plant, depending on the inherent mobility of the metal in the
13 xylem and phloem (Marschner, 1995).
14
15 4.4.2.4.2.1. Application. All plant species take up metals from soil through their roots via
16 various mechanisms (Raskin et al., 1994; Cataldo and Wildung, 1978). The default approach to
17 estimating exposure of plants to metal is measuring metal concentrations in bulk soil. However,
18 as with soil invertebrates, mis overestimates exposure because it does not account for differential
19 bioavailability that results from complexation. Furthermore, with time, the bioavailability of soil
20 metals may change due to dissolution or complexation; thus, "point-in-time" measurements of a
21 soil may not reflect the future bioavailability. For example, Pb sulfide spilled onto soil (relative
22 bioavailability 1-5%) with time would weather to Pb sulfate (relative bioavailability 50%),
23 which could further evolve to sorb onto Fe oxides or phosphates (relative bioavailability 10-15,
24 30-59%, respectively). Soil pH, organic matter, and cation exchange capacity are the most
25 important variables influencing bioavailability (see Section 4.5, Characterization of Ecological
26 Effects). However, CEC and clay content are not consistently reported in the literature and
27 therefore cannot be used to define relative bioaccessability and toxicity of metals. General
28 categories of uptake based on soil pH and organic matter are shown in Tables 4-17 and 4-18 in
29 Section 4.5 for ranges typically found in soils. These tables are a useful qualitative guide to
30 identify soils with increased (or decreased) metal bioavailability. It is very clear that strongly
31 acidic soils increase plant uptake of Zn, Cd, Ni, Mn, and Co and increase the potential for
32 phytotoxicity from Cu, Zn, and Ni. Alkaline soil pH increases uptake of Mo and Se, while Pb
33 and Cr are not absorbed to any significant extent at any pH (Chancy and Ryan, 1993).
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-116
-------
1
2
3
4
5
6
7
g
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
4.4.2.4.2.2. Limitations. Qualitative relationships between soil chemistry and bioavailability
are appropriate for national-scale application. However, for site-specific or metals-specific
applications, quantitative methods are preferred. It has been suggested that only uncomplexed,
free ionic species of cations can be taken up by roots, and this has been described using a FIAM
similar to the model used in aquatic systems (Parker and Pedler, 1997; Lund, 1990). Studies
have been conducted to generate models to predict free copper activity from total metal content,
pH, and organic matter content (Sauve et al., 1997a, b, 1995; McBride et al., 1997), and for lead,
empirical models were generated using only total metal levels and pH. However, exceptions to
the free-ion model have been identified. Ionic or organometallic complexes that increase the
total concentration of elements at the root surface have been correlated with increased uptake,
either through disassociated ions or through uptake of intact complexes (Parker et al., 2001;
McLaughlin et al., 1994). In addition, it is not clear how well plants can distinguish between
ions of similar size and charge. Plant uptake of macronutrients is much better understood than is
uptake of micronutrients or contaminants, with the primary work on uptake of micronutrients
focusing on iron (Welch, 1995). Different mechanisms have been identified that control
macronutrient uptake by plants, providing a means through which contaminants can enter root
tissue.
Exposure Pathway for
Terrestrial Wildlife
Food and the incidental
ingestion of soil are the two
most important exposure
pathways for terrestrial wildlife.
4.4.2.4.3. Wildlife. The relative importance of exposure pathways
and routes varies by species of animal as well as by metal,
although, in general, wildlife exposure is primarily through diet
and incidental ingestion of soils or sediments. The EPA has
concluded that there are certain chemicals and exposure situations
for which inhalation or dermal pathways are important, but under
most situations they can safely be considered to be insignificant contributors to total metal loads
(U.S. EPA, 2003c).
Wildlife food chain exposures for metals are controlled by bioavailability,
bioaccessibility, and bioaccumulation. The availability of metals in soils depends on whether
exposure occurs via pore water or other pathways external to the organisms. Bioaccessibility of
metals to animals and plants that live on or in the soils can be influenced by a number of soil
parameters, such as pH, CEC, and organic carbon. These soil factors tend to be less important
for soils that are incidentally ingested by wildlife species. For further review of soil
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-117
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
bioavailability factors that are important in wildlife exposure, see Section 3.4.6.3, Ecological
Effects (Wildlife).
4.4.2.4.3.1. Application, In the absence of site-specific information, the following
generalizations can be used;
• Incidental soil ingestion is a proportionally more important pathway for herbivores
than for carnivores or invertivores.
• Uptake into soil invertebrates (e.g., earthworms) is a proportionally more important
pathway for animals that feed on these organisms. (Note: This assessment reflects
work done with earthworms and may not apply to hard-bodied soil invertebrates such
as Colembolla.)
The relative importance of exposure pathways (soil vs. food chain) is dictated by the
fraction of metal-contaminated soil in the diet and the amount of accumulation of metal in food
items. Figure 4-7 provides a simple scheme for judging the relative contribution of food and soil
before accounting for bioavailability. Incidental ingestion of soil becomes proportionally more
important for exposure to wildlife when (1) the bioaccumulation factor (BAF) from soil to food
(e.g., to plants or soil invertebrates) is less than 1 and (2) the fraction of soil in the diet is greater
than 1%.
4.4.2.4.3.2. .Limitations, Experience at metals-contaminated sites indicates that the above
generalizations should be viewed with caution. As site-specific information is acquired, the
relative importance of pathways may change. For example, site-specific data may show that the
accumulation of a chemical into plants or soil invertebrates is much lower than indicated by the
default assumptions, hi such cases, incidental ingestion of soil would become proportionally
more important. The bioavailability of metals in incidentally ingested soil is also variable, as
discussed later. Therefore, when the exposure is being driven by incidental ingestion,
refinements of exposure estimates can benefit from a better understanding of bioavailability.
Attention should be paid to the bioavailability of metals for which incidental soil ingestion is the
predominant pathway and where ecological risk is indicated, although very little information is
available on this for most wildlife species.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer-review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-118
-------
4
5
6
7
8
9
10
II
12
13
14
15
16
Both the bioaccumulation factor and life histoiy components interact to control the
relative importance of soil or food as the primary metal transfer medium. In general, the
following statements can be made about dietary uptake of metals from soil versus from food:
• If bioaccumulation is low («1), importance of soil ingestion versus diet for metal
exposure increases.
• When bioaccumulation is greater (~1 or higher), the food pathway should dominate.
• The closer the association an animal has to the ground, the greater the importance of
soil ingestion. This association may be due to ground foraging, burrowing habits, etc.
• The looser the association with the ground (e.g., piscivores, aerial/arboreal"
insectivores, raptors), the lower the importance of soil ingestion.
100
-------
1 Extrapolation from models developed for estimation of bioavailability of metals in soils
2 for incidental human exposures may not be broadly applicable to all wildlife species owing to
3 the influence of differences in digestive physiology and anatomy across the broad and diverse
4 range
5 of mammalian and avian species (Menzie-Cura and TN&A, 2000). For example, metals present
6 in soils may be more or less bioavailable within the gut of an herbivore that relies on
7 fermentation as compared to the simpler gut of a carnivore that is designed to break down
8 proteins. These gut systems differ in chemistry (including pH) and residence time. For example,
9 ruminants such as deer, antelope, and other hoofed stock initially process food through microbial
10 digestion in the rumen, and their gut pH is general neutral or slightly acidic. Furthermore, end
11 products of microbial digestion in domestic sheep bind to copper and enhance its uptake in the
12 proximal intestine, resulting in a very low tolerance of these animals to dietary copper
13 (NAS/NRC, 1980). Hind gut fermentors such as horses, rabbits, and granivorous birds (e.g.,
14 grouse or pheasants) also rely on microbial processes for digestion of lignen and other plant parts
15 but have a more acidic foregut man do ruminants. Most insectivores and carnivores, on the other
16 hand, have relatively acidic digestive systems with a significant amount of protein present.
17 While the low pH may alter the metal speciation (or dissociate ions from attached ligands), most
18 metals require active transport to move through the gut wall and into the circulation. This may
19 be accomplished through binding with transporter proteins present either in the diet or within
20 gastrointestinel cells, which likely differ among the wildlife species (Hill, 1979). Taken
21 together, these physiological differences are significant and make it difficult to accurately
22 extrapolate dietary toxicity thresholds across species. However, thoughtful application of this
23 information will allow appropriate inferences to be made across species with similar physiology
24 and may explain instances where statistical projections of toxic thresholds (e.g., species
25 sensitivity distributions) appear to not be predictive of actual effects.
26
27 4.4.3. Food Chain Modeling for Wildlife
28 4.4.3.1. Application
29 Food chain modeling can be used to estimate the exposure of wildlife to metals based on
30 ingestion of soil, food, and water. The basic format of the model is the same as that for organic
31 substances and is shown in Figure 4-8; detailed explanations are available in several related
32 documents (e.g., U.S. EPA, 2003c [EcoSSLs]; U.S. EPA, 1997d [ECOFRAM]; Sample et al.,
33 1997). Measured, or predicted/estimated, concentrations of metals in soil, surface water, and
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-120
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
food items can be used in the model, or concentrations in food can be modeled using trophic
transfer factors. TRIM.FaTE (http://www.epa.gov/ttn/fera/human_apex.html can be used for
modeling media concentrations of metals as a result of aerial deposition. Information on diet,
foraging area, and the like can be found in CH2M HILL (2001) and Sample et al. (1997). The
absorbed fraction variable accounts for differences in relative bioavailability (RBA), and is
either 1 (default value) or an appropriate estimate. Note that there is no relative absorption
factor (RAF) included for food, because the default assumption is that site bioavailability equals
that from the toxic form used in the toxicity tests
This approach is the same as mat used in risk
assessments of organic substances except when trophic
transfer rates are used to model food concentrations only
on the basis of soil measurements (rather than using direct
measures of concentration of metals in food items), in
Trophic Transfer Values
Trophic transfer values for metals in
terrestrial systems are an inverse function
of soil concentrations; therefore, it is
inappropriate to use constants for this
term.
which case metal-specific transfer rates are required. As with aquatic organisms, trophic transfer
values for metals in terrestrial systems are an inverse function of soil concentrations. Therefore,
it is inappropriate to use constants for this term. Sources, use, and limitations of function of soil
where apparent uptake ratios are greater at the lowest and highest concentrations of metals in
soils as compared with the middle concentrations of metals in soils and where tissue metal
concentrations remain stable over a wide range of soil metal concentrations.
Sample et al. (1998a) developed uptake models to predict concentrations in earthworms
from soil concentrations. These models can be used to estimate the exposure of both the worms
themselves and of vermivorous wildlife (e.g., song birds, voles, and shrews). For selected metals
(arsenic, cadmium, copper, mercury, manganese, lead, and zinc), the best estimate of tissue
concentration in earthworms is a simple In-In regression. The addition of soil pH data to the
regression' model did not markedly improve fit. If soil calcium concentration was incorporated
into the regression model, a better fit could be obtained for cadmium and lead but not forother
metals. Tissue concentrations were inaccurately estimated for the transition metals nickel and
chromium by both simple and multiple regression models. For general estimates, log-linear
regression models may be used as bioaccumulation models for arsenic, cadmium, copper,
mercury, manganese, lead, and zinc in earthworms. For site-specific assessments, it is
recommended that location-relevant bioaccumulation models be developed through direct
measurements of local soil and tissue concentrations.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-121
-------
Oral intake (mg/kg/day) =
[Soilj x Ps x FIR x AFj*] + £ fl< x P« x FIR:
l/=i
where
P,=
FIR
AF =
N =
x AUF+[Waterjx WIR]
Concentration for contaminant (j) in soil (mg/kg dw)
Soil ingestion rate (proportion of FIR)
Food ingestion rate (kg/kg-body weight/day, dw)
Absorbed fraction of contaminant (j) in biota type (i)
Number of different types of biota in the diet
Contaminant concentration in biota type (i) (mg/kg dw)
Note: (Soilj x T~) can be substituted for Bi where T^ = soil-to-biota trophic transfer
factor (TTF) [as dry weight to dry weight] for contaminant (j) and biota type (I)
Proportion of biota type (i) in diet (unitless)
Area use factor; proportion of available habitat for a wildlife species within the area
of concern (%)
j = Concentration for contaminant (j) in water (mg/L)
WIR = Water ingestion rate (mL/kg/day)
p..
AUF
1 Figure 4-8. Wildlife Oral Exposure Model.
2 Source: U.S. EPA (2003c).
3
4 There is no compilation of plant biota/sediment accumulation factors (BSAFs), but U.S.
5 EPA (2003c) (EcoSSLs) provides some data on select metals. For national-level assessments,
6 uptake factors for plants provided by Efroymson et al. (2001) should be used. The highest
7 accumulation of metals in plants occurs in the roots, although other parts of the plant also
8 accumulate metals to varying degrees (Mitrofanov,1993; Greszta, 1982). With the exception of
9 a few hyperaccumulator species, most plant species do not bioconcentrate metals (i.e., BAFs <1).
10 Lead, arsenic, chromium, and cobalt are not taken up by plants in measurable quantities, and the
11 small amount that is taken up is mostly confined to root tissues (Chaney et al., 2000; McGrath,
12 1995; Chaney and Ryan, 1994; Xu and Thornton, 1985), Exceptions to this exist, notably
13 including, but not limited to, Se and Mo: uptake of these metals into the edible portion of plant
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. - It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-122
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
-22
23
24
25
26
27
28
29
30
31
32
33
tissues can be quite high, generally not sufficient to cause plant toxicities (McGrath, 1995;
Bingham et al., 1986; Foy et al., 1978), but may cause significant food chain toxicities.
In contrast, plants are quite sensitive to some metals (manganese., zinc, copper, for
example) and may die before achieving high levels that pose a threat to animals via food chain
transfer.
4.4.3.2. Limitations
Food chain modeling for wildife is similar for both metals and organic substances. All
estimates are limited by the information available about the receptors) of concern in terms of
dietary preferences, relative amounts consumed of various items, relative bioavailability between
laboratory and field, and other factors (see Figure 4-8). For metals, the largest uncertainty in the
dietary uptake model is in the estimates of trophic transfer factors. As discussed above, these
generally are not constants and therefore require the use of uptake equations (i.e., tissue
concentrations vary as a function of soil concentration). This necessitates the use of at least
quasi-probabilistic modeling, rather than a more simple determination of a single, deterministic
oral uptake value.
Use of either measured or modeled tissue metal accumulation levels as an indicator of
potential toxicity is limited by the requirement that the accumulated amount be related to a
benchmark effect level (i.e., Critical Body Residue, CBR). Very little information is available
for metal CBRs in terrestrial wildlife. This is particularly problematic when whole-body tissue
levels are reported because what really matters is the effective metal concentration at the site of
action of toxicity. If the concentration at the proximate site of action of toxicity is proportional
to the whole-body concentration, then this is a lesser concern. However, if the concentration at
the site of action is not proportional to the whole-body concentrations, then direct measures of
metal concentrations at the site of action is required, especially at higher assessment levels.
Furthermore, the absolute level of metal
accumulation is not as important as the rate of uptake
(Hook and Fisher, 2002; Hook, 2001; Roesijadi, 1992).
High uptake rates overwhelm the ability of organisms to
sequester or excrete the metal, leaving larger proportions of
If the concentration at the site of action is
not proportional to the whole-body
concentration, then direct measures of
metal concentrations at the site of action
is required, especially when conducting
detailed, higher level assessments.
the accumulated metal in a more bioavailable form. Uptake is believed to occur because of the
ability of some organisms to sequester metals that enter the cell (e.g., by inducing the synthesis
of metallothionein [MT] or granule formation). Adverse effects are avoided as long as the rate
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-123
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
of metal uptake does not exceed the rate at which the organism is able to bind the metal, thereby
preventing unacceptable increases in cytosolic levels of bioreactive forms of the metal. If the
rate of uptake is too great, the complexation capacity of the binding ligand (e.g., MT) could be
exceeded; cytosolic metal levels then become unacceptably high, and adverse effects could
ensue. Because measures of uptake rates are not available, static concentrations are used instead.
Measurement of the form of the metal that is present in a tissue may be a more
predictable indicator of the potential for effects than is total metal concentration (Mason and
Jenkins, 1995; Roesijadi, 1992). Metals also may be bound and sequestered by
organophosphorus granules, thereby rendering them nonavailable to bind with other intracellular
target enzymes (George, 1982; Coombs and George, 1978). Although a wealth of available data
exist on measured total tissue levels of metals, there are few data on intracellular speciation or
sequestration of the metals. Although models that can be used to perform this sort of evaluation
are currently under development, the ability of such models to be used as a tool in exposure
assessment remains to be demonstrated. Therefore, measurement of total metal in plant (or
animal) tissue remains the accepted default approach.
In sum, the following factors contribute significantly to uncertainty in food chain models:
• Soil ingestion rates are highly uncertain.
• Diet composition can be highly variable, and diet composition has a significant effect
on exposure.
•• Relative bioavailability from foods is completely unknown.
• Modeling over a number of trophic levels propagates uncertainty.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-124
-------
1 4.5. CHARACTERIZATION OF ECOLOGICAL EFFECTS
2 Metals are naturally occurring substances, and organisms have evolved mechanisms for
3 maintaining homeostasis in the presence of expected exposure levels. However, areas of metals
4 enrichment, particularly when generated by anthropogenic activities, can pose challenges to
5 organisms. Toxicity assessment for ecological receptors exposed to metals requires an
6 understanding of bom the natural mechanisms for tolerance for (or, in the case of micronutrients,
7 the use of) metals and the toxicological responses that occur when exposure exceeds the capacity
8 of the organism to regulate its body burdens. Interactions between metals in either their uptake
9 or toxicity (such as Cd/Ca/Zn, Hg/Se, Cu/Mo) also should be considered in toxicity assessments.
10 Risk assessments for metals are further complicated by the need to express the dose-response (or
11 concentration-response) functions of bioavailable units that are functionally equivalent to
12 measures of exposure. This section provides tools and approaches for addressing issues of
13 essentiality, appropriate toxicity tests, novel endpoints (e.g., gene expression), and acclimation
14 or adaptation to continued exposures.
15
16 4.5.1. Essentiality
17 Essentiality, or the requirement for normal i
. J Essentiality
18 organism metabolic function, of many metals is one of the _. .. ... .. . ._
J Essentiality, or the requirement for
19 primary factors,that differentiates risk assessment for
20 metals and metal compounds from that of synthetic organic
21 chemicals (Janssen and Muyssen, 2001). Some trace
synthetic organic chemicals.
22 elements, such as cobalt, copper, iron, manganese,
normal organism metabolic function, of
many metals is one of the primary factors
that differentiates risk assessment for
metals and metal compounds from that of
23 selenium, molybdenum, and zinc, are necessary for the normal development of plants and
24 animals. In many cases, these metals are added to animal feed and pharmaceutical products
25 (SRWG, 2002) or to plant fertilizers. Other metals, such as arsenic, cadmium, lead, and
26 mercury, are not known to be essential to plant and animal growth and development. Trace
27 elements can be divided into three groups:
28 .
29 - • Those known to be essential.
30
31 • Those that have beneficial metabolic effects but have not been shown to be essential.
32
33 • Those that occur widely in living organisms but seem to be only incidental
34 contaminants and are not known to be beneficial (Mertz, 1981).
35
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-125
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Table 4-14 classifies the metals addressed in this framework by their known essentiality
to organisms. The concept that many metals are required for organism health at one range of
concentrations and are toxic in quantities outside of that range has been referred to as the
"window of essentiality" (Hopkin, 1989) or the "optimal concentration range" for essential
elements (Alloway, 1995; Fairbrother and Kapustka, 1997; Van Assche et al., 1997). For
essential elements that exhibit biphasic dose-response curves (Figure 4-8, above), adverse effects
resulting from deficiency should be considered, as well as those that result from excessive
exposure. Recognition of the window of essentiality, as well as consideration of the biochemical
and physiological processes that regulate metals within living organisms, are both important
components of ecological effects assessment (Abernathy et al., 1993).
4.5.1.1. Application
The optimal concentration range (or safe intake range) for essential elements should
ensure that effects thresholds such as Toxicity Reference Values (TRVs) are not lower than the
nutritional requirements for the particular plant or animal species being evaluated. Where TRVs
or other effects concentrations (or doses) are intended as thresholds for detrimental effects due to
excessive intake), care should be taken to ensure that these toxicity thresholds for essential
metals are at the upper end of the optimum range or sufficiency range (at the point where toxic
effects begin to occur). If set too low (i.e., in the range where deficiency can occur), the
determination of risk will be erroneous. For wildlife, the literature on dietary requirements of
essential elements for livestock can be consulted. The NAS/NRC has published useful
summaries (NAS/NRC, 1994,1980), and McDowell
(2003) updates this information. Minimum
concentrations required for plant growth are
summarized in Marschner (1995).
4.5.1.2. Limitations
Because of differences in test conditions
among published studies, it may be difficult to
directly compare toxicity threshold values with
recommended dietary requirements of essential
elements. Extrapolation of data among species (e.g.,
from livestock to wildlife species) also may
Threshold Values
For essential elements, it is important to ensure
that effects thresholds, such as Toxicity
Reference Values (TRVs), are not lower than the
nutritional requirements for the plant or animal
species being evaluated. It may be difficult,
however, to directly compare toxicity threshold
values with recommended dietary requirements
because of differences in test conditions among
published studies.
In screening-level assessments, toxicity
threshold values are advised for application, if
they are not lower than estimated requirements.
Detailed, higher level assessments may require
additional bioassays to characterize the biphasic
dose-response curve and determination of both
required and excessive threshold levels.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-126
-------
1 add uncertainty to the effects assessment. Furthermore, addition of safety factors when deriving
2 protective values often results in concentrations significantly below required intake. Derived
3 toxicity threshold values should be used in screening-level risk assessments if they are not lower
4 than estimated requirements. Uncertainty in toxicity thresholds or estimated requirements
5 should be addressed as part of the risk management process. Higher level assessments, where
6 more accurate estimates of effects thresholds are expected, may require additional bioassays to
7 characterize the-biphasic dose-response curve and determination of both required and excessive
8 threshold levels. See the following sections for considerations of bioavailability factors,
9 mixtures of multiple metals, and interspecific extrapolations.
10
Table 4-14. Metals classified by their known essentiality
Metal
Aluminum (Al)
Antimony (Sb)
Arsenic (As)
Barium (Ba)
Beryllium (Be)
Cadmium (Cd)
Chromium (Cr)
Cobalt (Co)
Copper (Cu)
Lead (Pb)
Manganese (Mn)
Mercury (Hg)
Molybdenum
Nickel (Ni)
Selenium (Se)
Silver (Ag)
Strontium (Sr)
Thallium (Tl)
Vanadium (V)
Zinc (Zn)
Essential
(known requirement for
health and function)
Plants
X
X
X
X
X
Animals
X
X
X
X
X
X
X
X
Beneficial
(but not known to be
essential)
Plants
X
X
Animals
X
X
Nones sential
(and not known' to be
beneficial)
X
X
X
X
X
,
X
X
X
X
X
Source: Adapted from a table presented in SRWG (2002) and incorporating data from NAS/NRC (1980) and Barak
(1999). Fairbrother and Kapustka (1997) discussed the roots of essentiality of naturally occurring elements.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-127
-------
1 4.5.2. Acclimation and Adaptation
2 Organisms have developed various mechanisms to cope with variable background metal
3 concentrations, particularly for those metals that are essential elements (see Table 4-14). If the
4 amount is less than required, passive or active uptake mechanisms are used to enhance internal
5 concentrations, whereas during periods of excessive amounts exclusion mechanisms come into
6 play. These various approaches to homeostasis are discussed in detail in Kapustka et al. (2004).
7 Additionally, organisms can acclimate to suboptimal metal levels by changing various
8 physiological functions, or populations can undergo genetic adaptation and develop increased
9 tolerance to different levels (Rusk et al., 2004; Wallace and Srb, 1961).
10 The genetic makeup of an organism defines its
11 ability to cope with environmental conditions. Genes
12 can be expressed or remain "silent," and shifts in gene
13 expression can occur when the environment changes.
14 Furthermore, organisms use different portions of their
15 total array of genetic information in different life
16 stages. This shifting of tolerance within the genetically
17 defined limit of the organism is known as acclimation.
18 Physiological changes induced by acclimation may be
19 reversed if the environment reverts to the original
20 conditions (Posthuma and Van Straalen, 1993; Prosser,
21 1986). Tolerance acquired through physiological
22 acclimation processes is not always passed on to offspring; however, the same genetic
23 information that allowed acclimation to occur in the parents will be passed on, so the offspring
24 will retain the ability to acclimate in a similar fashion. If the offspring develop in the altered
25 environment, they will express the set of genes most appropriate for tolerance of those
26 conditions.
27 Genetic adaptation results from increased survival of tolerant genotypes and subsequent
28 changes in gene frequencies. However, linking these genetic changes to increased tolerance in
29 the field and identifying the specific mechanisms responsible has proven challenging.
30 Laboratory experiments conducted with Ft generations obtained from metal-contaminated
31 habitats provide the strongest evidence to support a genetic basis of tolerance (Klerks and
32 Levinton, 1993), and new methods in toxicogenomics (e.g., microarrays) are providing
33 additional insights.
Tolerance, Acclimation, and Adaption
Tolerance is the ability of an organism to
maintain homeostasis under a variety of
environmental conditions, such as variable
metal concentrations.
Acclimation is how an individual develops
tolerance during its lifetime, and it may be
gained or lost. Acclimation is also called
phenotypk plasticity.
Adaptation is a genetic change over
multiple generations as a response to natural
selection. Traits are not lost during single life
times. Adaptation is also known as genotypic
plasticity.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-128
-------
Costs of Metal Tolerance
Although considerable evidence supports
the hypothesis that previously exposed ,
populations will be tolerant to metals, both
physiological acclimation and adaptation to
contaminants may have specific costs. For .
example, induction of metal-binding proteins
increases metals tolerance but also uses energy
normally available for other metabolic
processes (e.g., growth, reproduction).
Similarly, genetic changes associated with
metals exposure might harm populations.
Reduced genetic diversity has been reported in
populations exposed to contaminants and may
result in population bottlenecks as well as
increased susceptibility to other stressors.
1 Evidence of adajptation (convergent evolution)
2 for metal tolerance in'plants comes from the fact that
3 plants of diverse taxonomic relationships grow on soils
4 high in metals. Metal-tolerant flora have been
5 described for soils high in zinc, nickel, chromium, and
6 copper (Antonovics et al., 1971; Brooks, 1972). These
7 reviews indicate that some species are restricted to the
8 high-metal soils, but other species exist across a broad
9 concentration range. These represent differences in
10 niche breadth (i.e., those restricted to high-metal soils
11 vs. those occurring in soils that have either high or low
12 concentrations of metals).
13 Metal adaptation in plants often is accompanied by metal adaptation in co-located
14 animals, as selection for metal tolerance is expected to improve fitness in exposed conditions
15 (Posthuma and Van Straalen, 1993). Compared with other environmental parameters, metal
16 exposure is regarded as a strong and stable selective force, and it can lead to rapid
17 evolution of tolerance (Posthuma and Janssen, 1995). Metal-tolerant animals tend to grow fast,
18 mature early, and have a high excretion efficiency (Posthuma and Janssen, 1995).
19 Metal adaptation in natural populations of terrestrial invertebrates has been demonstrated
20 conclusively for several animals: the terrestrial isopod Porcellio scaber and the springtails
21 Orchesella cincta, Isotoma notabilis, and Onychiurus armatus (Posthuma and Van Straalen,
22 1993). Metal tolerance also has been demonstrated in ticks and a fly species in response to the
23 application of a metal-based pesticide (Posthuma and Van Straalen. 1993).. There is evidence for
24 increased metal tolerance in other species, but acclimation and adaptation could not be
25 distinguished (Posthuma and Van Straalen, 1993).
26 Although there is considerable evidence to support the hypothesis that previously
27 exposed populations will be tolerant to metals, both physiological acclimation and adaptation to
28 contaminants may have specific costs. For example, although induction of metal-binding
29 proteins increases tolerance to subsequent metal'exposure, it also uses energy normally available
30 for other metabolic processes (e.g., growth, reproduction). Similarly, genetic changes associated
31 with exposure to contaminants might harm populations. Reduced genetic diversity has been
32 reported in populations exposed to contaminants and may result in population bottlenecks.
33 Furthermore, as tolerant genotypes are eliminated from a population, the reduced genetic
34 diversity may increase the susceptibility of this population to other stressors.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-129
-------
1 tolerance. For example, the replacement of sensitive species by tolerant species, termed
2 "interspecific selection" (Blanck et al., 1988), is a common response in polluted systems and one
3 of the most consistent indicators of metal pollution. Pollution-induced community tolerance
4 (PICT) has been proposed as an ecotoxicological tool to assess effects of contaminants on
5 communities (Blanck et al.,,1988). PICT is tested by comparing responses of communities
6 collected from polluted and reference sites to contaminant exposures under controlled
7 conditions. The increase in community tolerance at a polluted site that results from the
8 elimination of sensitive species is considered evidence that this restructuring was caused by the
9 pollutant.
10 Naturally occurring levels of metals play an important role in biogeographic distributions
11 of plants and animals and may, in fact, be limiting factors in species distributions or use of
12 landscapes. Therefore, it is difficult to make generalizations about effects levels that are
13 applicable and consistent to all organisms in all habitats, and it becomes very important to
14 clearly define the geospatial location of the area to which the assessment results will apply.
15 For site-specific assessments, the assessment results will be directly applicable to the entire
16 range of species that may be found on that site (although for assessments conducted over areas
17 larger than several square miles, it is possible that multiple soil types and other local landforms
18 may result in significant differences in metal bioavailability and plant communities.
19 For assessments conducted for regional or national assessments, criteria development, or
20 ranking purposes, it should be acknowledged that results will be based on organisms and soil
21 types that result in greatest bioavailability and sensitivity. Care should be taken, however, that
22 the organism-environment combinations that are assessed are, in fact, compatible with real-
23 world conditions. For example, benthic organisms generally associated with slow-moving,
24 warm waters would not be expected to tolerate conditions of high metal biovailability such as
25 those occur that in faster-moving, colder waters that have little organic matter. Thus, for site-
26 specific assessments, species tested and water (or sediment) used in the test system should be
27 similar to conditions at the site. In the absence of such information, data from standard test
28 species and conditions could be used, but uncertainty factors may be warranted to adjust the final
29 toxicity value accordingly.
30 More appropriately, single-result assessments for the entire country should be avoided.
31 Rather, such assessments should be subdivided into metal-related ecoregions known as
32 "metalloregions" (McLaughlin and Smolders 2001) so that protection levels, mitigation goals,
33 and ranking results will be appropriate for the suite of species naturally present within each type
34 of controlling environment. This is directly analogous to the use of ecoregions when
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-132
-------
Met allo regions
For national-scale assessments, the entire
country can be subdivided into metal-related
ecoregions (known as "metalloregions") to
help ensure that protection levels, mitigation
goals, and ranking results will be
appropriate for the suite of species naturally
present within each type of controlling
environment. However, the metalloregion
concept has not yet been applied across
1 establishing WQC (Griffith et al., 1999). The value of metalloregions is that they provide the
2 conceptual framework to account not only for the broad regional parameters affecting metal
3 availability in soils and waters, but also for the differences in organism response to added metal.
4
5 4.5.2.2. Limitations
6 Although there is considerable evidence to
7 support the hypothesis mat previously exposed
8 populations will be tolerant to metals, both physiological
9 acclimation and adaptation to contaminants may have
10 specific costs (Wilson, 1988). For example, although
11 induction of metal-binding proteins increases tolerance
12 to subsequent metal exposure, it also uses energy
13 normally available for other metabolic processes (e.g.,
14 growth, reproduction). Similarly, genetic changes
15 associated with exposure to contaminants might harm populations. Reduced genetic diversity
16 has been reported in populations exposed to contaminants and may result in population
17 bottlenecks. Furthermore, as tolerant genotypes are eliminated from a population, the reduced
18 genetic diversity may increase the susceptibility of this population to other stressors. There is
19 theoretical support for the hypothesis that populations adapted to contaminants have higher
20 metabolic costs or are more susceptible to other stressors (Hoffman and Parsons, 1997; Mulvey
21 and Diamond, .1991); however, few empirical studies have demonstrated increased costs.
22 One of the assumptions behind the use of PICT as an ecotoxicological tool is that
23 differences in tolerance among communities can be detected using short-term experiments. This
24 significantly constrains the application of PICT as an assessment tool. Although tolerance at the
25 population level can be assessed using a variety of species, logistical considerations will limit the
26 types of communities where tolerance can be investigated experimentally. Most of the original
27 research on PICT has been conducted using small organisms with relatively fast life cycles (e.g.,
28 benthic invertebrates, soil microbial communities). Therefore, extrapolation of these results to
29 communities of terrestrial plants and animals should be done with caution.
30 The metalloregion concept (McLaughlin and Smolders, 2001), although intuitively
31 appropriate, has not yet been fully developed for the United States. The country has been
32 divided into ecoregions for both aquatic and terrestrial systems (Bailey et al., 1994; Bailey,
33 1983). These are based on climactic and vegetation factors and form the basis of metalloregions.
34 EPA is still working to complete ecoregion maps at much finer scales for each state (see EPA
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-133
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Web site at http://www.epa.gov/wed/pages/ecoregions/ecoregions.htm). To complete the
metaUoregion concept, soil properties that affect bioavailability (e.g., pH, cation exchange
capacity (CEC), and OM) should be overlayed on the ecoregions, along with soil type (e.g.,
sandy loam, clay loam) and background metal concentratons of metals. Similar information is
needed for water bodies. Although this type of information is fairly current and available, soil
data have not been updated since the mid-1970s, which may limit their usefulness to some
extent. Nevertheless, work is under way to develop metalloregions, although it is likely to be
several years from the time of this writing before they are available for use in a decision-making
capacity.
4.5.3. Metals Mixtures
Mixtures of metals (including metalloids and other contaminants) are commonly
encountered in the natural environment as a result of anthropogenic inputs. Metal interactions,
according to Calamari and Alabaster (1980), occur at three levels:
•
• Chemical interactions with other constituents in the media;
• Interactions with the physiological processes of the organism during uptake; and
• Interactions at the site of toxic action.
Much of the difficulty in interpreting the available information on the toxic effects of
metal mixtures is due to differing measures and definitions of the bioavailable fraction of metals,
whether it is the fraction that is available for uptake from the environment or at the site of toxic
action. Some measure of the bioavailable metal fraction in the exposure media is needed to
accurately predict the effects of metals and metal mixtures (Di Toro et al., 2001; Sauve" et al.,
1998; Weltje, 1998; Posthuma et al., 1997; Ankley et al., 1996). Characterization of effects of
metal mixtures has also been reported to be concentration dependent (Mowat and Bundy, 2002;
Fargalova, 2001; Herkovits etal., 1999; Spehar and Fiandt, 1986).
4.5.3.1. Studies of Metal Mixtures
Few controlled studies exist on the toxicologic interactions of metals found in
environmental contamination scenarios. ATSDR has compiled and evaluated interaction studies
involving various metals: methyl mercury, arsenic, cadmium, chromium, lead manganese, zinc,
copper, cesium, cobalt, strontium, and uranium (draft interaction profiles available online at
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-134
-------
1 http://www.atsdr.cdc.gov/iphome.html). Few studies allowed any quantification of interaction
2 magnitude, whether using the authors' definitions of toxicologic interaction or EPA's definitions
3 based on dose and response addition. The summaries below indicate some of the qualitative
4 conclusions available on interaction potential.
\
5 A study of a mixture of cadmium, lead, and zinc in rats found slightly more marked
6 adverse hematological effects with the ternary mixture exposure than with the cadmium-lead,
7 cadmium-zinc, or lead-zinc mixtures (Thawley et al., 1977); inconsistencies in dietary levels of
8 calcium and vitamin D in this study, however, may have made comparisons problematic. A
9 well-controlled rat study has reported significant synergism between cadmium and lead
10 regarding testicular atrophy (Saxena et al.} 1989). That study also demonstrated protective
11 effects of high dietary levels of zinc, which effectively reduced the testicular effects of the
12 cadmium-lead mixture to control levels. No studies have been located that would allow
13 extrapolation of those high exposure results to more common, lower environmental levels. In
14 another study (Fowler and Mahaffey, 1978), a relatively wide range of endpoints were
15 investigated in studies that covered each metal singly and all possible binary and ternary
16 mixtures. Body weight gain was depressed equally by the ternary mixture and the cadmium-lead
17 mixture, and to a lesser extent by the arsenic-lead and cadmium-lead mixtures, whereas food
18 utilization was depressed more by the ternary and arsenic-cadmium mixtures than by the other
19 binary mixtures. In general, the biological parameters studied in this report indicated changes of
20 smaller magnitude and inconsistency in direction for binary mixtures compared with ternary
21 mixtures.
22 The data regarding interactions of environmental metals usually are not adequate for
23 predicting the magnitudes of interactions. Interaction profiles by ATSDR of metal-metal
24 interactions have considered the following combinations: arsenic, cadmium, chromium and lead;
25 lead, manganese, zinc, and copper; and cesium, cobalt, and strontium (see above Web site for
26 draft reports). Experimental efforts to identify and quantify interaction mechanisms among these
27 metals are needed. For some endpoints, the data are not robust enough to show even the
28 direction of interaction (i.e., whether the joint action will be dose additive or greater than or Jess
29 than additive). The animal studies discussed briefly in this report used commercial diets or semi-
30 purified diets that may have higher or lower levels of essential metals than human diets. Much
31 higher doses of the metals appear to be required to elicit effects when commercial diets are used
32 man when semi-purified diets are used. At the other extreme, effects are seen at very low doses
33 when deficient diets are used. Comparisons among studies are therefore problematic,
34 particularly when the diets are not specified.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-135
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
4.5.3,2. Mixtures
Similar to human health risk assessments, ecological risk assessments commonly involve
mixtures of metals. However, Agency guidance is less developed for how to estimate metal
mixture effects for wildlife. Binary interactions of dietary metals in livestock are discussed in
NAS/NRC (1994,1980) and McDowell (2003). These references can be used as sources of
information from which at least qualitative estimates of additivity, synergy, or antagonism can be
developed. Additionally, see section 4.3.6.3. for discussion of human health assessment for
mixtures, which provide information useful in consideration of wildlife effects.
4.5.3.3. Application
Two key questions should be addressed by effects assessments related to metal mixtures:
• To what extent does each metal contribute to any observed effect?
• Are the effects significantly greater than or lesser than the sum of the individual
component effects?
The answers to these questions also have the potential to affect water quality guidelines
(WQGs), EcoSSL values, cleanup targets, and other similar management decisions.
Methodologies (graphical and statistical) to predict impacts of metal mixtures and
interactions of individual metals within mixtures can be broadly classified as either
Concentration Addition models or Effects Addition models. Both models use metal water
concentrations to generate concentration-response curves for individual metals, and these data
are then used to generate specific critical concentrations for mixture models. Similar models can
be developed for soils or sediments.
In the Concentration Addition model, all metals in a mixture are added together to predict
toxicity; differing potencies are taken into account by converting chemical concentrations to an
equitoxic dose (e.g., Toxic Units (TUs) or Toxicity Equivalence Factors (TEFs), which convert
all metals to one metal concentration). In the Effects Addition model, differing potencies are
ignored, and the effect of each metal's concentration in a mixture is combined to predict mixture
toxicity. Only the Concentration Addition model allows detection of toxicity greater than
predicted (more than additive); the Effects Addition model can only predict strict additivity.
The TU approach involves deriving dimensionless units for each metal in a mixture by
dividing individual concentrations by their individual toxic concentrations (such as LC50 values).
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-136
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
The TUs for all the metals in the test mixture are then summed. A value of 1 ± 0. 1 in the case of
an LC50 would predict 50% mortality (or another effect if TUs are based on that effect rather than
on lethality), A value significantly greater than 1 would predict more than 50% mortality. A
value significantly less than 1 would predict less than 50% mortality.
The TU approach can be used with any endpoints, for instance LC50 or EC50 values.
However, it assumes the same mode of action for all the chemicals so will only predict additive
effects. The toxic concentration can be derived from guideline values (e.g., WQGs), from
literature toxicity data, or from specific experiments. This approach can be used when setting
WQC. It has, for example, been recommended for use as part of the Australia and New Zealand
WQGs (ANZECC and ARMCANZ, 2000). These guidelines employ a concentration addition
approach using WQG concentrations of metals as TUs:
where TTM is the predicted total toxicity of the mixture, C; is the concentration of me
component, and WQG; is the guideline for that component If TTM exceeds 1, then the mixture
has exceeded the water quality guideline. It is important to note that this has been developed for
water, and significant limitations may be associated with applying it to soil systems.
Norwood et al. (2003) conducted a literature review on the effects of metal mixtures.
Mixtures varied from 2 to 1 1 metals. The investigators determined that the TU approach is
presently the most appropriate model for predicting effects of metal mixtures, based on currently
available data (e.g., effect concentrations, ECx values). Effects addition models, especially if
based on body or tissue concentrations, might be more accurate in the future, but they require .
reliable dose-response and bioaccumulation curves for all single metals (not just ECx values)
and then careful testing of the models (research on tissue concentration effect levels is ongoing).
Application of HSAB. The QICAR approach (described in Section 4.1, Environmental
Chemistry) might also be applicable to predict the potential for interactions of metals in mixture.
Unsatisfied with the qualitative conclusions of Newman and McCloskey (1996), Ownby and
Newman (2003) fit binary metal mixture data derived from the Microtox assay to the following
model of joint independent action (Finney, 1947):
Predicted PA+B = PA + PB - PAPB
where:
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-137
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
PA+B = The biological response to the metal mixture, expressed as a proportion
PA and PB = The biological responses to A or B when present singly
This relationship would hold if the actions of the paired metals were independent.
However, if the metals were not independent, the PAPB term will deviate from its expected value
of zPAPB where z = 1. Assuming that paired metals with very similar binding tendencies are
more likely to interact with the same biological Ugands than are metals with very dissimilar
tendencies, deviations of the z coefficient from 1 would reflect departure from complete
independence of metal action. HS AB theory allowed prediction of metal interactions in this
model system. Although Microtox is considered to be a useful tool for organic contaminants, it
is very sensitive for metals (Willemson et al.} 1995).
4.5.3.4. Limitations
At present, the most appropriate approach for determining the toxicity of metal mixtures
and addressing the two key questions listed above is to use the TU approach as a screening-level
assessment. This approach cannot, however, be used beyond screening because it does not
provide enough certainly, Norwood et al. (2003) found that of 191 case studies examined, 70%
were additive or less than additive. Thus, this approach was primarily either appropriate or
overprotective, but 30% of the case studies indicated
that this approach would be underprotective. For
aquatic organisms acutely exposed to cationic metals,
the assumption of additivity is sufficient, particularly if
bioavaiiability adjustments are made using the BLM
(see Section 3.4.8).
Toxic Unit Approach
Currently, the most appropriate approach
for determining the toxicity of metal
mixtures is to use the Toxic Unit (TU)
approach as a screening-level assessment
tool. This approach cannot, however,
presently be used beyond screening because
it does not provide adequate certainty since
combined effects depend on the relative
amounts of each metal. Furthermore, the TU
approach is recommended for applications
with mixtures containing less than six
components.
Currently, there are no realistic means of ranking
mixtures of metals or individual metals within mixtures.
National criteria for mixtures also are not possible at
this time. Furthermore, the concentration addition (TU)
approach is recommended only for application to
mixtures with fewer than six components. This issue remains site specific because interaction
responses are dependent on both the actual metal mixture combinations (metals and ratios) and
the exposed organisms.
Researchers are presently attempting to predict the effects of metal mixtures on the basis
of critical body concentrations of metals. Ongoing research is attempting to integrate the effect
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-138
-------
I accumulation functions of each metal into a metal mixture model in which an effect addition
2 formula will be compared with a concentration addition formula, both based on body
3 concentration. Current work is focused on aquatic invertebrates, and similar research is required
4 for other organisms.
5 It is possible that the BLM may be expanded in the future to include mixtures. In theory,
6 if two metals compete for binding to the same site of toxic action on an organism, it should be
7 possible to model the total metal bound to that site and hence predict metal toxicity using a
8 mechanistic BLM approach in an Effects Addition model. Alternatively, if two metals do not
9 compete for the same binding site on the organism, then the BLM may provide more reliable
10 estimates of individual metal bioavailability, and these estimates can then be combined in more
11 accurate Effects Addition models. However, at present, these possibilities remain theoretical and
12 need testing. However, this possibility, while improving the ability to assess the effects of metal
13 mixtures, does not include temporal aspects (i.e., "time-to-response" versus concentration).
14
15 4.5.4, Background
16 Background is defined as the amount of metals
17 occurring in soils, water, or air as a result of
18 anthropogenic and natural processes. Anthropogenic
^. 19 contributions are limited to those that are not influenced
20 . by current, direct releases (i.e., emissions, discharges, or
21 disposal) from a source or site of concern. This
22 includes metals that may arise from manmade
23 substances (particularly metalloids) or from natural
24 substances (metallic ores) present in the environment as
25 a result of human activity that are not specifically
26 related to the release in question (U.S. EPA, 2003c).
27 Background should be defined in a specific spatial and
28 temporal aspect that is related to the scope of the
29 particular hazard or risk assessment. Background
30 concentrations can vary by as much as five orders of
31 magnitude, depending on soil type, geography, and other factors (Chapman and Wang, 2000).
32 Background may exacerbate lexicological effects and accumulations of metals from
33 direct emissions or other regulated sources or, conversely, it may result in adaptation of
34 organisms to higher metal concentrations and result in increased tolerance to emissions (see
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-139
Background
Background is defined as the amount of
metals occurring in soils, water, or air as a
result of anthropogenic and natural
processes. Anthropogenic contributions are
limited to those that are not influenced by
current, direct releases (i.e., emissions,
discharges, or disposal) from a source or site
of concern. This includes metals that may
arise from manmade substances (particularly
metalloids) or from natural substances
(metallic ores) present in the environment as
a result of human activity that are not
specifically related to the release in question.
Background can exacerbate toxicological
effects and accumulations from direct
emissions or other sources, or conversely it
may result in adaptation of organisms to
higher metal concentrations and result in
increased tolerance to emissions.
-------
I
2
3
4
-5
6
7
S
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
above section). Furthermore, because metals occur naturally, and some are essential macro- or
micronutnents, they are at least partially responsible for how plants and animals are distributed
within various ecoregions. The distribution of plants and animals, local species diversity
species survival, and the vitality of individuals can be profoundly affected by background levels
of metals m an area. Humans, on the other hand, are distributed throughout the world,
irrespective of naturally occurring levels of metals, so knowledge of background levels is not as
significant
4.5.4.1. Application
Concentrations of metals in soils and waters of the United States vary tremendously
Thus, use of a single number to represent all areas within the United States is to be discouraged
Statewide averages (US. EPA, 2003c) provide somewhat better resolution, but even these are
constrained by political boundaries, not by geochemical characteristics. Additional information
on concentration of metals in soils at smaller spatial resolutions is provided in Shacklette and
Boerngen (1984). Some metals (e.g., Fe, Cu, and Zn) are included in the State Soil Geographic
Database (STATSGO) available at www.nrcs.usda.gov/technial/techtools/stat_browser.html
These data can be grouped at whatever spatial scale is required, but they are not screened for
whether they represent true background concentrations. Similarly, data on water concentrations
can be retrieved from EPA's STORET database (www.epa.gov/storel/mdex.html). These data
should be used with caution, however, as this is a voluntary-entry database, and there is no
consistent method for measurement or for quality assurance/quality control (QA/QC) of the data
Like the STATSGO information, STORET data do not necessarily represent true background
levels; additionally, there is incomplete coverage across the United States.
Similar data for sediments are available from the EPA's National Sediment Quality
Survey database, which is available in the form of an MS Access 1997 database and can be
obtained on a CD from the EPA's Office of Water, Office of Science and Technology (EPA
OW, OST). It contains survey data from 1980 to 1999, including sediment chemistry data, tissue
residue data from selected organisms, and toxicity data (lethal and sublethal effects on various
test organisms). Overall, there are data from more than 50,000 stations and 4 6 million
analytical observations. The data were complied from a variety of sources, but mostly from state
and federal monitoring programs, and sampling and analysis strategies varied among sources
All data have "data qualifiers" associated with them in the database. However, this was not a
statistically designed survey, and it is heavily biased toward contaminated sites, so inferences to
areas that were not sampled should be made with caution.
11/24/2004 Peer Review Draft
4-140
-------
1 A more appropriate approach is to define average (and ranges) of background
2 concentrations for various ecoregions (Hargrove and Hoffman, 1999; Bailey, 1998; Omernick,
3 1986). Work is under way in this regard, but such values currently are not available. Therefore,
4 the recommended default is to use state averages where possible and always to define the range
5 that might be encountered within the spatial scale being considered. It is also important to
6 recognize that background concentrations in water are highly variable through time because
7 flooding or drought conditions substantially change the relative concentrations of metals in a
8 water body. Again, it is suggested that ranges rather than single number averages be used.
9 Several practical issues should be considered when evaluating the contribution of
10 background to hazard or risk and its implications for various risk management options when
11 conducting site-specific assessments. First, a physical and/or temporal boundary should be
12 defined for the analysis. Next, background should be described, estimated, or measured. U.S.
13 EPA (2002b, c) provides detailed guidance on how to estimate local background concentrations
14 and notes that locations of background samples should be areas that could not have received
15 contamination from the site but that have the same characteristics as the medium of concern (i.e.,
16 water, soil, or air).
17 Another reason to include background in the assessment is to evaluate the effect of the
18 remedial options. Although some areas of the site may have elevated concentrations for certain
19 metals, other metals may not be elevated. Remedial actions could cause these naturally
20 occurring metals to become more bioavailable, thereby resulting in unintended toxicity. The
21 HSAB theories and QICAR models (see Section 4.1.2) can be applied to derive useful estimates
22 of biological activity for metals to aid in making decisions regarding remedial actions.
23 Incorporating background. How the information on background concentrations is used
24 in the final risk assessment depends on how data on toxic responses were generated and the
25 relationship of the bioavailability characteristic of the naturally occurring material to the source-
26 related additional metal. For aquatic organisms, toxicity tests generally are conducted in waters
27 that are relatively low in background metals and are of moderate hardness. Therefore, the
28 toxicity thresholds described (e.g., in WQC) represent the total amount of metal in the water, not
29 the amount that can be added to the natural background levels. If, however, the tests are run on
30 site-specific waters where metal background and bioavailability may differ significantly from
31 standard waters and where organisms have been acclimated to such conditions, then the toxicity
32 threshold reported should include both the background and the amount that is added. When
33 using this approach, caution should be exercised, because the background levels of metals in the
34 aquatic system may be highly variable over short time periods. The primary consideration is that
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-141
-------
1 the bioavailable fraction of the natural background concentrations should be added to the
2 bioavailable fraction of the added metal.
3 In terrestrial systems, toxicity bioassays for soil organisms (plants and invertebrates) are
4 conducted either in artificial soils or in actual soils, both of which contain background
5 concentrations of metals (particularly of essential elements). Therefore, the toxic levels reported
6 are the amount added during the study. Unfortunately, the background amounts frequently are
7 not reported, so it may not be feasible to take a true "added risk" approach. Furthermore, the
8 amount of metal added to the test system generally is in a much more bioavailable form (e.g., a
9 metal salt) than is the background material. Therefore, for a site-specific release of a highly
10 bioavailable form of the metal, the background concentration may not contribute significantly to
11 total metal uptake. Thus, it becomes extremely important to measure exposure in terms of the
12 bioavailable fraction, so field exposures can be expressed in a manner comparable to the highly
13 bioavailable material used in toxicity tests for threshold setting.
14 For human health and wildlife assessments, the amount of metal in food material should
15 be taken into consideration in a manner similar to that discussed above for soil organisms.
16 Again, differences in bioavailability of food-incorporated metals and top-dressed metal salts
17 should be considered. Additionally, natural uptake and the amount of metals in forage and other
18 food items will vary, depending on the amount of metal in soil and the particular species of
19 plant/animal present in the area Site-specific assessments can take this into consideration. For
20 national or large-scale assessments, a default assumption can be made that food items contain
21 sufficient amounts of micronutrients to meet dietary needs and that the toxic threshold value
22 represents the bioavailable fraction that is added above these values.
23
24 4.5.4.2. Limitations
25 National databases of metal concentrations in soil or water do not currently differentiate
26 between naturally occurring levels and levels that are elevated owing to anthropogenic sources..
27 However, the databases can be screened to ascertain whether specific areas are affected by point
28 source emissions and so can be used as reasonable estimates for large-scale background levels.
29 More important, however, is that background concentrations are most frequently reported as total
30 amount of metal, without specifying the bioavailable fraction or chemical speciation or the data
31 (e.g., pH or redox potential) needed to estimate bioavailability. Information on probable mid- to
32 long-term changes in soil or water properties that might enhance bioavailability of background
33 metals also should be provided for accurate assessments of future risk.
34
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-142
-------
Indirect Effects
Indirect effects to organisms initiated
by metals toxicity can be negative
(density dependent or independent) or
positive (density dependent) and can
occur between species or within the same
species. Table 4-16 summarizes some of
these relationships.
1 4.5.5. Indirect Effects of Metals
2 Metals can initiate ecological changes by directly
3 affecting individual organisms (i.e., through toxic
4 responses). Organisms can die, fail to reproduce, or have
5 altered behavioral patterns. As a consequence of such
6 actions on individuals, other organisms within the
7 community will be indirectly affected through reduced
8 number of prey items or predators or changes in
9 competition for resources. Additionally, indirect interactions between organisms can and do
10 occur independently of initiating effects from toxicity due to metals exposures (described by Dill
11 et al., 2003). Although "initiators" may be biotic, physical, or chemical, in all cases there is an
12 effect to a species (the "transmitter") that has an effect on another species (the "receiver").
13 Indirect effects to organisms initiated by metals toxicity can be negative (density
14 dependent or independent) or positive (density dependent) and can occur between species or
15 within the same species. Some examples are summarized in Table 4-16; additional examples are
16 provided by Chapman et al. (2003). Interactive effects can also occur due to the combined
17 effects of environmental stressors and metals toxicity. For instance, Liess et al. (2001) found
18 that the toxicity of copper to an Antarctic amphipod was increased by food shortage and
19 excessive ultraviolet-B radiation. Some authors have considered such interactive effects to be
20 indirect; however, they can be addressed as direct effects in risk assessments and hence are not
21 considered further.
22
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-143
-------
1
2
3
Table 4-16. Examples of Indirect Effects of Metal Toxicit]
Transmitter organism
Prey populations reduced
Predator populations reduced
Competitor populations
reduced
Toxicity to some larvae of a
species
Receiver organism
Predator populations reduced
Prey populations increased
Competitor populations increased
Increased adult-to-larval survival,
increased growth and biomass of
adults of the same species
Comment
Negative indirect effects due to
reduced prey (food)
Positive indirect effects due to
reduced predation
Positive indirect effects due to
reduced competition
Positive indirect effects due to
density -dependent
compensation
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
4.5.5.1. Application
Functional redundancy (species having similar roles in ecosystem processes) is a well-
known but arguably not well-understood phenomenon in ecosystems. It is often assumed that if
an organism's primary prey item is reduced or eliminated, then the organism can switch to
another prey item due to functional redundancy. However, in the case of food web-mediated
effects, the following sequence can and does apply: chronic metal exposure reduces the food
abundance of certain dietary components; as a result there are increased energetic costs of
feeding and an associated reduced growth efficiency. It is important that all components of the
food web are exposed to metals; thus, the determination of the final effect should include the
concurrent changes to all elements.
A determination of whether indirect effects such as loss of preferred prey can occur
requires three components. First, appropriate conceptual diagrams should be developed in the
problem formulation phase of the risk assessment and subsequently refined. Such diagrams
should incorporate sufficient detail regarding key biotic interactions (e.g., competition and
predation) and the ecological context in which the species exist and pollution occurs. As noted
by Chapman et al. (2003), seasonal and life-stage changes in feeding patterns can occur and will
"require iterative temporal diagrams, showing the various reasonable possibilities." If the factor
is limiting, then the organism will exhibit a response; however, if it is not a limiting factor, the
likelihood that it will lead to organismal changes decreases.
Second, risk assessors should focus proactively on the first three possibilities in Table 4-
16 because it is extremely unlikely that indirect effects due to density-dependent compensation
will be detectable without extraordinary effort. Increased numbers of receiver organisms due to
reduced competition or predation are more likely to be detected and more likely to have overall
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-144
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
adverse implications for communities and/or ecosystems, despite their apparent positive aspects
(to Ihe particular receiver species).
Third, risk assessors should focus on any cases of reduced growth of individuals,
particularly where the individuals affected are relatively tolerant to metals toxicity and their
normal feeding patterns are disrupted. These are clear indications (although not conclusive
proof) of indirect effects of metals toxicity. An example of negative indirect effects due to
reduced prey is provided by Campbell et al. (2003). These authors summarize extensive field
research into yellow perch in metal-impacted lakes in eastern Canada. Yellow perch in lakes
shift from feeding on zooplankton to feeding on littoral macrobenthos during their second year
of growth and then begin to include a significant amount of fish in their diet during their third to
fifth year of growth. Fish in the most metal-impacted lakes did not undergo this normal
sequence of diet shifts. Instead, they continued to utilize smaller prey throughout their lives
owing to the loss of their primary prey species to direct metal toxicity. A bioenergetic
bottleneck developed because the perch's growth efficiency was reduced by the need to catch
and eat smaller prey. The perch were more tolerant to metals toxicity, and thus there were no
major direct effects on their survival (though multiple physiological effects were recorded).
However, the loss of their primary prey species resulted in smaller or stunted perch, a major
indirect effect of metals toxicity. Many methods exist in the ecological literature to gather
population, community, or ecosystem function data that may be achieved without significant
levels of effort. Some elegant analyses of population demographics can be accomplished within
a single sampling effort (e.g., Wilson et al., 1996; Heyer et al., 1994; Davis, 1982).
Presently there is no realistic means of
ranking metals on the basis of indirect effects,
nor are national criteria possible. This issue
remains a site-specific one because
interactions between the initiator (metal
toxicity), the transmitter, and receptor
organisms are dependent on both the level
and type of toxicity and the sensitivities of
individual organisms within structurally and
functionally unique populations,
communities, and food chains.
4.5.5.2. Limitations
It is difficult to predict natural (e.g.,
behaviorally mediated), indirect interactions in nature
(Dill et al., 2003). It is also difficult to predict indirect
interactions due to abiotic factors such as metals
toxicity. Accurate predictions require good
understanding of the functional interactions within and
between populations, particularly along food chains, as
well as density-dependent and density-independent
processes between and within species.
Presently there is no realistic means of ranking metals on the basis of indirect effects, nor
are national criteria possible. This remains a site-specific issue because interactions between the
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-145
-------
1 initiator (metal toxicity), the transmitter, and receptor organisms are dependent on both the level
2 and type of toxicity and the sensitivities of individual organisms within structurally and
3 functionally unique populations, communities, and food chains.
4 As noted by Campbell et al. (2003) and Dill et al. (2003), indirect effects of stressors on
5 organisms probably are not uncommon occurrences. However, there are few documented cases
6 of this occurring, and it is likely that most cases go unrecognized. It would be useful to review
7 previous detailed-level ecological risk assessments involving metals toxicity to determine and
8 document previous cases.
9 Indirect effects from metals toxicity are primarily associated with ecosystem function,
10 not structure. However, ecological risk assessments typically are focused on determining risk to
11 structure, not function. The assumption is made that measuring structure protects function
12 because of functional redundancy. Clearly, given the reality of indirect effects, this is-at best a
13 questionable assumption that needs to be tested.
14
15 4.5.6. Bioavailability in Terrestrial Systems
16 Unavailability of soil metals to terrestrial
17 organisms is closely linked to dynamic soil physical and
18 chemical parameters, and biotic processes. As discussed
19 in Section 4.1, there are qualitative and quantitative
20 methods and models for considering soil chemistry issues
21 and aging of metals in soil that modify metal
22 bioavailability and its toxicity in soil. Unless specific
23 bioavailability data exist, even with measures of soil
24 physical and chemical parameters such as soil loading
25 capacity, aging of metals, and speciation, accurate
26 estimates of exposure to the terrestrial biota cannot be made. In situations where information is
27 not available, bulk soil chemistry is typically used with a default of 100% relative
28 bioavailability.
29
30 4.5.6.1. Soil Organisms: Invertebrates and Microorganisms
31 Metal speciation is the primary consideration in assessing the bioavailability of metals to
32 soil invertebrates and microbes. Major assumptions regarding metal exposure in aquatic
33 systems, such as a the relatively homogeneus dissolution of metals in the exposure water, may
34 not be applicable or may apply at different scales in soil systems. Although soil microbes may
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-146
Unless specific bioavailability data
exist, even with measures of soil physical
and chemical parameters such as soil
loading capacity, aging of metals, and
speciation, accurate estimates of exposure
to terrestrial biota are not possible. In
situations where information is unavailable,
bulk soil chemistry is typically used with a
default of 100% relative bioavailability
(i.e., the bioavailability of the chemical in
soil is presumed to be the same as the
bioavailability in the tests used to develop
the toxicity data).
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
,21
22
23
24
25
26
27
28
29
30
31
32
be immersed in soil solution films surrounding soil particles, few invertebrates are exposed to
metals in this manner. Exposure usually consists of partial contact of soil solution films with the
surfaces of the invertebrates that are capable of absorbing metals (e.g., earthworm dermal
surfaces). Direct contact with membranes across which metal uptake can occur does not take
place for many hard-bodied soil invertebrates (e.g., arthropods), and metal uptake is almost
entirely through the ingestion of metal associated with particle matter or soil solution. For these
reasons, exposure and relative bioavailability cannot be expressed similarly for each organism in
the soil ecosystem, and an understanding of primary routes and mechanisms of metal exposure
should be established for species or groups of similar
organisms.
Often, direct toxicity testing of the soil of concern
is the best method for assessing bioavailability and
Exposure and relative bioavailability
cannot be expressed similarly for each
organism in the soil ecosystem, and an
understanding of primary routes and
mechanisms of metal exposure should be
established for species or groups of
similar organisms.
toxicily-to-soil biota (Fairbrother et al., 2002). Issues such
as spiking of metals solutions onto soils, aging, and
laboratory-to-field extrapolation should be considered. In
the ecological soil screening document (U.S. EPA, 2003c), published literature was evaluated
using primary soil parameters affecting lability of metals in soils in a matrix to qualitatively
indicate metal bioavailability. Further information on this topic and factors relating soil
chemistry to soil biota toxicity are discussed in Section 4.1.6. A terrestrial BLM method
currently under development (Allen, 2002) may provide a useful tool to link bioavailability, soil
chemistry, and toxicity to soil biota.
Ionic or organometallic complexes that
increase the total concentration of
elements at the root surface have been
correlated with increased uptake, either
through disassociated ions or through
uptake of intact complexes. Thus,
research into how to quantify metal
complexes in soils relative to their lability
is important
4.5.6.2. Plants
The most common route of metal exposure in
plants is through the roots. Ions and organic molecules
contact roots via the transpiration stream, diffusive
transport, and microbe-facilitated transport. At the root
surface, soluble contaminants have the potential to enter
the root tissue through the transpiration stream or through
a range of mechanisms designed to facilitate nutrient uptake. In general, it is thought that only
uncomplexed, free ionic species of cations and ions can be taken up by roots. This has been
described using a FIAM (Parker and Pedler, 1997; Lund, 1990). Sauve et al. (1998) put forward
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-147
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
a method to calculate free metal activity levels for copper and lead to derive a pH-dependent soil
criterion for soil biota. The soil criteria paper was based in part on previous studies (Sauve et al.,
1997a, b, 1995; McBride et al., 1997) to generate models to predict free copper activity from
total metal content, pH, and organic matter content or, in the case of lead, using only total metal
levels and pH. The equations were generated from more than 60 soils using ion selective
electrodes and standard methods for determining pH, organic matter content, and total metal
levels. However, many exceptions to the FIAM have been identified, such that it has been
abandoned in favor of its close cousin, the BLM. Ionic or organometallic complexes that
increase the total concentration of elements at the root surface have been correlated with
increased uptake, either through disassociated ions or through uptake of intact complexes (Parker
et al., 2001; McLaughlin et al., 1994). In addition, it is clear that plants do not distinguish well
between many pairs of ions of similar charge and size (e.g., As and P or Cd and Zn).
Plant bioassays can be used to measure the relative bioavailability of metals in various
soil types. Results can be used to determine either the direct or the indirect value of
bioavailability of contaminants in plants and to extrapolate an indirect estimate of relative
bioavailability to organisms that consume the plants (assuming a correlation between plant and
animal uptake). This type of testing has been routinely done in agriculture for decades and has
been used to validate extraction tests. Tests have most often focused on identifying plant
deficiencies of particular elements, but they are easily adapted to evaluate toxicities (Gettier et
al., 1985).
In the absence of test data, relative bioavailability can be estimated qualitatively based on
relative pH and organic matter content (Tables 4-17 and 4-18), although other soil factors can be
significant, most notably CEC. However, combinations of these two soil parameters and their
ranges are sufficient as a qualitative guide for assessors to identify soils where metals may have
increased (or decreased) availability to plants. This is particularly noted when soil pH exceeds 7
and is a mechanism to obtain essential elements under conditions of low availability.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally, disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-148
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Table 4-17. Qualitative bioavailability of metal cations in natural soils to plants and
soil invertebrates
Soil type
4 * Soil pH <; 5.5
5.5
-------
1 (Brown et al., 2000,1999a; Laperche et al., 1997). The behavior of plant species in response to
2 nutrient deficiencies varies, and this behavior can affect the uptake of metal elements
3 (Marschner, 1998), Because of these multiple confounding factors, the bioavailability of metals
4 in plants (as well as to consumers) is more accurately and reliably measured directly as the
5 edible plant tissue concentrations of the metal in association with soil metal concentrations in the
6 root zone (NAS/NRC, 2002).
7
8 4.5.6.3. Wildlife
9 For most metals, the dietary intake pathway is the main route of exposure for wildlife
10 (NAS/NRC, 2002; see also Section 4.3). However, the incidental ingestion of soil can often
11 contribute a large portion to the majority of the exposure to a wildlife consumer. Because many
12 inorganic metals do not readily accumulate in food (organometalic compounds are excluded
13 from discussion here), highly contaminated soil may result in higher exposures to metals through
14 activities such as grooming fur, preening feathers, consuming soiled prey or forage, burrowing,
15 and taking dust baths. However, canopy feeders would be anticipated to have less incidental soil
16 ingestion and therefore less exposure to inorganic metals than wildlife that consume food that is
17 in more intimate contact with the ground.
18 The relative importance of the dietary and incidental soil ingestion pathways is dictated
19 by (1) the fraction of total metal available in soil versus that in food and (2) the relative
20 bioavailability of the metal in the soil as compared to metal in food items. Figure 4-7 in Section
21 4.2.3. shows the relative contribution of food and soil to total.metal exposure before accounting
22 for bioavailability. Understanding of the bioavailability of metals in incidentally ingested soils
23 becomes necessary when there is a high amount of metal in the soil that is not taken up by soil
24 organisms (plants or invertebrates). However, the same variables that restrict uptake by plants or
25 other soil organisms act to reduce bioavailability to wildlife that ingest soil directly. Therefore, a
26 qualitative estimate of low relative bioavailability could be made for these soils. Furthermore,
27 data that are generated for human health studies could be used to estimate soil bioavailability for
28 wildlife, acknowledging the uncertainty inherent in such interspecies extrapolations.
29 Dietary bioavailability. Very little information is available on dietary bioavailability for
30 most wildlife species (see Menzie-Cura and TN&A, 2000), for a review, cited in NAS/NRC,
31 2002). One of the most significant challenges is that the bioavailability of metals may be
32 influenced by differences in digestive physiology and anatomy across the broad and diverse
33 range of mammalian and avian species. For example, metals present in soils or food may be
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-150
-------
1 more or less bioavailable within the gut of an herbivore that relies on fermentation as compared
2 to the comparatively simpler gut of a carnivore that is designed to break down proteins. These
3 gut systems differ in chemistry (including pH) and residence time. The potential differences
4 between species have not yet been explored rigorously. In general, ruminants absorb lower
5 amounts of metals than do monogastric animals such as rats (NAS/NRC, 1980). Some general
6 guidelines are provided for some metals (e.g., lead in NAS/NRC (1980)), and human-derived
7 values can be used as default values in the absence of species-specific data.
8 Critical body residues (CBRs). CBRs are internal concentrations of chemicals that are
9 correlated with the onset of a toxic response (Conder et al., 2002; Lanno et al, 1998). The use of
10 CBRs reduces uncertainties in ecological risk assessment procedures because they account for
11 site-specific bioavailability and multipathway issues (Van Straalen, 1996; Van Wensem et al.,
12 1994 ). CBRs can be based on whole-body residues (see below for discussion of this approach
13 in soil invertebates) or concentrations in specific tissues. Tissue-specific critical loads have been
14 established for several species of vertebrate wildlife for lead in liver, cadmium in kidney, and
15 selenium in eggs. See Beyer et al., 1996, for these figures. '
16
17 4.5.7. Bioavailability in Aquatic Systems
18 Many factors influence the bioavailability of inorganic metals in aquatic systems.
19 Abiotic (e.g., organic carbon, pH, cations) and biotic (e.g., uptake and metabolism) modifying
20 factors determine the amount of metal that interacts at biological surfaces (e.g., at the gill) and
21 subsequently is taken up. In the dissolved phase, metals can exist as free ions as well as in a
22 variety of complexed forms. These forms, or species, are of key importance in understanding
23 potential impacts because they have differing bioavailabilities, and therefore water quality can
24 dramatically influence the proportion of bioreactive forms of a metal. For many metals in
25 aquatic systems, it is the free ionic form which is believed to be responsible for toxicity. For
26 example, Cu2+ has been directly linked to toxicity in fish and invertebrates, while Cu complexed
27 by dissolved organic matter does not induce toxicity to the same degree (Ma et al., 1999;
28 Erickson et al., 1996) because bioavailability for uptake is reduced.
29 The relationship between speciation and bioavailability is expressed through the FIAM
30 (Campbell, 1995). However, the FIAM is not without limitations, as links between metal
31 speciation and toxicity are complicated and the free metal ion is not always the only bioreactive
32 form. For example, complexed metal, including Cu bound to DOC can be taken up and
33 contribute to toxic impacts and effects (McGeer et al., 2002; Erickson et al., 1996). Although
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-151
-------
1 the link between bioavailability of metals and factors influencing speciation (such as pH,
2 temperature, and organic and inorganic anionic complexation) are of prime importance, other
3 abiotic factors, particularly cations, influence metal bioaccumulation and toxicity. Dissolved
4 cations such as Na+, K+, Ca2+, and Mg2+ can competitively inhibit metal uptake. BLMs, the
5 recently developed integrated toxicity prediction models, have successfully combined abiotic
6 speciation, cationic competition, and bioaccumulation at the presumed site of toxic action
7 (reviewed by Paquin et al., 2002a). The influence of exposure conditions on metal
8 bioavailability is beginning to be understood for some metals, but speciation of metals within the
9 organism is generally less well understood. However, speciation within the organism is likely
10 important for storage, metabolism, elimination, and delivery to and bioavailability at the site of
11 toxic action (Mason and Jenkins, 1995).
12 It is important to account for bioavailaiblity in relation to exposure conditions when
13 comparing toxicity studies and relating these to natural environments. This includes knowing
14 the species of metal that is likely to be present in the environment. A number of other variables
15 also can influence the outcome of a laboratory toxicity assay, including the acclimation of test
16 animals to the culture conditions (see Section 4.5.2, Acclimation, for further discussion), the
17 natural background concentrations of the metals (including, but not limited to, the metal of
18 interest) in either the test water or the site of concern (see Section 4.5.4, Background), potential
19 interactions of the various metals (see Section 4.5.3, Mixtures), and the potential of the metal to
20 transform to a bioreactive species (see Section 4.1, Environmental Chemistry). The relative
21 compositions of the pretest culture medium and the underlying test medium also should be
22 examined because sudden changes in ionic composition have been shown to cause sufficient
23 stress to sensitize organisms and enhance toxicity (Taylor et al., 1990). This is a necessary
24 prerequisite for successful comparison of effect levels derived from laboratory tests to predicted
25 exposures.
26 The bioavailability of metals in aquatic systems can vary significantly among taxonomic
27 groups. For example, although most mayflies (Ephemeroptera) are generally sensitive to metals,
28 caddisflies (Trichopterd) and many stoneflies (Plecoptera) are relatively tolerant (Clements et
29 al., 1992). In fact, these species-specific differences in sensitivity to contaminants have
30 motivated the development of numerous indices of water quality that are based on composition
31 of benthic communities (Barbour et al., 1992; U. S. EPA 1989a; Hilsenhoif 1987). Because
32 responses of aquatic organisms to chemical disturbances are usually contaminant specific
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-152
-------
(Slooff, 1983), field assessments of metal effects may be difficult when more than one
contaminant is present, and caution is required when using biotic indices to assess effects of
3 complex effluents. In some situations, it may be necessary to either develop.chemical-specific
4 indices on the basis of sensitivity to specific classes of contaminants or calibrate metrics used in
5 individual field assessments. For example, metals discharged into a system dominated by highly
6 sensitive taxa (e.g., certain heptageniid mayflies) will have greater effects than the same effluent
7 discharged into a system dominated by tolerant organisms (e.g., certain chironomids). Kiffhey
8 and Clements (1996) showed that benthic communities from headwater streams were more
9 sensitive to metals than were communities from mid-elevation streams.
10
11 4.5.7.1. Application
12 A variety of methodologies (e.g., hardness adjustments, FIAM, Water Effect Ratio
13 (WER), and aquatic BLMs ) can be used to account for differences in concentrations of
14 bioavailable metal species when assessing the effects of metals in aquatic systems (U.S. EPA,
15 1999b). These bioavailability considerations are important in understanding the exposure
16 conditions of interest and how these relate to toxicity study results. WER determinations can
17 account for site-specific (i.e., dependent on the water quality at a specific location)
118 bioavailability (U.S. EPA, 1994c), although they require the development of animal test data.
19 Hardness adjustments were among the first computational methods to account for bioavailability
20 when applying WQC. FIAM approaches that produce speciation profiles of a metal in an aquatic
21 system provide insight into the relative bioavailabilities of the different forms of metal and the
,i
22 importance of complexation. Several models are available for the calculation of metal speciation
23 in natural waters, including MINEQL (Schecher and McAvoy, 1994; Westall et al., 1976),
24 MINTEQA2 (Brown and Allison, 1987), CHESS (Santore and Driscoll, 1995), WHAM
25 (Tipping, 1994) and PHREEQ (Parkhurst, et al., 1980). The BLM approach further extends
26 bioavialability considerations because it applies the latest information on the chemistry and •
27 physiological effects of metals in aquatic environments (Di Toro et al., 2001; Santore et al.,
28 2001; Paquin et al., 1999).
29 Quantitative Ion Character Activity Relationships (QICARs) can be used to extrapolate
30 from availability and/or toxic data for a metal for which there are data to a metal for which data
31 are limited. The approach is similar to the Quantitative Structure-Activity Relationships
32 (QSARs) mat have been used for organic compounds. The QICAR approach was recently
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
•Agency determination or policy.
4-153
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
integrated into the general context of modern QSAR practices by McKinney et al. (2000), and
Weltje (2002) successfully applied it to the lanthanides. The QICAR approach is best applied
with a full understanding of the system under study, and the approaches should be validated
before application. This approach is but one example of the types of tools that are available but
that still need further development to support metals risks assessments for aquatic systems.
In the BLM, chemical speciation is simulated as an equilibrium system that includes
complexation of inorganic ions and NOM using the CHemical Equilibria in Soils and Solutions
(CHESS) model (Santore and Driscoll, 1995). It includes a description of metal interactions
with NOM based on the WHAM (Tipping, 1994). The simulation of biological interactions is
based on the binding affinity characteristics measured in gill-loading experiments, originally
described by Playle et al. (1993a, b). The biotic ligand
(i.e. the gill) is represented as having a characteristic
binding site density and conditional stability constants
for each of the dissolved chemical species with which it
reacts. Predictions of metal toxicity are made by
assuming that the dissolved metal LC50, which varies
with water chemistry, is always associated with a fixed
critical level of metal accumulation at the biotic ligand.
This fixed level of accumulation at 50% mortality is
referred to as the LA50. The LA50 is assumed to be
constant, regardless of the chemical characteristics of
the water (Meyer et al., 1999,2002).
The HAM and BLM approaches both
consider:
• Chemical speciation of the metal and the
relative amounts of complexation to
inorganic and organic ligands.
• The relative bioavailability of metal
species (for example but not limited to
the free metal ion) and interactions with
other cations that may compete for
binding sites of uptake or toxicity.
• Accumulation of metal by the organisms
at the site of toxic action.
The BLM approach has been applied successfully to combine the influences of speciation
and cationic competition on metal toxicity. (Di Toro et al., 2001; Santore et al., 2001; McGeer et
al., 2000; de Schamphelaere and Janssen, 2002, 2004; de Schamphelaere et al., 2002,2003,
2004; Heijerick et al., 2002a, 2002b, also see example application for Cu below)., The model can
distinguish, at least conceptually, metal that will bioaccumulate and cause toxicity from the total
metal pools in an organism and the bioavailable metal pool in the exposure media Themodeling
approach has been extended to species such as algae (Heijerick et al., 2002; de Schamphelaere et
al., 2003) and Daphnia (de Schamphelaere and Janssen, 2002,2004; de Schamphelaere et al.,
2002, 2004), which are lexicologically relevant but more difficult to characterize in terms of
accumulation at me site of toxic action.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-154
-------
1 While the focus of currently existing BLMs is acute toxicity, the Ion Balance Model
2 (Paquin et al., 2002a, b,c) can be used to extend the applicability of the BLM to evaluate chronic
3 toxicity The Ion Balance model uses the BLM to predict accumulation levels at the biotic ligand
4 and then explicitly represent the degree of the physiological response of the organism to metal
5 exposure over time (i.e., disruption of ionoregulation and gradual loss of plasma sodium).
6 Although the approach, which was initially applied to silver, may ultimately provide a way to
7 predict effects due to metals over varying exposure durations, further development and testing is
8 required. Chronic toxicity models have recently been developed for the impacts of Cu on
9 Daphnia magna (de Schamphelaere and Janssen, 2004) and of Zn on fish, Daphnia and algae in
10 European Union risk assessments on zinc.
11 Overall, the BLM approach has wide application in terms of understanding
12 bioavailability in relation to toxicity. It incorporates the influence of speciation in the exposure
'13 - medium, bioaccumulation and toxic impacts in a robust approach, and has been applied in a
14 variety of contexts. For example, the BLM has recently been incorporated into draft revisions to
15 EPA's water quality criteria for Cu, it has been used in risk assessment and it is being applied as
16 an alternative to the Water Effect Ratio approaches for setting site specific discharge objectives.
17 When considering the application of this approach, as with all models, care should be taken to
18 understand and explicitly account for the assumptions and potential sources of uncertainty.
19
20 4.5.7.2. Limitations
21 Novel applications such as the FIAM and BLM have
22 been shown to offer dramatic improvements over traditional
23 approaches such as the hardness equations which depend on
24 the empirical relationship between water hardness and
25 toxicity and the correlation of water quality variables with
26 hardness. While these new approaches offer improvements,
27 there are still many unknowns and uncertainties in relation to bioavailability and the biotic and
28 abiotic influences on metal toxicity. For example, the role of natural dissolved organic matter in
29 bioavailablity and moderating toxicity is not well understood. There are considerable research
30 efforts ongoing currently, and it is likely that our understanding of metal bioavailability and the
31 method for integrating this knowledge into prediction models will improve quickly in the coming
32 years.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-155
EPA has developed a draft criteria
document for copper that incorporates
the BLM and will undergo peer
review. The draft criterion document
is available on the EPA/OST Web
page at www.epa.gov/waterscience.
-------
1 The BLMs that have been, and are being developed for a subset of metals (e.g. Cu, Ni,
2 Cd, Ag, Pb, Zn) are based on a limited number of species. BLMs are generally focused on
3 predicting acute toxicity, although there are a few examples of BLMs that predict chronic
4 toxicity, as discussed previously. The development of BLMs to predict chronic toxicity will
5 likely require new knowledge about the combined effects of dietary and waterborne exposure,
6 the mode of toxicity., and the accumulation of metal or dose delivered to the biological site of
7 toxic action. Similarly, acclimation responses, which can influence bioaccumulation at the site
8 of toxicity, are not well understood, and plant and animal species can differ considerably with
9 regard to the forms of metal taken up as well as their relative toxicities. Although these and
10 other issues add complexity to the evaluation, the BLM approach considers metal
11 biogeochemistry; and therefore represents a viable avenue toward understanding and predicting
12 the toxicity of metals.
13
14 4.5.7.3. Example: BLM Application to Development of Copper Aquatic Life Criteria
15 The BLM's ability to incorporate metal speciation reactions and organism interactions
16 allows for the prediction of metal toxicity to a variety of organisms over a wide range of water
17 quality conditions. Accordingly, the BLM is an attractive tool for deriving WQC in EPA's water
18 program. Application of the BLM may eliminate the need for site-specific criteria modifications,
•
19 such as water effect ratios, which are currently used to account for site-specific chemistry
20 influences on metal toxicity. EPA currently is using the BLM to develop a freshwater aquatic
21 life criteria criterion maximum concentration (CMC) for copper. The BLM accounts for
22 inorganic and organic ligand interactions of copper and also considers competitive interactions
23 that influence binding of copper at the site of toxicity. Although a new model is being used, the
24 criterion derivation is still based on the principles set forth in the 1985 guidelines (Stephen et al.
25 1985). To develop a BLM-based criterion, model predictions of critical accumulations on the
26 biotic ligand (LA50 values) and either LC50 or EC50 values are needed to calculate species
27 mean acute values (SMAVs) and genus mean acute values (GMAVs) as well as to derive a
28 species sensitivity distribution.
29
30 4.5.7.3.1. Model Input Parameters. Much of the aquatic toxicity literature reviewed for the
31 derivation of the copper criterion neither measured nor reported many of the key BLM input
32 parameters. In these cases, the input parameters were estimated. A detailed description of the
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-156
-------
1 methods used to obtain or estimate these input parameters is included in U.S. EPA (2003a).
2 Briefly, when critical water chemistry parameters were not available, a variety of strategies were
3 employed to find the additional or surrogate data (e.g., authors or lab personnel were contacted,
4 and alternative sources such as studies with similar water quality U.S. Geological Survey's
5 National Stream Quality Accounting Network (NASQAN; http://water.usgs.gov/nasqan/) and the
6 EPA STOrage and RETrieval (STORET; http://www.epa.gov/STORET/) were used). Where
7 sources could not be used to develop geochemical input parameters for the BLM, data were
8 generated using the reported water hardness and regression relationships constructed from
9 NASQAN data.
10
11 4.5.7.3.2. Quality Ranking of Water Chemistry Input Parameters. A ranking system of 1 to 6
12 was devised to evaluate the quality of the chemical characterization of the test water (but not the
13 overall quality of the study). Studies that included all of the necessary BLM input parameters,
14 based on measurements from either the test chambers or the source water, were assigned a
15 ranking of 1. Rankings of 2 to 4 were assigned to studies that did not measure all parameters but
16 provided reliable estimates of ion concentrations. Studies were assigned a ranking of 5 to 6
17 when one of the key parameters (DOC, Ca, pH, or alkalinity) was not measured and could not be
18 reliably estimated or if two or more key parameters (DOC, Ca, pH, or alkalinity) were not
19 measured. Only those studies with a rank of 1 to 4 were used to derive the criterion.
20 As with any modeling effort, the reliability of model output depends on the reliability of
21 the input. Although the input data have been extensively scrutinized and filtered, the reliability
22 of the BLM-derived values developed for copper in this project are subject to the limitations of
23 the input measurements/estimation procedures described above.
24
25 4.5.7.3.3. Criteria Generation. To calculate an acute criterion or CMC, reported acute toxicity
26 values fe.g., LC50s) and individual test water chemistry parameters were used to calculate LA50
27 values by running the model in the speciation mode. These LA50 values were then
28 "normalized" to a standard water condition by running the model in the toxicity mode and
29 specifying user-defined LA50s. These normalized LC50s were used to calculate SMAVs,
30 GMAVs, and a final acute value (FAV) pursuant to the 1985 guidelines procedure. The FAV
31 represents a hypothetical genus more sensitive than 95% of the tested genera. The FAV was
32 derived from the four GMAVs that have cumulative probabilities closest to the 5th percentile
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-157
-------
1 toxicity value for all the tested genera. Inputting this FAV as an LC50 concentration and
2 running the model in speciation mode determines the lethal accumulation associated with the
3 FAV in the standard test water.
4 the Cu BLM assumes that most of the parameters are invariant for different organisms.
5 Despite the complexity of the modeling framework, the thermodynamic constants used to
6 simulate the inorganic and organic chemical equilibrium reactions are determined by the
7 characteristics of the metal and available ligands. As such, the constants do not change for
8 simulations involving different organisms. Although most BLM parameter values (including the
9 biotic ligand binding constants and site densities) are consistent across organism types,
10 differences in sensitivity across organisms types should still be accounted for. This is
11 accomplished by adjusting the critical biotic ligand concentration (e.g., LA50) values for each
12 species.
13 This criterion LA50 is programmed into the model as a constant. To derive a criterion
14 for a specific site, the site water chemistry data are input to the model. The model then uses an
15 iterative approach to determine the dissolved copper concentration needed to achieve a Cu-biotic
16 ligand concentration equal to the criterion LA50. This dissolved Cu concentration is, in effect,
17 the FAV based on site water chemistry. The site-specific CMC is this predicted dissolved metal
18 concentration (or criterion FAV) divided by two. The site-specific CCC is the CMC divided by
19 the final acute-chronic ratio.
20
21 4 J.7.3.4. Next Steps. EPA has developed a draft criteria document that will undergo peer
22 review. When EPA solicits scientific views from the public on the draft criterion document, the
23 model will be made available on the Office of Water web page at www.epa.gov/waterscience.
24 Until the peer review is completed and a final copper criteria document is published by EPA, the
25 procedures described here are draft and subject to change, and the criteria are not considered to
26 be available for use. After completing the copper criteria update, EPA will consider
27 incorporating the BLM into derivation procedures for other metal criteria, such as for silver,
28 cadmium, and zinc. Although the BLM is currently appropriate for use in deriving an updated
29 freshwater copper CMC, further development is required before using it to evaluate a saltwater
30 copper CMC, a CCC, or a chronic value.
31 Alternative approaches may be considered to establish the database of input parameters
32 for the acute toxicity studies, because the approach described in this document is labor intensive.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by Ihe EPA and should not be construed to represent any
Agency determination or policy.
4-158
-------
1 Alternatives may include either developing a few high-quality data sets that satisfy the minimum
2 data requirements of the guidelines for a limited set of organisms or developing data sets for
3 known sensitive species. Estimating missing input parameters by relying on statistical
4 techniques or Monte Carlo approaches may also be explored.
5
6 4.5.8. Bioaccumulation and Bioconcentration in Aquatic Organisms
7 Bioaccumulation can be defined as the net accumulation of a metal in a tissue of interest
8 or a whole organism that results from exposure from all relevant sources (e.g. water, food, and
9 sediment). Metal bioaccumulation can apply to the entire organism, including both metal
10 adsorbed to surfaces or absorbed by the organism or to specific tissue. It is usually expressed on
11 a weight (dry or wet)- adjusted basis. Bioaccumulation that occurs under steady-state conditions
12 (i.e., where accumulation remains relatively constant due to uptake being offset by elimination )
13 is often of primary concern in risk assessments. Bioaccumulation of metals is a concern with
14 respect to the accumulating organism when it occurs in the toxicological form(s) and
15 concentrations at the site(s) of toxic action. It also can be indicative of dietary exposure to
16 aquatic organisms at higher trophic levels. While inorganic metals can bioaccumulate in animal
17 and plant tissues, biomagnification across three or more trophic levels is rare.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or poEcy.
4-159
-------
1 The simplest tools for estimating
2 bioaccumulation include direct measures of
3 tissue levels and the derivation of simple
4 relationships between tissue levels and
5 environmental concentrations. These
6 simple empirical relationships have often
7 been expressed as bioaccumulation factors
8 (BAF) or bioconcentration factors (BCF)
9 and various data bases summarize values
10 reported in the literature. It is well
11 recognized, however, that there can be
12 considerable uncertainty associated with
13 the application of literature-derived,
14 bioaccumulation or bioconcentration
15 factors to specific risk assessment
16 situations. In other cases, efforts have been
17 made to establish more detailed
18 mathematical relationships between
19 exposure concentrations and tissue levels.
20 For example, empirical approaches that
21 extend beyond simple factors include the
22 use of regression equations that describe
23 the relationship between exposure and
24 accumulation. Significant advances are
25 also being made on the application of
26 kinetic or steady-state uptake models for
27 describing and predicting bioaccumulation
28 (Reinfelder et al, 1998; Chang and
29 Reinfelder, 2002; Kahle and Zauke, 2003;
30 Wang and Zauke, 2004; ). These offer
31 promise for understanding the contribution of differing routes of exposure (e.g., water vs. diet).
32 There is considerable experience in using mechanistic bioaccumulation models for estimating
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-160
Biomagnification is defined as an increase in
the concentration in an organism from a lower
trophic level to a higher trophic level within the
same food web due to bioaccumulation from the
diet (also see trophic transfer). Biomagnification is
expressed as the ratio of the concentration in the
organisms of the higher trophic level to the
concentration in the organisms of the lower trophic
level (i.e., biomagnification factor). Note,
however, that biomagnification that is quantified
using field data makes an explicit assumption that
bioaccumulation results from the diet only, when
actually multiple sources may be involved (e.g.,
water or sediment). Biomagnification factors
greater than one indicate biomagnification, while
factors less than one indicate no biomagnification
or biodilution (see below). Inorganic forms of
metals rarely biomagnify across three or more
trophic levels.
Biodilution, a decrease in organism
concentration with increasing trophic level has
generally been more commonly observed when
concentrations are evaluated across three or more
trophic levels.
Trophic transfer. The transfer of a
bioaccumulated substance in a prey species to a
predator species via dietary exposure. When the
concentration in the predator species is increased
relative to the prey, trophic transfer is a form of
biomagnification. Trophic transfer is important
due to its relationship to dietary toxicity (also
called secondary poisoning) in which toxicity is
manifest through accumulation in prey species and
subsequent dietary exposure to predatory species.
Importantly, however, dietary toxicity cannot be
directly inferred from an evaluation of trophic
transfer alone, but rather the combination of trophic
transfer and toxicity information for predator/prey
species within an ecologically-Hnked food web.
-------
1 risks of organic chemicals in aquatic systems and it is anticipated that such models will have an
2 increasing role for risk assessments of metals.
3 Because the simpler empirical approaches such as the BCF and BAF have received much
4 use and attention in the past, they are discussed in further below with respect to applications and
5 limitations. However, the kinetic and steady-state uptake and bioaccumulation models will
6 likely emerge as important approaches.
7
8 4.5.8.1. Scientific Issues mthBCF/BAF
9 The BCF is the ratio of contaminant concentration in an organism to its concentration in
10 water at steady state conditions and water-only exposures. Metal concentrations are usually
11 expressed on a weight-adjusted, whole-organism basis, and waterbome metals may be expressed
12 as total or dissolved metals. BCFs have been developed primarily with hydrophobic organic
13 chemicals in aquatic systems, but similar such accumulation factors have been applied to other
14 matrices (e.g., sediment and soils) for both organic chemicals and metals. Strictly speaking,
15 metal bioconcentration in sediment and soil systems is the net accumulation of a metal in or on
16 an organism from pore water only. Hence, in sediment and soil, the denominator for the BCF
17 ratio should comprise the pore water concentration of metal, not the total metal concentration in
18 the sediment or soil. In the broadest context, the BAF is the ratio of a contaminant concentration
19 in an organism to that in a specified medium at steady state in situations where the organism is
20 exposed to multiple environmental media.
21 Although BAFs and BCFs are calculated in a similar manner, the interpretation is slightly
22 different with metal accumulation in organisms arising from water only for BCFs and from both
23 water and dietary sources for BAFs. For aquatic organisms, BAFs are derived from
24 measurements in natural environments, and BCFs are nearly always measured under laboratory
25 conditions where exposure can be effectively limited to the water column. Unless metal
26 concentrations in pore water serve as the denominator for the ratio, soil and sediment BAFs are
27 usually termed BSAFs. Concentrations are usually measured on a total-metal and weight-
28 adjusted whole-organism (or tissue) basis.
29 Toxicological bioaccumulation is the fraction of the metal that bioaccumulates and is
30 distributed to receptors at sites of toxic action (Figure 2-2). For metals, this would include
31 reactions with target proteins or other receptors that result in toxicity but not interactions with
32 metallothionein and other metal-binding ligands or incorporation into granules that make metals
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-161
-------
1 unavailable for interactions with target molecules. This fraction is conceptual in nature and
2 difficult to measure in practice; it is akin to the minimal effective dose measured in blood that is
3 often used in medicine for assessing therapeutic effects. It could be conceptually defined as a
4 toxicological bioaccumulation fraction or the ratio of total metal concentration in an organism to
5 the metal concentration at the site(s) of toxic action.
6 As discussed in more detail in McGeer et a! 2004, there are fundamental differences in
7 the physical, chemical and toxicological properties of inorganic metal substances and organic
8 substances, such that the BAF/BCF model would not apply to the former. The success of the
9 BAF/BCF model as a valid indicator of the environmental and toxicological behavior of neutral
10 organic substances is due to their hydrophobic/lipophilic chemical properties, and this has
11 important consequences for application to inorganic metals. Many of the assumptions and
12 characteristics of the BAF/BCF model openly conflict with the physical, chemical, biological,
13 and toxicological realities associated with inorganic metal substances. For example, the
14 BAF/BCF model works well for neutral organic substances because uptake of lipophilic
15 substances into biota occurs via simple passive diffusion. However, uptake of the vast majority
16 of inorganic metals is a physiological process, which occurs via a number of specific routes,
17 most of which involve saturable transport kinetics. The degree of uptake and ultimate internal
18 fate of inorganic metals is strongly influenced by ligand binding and receptor site competitive
19 interactions which control metal availability and/or transfer processes, and cannot be reliably
20 represented by a simple partitioning between water and organism. As a result of the differences
21 explained below, simple application of these models is meaningless in assessing acute and
22 chronic toxicity (adapted from McGeer et al., 2004):
23
24 • The principal theoretical features of the BAF/BCF model that make it applicable to
25 . neutral organic substances also make it inapplicable to inorganic metal substances.
26 These factors produce an inverse relationship between BAF/BCF and exposure
27 concentration; this has been observed for both essential metals and nonessential
28 metals.
29
30 • The approach of using one simplified bioaccumulation model (BCF and BAF) and
31 applying it to inorganic metals ignores the basic physical and chemical differences
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-162
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
between organic and inorganic substances and is not supported by theoretical and
empirical weight of evidence.
• Based on the inherent assumptions of the BCF and BAF model and on the
environmental and toxicological behavior of the organic substances from which they
were developed and validated, for the vast majority of inorganic metals evaluated, the
scientific basis for broad application of the BAF/BCF model is lacking in the context
of hazard assessment.
• The complexity of assessing and predicting the bioavailability and bioaccumulation
of metals in aquatic systems arises from many factors including:
- Essentiality of some metals resulting in a "U"-shaped dose response curve.
- Variation in assimilation efficiency for different species of metal, for different
biota and at different sites of uptake.
- Ability to modulate uptake at the various sites of uptake.
Contributions of different routes of entry to the metal body burden and effects.
Ability to sequester, store, detoxify and eliminate bioaccumulated metals.
- All metals will bioaccumulate to some degree without impacts as a result of
exposure to natural background concentrations
• Based on reviews (e.g., McGeeretal.,2003), it would appear that for the vast
majority of the metal/taxonomic group combinations assessed, the assumptions
regarding the independence of BCF/BAF with exposure concentration and
' proportionality of hazard with increasing BCF/BAF do not hold true.
* The latest scientific data on bioaccumulation do not currently support the use of BCF
and BAF data when applied as generic threshold criteria for the hazard potential of
inorganic metals (e.g., for classification as a "PBT" chemical).
• In cases where the use of BCF/BAF data are being considered, a careful evaluation of
existing data for metals that formally documents the extent to which values represent
bioaccumulation linked to toxic impact (both direct and indirect) and the resulting
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-163
-------
1 uncertainty introduced by concentration dependency and other aspects of the
2 BCF/BAF data should be done.
3
4 • Moving from the current situation to a revised, improved, and validated set of criteria
5 and/or methodologies for assessing metal hazards and risks presents a number of
6 options and challenges, as currently there would appear to be no clear alternatives
7 ready for application.
8
9 • BCFs for metals will not be absolute values, but will vary widely with both the
10 specific exposure circumstances/conditions as well as the status/age/condition of the
11 particular organism measured. This applies to all substances, organic and inorganic,
12 although the relative s cale of the uncertainty added is greater for inorganic
13 substances.
14 The Organization for Economic Co-operation and Development (OECD) has published
15 guidance for the hazard classification of chemical substances (OECD, 2001). The hazard
16 classification schemes presented in the OECD guidance incorporates, among other parameters,
17 evidence of bioaccumulation as a basis for hazard ranking. The OECD guidance recognizes the
18 shortcomings associated with the use of B AF/BCF data as a surrogate for the hazard potential of
19 metals, and therefore cautions that metals should be assessed on a case-by-case basis rather than
20 recommending the application of the simple bioaccumulation model (BAF/BCF). The principle
21 features of the BAF/BCF model that make it applicable as a surrogate for acute and chronic
22 toxicity to neutral organic substances also make it inapplicable to inorganic metal substances as
23 discussed above and in McGeer et al. (2004).
24 The phenomena described above regarding the limitations of the BCF/BAF model also
25 have a bearing on the use of BCFs and BAFs in national-level assessments (e.g., water quality
26 criteria and national-scale risk assessments) in addition to site-specific assessments. For
27 example, variability in BCFs and BAFs caused by an inverse relationship of BCF with exposure
28 concentration can confound the application of BCFs in risk assessments, the extent to which
29 depends in part on the magnitude of this relationship (i.e., its slope) and the extent of
30 extrapolations made (i.e., extrapolation of BCFs across differing exposure concentrations and
31 species). For organic chemicals, generic normalizing factors have been shown to reduce
32 variance in BAFs due to bioavailability differences (e.g., normalizing to organic carbon and
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-164
-------
1 lipid; see Burkhard et al, 2003). However, analogous normalizing factors have yet to be
2 developed and widely applied in the context of BAFs or BCFs for metals, although some
3 progress has been made with relating accumulation parameters to body weight (Hendriks and
4 Heikens, 2001) or permeable surface area in the case of the amphipod, Gammarus zaddachi
5 (Wang and Zauke, 2004). Simple normalizing factors can be confounded by the fact that
6 animals and plants have evolved physiological and anatomical means for regulating and or
7 storing internal concentrations of metals and the species-dependent specificity of these
8 mechanisms. Regulation of metals by biota can actually result in decreased apparent BAF/BCF
9 factors with increasing exposure concentrations. Thus, a fundamental understanding of such
10 mechanisms is important for reducing the uncertainties associated with interpreting tissue
11 residues. The mere presence of a metal in a plant or animal does not mean it is at a site of toxic
12 action. In summary, caution should be exercised when using published BAF and BCF values to
13 make judgements about resultant tissue levels of metals in animals and plants. Where this is
14 found to be necessary as an initial, screening level of assessment, the associated uncertainties
15 should be identified and discussed with respect to their potential impact on the risk estimates and
16 resultant decisions.
17 In situations where a decision has been made to employ a BAF or BCF factor, there may
18 ways to reduce some of the uncertainty. One idea involved subtracting "normal" accumulation
19 from the calculation of the BCF. This involves separating the portion of metal that
20 bioaccumulates from exposure under "normal" or background conditions from that portion that
21 occurs as a result of exposure to elevated levels of metals (McGeer et al., 2003). The ACF value
22 also accounts for the accumulation of essential metals required for physiological function. These
23 ACF values were dramatically lower than BCF values (illustrating the importance of normal
24 bioaccumulation) for some metals, particularly essential metals. However, for most metals
25 ACFs still varied with exposure concentration and were not viewed as a sufficient replacement to ,
26 the BCF (McGeer et al., 2003).
27 Another approach that may reduce uncertainty caused by concentration dependency of
28 . BCFs and BAFs involves limiting the selection and application of BCF and BAF values to
29 concentration ranges that more closely correspond to toxicological thresholds. For example,
30 limiting the calculation of BAF and BCF values to concentrations that approximate the
31 applicable water quality criterion has also been suggested as a method for reducing the
32 uncertainty around BCF and BAF values in situations where concentration dependency is
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
. 4-165
-------
1 evident. This would account for bioaccumulation at an exposure level where concern over
2 bioaccumulation might be expected. This approach has limitations (McGeer et al., 2004) and
3 does not appear to reduce the overall variability associated with BCF and BAF measurements
. 4 when evaluated across broad classes of organisms. An additional issue for mis approach is that
5 WQC reflect some of the more sensitive organisms, whereas the BCF and BAF measurements
6 are not necessarily from these same organisms and include data from biota that may not be as
7 sensitive to chronic effects. Therefore, as a modifier for broad-based application, mis variation
8 of the BCF/BAF methodology does not appear to explain variability. However, on a site-
9 specific basis where toxicity thresholds and species are better characterized, this approach may
10 have value in reducing uncertainty. As already noted, caution should be exercised with any of
11 the BAF or BCF applications.
12
13 4.5.8.2. Bioaccumulation in Relation to Dietary Toxicity
14 Discriminating between metals that have the potential to cause effects via trophic transfer
15 and metals that do not is another approach that might be useful in distinguishing between metals.,
16 based on bioaccumulation and impacts. Metals taken up and stored within an organism may not
17 cause direct effect to that organism but may be bioavailable to organisms in the next trophic
18 level that feed on it. Bioaccumulation of metals in prey organisms may be quite high, especially
19 in organisms such as high volume filter feeders and those that accumulate elevated levels of
20 metals, for example, those that store detoxified forms. Metals that bioaccumulate to levels in
21 prey organisms that cause impacts in predatory organisms are clearly important issues to address
22 however, there is a general lack of understanding of the potential for toxicity.
23 A prey organism with a high concentration of a particular metal represents a potential
24 opportunity for the trophic transfer of the metal from an enriched source to a predator at the next
25 trophic level. The form of detoxified storage of that accumulated metal in the prey species has a
26 significant effect on the potential assimilation of that metal by the predator (Wang and Fisher, •
27 1999). For example, Nott and Nicolaidou (1990) have shown mat the bioavailability to
28 neogastropod mollusc predators of metals present in detoxified metalliferous granules in prey
29 varies among metals and with type of granule; thus the zinc-rich pyrophosphate granules
30 accumulated in barnacles are not digested in the digestive tract of the predator Nucella lapillus
31 and are therefore not bioavailable to that predator. Similarly the physico-chemical form of
32 accumulated cadmium in the oligochaete Limnodrilus hoffmeisteri is critical in the assimilation
11/24/2004 Peer Review D raft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-166
-------
1 of cadmium by a predator, in this case the decapod Palaemonetes pugio (Wallace et al., 1998;
2 Wallace and Lopez, 1997). Other studies have indicated that lead bioaccumulated in mussels
3 and stored in a detoxified form within the soft tissues of the mussels is bioavailable and may
4 cause impacts (Regoli and Orlando, 1994). The potential effects of food preparation (cooking)
5 on metal bioavailability should be considered when assessing risks to human consumers.
6 Currently there are no standardized tools for incorporating the potential impact of dietary
7 . exposure into assessments. Research efforts are focusing on the relative importance of: dietary
8 pathways in relation to waterbome exposures, transfer from sediments into food chains, the
9 potential of bioaccumulated to cause impacts in consumer organisms. One example of an
10 approach that illustrates how exposure, bioaccumulation and trophic transfer can be linked is
11 illustrated below.
12 A general approach to account for potential impacts arising from trophic transfer would
13 be to link bioaccumulation in prey items to exposure in the water column as well as potential
14 impacts in consumers. One methodology to achieve this is to integrate tissue burden to toxicity
15 relationships (prey: predator interactions) with exposure to bioaccumulation relationships (e.g.,
16 BCF and BAF values). A theoretical example of this approach, developed by Brix et al.
17 (www. epa.gov/ncea/rafypdfs/metals/sumryrprt_metals.pdf) is shown in Figure 4-9.
18 Within the context of this approach, it would be necessary for relevance to ensure that the
19 species being considered were linked within trophic food webs (i.e., the exposure to
20 bioaccumulation relationship was for prey items being consumed by the predators that are
21 sensitive to trophic transfer). Some data already exists to begin these evaluations. For example,
22 it is possible to derive the water concentrations necessary to produce impacts via dietary
23 exposure (see Brix et al. reference above) and site-specific case studies would be valuable in
24 illustrating, testing, and validating these relationships. The potential benefits from this approach
25 would be reducing the uncertainty from extrapolations across exposure concentrations currently
26 being made with metals BCF/BAF data. Also, there would be an ability to link impacts through
27 to waterborne metal concentrations. However, particularly if BCF and BAF data is used, this
28 approach would bring with it the inherent uncertainty associated with predicting tissue metals
29 burdens (e.g., the high variability) and the inability to account for geochemical influences on
30 uptake and accumulation. Ensuring that the local site specific factors that influence the
31 bioaccumulation and the expression of impacts are understood and recognized would be
32 important, to ensure for example that criteria and guideline values were not set at existing
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-167
-------
1
i
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
background levels. Two regression relationships, one for exposure and bioaccumulation (green
line) and the other for bioaccumulation to dietary toxicity thresholds (blue line) are used to line
exposure and dietary impacts.
Wildlife dietary
threshold
(mg/kg dw)
100
50
10 -
5-
1-
1-
5
10
50
100
waterborne
[metal] (jig/L) 50o
-1000
dietary (prey)
[metal] (mg/kg dw )
—I 1
10
50 100
derived from the
inverse BCF relationship
(bivalve or fish)
Figure 4-9. Linkages between dietary toxicity threshold, bioaccumulation in prey
organisms and waterborne exposure.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-168
-------
1 4.5.8.3. Alternatives to Tissue Burdens and Bioaccumulation
2 A key parameter that a bioaccumulation measure should be validated against is chronic
3 toxicity. Because bioaccumulation criteria within the context of persistent, bioaccumulative,
4 toxic substances are used as indicators of chronic toxicity (Franke et al., 1994; OECD, 2001),
5 validation of linkages to chronic metal toxicity would provide confidence in their use and
6 application. A number of key issues should be addressed when considering bioaccumulation of
7 metals in relation to the potential for chronic impacts, and these add uncertainty to the
8 interpretation of data. However, unlike the substances that the PBT concept was originally
9 developed for, there is often substantial information on the chronic toxicity of metals.
10 In some regards, our ability to understand and interpret chronic metal toxicity is as advanced,
11 or possibly more advanced, than metal bioaccumulation. Therefore, rather than trying to derive
12 and validate a surrogate for the chronic impacts of metals, it might, in some cases, be feasible to
13 eliminate bioaccumulation and only consider chronic toxicity data. The development of a
14 criterion based on chronic toxicity could be based on a variety of approaches,«all of which will
15 require a modified framework for consideration as BCFs, BAFs, and bioaccumulation would be
16 replaced. Despite the challenges associated with this degree of change, it is worthy of
17 consideration.
18
19 4.5.9. Bioaccumulation in Terrestrial Organisms
20 For terrestrial ecosystems, the concept of bioaccumulation is intended to capture the potential
21 for two ecologically important outcomes: (1) direct toxicity to plants and wildlife and (2)
22 secondary toxicity to animals feeding on contaminated plants and animals. This approach
23 captures the potential for trophic transfer of metals through the food web, so total exposure can
24 be calculated, including dietary intake as well as intake from contaminated environmental media
25 (soil and water). For vegetation, the BAF (or BSAF) is defined as field measurements of metal
26 concentration in plant tissues divided by metal concentration in soil (or soil solution); the BCF is
27 defined as the same measurement carried out in the laboratory (Smolders et al., 2003).
28 Data applicability is directly related to which tissue is sampled and how it is processed.
29 BAFs for plants include metals aerially deposited on leaves as well as those in soil particles
30 adhering to roots. Such metals will not be part of BCFs, which frequently are determined in
31 hydroponic culture. Similar differences between BCFs and BAFs apply for earthworms exposed
32 in laboratory studies using the filter paper substrate protocols. Furthermore, BCFs with
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-169
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
earthworms may not include additional feeding of the animals during the study.. Additionally,
field studies are reflective of chronic exposures, whereas BCFs may be calculated from shorter
time frames. For birds and mammals, whole-body BAFs generally are not calculated, except for
small mammals such as rodents (Sample et al., 1998b). Rather, concentrations in target tissues
are measured for comparison with critically toxic levels (Beyer et.al., 1996).
The BAFs for metals by soil invertebrates and most plants are typically less than 1, although
they usually are based on the total metal in soil and tissue. BSAFs expressed in this manner may
be suitable for comparisons of metal uptake within the same soil type, but they would be
misleading if soil bioavailability factors (e.g., pH and organic carbon) differ. Unfortunately, the
literature database is populated almost entirely with BAFs derived from measurements of total
metal. Furthermore, bioaccumulation is not a simple linear relationship. Uptake is nonlinear,
increasing at a decreasing rate, as medium concentration increases. Reliance on incorrect
relationships of the bioavailable portion will be trivial compared to the error associated with
BAFs at high concentrations. In the future, a ratio of total metal in the organism to some
measure of the bioavailable fraction of metal in the soil (e.g., free ion concentration or weak salt
extractable) should be used for expressing a BSAF that allows comparison among different soils.
An alternative approach currently under study is to use a multivariate statistical model to look
for patterns of uptake of multiple metals to predict the potential bioconcentration of one metal of
particular interest (Scott-Fordsmand and Odegard, 2002).
4.5.9.1. Models for Bioaccumulation in Soil Inverterbrates
4.5.9.1.1. Application
4.5.9.1.1.1. Univariate models. The bioaccumulation of metals in soil organisms cannot
reasonably be modeled from information based
solely on soil concentrations. Therefore, models for
the prediction of metal bioaccumulation by soil
invertebrates are primarily empirical in nature,
describing relationships between metal body burdens
in oligochaetes and collembola, soil metal
concentrations, and soil physical/chemical
characteristics. Statistical relationships have been
established using univariate and multiple regression
The bioaccumulation of metals in soil
organisms cannot be modeled from information
based solely on soil concentrations. Two types
of models are available to help describe the
relationships between metals body burden and
soil parameters: (1) univariate models that
describe statistical relationships between body
burdens, soil metal concentrations, and soil
physical and chemical properties; and (2)
multivariate models that explain BAFs as a
function of soil properties.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-170
-------
1 approaches. Sample et al. (1998a) and Peijnenburg et al. (1999b) have each developed
2 , univariate uptake models for earthworms that are based on empirical data (metal concentrations
3 in worms vs. the natural log of amount of metal in soils) that are widely used as a first
4 approximation. However, these models are not specific to soil type and therefore do not account
5 for bioavailability factors such as pH, clay content, or cation exchange. Furthermore, they do
6 not adequately predict Cr or Ni uptake.
7
8 4.5.9.1.1.2. Multivariate models. Multivariate models also are available (Peijnenburg et al.,
9 1999a, b) for Eisenia andrei and Enchytraeus crypticus that explain BAF as a function of soil
10 characteristics. The soil parameters that generally contributed the most to explaining the
11 variance between uptake rate constants and BAFs were pH (for Cd, Zn) and also CEC (for Pb)
12 and clay content (for Cd). Similar studies are needed for describing relationships between soil
13 physical/chemical characteristics and metal bioaccumulation in other groups of soil invertebrates
14 such as collembola and isopoda. Until these are available, the models for earthworms and
15 enchytrids can be applied to other groups, although the added uncertainty should be
16 acknowledged.
17 Path analysis has been suggested as an alternative for multiple regression in describing these
18 relationships. It partitions simple correlations into direct and indirect effects, providing a
19 numerical value for each direct and indirect effect and indicates the relative strength of that
20 correlation or causal influence (Basta et al., 1993). Bradham (2002) used path analysis and
21 backwards stepwise regression analysis to derive statistical models capable of predicting uptake
22 and effects of As, Cd, Pb, and Zn in earthworms as a function of soil properties.
23 Saxe et al. (2001) described a model for predicting whole-body concentrations of Cd, Cu, Pb,
24 and Zn in Eisenia andrei as a function of pH, soluble metals in the soil at gut and environmental
25 pH, and soluble organic carbon in soil extracts. The model also includes parameters that
26 characterize the ability of worms to regulate the metal body burden, whether metal uptake is via
27 the epidermal or gut surface and whether the metal is essential. The model has been validated
28 against a series of Dutch soils and is very good at correctly predicting metal accumulation.
29
30 4.5.9.1.2. Limitations, Bioaccumulation of organic substances is typically modeled using a
31 One-Compartment, First-Order Kinetics (1CFOK) model. However, most of the assumptions of
32 the model are violated when applied to bioaccumulation of metals by soil invertebrates. Soil
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the-purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-171
-------
1 invertebrates are not exposed to a constant concentration of metals in the soil over space and
2 time, making it difficult to accurately define exposure. Sufficient data exist on the metabolism
3 of metals to show that all pools of metal taken up by the soil invertebrates are not equally
4 available for depuration (some are actually never depurated and are released only when the.
5 organism dies), making a 1CFOK model an inaccurate approximation.
6 Metal concentrations do not reach a steady state that is proportional to external (i.e., soil)
7 concentrations. This is similar to the situation in aquatic systems where metal bioaccumulation
8 is a function of water concentration. Internal essential metal concentrations are regulated and
9 remain relatively constant over a wide range of soil metal concentrations. Only when normal
10 regulatory mechanisms are overwhelmed do internal levels of essential metals increase.
11 Accumulation of nonessential metals also violates the assumption of steady state, as organisms
12 have evolved mechanisms for the detoxification of nonessential metals that involve the internal
13 accumulation of the metal in forms that are not toxic to the organism (e.g., incorporation into
14 inorganic granules or binding to organic molecules to form metal ligands such as
15 metallothioneins).
16
17 4.5.9.2. Critical Body Residues
18 CBRs are an extension of the concept of
19 bioaccumulation to internal concentrations of
20 metals that are correlated with some toxic response
21 and hence represent lexicological bioavailability
22 (Conder et al., 2002; Lanno et al., 1998). Use of
23 CBRs in appropriate species may reduce
24 uncertainties in ecological risk assessment
25 procedures (Van Straalen, 1996; Van Wensem et
26 al., 1994). However, only a few CBRs have been
27 developed in soil invertebrates for metals.
28 Crommentuijn et al. (1997,1994) and Smit (1997)
29 established CBRs for sublethal effects for Cd and
30 Zn, respectively, in the springtail (Folsomia
31 Candida). Conder et al. (2002) demonstrated that
32 effects of Cd in earthworms (Eiseniafetida) are
Knowing the species of metal that is likely to
be present in the environment is important for
extrapolating results from laboratory tests to
natural settings. Laboratory studies are often
done under simplified environmental conditions
(compared to the natural environment) that may
enhance aquatic bioavailability. Conversely,
these studies frequently do not include potential
effects of dietary metal exposure.
In addition to abiotic factors that can influence
metal bioavaiability, other variables may
potentially affect the outcome of laboratory
toxicity assays and should be considered in
interpretation of data. These include acclimation
of test animals to culture conditions, natural
background concentrations of the metal
(including, but not limited to the metal of
interest), potential interactions of the various
metals, and potential transformations to
bioreactive species. The similarity of the pretest •
culture medium to the underlying testing medium
also should be considered.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-172
-------
1 correlated with concentrations of the metal in the solid phase of the worms (i.e., the pellet
2 fraction, following homogination and centrifugation). If future research can isolate the fraction
3 of an invertebrate that represents lexicological bioavailability, then it may be possible to
4 estimate a lexicological BSAF representing a relationship between a specific fraction of metal
5 that accumulates in the organism and a measure of chemical bioavailability in the soil. Until
6 then, CBRs based on whole-organism analyses are a reasonable approximation for use in
7 ecological risk assessments.
8
9 4.5.10. Sediment Toxicity - Equilibrium Partitioning Approach for Metals
10 4.5.10.1. Rationale for Use of EqP Benchmarks
11 Toxic pollutants in bottom sediments of the nation's lakes, rivers, wetlands, estuaries,
12 and marine coastal waters create potential for continued environmental degradation even where
13 water column concentrations comply with established human health and aquatic life WQC. In
14 addition, contaminated sediments can be a significant pollutant source that may cause water
15 quality degradation to persist even when other pollutant sources are stopped (U.S. EPA
16 1997d,e,f; Larsson, 1985, Salomons et al., 1987; Burgess and Scott, 1992).
17 Because of their widespread use and associated environmental releases, metals such as
18 cadmium, copper, lead, nickel, silver, and zinc are commonly elevated in aquatic sediments.
19 Various types of sediment guidelines have been proposed for assessing the potential effects of
20 these metals on benthic invertebrate communities. Many of these involve empirical correlations
21 of metal concentrations in sediment with associated biological effects (e.g., sediment toxicity);
22 these include ER-M, ER-L, TEL and PEL, SLC, AET and others (Sullivan et al., 1985; Persaud
23 et al., 1989; Long and Morgan, 1990; Ingersoll et al., 1996; MacDonald et al., 1996). Most of
24 these approaches use measures of total metal in sediment, and do not account for differences in
25 bioavailability of metals among sediments. Nevertheless, when examined across large data sets,
26 these empirical guidelines do show relationships between concentrations of total metal and
27 biological effects. However, a limitation to these empirical approaches is that the causal linkage
28 between the measured concentration of metals and the observed toxicity cannot be established, in
29 part because of the procedures used to derive correlative values and because values derived are
30 based on total rather than bioavailable metal concentrations. That is, for any given total metal
31 concentration, adverse toxicological effects may or may not occur, depending on the
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-173
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
physicochemical characteristics of the sediment of concern (Tessier and Campbell, 1987;
Luoma, 1989; Di Toro et al., 1990).
An alternative to empirical approaches is equilibrium partitioning (EqP) theory. The EqP
approach is intended to predict concentrations that will or will not cause adverse effects based on
an understanding of the factors that control metal bioavailability in sediments and the
relationship between that bioavailability and biological effects. For cationic metals (e.g.,
cadmium, copper, lead, nickel, silver, and zinc), these factors include the presence of acid
volatile sulfides (AVS), which form insoluble metal sulfides that are believed to have low
biological availability. Beyond AVS, particulate organic carbon (POC) and iron and manganese
hydroxides in sediment, and dissolved organic matter (DOM) in interstitial water are also
believed to influence metal bioavailability in sediment. By recognizing differences in
bioavailability among sediments, the intent of the EqP approach is to produce guideline values
that are applicable across a wide variety of sediments and represent causal relationships between
specific chemicals and sediment toxicity.
In 1987, EPA reviewed proposed approaches to developing numerical sediment quality
guidelines (Chapman, 1987). All of the approaches reviewed had strengths and weaknesses, and
no single approach was found to be applicable for the derivation of sediment benchmarks in all
situations (U.S. EPA, 1989d). The EqP approach was selected for further development because
it presented the greatest promise for generating defensible national chemical-specific sediment
benchmarks applicable across a broad range of sediment types. While the Agency has never
adopted formal sediment quality criteria, the technical development of sediment quality
guidelines was pursued and has resulted in the publication of several numerical guidelines (See
http://www.epa.gov/nheerl/publications/). These guidelines are called "EqP sediment
benchmarks" (ESBs) and are numerical concentrations for individual chemicals or mixtures of
chemicals derived using the EqP approach. As described below, these are intended to be
concentrations below which there should not be direct toxicity to most benthic organisms as a
result of the chemical or chemicals addressed in the guideline.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-174
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
4.5.10.2. Application of the EqP Approach
As originally proposed, the EqP approach for copper, cadmium, lead, zinc, nickel, and
silver focused on AVS as the principal partitioning phase. Because these metals form sulfides
that are highly insoluble, toxicity from these metals is not expected if there is sufficient sulfide
available to bind all available metal. Because some metals in sediment are present in mineral
forms that are not highly reactive, the metal concentrations to which the sulfide concentration is
compared is not the total metal in sediment, but the metal extracted simultaneously with the
sulfide, and is therefore referred to as simultaneously extracted metal (SEM). Because cadmium,
copper, lead, nickel, and zinc are divalent metals, 1 mol of each metal can bind with 1 mol of
AVS. The molar concentrations of these metals are compared with AVS on a one-to-one basis.
Silver, however, exists predominantly as a monovalent metal, so that silver monosulfide (Ag2S)
binds 2 mols of silver for each mol of AVS. Therefore, SEM Ag will represent the molar
concentration of silver divided by two, [Ag]/2, which is compared with the molar AVS
concentration. Thus, the solid-phase AVS ESB is defined as:
where:
£ [SEMJ = [SEMCJ
+ [SEMpJ + [SEMNJ + [SEMJ +
When this sum is less than the molar concentration of AVS, no toxicity is expected
because there is sufficient sulfide to bind all SEM. Results of calculations using chemical
equilibrium models indicate that metals act in a competitive manner when binding to AVS. That
is, the six metals will bind with AVS to form their respective sulfides in the order of their
increasing solubility: silver, copper, lead, cadmium, zinc, and nickel. Therefore, they should be
considered together. There cannot be a guideline just for nickel, for example, because all the
other metals may be present as metal sulfides and therefore, to some extent, as AVSs. If these
other metals are not measured as a mixture, then the £SEM will be misleadingly small, and it
might appear that £[SEM]<[AVS] when in fact this would not be true if all the metals were
considered together. It should be noted that EPA currently restricts this discussion to the six
metals listed above; however, in situations where other sulfide-forming metals (e.g., mercury) '
are present at high concentrations, they also should be considered.
1 1/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-175
-------
1
2
3
4
5
6
7
8
.9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
While the SEM-AVS comparison has been found effective for predicting when metals in
sediment will not cause toxicity, the presence of excess SEM above AVS does not always
portend toxicity. This is, at least in part, because other components of sediments can also reduce
the bioavailability and toxicity of metals. To address these additional factors, a second
component of the metal mixture ESB was developed: the interstitial water benchmark. The
concept behind the interstitial water component is that even if there is excess SEM above AVS in
a sediment, if there is sufficient additional binding available to maintain non-toxic
concentrations of metals in the intersititial water, then the sediment should not be toxic, despite
the excess SEM. The aggregate toxicity of metals in the interstitial water is assessed by dividing
the concentration of dissolved metal in interstitial water by the final chronic value from the EPA
water quality criterion for the protection of aquatic life. This ratio is called the Interstitial Water
Benchmark Unit (IWBU), and these are summed across all of the metals (using an assumption of
addivity, as^described in Section 4.5.3); if the sum is less than 1, then the interstitial water-phase
ESB is met, and toxicity is not expected. Mathematically, this equates to:
where:
Ii([MNi,d]/FCVNl>(1)
For freshwater sediments, the FCVs are hardness dependent for all of the divalent metals
under consideration, and thus should be adjusted to the hardness of the interstitial water of the
sediment being considered. Because there are no FCVs for silver in freshwater or saltwater, this
approach is not applicable to sediments containing significant concentrations of silver (i.e.,
SSEM>AVS). Because silver has the smallest solubility product and the greatest affinity for
AVS, it would be the last metal to be released from the AVS or the first metal to bind with AVS.
Therefore, it is unlikely that silver would occur in the interstitial water of any sediment with
measurable AVS (Berry et al., 1996).
When used together, the overall metal mixture ESB is met if either the SEM-AVS or
interstitial water ESBs are met. Further, both guidelines are no-effect guidelines, in that they
predict the absence of toxicity if they are met. Though the probability of toxicity is higher if
both components of the ESB are exceeded, and this probability increases as the magnitude of
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-176
-------
1 exceedance increases, failing to meet the ESB does not imply that toxicity will necessarily occur.
2 This is because neither the SEM-AVS or interstitial water components address all factors that
3 may reduce the bioavailability and toxicity of metals in sediment. Site-specific toxicity testing
4 may be advisable to refine assessments for specific sites.
5
6 4.5.10.3. Enhancements to the Metals ESB
7 The AVS guideline is a "no effects" guideline; that is to say that if the guideline is not
8 exceeded, the sediments should not be toxic due to the metals included in the guideline, but an
9 exceedence of the guideline does not necessarily mean that the sediments will be toxic due to the
10 presence of those metals. One way to reduce the uncertainty of a prediction of toxicity is to
11 normalize for the fraction of organic carbon in the sediment and use £SEM-AVS)/foc for the
12 solid-phase guideline. The use of this guideline is described in the draft metals ESB (EPA
13 2002f).
14 A refinement of this organic carbon correction has been proposed by Di Toro et al. (2004,
15 submitted). In this analysis, organic carbon partition coefficients are calculated for each metal
16 and used to estimate combined AVS-organic carbon partitioning. Although this approach has
17 not been widely applied as yet, it has a strong theoretical basis, and initial applications suggest
18 that it can strengthen the ability of EqP to predict metal toxicity (rather than just non-toxicity) in
19 sediments.
20 The metals ESB currently applies to only six metals: cadmium, copper, lead, nickel,
21 silver, and zinc. However, AVS can also be used to predict the lack of toxicity in sediments due
22 to chromium because the presence of AVS in sediments is indicative of reduced sediments.
23 Chromium is present primarily in its reduced form (CrIII) in reduced sediments, and in.this form
24 it is much less toxic and bioavailable than the oxidized form of chromium (CrVI) (Berry et al., in
25 prep.). The use of AVS in the prediction of the lack of chromium toxicity in sediments will be
26 described in an addendum to the metals ESB.
27
28 4.5.10.4. Endpoints of Concern
29 The metals ESB should be viewed as an approach, as opposed to a specific number. The
30 ESB calculates a concentration in sediment that will correspond to a given toxicological
31 endpoint using partitioning theory. However, the correspondence of the benchmark is variable,
32 depending on which toxicological endpoints are selected. The FCVs from the metals WQC are
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-177
-------
1 used in the ESB as examples because they are familiar and provide a widely accepted level of
2 protection. Other toxicological endpoints could be chosen, however, if a different level of
3 protection was desired. For example, the water-only, chronic no-effect concentration for a
4 threatened or endangered species might be used as the toxicological endpoint. EqP theory could
5 then be used to calculate a sediment concentration that would be protective of that species.
6
7 4.5.10.5. Limitations
8 It should be emphasized that the metals ESBs are intended to protect benthic organisms
9 from the direct effects of these six metals in sediments that are permanently inundated with
10 water, intertidal, or inundated periodically for durations sufficient to permit development of
11 benthic assemblages. They do not apply to occasionally inundated soils containing terrestrial
12 organisms. These benchmarks do not address the possibility of bioaccumulation and transfer to
13 upper-trophic-level organisms or the synergistic, additive, or antagonistic effects of other
14 substances.
15 The EqP approach to metal assessment has been questioned on several technical issues,
16 such as whether it applies to organisms that live in oxygenated burrows, whether it adequately
17 accounts for ingestion of sediment, and whether it is appropriate to use bulk sediment chemistry
18 to represent responses to microenvironments that exist in bedded sediment. Many of these issues
19 are difficult to address in a comprehensive manner, and more study will be required to fully
20 evaluate them. In the interim, it should be noted that there is a considerable body of .
21 experimental data supporting the EqP approach to metals (Berry et al. 1996; Hansen el al. 1996;
22 Di Toro et al. 2004; EPA 2002f), and these data include tests with organisms with varying life
23 histories, that use irrigated burrows, and that ingest sediment. While this does not prove that the
24 criticisms above are not legitimate, it provides support for continuing to apply the EqP approach
25 as a reasonable representation of metal toxicity in sediment, at least until a superior approach is
26 developed.
27 Because AVS is a product of microbial activity, AVS can vary seasonally with changes
28 in microbial activity. The degree to which such cycling can affect the potential risk from metals
29 in sediments is unclear but should be considered in sampling programs designed to assess SEM-
30 AVS.
31 Because metals bound to sulfide do not appear to be sufficiently bioavailable to cause
32 toxicity in sediment toxicity tests, one would not expect bioaccumulation of metals in sediments
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-178
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
with more AYS than SEM. However, several studies have found bioaccumulation of metals
even when excess AVS is present (Ankley et al., 1996). This has caused considerable debate
about the appropriateness of SEM-AVS for assessing metal toxicity in sediments (e.g., Lee et al.,
2000), because it suggests that metal bioavail ability may not be effectively represented by SEM-
AVS analysis. However, there is a large number of studies indicating that toxic effects of metals
are absent in sediments when SEM is less than AVSs, even when bioaccumulation is observed.
This suggests that the bioaccumulated metals may not be toxicologically available or of
sufficient concentration in the organism to cause effects. In addition, these metals do not
biomagnify to higher trophic levels in aquatic ecosystems (Suedel et al., 1994). Therefore, an
ESB based on the difference between the concentrations of SEM and AVS still appears
appropriate for protecting benthic organisms from the direct effects of sediment-associated
metals, but not for protecting against metal bioaccumulation.
4.5.10.6. Use and Implementation
In practice, the sediment benchmarks for these six metals are not exceeded, and benthic
organisms are sufficiently protected (defined in this case as the level of protection afforded by
the WQC), if the sediment meets either one of the following benchmarks:
• The solid phase benchmark: ^[SEM J < [AVS]; in other words, if metal, measured
as SEM, does not exceed AVS; or
• The interstitial water benchmark: Ei([Mj J/FCVid) < 1; in other words, if the sum of
the contributions of all six metals in the interstitial water would not be expected to
cause chronic effects at the level of protection afforded by the WQC.
If the AVS or interstitial water ESBs is exceeded, there is reason to believe that the
sediment might be unacceptably contaminated by these metals. Further evaluation and testing
would, therefore, be necessary to assess actual toxicity and its causal relationship to the metals of
concern. If data on the sediment-specific SEM, AVS, and organic carbon concentrations are
available, the uncertainty bounds for (£,SEM-AVS)/foc described in the metals ESB could be
used to further classify sediments as those in which metals are not likely to cause toxicity, metal
toxicity predictions are uncertain, or metal toxicity is likely. For sediments in which toxicity is
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be constiued to represent any
Agency determination or policy.
4-179
-------
1 likely or uncertain, acute and chronic tests with species that are sensitive to the metals suspected
2 to be of concern, acute and chronic sediment toxicity identification evaluations (TIEs), in situ
3 community assessments, and seasonal and spatial characterizations of the SEM, AVS, and
4 interstitial water concentrations would be appropriate (Ankley et al., 1994).
5
6 4.5.11. Soil Toxicity
7 Variability among soil toxicity test results is due in part to the influence of soil properties
8 on bioavailablity of metals (e.g., pH, organic matter and CEC). See Sections 3.1, 3.4.6, and 3,2,
9 covering environmental chemistry bioavailability, terrestrial bioavailability, terrestrial exposure
10 issues. Additionally, incorporation of sparingly soluble substances, such as many environmental
11 forms of metals, into the soil matrix is difficult, and acclimation/adaptation of test organisms can
12 further complicate test results. Use of soluble metal salts with the addition of organism to the
13 test matrix immediately after mixing is not representative of most environmental situations,
14 where aging and other physical/chemical processes affect metal speciation and uptake.
15 Furthermore, testing of soil microbial function is particularly problematic because the test
16 substance is added to soils with the microbial population already in place and that contain
17 background amounts of metals.
18
19 4.5.11.1. Application
20 Modifications to standard toxicity bioassays for plants and soil organisms to account for
21 properties of metals were discussed in an expert workshop and subsequently provided in
22 Fairbrother et al. (2002). These modifications include directions on type of soil matrices to use,
23 mixing and aging of metals into the soil, and cautions about acclimation of test organisms.
24 Future studies conducted specifically for development of toxicity endpoint values for metals can
25 follow these suggested protocols and circumvent many of the past problems.
26 There is, however, a large body of literature on toxicity of metals to soil organisms that
27 has already been developed, although often the objectives were to understand processes rather
28 than to develop defensible toxicity thresholds. The challenge, therefore, lies in how to use these
29 data, taking into account the test-to-test variability in soil chemistry parameters, and how to
30 develop a technically defensible means of extrapolating toxicity responses across soil type—in
31 • other words, how to adjust the toxicity threshold values for bioavailability differences in test
32 conditions. Ideally, an aquatic BLM could be extended to terrestrial setting to account for
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-180
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
differences in bioavailability due to environmental chemistry, particularly for plants and soft-
bodied soil invertebrates. Studies have been initiated to develop what has been referred to as a
terrestrial BLM, or tBLM. The conceptual approach to development of a tBLM is very much the
same as for the variations of the BLMs that have been developed for aquatic settings (Allen,
2002). Although not currently developed to the point of being of practical use, experimental
testing and model development programs are fully underway in the hope of providing a tool that
will be of great practical utility in the relatively near future.
Another approach to addressing soil variability in soil toxicity tests is to normalize test
results by dividing the LC50 by percent organic matter (van Gestal, 1992). This approach is
based on observed correlations between the LC50 of copper to earthworms and soil organic
matter content. Most recently, CEC has been shown to be the most important factor modifying
zinc bioavailability in soils for both invertebrates and plants, and it will presumably show a
similar relationship with other cationic metals (Smolders, 2003 presentation in DC; based on
work by C. Janssen, Univ. Ghent). Therefore, effect responses in different soil types can be
normalized on the basis of relative CEC. It should be remembered that CEC is a function, at
least in part, of soil pH. Therefore, normalization can be done only among soils of similar pH
ranges. However, comparison of field data with laboratory toxicity response information is best
done through measuring metals in soil pore water from field assessments and comparing such
data to spiked laboratory soils.
Several studies of plants and invertebrates have suggested that consideration of aging
may result in significant overestimates of field effects using data laboratory studies (e.g., Smit
and Van Gestel, 1998). Estimates of effects of aging (see Section 4.1.7.2., Environmental
Chemistry for an aging discussion) range from 3 to 8-fold differences between laboratory and
field results.
Because most soil properties are correlated to some degree, isolating individual soil
parameters and relating them to soil biota toxicity is difficult. Treating the soil like a black box
and using sequential extraction techniques has been useful in determining what fraction of the
metals are in a labile form (Kabata-Pendias and Pendis, 2000). However, some scientists have
questioned the usefulness of this process (Morgan and Morgan, 1988).
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
{ ' 4-181
-------
1 4.5.11.2. Limitations
2 Only two methods have been proposed in the literature for normalizing toxicity data
3 across soils to account for differences in bioavailability: adjusting of endpoint values by percent
4 organic matter in the soil or as a function of organic matter plus clay content. Both methods fail
5 to incorporate either pH or cationic exchange capacity, both of which are of critical importance
6 in determining bioavailability. Aging of metals in soils also is not included in these approaches.
7 Furthermore, the data sets used to generate the relationships were not sufficiently robust to make
8 generalizations possible across all soils and all organisms. The development of a tBLM shows
9 promise as a method that will overcome these limitations; however, it likely will not be
10 completed for at least 2 years after the publication of this Framework.
11 The use of a tissue residue approach has been suggested as another method to address
12 soil chemistry and metal toxicity issues, suggesting that a metal concentration must reach a
13 threshold value in the organism or at the target site before effects begin to occur (McCarthy and
14 Mackay, 1993;.Lanno and McCarty, 1997). For essential elements in plants,
15 deficiency/sufficiency concentrations in foliage have been developed. However, the relationship
16 between toxicity and tissue residues is complex and varies depending on tissue type (roots vs.
17 shoots), plant species, and metal. Little information is available for soil invertebrates, so
18 relationships between tissue concentration and toxic response cannot yet be developed.
19 Therefore, this approach, although conceptually sound, requires significant research before
20 critical tissue levels can be established.
21 For plants, a large proportion of the toxicity literature was developed in support of
22 understanding potential toxicity and metal uptake from biosolids (e.g., sewage sludge). It is
23 difficult to determine single-species, single-metal thresholds from this database for several
24 reasons. First, biosolids tend to contain a mixture of metals, so any response observed cannot be
25 attributed to a single metal and should account for potential antagonism or synergies that might
26 occur. Second, biosolids are, by their nature, high in organic matter, which significantly affects
27 bioavailability of the metals. Until a robust method is developed to adjust toxicity endpoints for
28 the influence of organic matter (and other bioavailability factors), it will remain difficult to apply
29 such results to unamended soils. The additional organic matter also provides excess nutrients to
30 the plants, which further confounds possible metal effects. The guidance for development of
31 Ecological Soil Screening Levels or EcoSSL (U.S. EPA, 2003c) can be used to judge the
32 applicability of literature studies to plant or soil invertebrate toxicity threshold determinations.
33
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-182
-------
1 4.5.12. Food Chain (Wildlife) Toxicity
2 Toxicity in wildlife from metals exposures is generally poorly understood and is rarely
3 quantified in field settings. A few notable exceptions are those mechanisms described in avian
4 waterfowl exposure to selenium (Adams et al., 2003), exposure of waterfowl to lead-
5 contaminated sediments (Beyer et al., 1998; Blus et al., 1991; Henny et al., 2000), and white-
6 tailed ptarmigan exposure to cadmium in vegetation (Larison etal., 2000). Most metals express
7 multiorgan toxicity, resulting in a general decrease in overall vigor, as opposed to well-defined
8 mechanisms of action documented from organic xenobiotics such as pesticides. Typically,
9 lexicological data used to assess the risk of many metals to wildlife are derived from laboratory
10 species such as rats or mice or domestic livestock species (e.g., cattle and chickens) exposed to
11 soluble metal salts. Extrapolating the results of such tests to evaluate toxicity to wildlife is
12 necessary because of the paucity of data on the toxicity of metals to these receptors. However,
13 extrapolation of results should be approached with caution due to the large amount of uncertainty
14 that could be introduced into the risk assessment process (Suter, 1993).
15 Laboratory and domestic species may be more or less sensitive to chemicals than is the
16 selected receptor. Toxicological responses among species vary because of many physiological
17 factors mat influence the toxicokinetics (absorption, distribution, and elimination) and
18 toxicodynamics (relative potency) of metals after exposure has occured. For example,
19 differences in gut physiology, renal excretion rates, and egg production influence the
20 toxicokinetics of metals. The ability of some species to more rapidly produce protective proteins
21 such as metallothionein after exposure to metals are toxicodymamic features leading to
22 interspecific extrapolation uncertainty. For example, mammal studies should not be extrapolated
23 • to birds, and extrapolation of data from rats (simple, monogastric digestive physiology) to
24 ruminants introduces more uncertainty than does extrapolation from rats to canids, and so on, In
25 the case of metals, which some species are able to regulate or store in their tissues without
26 experiencing toxic effects (i.e., biota-specific detoxification), extrapolations between species
27 used to assess bioaccumulation and toxicity can be especially problematic.
28
29 4.5.12.1. Application
30 Methods for extrapolating metal effects data among species are not unique to metals risk
31 assessment, with the exception of understanding different requirements for essential elements.
32 Some of the methods for extrapolating effects data among species include body weight
33 normalization (Sample and Arenal.,1999), distribution-based approaches (Van Straalen, 2001),
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-183
-------
1 or the use of uncertainty factors (e.g., Calabrese and Baldwin, 1994). All of these approaches
2 suffer from a lack of an underlying physiological basis quantifying toxicokinetic and
3 toxicodynamic responses among species. A review of potential extrapolation methodologies can
4 be found in Kapustka et al. (2004).
5 Currently, the best sources of information on metal toxicity thresholds are NAS/NRC
6 (1980; 1994), McDowell (2003), and the documentation supporting development of EcoSSLs
7 values (U.S. EPA, 2003c). The EcoSSL document also includes a general approach for
8 screening studies for acceptability for use in derivation of toxicity thresholds for risk
9 assessments that can be used for deriving site-specific TRVs for the most applicable endpoints.
10 These endpoints should then be extrapolated to species with similar physiology, particularly of
11 the digestive system, due to the predominance of the dietary exposure pathway (e.g., cow data
12 can be applied to wild bovids such as bison and possibly to other ruminants such as deer or elk).
13 Uncertainty factors can be carefully applied if there is concern for extrapolation of data to
14 species in a different taxonomic category (e.g., genus, family or class). General summaries for
15 some metals are available in Beyer et al. (1996) and Fairbrother et al. (1996).
16
17 4.5.12.2. Limitations
18 Information on toxicity of metals to wildlife under field conditions is severely limited,
19 focusing on only a few species and a few metals. Deriving TRVs for metals in wildlife is
20 problematic because the administered form of metal is typically not found in most applied
21 settings. For example, lead has one of the largest databases for laboratory exposures but has
22 been limited by the form of the metal studied. Until recently, almost all field studies were
23 conducted in support of toxic effects of lead shot (i.e., pure elemental lead), whereas almost all
24 laboratory studies have administered lead to test subjects as lead acetate. These extreme forms
25 of lead relative bioavailability make extrapolations to dietary exposures difficult for this
26 substance. The best approach in this case is the use of critical tissue residues, because liver lead
27 levels indicative of lead poisoning are well established (Beyer et al., 1996). Selenium is another
28 example where tissue-based toxicity thresholds (in this case, in the avian egg) may be most
29 appropriate (e.g., Adams et al., 2003).
30 Cross-species extrapolations should be conducted with some knowledge of animal
31 physiology and specific responses to metals. Digestive physiology is the most important
32 distinction, because most metal exposures in wildlife are by the dietary route. However, other
33 specific organismal responses should be understood as well. For example, sheep are much more
11/24/2004 Peer Review Draft
DISCLAIMER: This infoimation is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-184
-------
1 sensitive to copper than are other ruminants because of the particular nature of their gut flora
2 (NAS/NRC, 1980), Therefore, extrapolation of sheep data to other ruminants would be highly
3 overconservative. On the other hand, pigs are extremely tolerant to copper, possibly due to low
4 gut uptake rates (NAS/NRC, 1980), so extrapolations of such data to other monogastric animals
5 would not be protective.
6 Interactions of metals (see Section 3.3. Human Health) also should be taken into account
7 when analyzing metal toxicity data for wildlife. As noted above, values that may be protective
8 for a particular metal within a certain animal may not be so if other metals are present or
9 deficient (or vice versa). Therefore, dietary studies should be examined to understand the
10 presence of other metals and to ascertain the sufficiency of essential elements. Application of
11 single-metal thresholds to field situations, whether in a site-specific context or on a national
12 scale, should make provision for adjustment to account for interactive effects.
13
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
4-185
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
5. METALS RESEARCH NEEDS
The field of metals risk assessment has been advancing rapidly for more than 5 years, as
reflected in the large amount of information described in this document. However, significant
uncertainties and gaps remain that require additional research if accurate assessments are to be
conducted. This chapter briefly reviews major, ongoing research programs within and outside
EPA and provides a list of recommendations for future research endeavors. This chapter is not
intended to outline a research strategy for metals risk assessment. Rather, it discusses the current
direction of metals-related research conducted by EPA, external institutions, and academia as it
relates to uncertainties and gaps in metals risk science.
5.1. U.S. EPA RESEARCH
EPA has devoted significant resources to researching metals-related topics, and a variety
of metals-related endeavors are planned and under way in an ongoing effort to better understand
the behavior and effects of metals in humans and the environment and to advance the field of
metals risk assessment Specific summaries of metals-related research at EPA are available for
viewing on the Agency's Science Inventory, a searchable catalog of EPA research available at
http://cfpub.epa.gov/si. In addition, EPA's Office of Research and Development (ORD) has a
multiyear planning effort to guide the direction of its research program. The purpose of the
multiyear plans (MYPs) is to provide a framework that integrates research across ORD's
laboratories and centers. The MYP for contaminated sediments describes a number of projects
planned and under way that involve current topics and research needs. The most recently
available MYPs are available for review at http://www.epa.gov/osp/myp.htm. Examples of
planned and ongoing Agency research and assessment projects involving metals topics include
the following:
• Ecological effects of selenium on soil invertebrates to support the development of
soil-screening limits for selenium (fiscal year [FY] 2004).
• Ecological assessment of the risks associated with ground water contamination and
exposures (FY 2005).
• Characterization and assessment of the impact of metals speciation on ecological
receptors (FY 2006).
• Evaluation of stabilization of metals in sediments (FY 2007).
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
5-1
-------
1
2 • Evaluation of perturbation on metals speciation and ecological receptors (FY 2008).
3
4 5.2. EXTERNAL RESEARCH
5 The Metals in the Environment Research Network (MITE-RN) is a network of
6 collaborating institutions with participants from academia, government, and industry, formed in
7 . 1998 with the aim of developing a better understanding of sources of metals in the environment,
8 how metals move and transform within the environment, and how they can affect ecosystems
9 and human health (www.mite-m.org). MITE-RN has been funded at approximately $7 million
10 for the period 1999-2004 and has been refunded for the next 5 years (2005-2010). Funding
11 consists of contributions from the Mining Association of Canada; National Sciences and
12 Engineering Research Canada; the Ontario Power Generation Inc.; "in-kind" contributions from
13 Environment Canada, Fisheries and Oceans Canada, and Natural Resources Canada; and funding
14 support from the International Lead Zinc Research Organization, International Copper
15 Association, and Nickel Producer's Environmental Research Organization. MITE-RN has
16 published many research articles covering all of the topics discussed in the framework
17 (www.mite-m.org/files/mite-m_pubs.pdf). The most recent compilation of papers may be found
18 in Human and Ecological Risk Assessment, vol. 9 (4).
19 Several universities have established multidisciplinary research centers for metals. For
20 example, under a grant from the National Institutes of Environmental Health Sciences (NIEHS),
21 Harvard University has established the Metals Research Core, which promotes innovative
22 research among investigators who are studying the environmental fate and health effects of
23 exposure to metals and related fields with emphasis on potential gene-metal, metal nutrient, and
24 metal-metal interactions (www.hsph.harvard.edu/niehs/metals.html). The Agency funds amulti-
25 institutional Center for Study of Metals in the Environment, coordinated out of the University of
26 Delaware and including Colorado School of Mines, Manhattan College, McMaster University,
27 Ohio State University, Oklahoma State University, University of Wyoming, and University of
28 Missouri at Rolla (www.ce.udel.edu/CSME/Index.html). The Center for Air Toxic Metals®
29 (CATM®) at the University of North Dakota Energy and Environmental Research Center
30 (EERC), established in 1992 by the U.S. EPA's Office of Environmental Engineering and
31 Technology, develops information on trace elements that can be used to develop pollution
32 , prevention strategies (www.eerc.und.nodak.edu/catm/). Dartmouth's Toxic Metals Research
33 Center is an interdisciplinary group that studies how arsenic and other metals affect human
34 health and the environment (http://www.dartmouth.edu/~toxmetal/). The Agency also funds the
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
5-2
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Hazardous Substance Research Centers (HSRCs), which address metals-linked topics, such as
mining, contaminated sediments, and ground water contamination. Web sites for HSRCs may be
accessed at:
http://www.engr.colostate.edu/hsrc/new.html
http://www.hsrc.org/hsrc/html/ssw/newsletter/sswnews.html
http://wrhsrc.oregonstate.edu/publications/index.htm
The metals industry also sponsors research conducted at both public and private
institutions worldwide. Sponsors include the International Lead Zinc Research Organization
(www.ilzro.org), the International Copper Association (www.ica.org), the Nickel Producer's
Environmental Research Organization (www.nipera.org), the International Zinc Association
(www.iza.org), Eurometeaux, the International Cobalt Association, and the International Council
on Mining and Metals (www.icmm.com).
5.3. SPECIFIC RECOMMENDATIONS
The information provided here is a summary of research recommendations provided in
the metals issue papers (http://cfpub2.epa.gov/ncea/raf/recordisplay.cfm?deid=59052) and
additional comments provided by reviewers of this framework.
5.3.1. Environmental Chemistry
In general, environmental chemistry of metals research could benefit from:
• The development of more routine chemical-species-specific analytical methods.
• The development of extraction techniques that have general utility in assessing
bioavailability and/or mobility.
• The validation of geochemical and chemical-specific environmental fate and transport
models.
• Additional research on metal mobility and how to apply the Diffuse Layer (DL)
adsorption model to metal behavior in soils and sediments using estimated values.
• Increased understanding of metal chemistry in sediments, including the redox
behavior metals.
• Research to understand the chemical and physical forms of metals in the primary
media of exposure.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
5-3
-------
1
2
3
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
• Improved reliability of aging predictions.
53.2. Bioaccumulation and Unavailability
5.3.2.1, Aquatic
• Evaluation of the bioaccumulation of metals bound to colloidal material in ambient
water.
• More thorough evaluation of the efflux rates of metals from different animals,
including specific tissues, following bioaccumulation from the dissolved phase and
from the dietary pathway.
• Evaluation of metal bioaccumulation in aquatic bacteria, which may influence the
fluxes of certain metals in aquatic systems and which may introduce metals into
bacteria-based food chains.
• A more detailed knowledge base is 'required on the basic physiology and ecology of
organisms that are used or at least have the potential to serve as bioindicator
organisms. Furthermore, monitoring programs could focus on key biomarkers of
exposure and effects and would bewise to develop an algorithm to calculate an
integrated stress index.
• New approaches to evaluate the bioaccumulation of metals from waters in which
there are numerous contaminants (such as would be found in most contaminated
harbors or rivers) to assess synergistic and antagonistic effects.
* With regard to the BCF/B AF model, development of additional guidance should be
directed at reducing uncertainty and consideration should be given to:
- Articulating the limitations of the BCF/B AF approach.
- When the BCF/BAF approach is or is not applicable.
- How the BCF/BAF approach could be modified.
- What alternative measures and criteria could be used to better account for metal
bioaccumulation in relation to toxicity potential.
53.2.2. Terrestrial Soil Organisms
i
• Development and validation of empirical and mechanistic models linking soil
physicochemical characteristics, metal speciation, and toxic effects and
bioaccumulation in soil invertebrates (e.g., Biotic Ligand Model (BLM) for soil
organisms).
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
- ' 5-4
-------
• Development and validation of kinetics models describing metal bioaccumulation in
soil invertebrates.
• Basic research on the physiology of metal metabolism in various groups of soil
invertebrates; evaluation of the relevance of soil pore water or diet in exposure and
partitioning of metals in soil invertebrates.
• Identification of the risks for predators associated with the consumption of soil
invertebrates containing metals; evaluation of the risk to predators of consuming
metal partitioned to different fractions in soil invertebrates (e.g., storage granules
versus metallothionein).
• Development of metal-specific biomarkers capable of quantitatively detecting
magnitude and species of metal exposure.
•
5.3.3. Exposure
Research on exposure issues is best defined with regard to the particular receptors; hence,
this section is divided into discussions of exposure research for human, terrestrial, and aquatic
organisms.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21 5.3.3.1. Human Health Receptors
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
Research to improve sampling, measurement approaches, and exposure models.
Data on rates of soil and surface dust ingestion, including estimates of central
tendencies, both short-term and long-term, inter- and intraindividual variability (e.g.,
within-age and across ages), and relative contributions of surface dust and soil.
Information about the types and frequencies of activities that place children in contact
with contaminated soils, dusts, or surfaces (e.g., hand-to-mouth behavior, rates of
contact with surfaces).
Methods to predict concentrations in surface dust, a primary medium of contact, from
measurements made in surface soil samples, surface dust wipe samples, and surface
dust vacuum samples.
Better estimates of dietary intakes of metals.
Information on the contribution of locally harvested foods to metal intakes (e.g.,
uptake of metals from soil, intakes of home-grown or home-harvested foods).
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
5-5
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
53.3.2. Aquatic Receptors
• Define the best ways of expressing exposure concentration.
• Accommodate differences in exposure durations.
• Combining exposure concentrations when exposure involves metal mixtures.
• Assess the relative merits of different methods used to express exposure
concentrations in sediments and suspended solids.
• Compare the simple extrapolation methods to richer survival or lime-to-event models
is essential.
53.3.3. Terrestrial Receptors
• Conduct further work to define or reduce the associated uncertainty of using
generalized wildlife exposure models.
• Develop laboratory test data for soil systems that better reflect the actual forms of
metals in field soil..
• Develop data for terrestrial receptors on the j oint effect of metals in mixtures.
53.4. Human Health Effects
»
• Research should be conducted concerning the potential interactions between essential
metals and nonessential metals and between nonessential metals (i.e., metal
mixtures).
• Research should be conducted concerning the applicability of toxicokinetic and
toxicodynamic models for risk assessment for metals and inorganic metal
compounds. Consideration should be given to differences in models for essential and
nonessential metals, for low versus high dose exposures, for multipathway exposures,
and for mixtures of metals with multiple modes of action.
• Research is needed on methods to assess mixed (or multiple) exposures, including
occupational exposures. t .
• There should be further research and development regarding the use of biomarkers,
especially genomic, transcriptomic, and proteomic methodologies as endpoints and
their relationship to frank effects on human health endpoints used in risk assessment.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
5-6
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
• Speciation of metals in tissues of target organs should be determined. Research
should be conducted on mechanisms of toxicity, including carcinogenicity, and on
whether carcinogenicity of specific metals is a threshold or nonthreshold event.
• Research is needed to meet the needs of sensitive individuals on the basis of genetic
and developmental factors.
• Research should be conducted to determine the potential essential or beneficial
effects of metals and inorganic metal compounds (especially as these effects impact
low-dose extrapolation).
5.3.5. Characterization of Ecological Effects
Because of significant differences in ecology, physiology, and toxicology of aquatic and
terrestrial organisms, research recommendations are provided separately for each group.
53.5.1. Toxicological Research Needs
• Test matrix. Perform tests under conditions that are representative of
environmentally relevant exposure conditions to form a basis for development of
methods to extrapolate from laboratory to field conditions, and from site to site in the
field in situations where the water-,, sediment-, or soil-quality characteristics are quite
variable.
• Model development. Develop descriptive statistical models that establish the
boundaries for which the results are applicable. Develop predictive speciation and
uptake models, such as the Free Ion Activity Model (FIAM) or Terrestrial Biotic
Ligand Model (tBLM).
• Test design. Conducting a concerted effort to generate reasonably complete
concentration-response surfaces for metals in major soils representative of larger
areas of the continent would be very useful. Test designs should be based on
regression models and strive to depict the range of responses from relatively low
concentrations to relatively high concentrations.
• Measurement endpoints. Multiple endpoints should be scored over the course of
the in-life portion of tests. Data on growth parameters, overall healthiness of test
organisms, and behavioral and reproductive endpoints should be explored. Such data
could be useful in developing descriptive statistical models, including multiple
regressions and clustering analyses. In addition, such data could be used to calibrate
predictive models that attempt to relate effects'across exposure periods.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
5-7
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
• Interspecies extrapolation. Studies designed to develop interspecies extrapolation
models, especially for terrestrial organisms, are necessary.
* Interactions among metals. Research to understand interactive effects, for both
acute and chronic exposures, is necessary. Methods for representing the joint effects
of metal mixtures on the organism, in the context of the BLM, have been proposed,
but these methods have not been implemented or tested to date.
t
5.3.5.2. Ecological Research Needs
* Extrapolations from laboratory to field. Studies documenting the correspondence
or lack of correspondence between simple laboratory toxicity tests and field
assessments are necessary. In situations where laboratory and field results are
inconsistent, research is necessary to identity factors mat contribute to these
differences.
• Indirect effects of metals and species interactions. Studies of the effects of metals
on species interactions and their ecological significance are needed.
* Acclimation and adaptation to metals. Additional research is necessary to
understand the cost of tolerance and adaptation to metals and the potential
consequences with regard to exposure to multiple stressors.
5.3.6. Unit World Model for Metals
Recognizing the continuing need to overcome limitations in understanding and, therefore,
adequately predicting metal behavior in the environment, it has been proposed that use of a suite
of evolving computational modeling tools could provide the basis for metals assessments,
particularly for hazard ranking and screening-level risk assessments. The model would be run
for a generic environment (the "Unit World"), giving output in the form of substance-specific
loadings that would result in accumulations in target compartments that equal specified toxicity
thresholds (e.g., LC50, EC25), known as the "critical load." It is anticipated that such outputs
could be used for both classification and priority ranking as well as for regional (or national)
screening assessments. For screening assessments, an attempt is made to describe how the
substance will behave in the environment, to which media it will partition (e.g., air, soil, or
water), how long it will persist, in which media it will primarily degrade, and ultimately, what
mass input will result in media that will cause a toxic effect. These assessments introduce an
illustrative or hypothetical release into a hypothetical, evaluative region of interest. The model
yields calculated masses, concentrations, and rates of transport and transformation that are
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
5-8
-------
1 entirely fictitious. Features that do not depend on the quantity of chemical or specific
2 environmental characteristics, such as environmental persistence, relative importance of
3 degradation and transport pathways, and relative concentrations among media, can be deduced.
4 It should be noted that the decision point for whether an assessment of toxiciry is required is
5 derived through an evaluation of environmental fate processes.
6 This approach also could be used to conduct a risk assessment in which real emission
•
7 rates are introduced into a model construct of a real environment in an attempt to calculate real
8 masses and concentrations. These predicted environmental concentrations (PECs) can be
9 compared with measured values to evaluate the performance of the model and with
10 concentrations below which adverse toxicological effects are not expected to occur (e.g.,
11 PNECs). This comparison results in an assessment of the risk of an adverse effect at the
12 predicted concentration.
13 A "critical load" is defined as the mass per unit time
14 of a substance that, when introduced into the environment,
15 results in accumulations in environmental media that reach
16 specified toxicity thresholds. Key inputs are toxicity data
17 (acute or chronic thresholds) for individual environmental
18 compartments and the physical and chemical properties of the
19 substance. Fate models may be envisioned as being run in
20 reverse to back-calculate the loading that results in achieving
21 the specified toxicity concentration. Processes that affect fate and potential exposure of
22 organisms, such as intercompartmental transfer, complexation, and adsorption and precipitation
23 reactions, are included. Since many of the fate processes that affect metal ions and organic
24 compounds are similar, a common modeling framework may be envisioned for organic
25 substances and metals, with processes (e.g., biodegradation) turned on or off as required,
26 depending on the nature of the substance.
27- The Unit World approach uses models derived from previous modeling efforts for metals
28 for aquatic systems (e.g., Di Toro, 2001; Bhavsar et al., submitted). The aim is to produce a
29 model that is similar but that includes the processes that are necessary to describe the behavior of
30 metals in both aquatic and terrestrial environments simultaneously. The model is not intended to
31 represent a specific location but, rather, a representative setting that is typical of the class of
32 environments being evaluated. It also is not intended to be a complete description of metal fate
33 and transport. Rather, it focuses on the primary processes that affect the toxicity and long-term
Unit World Model
The Unit World Model calculates
the mass loading needed to achieve
the critical load of metal in each
environmental compartment (soil,
sediment, water). A critical load is
the amount of substance in the
environment that results in specified
toxicity thresholds.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
5-9
-------
1 fate of metals. It is designed to be used for evaluative purposes, not for detailed site-specific
2 evaluation.
3 The model framework, presented in Figure 5-1, is composed of an aquatic and terrestrial
4 sector. The model is formulated as a series of mass balance equations that are formulated
5 assuming that the rates of adsorption and desorption are fast relative to the other processes (i.e.,
6 the local equilibrium assumption). By contrast, the kinetics of metal sulfide precipitation and
7 dissolution are formulated as kinetic processes. The concentrations and characteristics of the
8 necessary water column and paniculate partitioning phases will be established to represent the
9 "unit worlds" to be used in the evaluation.
Terrestrial
Metal Input
Upper
Soil Horizon
Lower
Soil borfeon
i
Aquatic
Metal input
1
11
Deep Percolation
.1
I
I
Water
Column
Outlaw
Aerobic
Layer
Anaerobic
Uyer
Q S
Figure 5-1. Schematic representation of the Unit World Model for Metals.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
5-10
-------
1 Partitioning in the water column and aerobic sediment layer is computed using
2 WHAM6/SCAMP (Tipping, 1998; Lofts and Tipping, 1998). These models have been
3 calibrated with laboratory data and have parameters for many of the metals that are of interest.
4 Some field testing has also been performed with reasonable results (Bryan et al., 2002; Lofts and
5 Tipping, 1998). The aqueous phase speciation includes dissolved organic carbon (DOC)
6 complexation. The particulate partitioning phases are organic carbon; the oxides of Al, Si, Mn,
7 and Fe; and a mineral cation exchanger. The concentrations of these particulate phases are
8 specified externally as part of the input parameters. SCAMP assumes that the partitioning to
9 these phases is additive. The importance of metal sulfide precipitation in the anaerobic layer and
10 subsequent oxidation in the aerobic layer is well known, and models of these phenomena have
11 been developed (e.g., Di Toro et al.} 1996; Boudreau, 1991).
12 The soil model comprises two soil horizons, containing solids and solution. The upper
13 horizon receives the metal of interest in soluble form. The soil solution flows from the upper
14 horizon to the lower or directly to a surface water. Soil solution from the lower horizon flows to
15 the surface water or is lost, together with dissolved metal and metal bound to suspended
16 particulate matter (SPM), to deep percolation. The physical and chemical conditions are
17 specified for each horizon; the upper horizon has a higher organic matter content than the lower.
18 The processes governing metals in soils include solution speciation (described with WHAM6)
19 and solid-solution partitioning (described using Kj values, characterized by multiple regression
20 equations with pH and soil organic matter (Sauve et al., 2000,1998), and take into account
21 competition for dissolved organic matter binding sites by Al and Fe(III) species (Tipping et al.,
22 2002), particle aging (to be characterized for different metals on the basis of experimental
23 information, taking account of soil pH), and the input flux of background metals from
24 weathering. The value of Fin (moles m"2 a"1) that is sought from the model corresponds to the
25 Critical Load.
26 There are very few key processes that most significantly influence the outcome of the
27 model (i.e., the critical load). These are the input terms of ecotoxicity and bioaccumulation, the
28 speciation and partitioning reactions between water, sediment, and soil that determine both the
29 fate—the compartments where the metal finally resides—and the bioavailability. Additionally,
30 the transfer of metals into nonbioavailable forms has a significant impact on the critical load. In
31 this model, these transfer processes are aging reactions in soils and acid-volatile sulfide
32 (AVS)-binding of the metal. The approach assumes that the soluble metal form is introduced to
33 the air compartment and is distributed appropriately to the soil and water compartments.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
5-11
-------
1 Calculation of the fraction of soluble metal relative to the parent compound, or "transfer"-factor,
2 is done separately,
3 The Unit World Model approach is based on well-established principles governing the
4 environmental behavior of metals. However, at present, the proposed approach should be
5 considered as a conceptual framework embodying components that are at different stages of
6 development and evaluation. It will therefore be necessary to undertake a series of
7 well-integrated activities to move forward from the conceptual stage to a fully implemented and
8 accepted evaluative method that is capable of tracing the significant fate and transport processes
9 of a wide range of metals and predicting both the concentration and speciation at the exposure
10 point with a sufficient degree of accuracy to reflect the objectives of the assessment
11 (classification, ranking, or screening-level risk assessment).
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
5-12
-------
6. GLOSSARY OF TERMS
Absolute Bioavailabiiity (or ABA): Conventionally expressed as the fraction of the externally
administered amount of a metal substance that is absorbed and reaches the systemic circulation
or central compartment of the receptor.
Acclimation: How an individual organism develops tolerance during its lifetime, and it may be
gained or lost. Acclimation is also calledphenotypicplasticity.
Adaptation. A genetic change over multiple generations as a response to natural selection.
Traits are not lost during single life times. Adaptation is also known as genotypicplasticity.
Additivity: When the effect of the combination is estimated by the sum of the exposure levels or
the effects of the individual components.
Adsorption: Adsorption is the adhesion of molecules of gas, liquid, or dissolved solids to a
surface.
Antagonism: When the effect of the combination is less than that suggested by the component
toxic effects.
Background: The amount of metals occurring in soils, water, or air as a result of anthropogenic
and natural processes.
Bioaccessibility: The portion of total metal in soil, sediment, water, or air that is available for
physical, chemical, and biological modifying influences (e.g., fate, transport, bioaccumulation) is
termed the environmentally available fraction. Also referred to as Environmental Availability.
Bioaccumulation: The net accumulation of a metal in tissue of interest or the whole organisms
that results from exposure from all environmental sources, including air, water, solid phases (i.e.,
soil, sediment), and diet, and representing a steady-state balance of losses from tissue and the
body.
•
Bioaccumulation Factor (or BAF): The ratio of the metal concentration in an organism to that
in the surrounding medium at steady state. Metal accumulation in organisms is derived from all
routes of exposure.
Bioconcentration Factor (or BCF):'The ratio of metal concentration in an aquatic organism to
the metal concentration in water at steady state. Metal accumulation in aquatic organisms is .
derived from water only.
Bioavailabiiity of Metals: The extent to which bioaccessible metals adsorb onto or absorb into
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
6-1
-------
and across biological membranes of organisms, expressed as a fraction of the total amount of
metal the organism is proximately exposed to (at the sorption surface) during a given time and
under defined conditions.
Biomagnification The tendency of a chemical to accumulate to higher concentrations at higher
levels in the trophic system through dietary accumulation.
Biomarker: Biochemical, physiological, and histological changes in organisms that can be used
to estimate either exposure to chemicals or the effects of exposure to chemicals.
Biomonitoring: Use of living organisms as "sensors" in environmental quality surveillance to
detect changes in environmental conditions that might threaten living organisms in the
environment.
Cation Exchange Capacity (CEQ: A measure of the soil's ability to adsorb or release cations,
which is proportional to the number of available, negatively charges sites. The CEC is one of
the important parameters in controlling the potential bioavailability of metals in soils.
Community: An assemblage of populations of different species within a specified location in
space and time.
Conceptual model: A conceptual model in problem formulation is a written description and
visual representation of predicted relationships between ecological entities and the stressors to
which they may be exposed.
Delft 3D model: Software package that simulates two- and three-dimensional flow, waves, water
quality, ecology, sediment transport, and bottom morphology.
Donald J. O'Connor model: Algorithms for inclusion in the Simplified Lake and Stream
Analysis [SLSA] model.
Dose: The amount of a substance available for interaction with metabolic processes or
biologically significant receptors after crossing the outer boundary of an organism. The potential
dose is the amount ingested, inhaled, or applied to the skin. The applied dose is the amount of a
substance presented to an absorption barrier and available for absorption (although not
necessarily having yet crossed the outer boundary of the organism). The absorbed dose is the
amount crossing a specific absorption barrier (e.g., the exchange boundaries of skin, lung, and
digestive tract) through uptake processes. Internal dose is a more general term denoting the
amount absorbed without respect to specific absorption barriers or exchange boundaries. The
amount of the chemical available for interaction by any particular organ or cell is termed the
delivered dose for that organ or
Dose-response Curve: A graphical representation of the quantitative relationship between
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
6-2
-------
administered, applied, or internal dose of a chemical or agent, and a specific biological response
to that chemical or
Dose-Response Relationship: The relationship between a quantified exposure (dose) and the
proportion of subjects demonstrating specific biologically significant changes in incidence
and/or in degree of change (response).
ECSO: A statistically or graphically estimated concentration that is expected to cause one or more
specified effects in 50% of a group of organisms under specified conditions (1996). (Pertains to
ecological assessments)
Ecosystem: All the living (e.g., plants, animals, and humans) and nonliving (rocks,
sediments,soil,water,andair)material interacting within a specified location in time and space.
Endpoint: A measured response of a receptor to a stressor. An endpoint can be measured in a
toxicity test or in a field survey.
Environmental Availability: See Bioaccessibttity.
Essential Metals: Trace metals present in all healthy tissue of humans, whereby their withdrawal
from the body induces physiological, biochemical, and structural abnormalities and their
addition either reverses or prevents these abnormalities.
Exposure: Contact of a pollutant with the outer boundary of an organism; exposure is quantified
as the concentration of the agent in the medium over time.
Exposure Pathway: The course a chemical or physical agent takes from a source to an exposed
organism. An exposure pathway describes a unique mechanism by which an individual or
population is exposed to chemicals or physical agents at or originating from a site.
Exposure Route: The mechanism for which a chemical or physical agent comes in contact with a
person (e.g., by ingestion, inhalation, dermal contact).
Fugacity: Tendency for a metal to transfer from one medium to another.
Ground Water: Water in a saturated zone or stratum beneath the surface or land or water.
HST3D model: A 3-Dimensional ground water flow, heat and solute transport model.
HYDRAQL model: Computer-based speciation model,
Hydrophilic: Denoting the property of attracting or associating with water molecules;
characteristic of polar or charged molecules.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
6-3
-------
Hydrophobic. With regard to a molecule or side group, tending to dissolve readily in organic
solvents, but not in water, resisting wetting, not containing polar groups or sub-groups.
Indirect Effects: Changes in a resource that are due to a series of cause-effect relationships
rather than to direct exposure to a contaminant or other stressor. As a consequence of potential
direct effects of metals on organisms (e.g., mortality, reduced fecundity), other organisms in the
community can be indirectly affected (e.g., reduced prey items, predators).
Inhibition: When one substance does not have a toxic effect on a certain organ system, but when
added to a toxic chemical, it makes the latter less toxic.
Integrated Risk Information System (IRIS): IRIS is an electronic database that contains EPA's
latest descriptive and quantitative regulatory information about chemical constituents. Files on
chemicals maintained in IRIS contain information related to both noncarcinogenic and
carcinogenic health effects.
LCSO: A statistically or graphically estimated concentration that is expected to be lethal to 50% of
a group of organisms under specified conditions
Margin of Exposure: The ratio of the critical NOAEL to the expected human exposure level.
Mechanism of Action: The complete sequence of biological events that must occur to produce
a toxic effect.
Metal and Metalloid: An element that acts as a cation in chemical reactions, forms a base with
the hydroxyl radical, and can replace the hydrogen of an acid to form a salt. Antimony, arsenic,
molybdenum, selenium, and vanadium generally occur as oxyanions in waters and soils, and not
as cations. These elements are sometimes described as metalloids.
MIKE21 model. Modeling tool for rivers, estuaries, and coastal waters for two-dimensional free-
surface flows.
MINTEQ model. Computer-based speciation model for modeling metal partitioning in
discharge; the model usually must be run in connection with another fate and transport model.
MINEQL model: Computer-based speciation model.
MINTEQA2 model: An equilibrium speciation model that can be used to calculate the
equilibrium composition of dilute aqueous solutions in the laboratory or in natural aqueous
systems (updates MINTEQ).
Mode of Action: A less-detailed description of the mechanism of action in which some, but not
all, of the sequence of biological events leading to a toxic effect is known.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
6-4
-------
NICA model. Non-Ideal Competitive Absorption model. Computer program for studies of
metals in soil moisture.
Nonthreshold Effect; An effect for which it is assumed that there is no dose, no matter how
low, for which the probability of an individual's responding is zero.
Organometattics. Compounds that have a metal/metalloid-carbon bond.
Physiologically-Based Pharmacokinetic (PBPK) Model: A model that estimates the dose to a
target tissue or organ by taking into account the rate of absorption into the body, distribution
between target organs and tissues, metabolism, and excretion.
PHREEQC model. Speciation and reaction path calculations for freshwaters and brines.
Pollution-Induced Community Tolerance (or PICT): A tool to assess effects of pollutants on
ecological communities by comparing responses of communities collected from polluted and
reference sites to contaminant exposures under controlled conditions. An increase in community
tolerance at a polluted site that results from the elimination of sensitive species is considered
evidence that this restructuring was caused by the pollutant.
(
Receptors: Organisms, populations, or ecosystems that are exposed to a contaminant or other
stressor.
Recommended Dietary Allowance (or EDA): The dietary level of intake of essential nutrients
considered to be adequate to meet the known nutritional
REDEQL2 model: Chemical reaction model; computer programs [REDEQL and MINEQL
series] to calculate chemical equilibrium in complex systems, including natural waters and
manmade chemical systems.
Reference Site: A relatively uncontaminated site used for comparison to contaminated sites in
environmental monitoring studies, often incorrectly referred to as a control.
RESRAD model. Family of computer risk codes developed to calculate site-specific RESidual
RADioactive material guidelines as well as radiation dose and excess lifetime cancer risk to a
chronically exposed on-site resident.
Reference Concentration (orRfC). An estimate (with uncertainty spanning perhaps an order of
magnitude) of a continuous inhalation exposure to the human population (including sensitive
subgroups) that is likely to be without an appreciable risk of deleterious effects during a lifetime.
It can be derived from aNOAEL, LOAEL, or benchmark concentration, with uncertainty factors
generally applied to reflect limitations of the data used.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
6-5
-------
Reference Dose (orR/D): An estimate (with uncertainty spanning perhaps an order of
magnitude) of a daily oral exposure to the human population (including sensitive subgroups) that
is likely to be without an appreciable risk of deleterious effects during a lifetime. It can be
derived from a NOAEL, LOAEL, or benchmark dose, with uncertainty factors generally applied
to reflect limitations of the data used.
Regression Analysis: Analysis of the functional relationship between two variables; the
independent variable is described on the X axis and the dependent variable is described on the Y
axis (i.e. the change in Y is a function of a change in X).
i
Relative Absorption Factor (or RAF): Similar to the RBA, the RAF more specifically refers to
the fraction or percentage of a metal that is absorbed across a biological membrane. The RAF is
one of the more common measures of uptake of metals into the body from environmental
exposure media.
Relative Bioavattabttity (or RBA): The ratio (fraction or percentage) of the amount of a metal
substance of interest that is adsorbed or absorbed under defined conditions (e.g., metal salt type,
specified vehicle or matrix, differing test doses, different physiological states of the receptor) as
compared to a reference metal substance tested under standard conditions.
Risk: The expected frequency or probability of undesirable effects resulting from exposure to
known or expected stressors.
Risk Assessment: Qualitative or quantitative evaluation of the risk posed to human health and/or
the environment by the actual or potential presence or release of hazardous substances, pollutants
or contaminants.
Risk Characterization: A phase of risk assessment that integrates the results of the exposure and
effects analyses to evaluate the likelihood of adverse effects associated with exposure to the
stressor. The ecological significance of the adverse effects is discussed, including consideration
of the types and magnitudes of the effects, their spatial and temporal patterns, and the likelihood
of recovery.
Secondary Effect. An effect where the stressor acts on supporting components of the ecosystem,
which in turn have an effect on the ecological component of interest (synonymous with indirect
effects; compare with definition for primary effect).
Sediment: Particulate material lying below water.
Simultaneously extracted metals (SEM). Divalent metals, commonly cadmium, copper, lead,
mercury, nickel, and zinc, that form less soluble sulfides than do iron or manganese and are
solubilized during the acidification step (0.5m HC1 for 1 hour) used in the determination of acid
volatile sulfides in sediments.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
6-6
-------
Spedation: Refers to the distribution of an element among defined chemical species, which are
the specific form of an element defined as to isotopic composition, electronic or oxidation state,
and/or complex or molecular structure.
Stressor: Any physical, chemical, biological entity that can induce an adverse
Synergism: When the effect of the combination is greater than that suggested by the component
toxic effects.
Susceptibility: Increased likelihood of an adverse effect, often discussed in terms of relationship
to a factor that can be used to describe a human subpopulation (e.g. life stage, demographic
feature, or genetic characteristic).
/
Surface Water: Surface water is all water naturally open to the atmosphere, such as rivers,
lakes, reservoirs, streams, and seas.
Susceptible Subgroups'. May refer to life stages, for example, children or the elderly, or to other
segments of the population, for example, asthmatics or the immune-compromised, but are likely
to be somewhat chemical-specific and may not be consistently defined in all cases.
Threshold Effect: An effect for which there is some dose below which the probability of an
individual's responding is zero.
Tolerance: The ability of an organism to maintain homeostasis under a variety of environmental
conditions, such as variable metal concentrations.
i
Toxicity: Deleterious or adverse biological effects elicited by a chemical, physical, or biological
agent.
Toxicodynamics: The determination and quantification of the sequence of events at the
cellular and molecular levels leading to a toxic response to an environmental agent (also called
pharmacodynamics).
Toxicokinetics: The determination and quantification of the time course of absorption,
distribution, biotransformation, and excretion of chemicals (also called pharmacokinetics).
Trophic Levels: A functional classification of taxa within a community that is based on feeding
relationships (e.g., aquatic and terrestrial green plants make up the first trophic level and
herbivores make up the second).
WATEQ4F model: Program model for calculating speciation of major, trace, and redox elements
in natural waters.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
6-1
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
7. REFERENCES
Abemathy, CO; Cantilli, R; Du, JT. (1993) Essentiality versus toxicity; some considerations in the risk assessment
of essential trace elements. In: Saxena, J, ed. Hazard assessment of chemicals. Washington, DC: Taylor and Francis,
pp. 81-113.
Adams, WJ. (2000) Hazard identification for metals and metal compounds: an alternative to the U.S. EPA strategy
for identifying hazards associated with metals and metal substances. Presentation at the expert workshop: state of the
science regarding PBT concepts and metals and metal compounds. U.S. EPA, EPRI, ICA, ILZRO, NiPERA and
ICME. January 19. Arlington, VA.
Adams, SM; Crumby, WD; Greeley, MS; et al. (1992) Relationships between physiological and fish population
responses in a contaminated stream. Environ ToxicolChem 11:1549-1557. .
Adams, WJ; Conard, B; Ethier, G; et al. (2000) The challenges of hazard identification and classification of
insoluble metals and metal substances for the aquatic environment. Hum Ecol Risk Assess 6:1019-1038.
Adams, WA; Brix, KV; Edwards, M; et al. (2003) Analysis of field and laboratory data to derive selenium toxicity
thresholds for birds. Environ Toxicol Chem 22(9):2020-2029.
AIHA (American Industrial Hygiene Association). (2003) The occupational environment: its evaluation, control,
and management. 2nd Edition. Ed. Salvatore Dinardi. Fairfax, VA; AIHA Press, American Industrial Hygiene
Association: pp 37-38,356.
Allen, HE. (2002) Bioavailability of metals in terrestrial ecosystems: Importance of partitioning for bioavailability to
invertebrates, microbes and plants. ISBN 1-880611 -46-5. Society of Environmental Toxicology and Chemistry.
Pensacola, FL: SETAC Press. 158 pp.
Allison, JD; Brown, DS; Novo-Gradac, KJ. (1991) MINTEQA2/PRODEFA2, A geochemical assessment model for
environmental systems: version 3.0, user's manual, USEPA ERL ORD, Athens, GA. EPA/600/3-91/021.
Alloway, BJ. (1995) Metals in soils, 2nd ed. New York: Blackie Academic & Professional, Chapman & Hall.
Ambrose, RB; Wool, TA; Martin, JL. (1993) The water quality analysis simulation program, WASPS. Version 5.10.
Part A: Model documentation. U.S. Environmental Protection Agency, Athens, GA.
Ambrose, T.M., M. Al-Lozi, and M.G. Scott. (2000) Bone lead concentrations assessed by in vivo x-ray
fluorescence. Clin. Chem. 46:1171-1178.
Andersen, M.E, H.J. Clewell, M.L. Gargas, F.A. Smith, and R.H. Reitz. (1987) Physiologically based
pharmacokinetics and the risk assessment process for methylene chloride. Toxicol. Appl. Pharmacol. 87:185-205.
Andersen, ME. (1995) Development of physiologically based pharmacokinetic and physiologically based
pharmacodynamic model for applications in toxicology and risk assessment. Toxicol Lett 79:35
Ankley, GT; Phipps, GL; Leonard, EN; et al. (1991) Acid-volatile sulfide as a factor mediating cadmium and nickel
bioavailability in contaminated sediments. Environ Toxicol Chem 10:1299-1307.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-1
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
Ankley, GT; Mattson, V; Leonard, E; et al. (1993) Predicting the acute toxicity of copper in freshwater sediments:
evaluation of the role of acid volatile sulfide. Environ Toxicol Chem 12:312-320.
Ankley, GT; Phipps, GL; Leonard, EN; et al. (1991) Acid-volatile sulfide as a factor mediating cadmium and nickel
bioavailability in contaminated sediments. Environ Toxicol Chem 10:1299-1307.
Ankley, GT; Di Toro, DM; Hansen, DJ; et al. (1996) Technical basis and proposal for deriving sediment quality
criteria for metals. Environ Toxicol Chem 15:2056-2066.
Antonovics, J; Bradshaw, AD; Turner, RG. (1971) Metal tolerance in plants. Adv Ecol Res 7:1-85.
ANZECC and ARMCANZ (Australian and New Zealand Environment and Conservation Council and Agriculture
and Resource Management Council of Australia and New Zealand). (2000) Australian and New Zealand Guidelines
for Fresh and Marine Water Quality, vols. 1 and 2.
Archer, D; Emerson, S; Reimers, C. (1989) Dissolution of calcite in deep-sea sediments: pH and oxygen
microelectrode results. Geochim Cosmochim Acta 53:2831-2845.
Arnold, PT; Kapustka, LA. (1993) Comparative uptake kinetics and transport of cadmium and phosphate inPhleim
pratense-Glomus deserticolum associations. Environ Toxicol Chem 12:177-186.
Arnold, M.L., F.E. McNeill, and D.R. Chettle. (1999) The feasibih'ty of measuring manganese concentrations in
human liver using neutron activation analysis. Neurotoxicology 20:407-412.
ATSDR (Agency for Toxic Substance and Disease Registry). (1995) Multiple lead and cadmium exposure study
with biological markers incorporated. Atlanta, GA.
ATSDR (1999a) Toxicolbgical profile for Lead. Atlanta, GA
ATSDR (1999b) Toxicological profile for Mercury. Atlanta, GA.
ATSDR (2003) Toxicological profile for Nickel, Draft for Public Comment, Atlanta, GA.
ATSDR (2004) Evaluation of the toxicology of chemical mixtures commonly found at hazardous waste sites. Draft
Atlanta, GA.
AWWA (American Water Works Association). (1999) Water Quality and Treatment A Handbook of Community
Water Supplies. 5th Edition. 1233pp. ISBN 0-07-001659-3.
Azar, A; Snee, RD; Habibi, K. (1975) An epidemiologic approach to community air lead exposure using personal air
samplers. In: Griffin, TB; Knelson, JH, eds. Lead. Stuttgart: Georg Thieme Pub; pp. 254-290.
Azenha, M; Vasconcelos, MT; Cabral, JPS. (1995) Organic ligands reduce copper toxicity in Pseudomonas syringae.
Environ Toxicol Chem 14:369-373.
Baas-Becking, LGM; Kaplan, IR; Moore, D. (1960) Limits of the natural environment in terms of pH and oxidation-
reduction potentials. J. Geol. 68:243-284.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-2
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
Babukutty, Y; Chacko, J, (1995) Chemical partitioning and bioavailability of lead and nickel in an estuarine system.
Environ Toxicol Chera 14:427-434.
Bailey, RG. (1998) Ecoregions: the ecosystem geography of the oceans and continents. New York: Springer.
Bailey, MR; Roy, M. (1 994) Clearance of particles from the respiratory tract Annals of the ICRP (Publication No.
66)24(1-3):301-413.
Ball, JW; Runkel, RL; Nordstrom, DK. (1999) Transport modeling of reactive and nonreactive constituents from
Summittville, Colorado: Preliminary results from the application of the OTIS/OTEQ model to the Whitman
Fork/Alamosa River System. In: Proc. Technical Meeting, Charleston, SC, Mar. 8-12, 1999. Vol. 1 of 3.
Contamination from hard-rock mining. U.S. Geological Survey Water Resources Investigations Report 99-401 8A.
Ballatori, N. (2002) Transport of toxic metals by molecular mimicry. Environ Health Perspect 1 10(Suppl.
Barak, P. (1999) Essential elements for plant growth. Department of Soil Science, University of Wisconsin-Madison.
Available at: http://www.soUs. wisc.edu/~barak/soilscience326/essentl.htm.
Barbour, MT; Plafkin, JL; Bradley, BP; et al. (1992) Evaluation of EPA' s rapid bioassessment benthic metrics:
metric redundancy and variability among reference stream sites. Environ Toxicol Chem 1 1 :437-449.
Barltrop, D. (1966) The prevalence of pica. Am J Dis Child 1 12:1 16-123.
\
Barltrop, D; Meek, F. (1979) Effect of particle size on lead absorption from the gut. Arch Environ Health
34:280-285.
Barnes RM. (1 990) Childhood soil ingestion: how much dirt do kids eat? Anal Chem 62:1023-1 033.
Basta, NT; Pantone, DJ; Tabatabai, MA. (1993) Path analysis of heavy metal adsorption by soil. Agron J
85:1054-1057.
Beck, BD; Mattuck; RL; Bowers, TS; et al. (2001) The development of a stochastic physiologically-based
pharmacokinetic model for lead. Sci Total Environ 274: 1 5-1 9 .
Beckett, PHT. (1989) The use of extractants in studies on trace metals in soil, sewage sludges, and sludge-treated
soils. Adv Soil Sci 9:143-176.
Bell, PF; Adamu, CA; Mulchi, CL; Mclntosh, M; Chaney, RL. (1988) Residual effects of land applied municipal
sludge on tobacco. I: Effects on heavy metals concentrations in soils and plants. Tob. Sci. 32:33-38.
Benjamin, MM; Leckie, JO. (1981) Competitive adsorption of Cd, Cu, Zn and Pb on amorphous iron hydroxides. J.
Colloid Interface Sci. 83:410-419.
Berding, V; Schwartz, S; Matthies, M. (1999) Visualisation of the complexities of BUSES. Environ Sci Pollut Res
6:37-43.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-3
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
Berglund, M., A. Askesson, B. Nermell, and M. Vahter. (1994) Intestinal absorption of dietary cadmium in women
depends on body stores and fiber intake. Environ. Health Perspect. 102:1058-1065.
Berry, WJ; Hansen, DJ; Mahony, JD; et al. (1996) Predicting the toxicity of metal-spiked laboratory sediments using
acid-volatile sulfide and interstitial water normalizations. Environ Toxicol Chem 15:2067-2079.
Beyer, WN; Heinz, OH; Redmon-Norwood, AW. (1996) Environmental contaminants in wildlife: intepreting tissue
concentrations. Boca Raton, FL: Lewis Publishers.
1
Beyer, WN; Audet, DJ;, Morton, A; et al. (1998) Lead exposure of waterfowl ingesting Coeur d'Alene River Basin
sediments. J EnvironQual 27:1533-1538.
Biesinger, KE; Christensen, GM; Fiandt, JT. (1986) Effects of metal salt mixtures on Daphnia magna reproduction.
Ecotoxicol Environ Saf 11:9-l 4.
Binder, S; Sokal, D;Maugham, G. (1986) Estimating soil ingestion: the use of tracer elements in estimating the
amount of soil ingestion by young children. Arch Environ Health 41:341-345.
Bingham, FT; Sposito, G; Strong, IE. (1986) The effect of sulfate on the availability of cadmium. Soil Sci
141:172-177.
Bjerregaard, P., S. Top9uoglu, N.S. Fisher and S.W. Fowler. (1985) Accumulation and retention of 2}7Pu and 2<1Am
in the mussel Mytilus edulis. Mar. Ecol. Prog. Ser. 21:99-111.
Blanck, H; Wangberg, SA; Molander, S. (1988) Pollution-induced community tolerance: a new ecotoxicological
tool. In: Cairns, J, Jr; Pratt, JR,eds. Functional testing of aquatic biota for estimating hazards of chemicals. ASTM
STP 998. American Society for Testing and Materials, Philadelphia, PA; pp. 219-230.
Blumberg, AF; Mellor, GL. (1987) A description of a three-dimensional coastal ocean circulation model. In: Heaps,
N, ed. Three-dimensional coastal ocean models. American Geophysical Union, Washington, DC.
Blus, LJ; Henny, CJ; Hoffman, DJ; et al. (1991) Lead toxicosis in tundra swans near a mining and smelting complex
in northern Idaho. Arch Environ Contam Toxic 21:549-555.
Bodek, IB; Lyman, WJ; Reehl, WF; Rosenblatt, DH, eds. (1988) Environmental inorganic chemistry: Properties,
processes, and estimation methods. SETAC Spec. Publ. Ser. New York: Pergamon Press.
Bonn, HL; McNeal, BL; O'Connor, GA. (1985) Soil chemistry, 2nd ed. New York: John Wiley & Sons.
Bogumil, R; Faller, PI; Binz, PA; Vasak, M; Charnock, JM; and Gamer, CD. (1998) Structural characterization of
Cu(I) and Zn(II) sites in neuronal-growth-inhibitory factor by extended X-ray absorption fine structure (EXAS).
Eur.J. Biochem. 255:172-177.
Bolger, PM; Yess, NJ; Gunderson, EL; Troxell, TC; Carrington, CD. (1996) Identification and reduction of sources
of dietary lead in the United States. Food Additives Contam. 13:53-60.
Bossuyt, BT; Janssen, CR. (2002). Influence of bioavailability on acclimation of Daphnia magna to copper.
Mededelingen Faculteit Landbouwkundige en Toegepaste Biologische Wetenschappen 67(4). 2002. 133-136.
f
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-4
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
Bowers, TS; Mattuck, RL. (2001) Further comparisons of empirical and epidemiological data with predictions of the
integrated exposure uptake biokinetic model for lead in children. Hum Ecol Risk Assess 7(6): 1699-1713.
Bowers, TS; Beck, BD; Karam, HS. (1994) Assessing the relationship between environmental lead concentrations
and adult blood lead levels. Risk Anal 14(2): 183-189.
Bradham, KD. (2002) Effect of soil properties on the bioavailability and toxicity of metals to Eisenia andrei. Ph.D.
dissertation. Dept. of Zoology, Oklahoma State University.
Brady JA, Gemmiti-NunnD, Polan AK, Mitchell D, Weil R, Vianna NJ. (1980) The relationship of dental practice .
characteristics to blood mercury levels. NY State Dental J 46:420-424.
Brix, KV; DeForest, DK. (2000) Critical review of the use of bioconcentration factors for hazard classification of
metals and metal compounds. Parametrix, Inc., Kirkland, WA.
Brooks, RR. (1972) Geobotany and biogeochemistry in mineral exploration. New York: Harper & Row.
Brown, DS; Allison, JD. (1987) MINTEQA1. An Equilibrium Metal Speciation Model. User's Manual. U.S.
Environmental Protection Agency, Athens, GA. EPA-600-3-87-012.
*
Brown, PL; Markich, SJ. (2000) Evaluation of the free ion activity model of metal-organism interaction: extension of
the conceptual model. Aquat Toxicol 51:177-194.
Brown, GE, Jr; Foster, AL; Ostergren, JD. (1999a) Mineral surfaces and bioavailability of heavy metals: a
molecular-scale perspective. Proc Natl Acad Sci 96:3388-95.
Brown, SL; Chancy, RL; Berti, W. (1999b) Field test of amendments to reduce the in situ availability of soil lead. In:
Wenzel, WW; Adriano, DC; Alloway, B; et al., eds. Abstracts of the 5th international conference on biogeochemistry
of trace elements. International Society for Trace Element Research. July 11-15, Vienna, Austria.
Brown, SL; Henry, CL; Compton, H; et al. (2000) Using municipal biosolids in combination with other residuals to
restore metal-contaminated mining areas. In: Proceedings of a symposium on mining, forest and land restoration: the
successful use of residual/biosolids/organic matter for reclamation activities. Rocky Mountain Water Environment
Association. July 17-20, Denver, CO.
Brumbaugh, WG; Ingersoll, CG; Kemble, NE; May, TW; Zajicek, JL. (1994) Chemical characterization of
sediments and pore water from the Upper Clark Fork River and Milltown Reservoir, Montana. Environ Sci Chem
13:1971-1983.
Bullock, O.R. (2000) Current methods and research strategies for modeling atmospheric mercury. Fuel Processing
Technology 65-66:459-471; EPA/600/J-00/261. ~
Bullock, O.R. (2000) The importance of emissions speciation to the atmospheric transport and deposition of
mercury. Presented at: EPA Conference on Assessing and Managing mercury from Historic and Current Mining
Activities, San Francisco, CA, November 28-30, 2000.
Bullock, O.R., and Brehme, K.A. (2002) Atmospheric mercury simulation using the CMAQ model: formulation
description and analysis of wet deposition results. Atmospheric Environment 36 (13):2135-2146; EPA/600/J-03/227.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-5
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
Burgess, RM; Scott, KJ. (1992) The significance of in-place contaminated marine sediments on the water column:
processes and effects. In Burton GA, Jr., ed. Sediment toxicity assessment Boca Raton, FL: Lewis Publishing; pp
129-154.
Cabana, G; Tremblay, A; Kalff, J; et al. (1994) Pelagic food chain structure in Ontario lakes: a determinant of
mercury levels in lake trout (Salvelinus namaycush). Can J Fish Aquat Sci 51:381-389.
Calabrese, EJ; Baldwin, LA. (1994) A toxicological basis to derive a generic interspecies uncertainty factor. Environ
Health Perspect 102:14-17.
Calabrese, EJ; Stanek, EJ. (1993) Soil pica:nNot a rare event. J Environ Sci Health (Part A) 28:373-384.
Calabrese, EJ; Barnes, R; Stanek, EJ, III; et al. (1989) How much soil do young children ingest: an epidemiologic
study. Regul Toxicol Pharmacol 10:123-137.
Calabrese, EJ; Stanek, EJ; Gilbert, CE; et al. (1990) Preliminary adult soil ingestion estimates: results of a pilot
study. Regul Toxicol Pharmacol 12:88-95.
Calabrese, EJ; Stanek, EJ, III; Pekow, P; et al. (1997) Soil ingestion estimates for children residing on a Superfund
site. Ecotoxicol Environ Safety 36:258-268.
Calamari, D; Alabaster, JS. (1980) An approach to theoretical models in evaluating the effects of mixtures of
toxicants in the aquatic environment. Chemosphere 9:533-538.
Campbell, PGC. (1995) Interactions between trace metals and aquatic organisms:a critique of the free-ion activity
model. In: Tessier, A; Turner, DR, eds. Metal speciation and bioavailability in aquatic systems. Chichester, UK:
Wiley; pp. 45-102.
Campbell, PGC; Hansen, HJ; Dubreuil, B; et al. (1992) Geochemistry of Quebec North Shore salmon rivers during
snowmelt: organic acid pulse and aluminum mobilization. Can J Fish Aquat Sci 49:1938-1952.
Campbell, PGC; Hontela, A; Rasmussen, JB; Giguere, A; Gravel, A; Kraemer, L; Kovesces, J; Lacrpix, A;
Levesque, H; Sherwood, G. (2003) Differentiating between direct (physiological) and food-chain mediated
(bioenergetic) effects on fish in metal-impacted lakes. Human Ecol Risk Assess 9: 847-866.
Capar, S.G. and W.C. Cunningham. (2000) Element and radionuclide concentrations in food: FDA Total Diet
Study 1991-1996. J. AOAC International. 83:157-177.
Carbanaro, R. (1999). Modeling metal sulfide fluxes from sediments. MS thesis. Department of Environmental
Engineering. Manhattan College, Riverdale, NY.
Carignan, R; Tessier, A. (1985). Zinc deposition in acid lakes: the role of diffusion. Science, 228,1524-1-526.
Carlisle, JC; Wade, MJ. (1992) Predicting blood lead concentrations from environmental concentrations. Regul
Toxicol Pharmacol 16:280-289.
Carlisle, J., J. Christopher, B. Davis, K. Klein, B. Renzi, and M. Wade. (2000) Updated version of the California
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-6
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
EPA lead risk assessment spreadsheet model for predicting blood lead in children and adults. Available at:
http://www.dtsc.ca.gov/docs/sppt/herd/ledspred.html
1
Carlson, AR; Phipps, GL; Mattson, VR; et al. (1991) The role of acid-volatile sulfide in determining cadmium
bioavailability and toxicity in freshwater sediments. Environ Toxicol Chem 10:1309-1319.
Casas, AM; Crecelius, EA. (1994) Relationship between acid volatile sulfide and the toxicity of zinc, lead and
copper in marine-sediments. Environ Toxicol Chem 13:529-536.
Cataldo, DA; Wildung, RE. (1978) Soil and plant factors influencing the accumulation of heavy metals by plants.
Environ Health Perspect 27:149-159.
Cederberg, GA; Street, RTL; Leckie,JO. (1985) A groundwater mass transport and equilibrium chemistry model for
multicomponent systems. Water Resour. Res. 21:1095-1104.
CDC (Center for Disease Control and Prevention) (1991). Preventing lead poisoning in young children. Atlanta,
GA, U.S. Department of Health and Human Services, Public Health Service, Centers for Disease Control and
Prevention.
CDC (2003) National Report on Human Exposure to Environmental Chemicals: Second Report) available at
www.cdc. gov/exposurereport
Chaney, RL. (1980) Health risks associated with toxic metals in municipal sludge. In: Bitton, G.,
B.L. Damron, G.T. Edds, and J.M. Davidson, eds. Sludge—Health risks of land application. AnnArbor, Ml: Ann
Arbor Sci. Publ. Inc., pp. 59-83.
Chaney, RL; Mielke, HW; Sterrett, SW. (1988) Speciation, mobility and bioavailability of soil lead. In: Davis, BE;
Wixon, BG, eds. Lead in soil: issues and guidelines. Northwood: Science Reviews; pp. 104-129.
Chaney, RL; Ryan, JA. (1993) Heavy metals and toxic organic pollutants in MSW-composts: Research results on
phytoavailability, bioavailability, etc. In: Hoitink, A.J. and H.M. Keener, eds. Science and engineering of
composting: Design, environmental, microbiological and utilization aspects. Columbus, OH: Ohio State University,
pp. 451-506.
Chaney, RL; Ryan, JA. (1994) Risk-based standards for arsenic, lead and cadmium in urban soils. Frankfurt,
Germany: DECHMEA.
Chaney, RL; Mielke, HW; Sterrett, SW. (1988) Speciation, mobility and bioavailability of soil lead. In: Davis, BE;
Wixon, BG, eds. Lead in soil: issues and guidelines. Northwood: Science Reviews; pp. 104-129.
Chaney, RL; Brown, SL; Angle, JS; et al. (2000) In situ remediation/reclamation/
restoration of metals contaminated soils using tailor-made biosolids mixtures. In: Proceedings of mining, forest and
land restoration symposium/workshop. Rocky Mountain Water Environment Association Biosolids Committee. July
17-19, Golden, CO.
Chang, SI; Reinfelder, JR. (2000) Bioaccumulation, subcellular distribution and trophic transfer of copper in a
coastal marine diatom. Environ. Sci. Technol. 34:4931-4935.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-7
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
Chapman, OA. (1987) Establishing sediment criteria for chemicals: regulatory perspective. In: Dickson, KL; Maki,
AW; Brungs, WA, eds. Fate and effects of sediment-bound chemicals in aquatic systems.New York: Pergamon
Press; pp 355-376.
Chapman, PM; Wang, F. Undated, Issues in ecological risk assessments of inorganic metals and metalloids.
Available at: http://www.uoguelph.ca/cntc/fyes/era-metals.pdf
Chapman, PM; Wang, F. (2000) Issues in ecological risk assessment of inorganic metals and metalloids. Hum Ecol
Risk Assess 6(6): 965-988.
Chapman, PM; Dias, GM (2003) [DRAFT] Topic: Assessing Atmospheric Metals Deposition. ORD Order No. 3 W-
0748-NASX.
Chapman, PM; Wang, F; Janssen, CR; et al. (2003) Conducting ecological risk assessments of inorganic metals and
metalloids: current status. Human Ecol Risk Assess 9:641-697.
Charlatchka, R; Cambier, P; Bourgeois, S. (1997) Mobilization of trace metals in contaminated soils under
anaerobic conditions. In: Prost, R, ed. Contaminated soils. Proceedings of the 3rd international conference on
biogeochemistry trace elements. May 15-19, Paris, France.
Chen, CY; Folt, CL. Bioaccumulation and diminution of arsenic and lead in a freshwater food web. Environmental
Science & Technology. 34(18):3878-3884.
Chen, Z; Mayer, LM. (1998) Mechanisms.of Cu solubilization during deposit feeding. Environ Sci Technol
32:770775.
Cherian, MD; Goyer, RA. (1995) Part Three, Chapter 9, Section A, In: Berthon, G. ed. Handbook of metal-ligand
interactions in biological fluids, Vol. 1. New York: Marcel Dekker, Inc., pp. 648-654.
Chiodo LM; Jacobson, SW; Jacobson, JL. (2004). Neurodevelopmental effects of postnatal lead exposure at very
low levels. Neurotoxicol Teratol. 26(3):359-71.
Choudhury, H; Harvey, T; Thayer, WC; et al. (2001) Urinary cadmium elimination as a biomarker of exposure for
evaluating a cadmium dietary exposure-biokinetics model. J Toxicol Environ Health 63:101-130.
Cia, L; Tsiapalis, G; Cherian, MG. (1998) Protective role of zinc metallothionein on DNA damage in vitro by ferric
nitriloacaetate (Fe-NTA) and ferric salts. Chem-Biol Interact 115:141-151.
Clark TP, Chieffo C, Huhn JC, Nimz EL, Wang C, Boy MG. (2003) The steady-state pharmacokinetics and
bioequivalence of carprofen administered orally and subcutaneously in dogs. J Vet Pharmacol Ther. 2003
Jun;26(3): 187-92.
Clarke, R.H. (1995) ICRP recommendations applicable to the mining and minerals processing industries and to
natural sources. International Commission on Radiological Protection. Health Phys. 69:454-60.
Clarkson, TW. (1986) Effects—general principles underlying the toxic action of metals. In: Friberg, L; Nordberg,
GF; Vouk, V. eds. Handbook on the toxicology of metals, 2nd ed., Vol. 1. Amsterdam: Elsevier, pp. 85-127.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-8
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
Clarkson, TW. (1993) Molecular and ionic mimicry of toxic metals. Ann Rev Pharmacol Toxicol 32:545-571.
Clausing, P; Brunekreef, B; van Wijen, JH. (1987) A method for estimating soil ingestion by children. Int Arch
Occup Environ Health 59:73-82.
Clayton, CA; Pellizzari, ED; Whitmore, R; et al. (1999) National Human Exposure Assessment Survey (NHEXAS):
distributions and associations of lead, arsenic and volatile organic compounds in EPA Region 5. J Exp Anal Environ
Epidem 9:381-392.
Clayton, CA; Pellizzari, ED; Quackeriboss, JJ. (2002) National Human Exposure Assessment Survey: analysis of
exposure and routes for arsenic and lead in EPA Region 5. J Exp Anal Environ Epidem 12:29—43.
Clements, WH; Cherry, DS; VanHassel, JH. (1992) Assessment of the impact of heavy metals onbenthic
communities at the Clinch River (Virginia): evaluation of an index of community sensitivity. Can J Fish Aquat Sci
49:1686-1694.
Clewell, H J. (1995) The application of physiologically based pharmacokinetic modeling in human health risk -
assessment of hazardous substances. Toxicol Lett 79:207-217.
Cohen, J.T., B.D. Beck, T.S. Bowers, R.L. Bomschein, and E.J. Calabrese. (1998) An arsenic exposure model:
Probabilistic validation using empirical data. Hum. Ecol. Risk Assess. 4(2):341-377.
Conder, JM; Seals, LD; Lanno, RP. (2002) Method for determining toxicologically relevant cadmium residues in the
earthworm Eiseniafetida. Chemosphere 49:1-7.
Coombs, TL; George, SG. (1978) Mechanisms of immobilization and detoxication of metals in marine organisms.
In: McLusky, DS; Berry, JA, eds. Physiology and behavior of marine organisms. Oxford: Pergamon Press; pp.
179-187.
Crecelus, EA; Hardy, JT; Bobson, CI; et al. (1982) Copper bioavailability to marine bivalves and shrimp:
Relationship to cupric ion activity. Mar Environ Res 6:13-26.
Cress, WA; Throneberry, GO; Lindsey, DI..(1979) Kinetics of phosphorus absorption by mycorrhizal and
nonmycorrhizal tomato roots. PlanPhysiol 64:484-487. '
Crist, ME; Oberhoiser, K; Schwartz, D; et al. (1988) Interactions of metals and protons with algae. Environ Sci
Technol 22:755-760.
Crommentuijn, T; Doodeman, CJAM; Van der Pol, JJC; et al. (1994) Lethal body concentrations and accumulation
patterns determine time-dependent toxicity of cadmium in soil arthropods. Environ Toxicol Chem 13:1781-1789.
Crommentuijn, T; Doodeman, CJAM; Van der Pol, JJC: et al. (1997) Bioavailability and ecological effects of
cadmium on Folsomia Candida (Willem) in an artificial soil substrate as influenced by pH and organic matter. Appl
Soil Ecol 5:261-271.
Dabeka RW; McKenzie, AD. (1995) Survey of lead, cadmium, fluoride, nickel, and cobalt in food composites and
estimation of the dietary intakes of these elements by Canadian adults and children. Journal of AOAC International
78: 897-909.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-9
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
Dabeka RW; McKenzie, AD; Lacroix, GM; Cleroux, C; Bowe, S; Graham, RA; Conacher, HBS; Verdier, P. (1993)
Survey of arsenic in total diet food composites and estimation of the dietary, intake of arsenic by Canadian adults and
children. Journal of AOAC International 76:14-25.
Davison, AG; Taylor, A; Darbyshire, J. (1988) Cadmium fume inhalation and emphysema. Lancet 26:663-667.
Davidson, W; Zhang, H. (1994) In situ speciation measurements of trace components in natural waters using thin-
film gels. Nature 367:546-548.
Davis, S; Waller, P; Buschbom, JR; et al. (1990) Quantitative estimates of soil ingestion in normal children between
the ages of 2 and 7 years: population-based estimates using aluminum, silicon, and titanium as soil tracer elements.
Arch Environ Health 45:112-122.
Dawson, DC; Ballatori, N. (1995) Membrane transporters as site of action and routes of entry for toxic metals. In:
Goyer, RA; Cherian, MG, eds. Toxicology of metals: biochemical aspects. New York: Springer-Verlag; pp. 53-76.
Delos, CG; Richardson, WI; DePinto, JV; et al. (1984) Technical guidance manual for performing waste load
allocations: Book II. streams and rivers. U.S. Environmental Protection Agency, Office of Water, Washington, DC.
de Schamphelaere, KAC; Janssen, CR. (2002) A biotic ligand model predicting acute copper toxicity for Daphnia
magna: The effects of calcium, magnesium, sodium, potassium and pH. Environ. Sci. Techno! 36:48-54.
de Schamphelaere, KAC; Janssen, CR. (2004) Development and field validation of a biotic ligand model predicting
chronic copper toxicity to Daphnia magna. Environ. Toxicol. Chem. 23:1365-1375.
de Schamphelaere, KAC; Heijerick, DG; Janssen, CR. (2002) Refinement and Field validation of a biotic ligand
model predicting acute copper toxicity to Daphnia magna. Special issue: The biotic ligand model for
metals—current research, future directions, regulatory implications. Comp. Biochem. Physiol. C 133:243-
DeVoto, E; Yokel, RA. (1994) The biological speciation and toxicokinetics of aluminum. Environ Health Perspect
102:940-51.
Diamond, ML. (1995) Application of a mass balance model to assess in-place arsenic pollution. Environ. Sci.
Technol. 28(1): 29-42.
Diamond, ML; Mackay, D;etal. (1990) A model of the exchange of inorganic chemicals between water and '
sediments. Environ Sci Technol 24:713-722.
Diamond ML; Mudroch A. (1990) Review of techniques for quantifying the transfer of contaminants and nutrients
from bottom sediments. National Water Research Institute Contribution No. 90-43, National Water Research
Institute, Environment Canada, Burlington, Ontario, Canada.
Diamond ML; Mackay, D; Comett, RJ; Chant, L. (1990) Amodel of the exchange of inorganic chemicals between
water and sediments. Environ. Sci. Tech. 24: 713-722.
Dias, GM; Edwards, GC. (2003) Differentiating natural and anthropogenic sources of metals to the environment..
Human Ecol Risk Assess 9:699-721.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
' 7-10
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
Dill, LM; Heithaus, MR; Walters, CJ. (2003) Behaviorally mediated indirect interactions in marine communities and
their conservation implications. Ecology 84:1151-1157.
Di Toro, DM. (1984) Probability model of stream quality due to runoff. J Environ Engineer ASCE 110:607-628.
Di Toro, DM. (1987) Modeling the fate of toxic chemicals in surface waters. Short course notes from 8th annual
SETAC meeting. Pensacola, FL.
Di Toro, DM. (2001). Sediment flux modeling. New York; J. Wiley and Sons; pp 624.
Di Toro, DM; Fitzpatrick, JJ; Thomann, RV. (1981) Water quality analysis simulation program (WASP) and model
verification program (MVP). Hrydroscience Inc. for U.S. EPA. Grosse He Laboratory, Grosse lie, MI (revised in
1983).
Di Toro, DM; Mahony, JD; Hansen, DJ; et al. (1990) Toxicity of cadmium in sediments: the role of acid volatile
sulfide. Environ Toxicol Chem 9:1487-1502.
Di Toro, D M; Hallden, JA; et al. (1991) Modeling ceriodaphnia toxicity in the Naugatuck River: II. copper,
hardness and effluent interactions. Environ Toxicol Chem 10:261-274.
Di Toro, DM; Mahony, JD; Hansen, DJ; et al. (1992). Acid volatile sulfide predicts the acute toxicity of cadmium
and nickel in sediments. Environ Toxicol Chem 26:96-101.
Di Toro, DM; Mahony, JD; Hansen, DJ; et al. (1996) A model of the oxidation of iron and cadmium sulfide in
sediments. Environ Toxicol Chem 15:2168-2186.
Di Toro, DM; Hansen, DJ; McGrath, JA; et al. (1999) Predicting the toxicity of metals in sediments. In: Integrated
approach to assessing the bioavailability and toxicity of metals in surface waters and sediments. Presented to the
EPA Science Advisory Board, Office of Water, Office of Research and Development, Washington, DC. EPA-822-E-
99-001.
Di Toro, DM; Allen, H; Bergman, H; et al. (2000) Biotic ligand model of the acute toxicity of metals: I. technical
basis. Environ Toxicol Chem 20(10):2383-2396.
Di Toro, DM; Allen, HE; Bergman, HL; Meyer, JS; Paquin, PR; Santore, RC. (2001) A biotic ligand model of the
acute toxicity of metals. I: Technical basis. Environ. Toxicol. Chem. 20:2383-2396.
Di Toro, DM; McGrath,JA; Hansen, DJ; Berry, WJ; Paquin, PR; Mathew, R; Wu, KB; Santore, RC. (2004)
Predicting sediment metal toxicity using a sediment biotic ligand model: (1) single metals. Submitted.
Dlugaszek, M; Fiejka, MA; Graczy, A; et al. (2000) Effects of various aluminum compounds given orally to mice on
Al tissue distribution and tissue concentrations of essential elements. Annu Rev Pharmacol Toxicol 86:135-13 9.
Domingo, JL. (1995) Reproductive and developmental toxicity of aluminum: a review. Neurotoxicol Teratol
17:515-521.
Duggan, MJ; Inskip, MJ. (1985) Childhood exposure to lead in surface dust and soil: a community health problem.
Public Health Rev 13:1-54.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-11
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
Duggan, Ml; Inskip, MJ; Rundle, SA; et al. (1985) Lead in playground dust and on the hands of schoolchildren. Sci
Total Environ 44(1):65-80.
Dzombak, DA; Ali, MA. (1993) Hydrochemical modeling of metal fate and transport in freshwater environments.
Watr Pollut Res J Can 28(1):7-50.
Dzombak, DA; Morel, FMM. (1990) Surface complexation modeling: Hydrous ferric oxide. New York: Wiley
InterScience.
EC (European Commission). (1996) EUSES documentation: the European Union system for the evaluation of
substances. National Institute of Public Health and the Environment (RIVM), The Netherlands.
Efroymson, RA; Sample, BE; Suter GW, II. (2001) Bioaccumulation of inorganic chemicals from soil by plants:
regressions of field data. Environ. Toxicol. Chem 20:2561-2571.
Egan, SK; Tao, SS; Pennington, JA; et al. (2002) U.S. Food and Drug Administration's total diet study: intake of
nutritional and toxic elements, 1991-1996. Food Addit Contam 19:103-125.
Ellis, KJ; Vartsky, D; Zanzi, I; et al. (1979) Cadmium: in vivo measurement in smokers and nonsmokers. Science
205:323-325.
Enserink, EL; Maas-Diepeveen, JL; Van Leeuwen, CJ. (1991) Combined effects of metals: an ecotoxicological
evaluation. Wat Res 25:679-687.
Epstein, E. (1965) Mineral metabolism. In: Bonner, J; Vamer, JE,,eds. Plant biochemistry. London: Academic Press;
pp. 438-466.
Epstein, E. (1972) Mineral nutrition of plants: principles and perspectives. New York: John Wiley & Sons.
(EC Health and Consumer Protection) European Commission Health & Consumer Protection Directorate-General
Directorate C - Public Health and Risk Assessment C7 - Risk Assessment Brussels, C7/GF/csteeopAVFD/280504
D(04) Opinion of the Scientific Committee on Toxicity, Ecotoxicity and the Environment (CSTEE) On "The Setting
of Environmental Quality Standards for the Priority Substances included in Annex X of Directive 2000/60/EC in
Accordance with Article 16 thereof " Adopted by the CSTEE during the 43rd plenary meeting of 28 May 2004
EU (European Union) (1996). Technical Guidance Document in support of commision directive 93/67/EEC on risk
assessment for new notified substances and commission regulation (EC) No 1488/94 on risk assessment of existing
substances. Part II. Brussles, European Union. ISBN 92-827-8012-0.
Fairbrother, A. ed. (2002) Test Methods for Hazard Determination of Metals and Sparingly Soluble Metal
Compounds in Soils. SETAC Press, Pensacola, FL.
Fairbrother, A. (2003) Lines of evidence in ecological risk assessment. Human Ecol. Risk Assess. 9:1475-1491.
Fairbrother, A; Kapustka, LA. (1997) Hazard classification of inorganic substances in terrestrial systems: A
discussion paper. International Council on Metals and the Environment, Ottawa, Canada.
Fairbrother, A., L. Locke, and G. L. Hoff, eds. (1996) Non-infectious diseases of wildlife. 2nd Ed. Iowa State
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-12
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
University Press, Ames, I A.
Fan, W; Wang, W-X. (2001) Sediment geochemical controls on Cd, Cr, and Zn assimilation by the clam Ruditapes
philippinarum. Environ Toxicol Chem 20:2309-2317.
FargaSova, A. (2001) Winter third- to fourth-instar larvae of Chiranamus plumosus as bioassay tools for assessment
of acute toxicity of metals and their binary combinations. Ecotoxicol Environ Saf 48:1-5.
Fetmy, AR; Brown, SM; Qnishi, Y; et al. (1984) MEXAMS: the metals exposure analysis modeling system. U.S.
Environmental Protection Agency, Athens, GA. NTIS No. PB84 157155.
Fergusson, JE; Ryan, DE. (1984) The elemental composition of street dust from large and small urban areas related
to city type, source and particle size. Sci Total Environ 34:101-116.
Field, L; Macdonald, D; et al. (2002). Predicting amphipod toxicity from sediment chemistry using logistic
regression models. Environ Toxicol Chem 21(9):1993-2005.
Finney, DJ. (1947) Probit analysis: a statistical treatment of the sigmoidal response curve, Cambridge, UK:
Cambridge University Press.
FIU (Florida International University), Dietary Reference Intakes (DRIs)/Recommended Dietary Allowances
(RDAs) (2004) Dietary Reference Intakes for Older Adults (updated 03/19/04), compiled by the National Policy and
Resource Center on Nutrition and Aging, FIU, Miami, FL.
Fisher, NS. (1986) On the reactivity of metals for marine phytoplankton. Limnol Oceangr 31:443-447.
Fisher, NS; Teyssie, J-L; Fowler, SW; Wang, W-X. (1996) Accumulation and retention of metals in mussels from
food and water: a comparison under field and laboratory conditions. Environ. Sci. Technol. 30:3232-3242.
Fleming, D.E.B., D.R. Chettle, J.O. Wetmur, R.G. Desnick, J. Robin, D. Boulay, N.S. Richard, C.L. Gordon, and
C.E. Webber. (1998) Effect of the d-aminolevulinic dehydratase polymorphism on the accumulation of lead in bone
and blood in lead smelter workers. Environ. Res. 7,7:49-61.
FNB (Food and Nutrition Board). (2000) Dietary Reference Intakes for Vitamin A, Vitamin K, Arsenic, Boron,
Chromium, Copper, Iodine, Iron, Manganese, Molybdenum, Nickel, Silicon, Vanadium, and Zinc. Washington:
National Academies Press. Summary Pp 1-28.
Forstner, U. (1987) Sediment-associated contaminants: an overview of scientific basis for developing remedial
options. Hydrobiologia 149:221-246.
Forstner U. (1995a) In: Biogeodynamics of Pollutants in Soils and Sediments. W Salomons, WM Stigliani (Eds).
Springer, Berlin.
Forstaer, U. (1995b) Land contamination by metals: Global scope and magnitude of the problem. In: Allen, H. Et
al., eds. Metal speciation and contamination of soil. Boca Raton, FL: Lewis Publishers.
Fowler BA. ed. (1983) Biological and environmental effects of arsenic. New York: Elsevier Science Publishing
Company.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-13
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
Fowler, BA; Mahaffey, KR. (1978) Interactions among lead, cadmium and arsenic in relation to porphyrin excretion
patterns. Environ Health Perspect 25:87-90.
Foulkes, EC. 2000. Transport of toxic heavy metals across cell membranes. Proc. Soc. Exp. Biol. Med 223:234-40.
Fowler, B.A., and K.R. Mahaffey. 1978. Interactions among lead, cadmium and arsenic in relation to porphyrin
excretion patterns. Environ. Health Perspect. 25:87-90.
Foy, CD; Chaney, RL; White, MC. (1978) The physiology of metal toxicity in plants. Ann Rev Plant Physiol
29:511-566.
Franke, C; Studinger, G; Berger, G; et al. (1994) The assessment of bioaccumulation. Chemosphere 29:1501-1514.
Freeman, GB; Johnson, JD; Killinger, JM; et al. (1993) Bioavailability of arsenic in soil impacted by smelter
activities following oral administration in rabbits. Fund Appl Toxicol 21: 83-88,
Freeman, G.B., RA. Schoof, M.V. Ruby, A.O. Davis, J.A. Dill, S.C. Liao, C.A. Lapin, and P.D. Bergstrom. (1995)
Bioavailability of arsenic in soil and house dust impacted by smelter activities following oral administration in
cynomolgus monkeys. Fundam. Appl. Toxicol. 28:215-22.
Freeze, SA; Cherry, JA (1979) Groundwater. Engiewood Cliffs, NJ: Prentice Hall, Inc.
Friis, L; Petersson, L; Edling, C. (1998) Reduced cadmium levels in human kidney cortex in Sweden. Environ
Health Perspect 106:175-178.
George, SG. (1982) Subcellular accumulation and detoxication of metals in aquatic animals. In: Vemberg, WB;
Calabrese, A; Thurberg, FP; et al., eds. Physiological mechanisms of marine pollutant toxicity. Oxford: Academic
Press.
George, SG. (1990) Biochemical and cytological assessments of metal toxicity in marine animals. In: Heavy metals
in the marine environment. Furaess, RW; Rainbow, PS, eds. Boca Raton, FL: CRC Press; pp. 123-142.
George, E; Romheld, V; Marscbner, H. (1994) Contribution of Mycorrhizal Fungi to Micronutrient Uptake by
Plants. IN: Manthey, J.A., D.E. Crowley, and D.G. (eds). Biochemistry of metal micronutrients in the rhizosphere
Lewis Publishers, Boca Raton, FL pp.93-110.
Gerbel, T. (1998) Suppression of arsenic induced-chromosome mutagencity by antimony. Mut. Res. 412:213-218.
Gettier, SW; Martens, DC; Donohue, SI. (1985) Soybean yield response prediction from soil and tissue manganese
levels. AgronJ 77:63-67.
Goodrum, PE; Diamond, GL; Hassett, JM; et al. (1996) Monte Carlo modeling of childhood lead exposure:
development of a probabilistic methodology for use with the U.S. EPA IEUBK Model for Lead in Children. Hum
Ecol Risk Assess 2:681-708.
Gooddy, DC; Shand, P; Kinniburgh, DG; Van Riemsdijk, WH (1995) Field-based partition coefficients for trace
elements in soil solutions. Eur. J. Soil Scie. 46(2):265-286.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-14
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
Gorsuch, JW; Janssen, C; Lee, CM; et al. (eds). (2002) The biotic ligand model for metals: current research, future
directions, regulatory implications. Comp Biochem Physiol C 133:343.
Goyer, R.A. 1995. Transplacental transfer of lead and cadmium. In: Goyer, R.A. and M.G. Cherian, eds. Toxicology
of metals. New York: Springer-Verlag, pp. 1-13.
Goyer, RA; Clarkson, TM (2001) Toxic effects of metals. In: Klaassen, CD, ed. Casarett & DoulPs toxicology. New
York: McGraw-Hill; pp. 811-868.
Goyer, RA; Golub, M; Choudhury, H; Hughes, M; Kenyon, E; Stifelman, M. (2004) Issue paper on the human
health effects of metals. September 2004. Available at:
http://cfpubl.epa.gov/ncea/raf/recordisplay.cfm?deid=86119.
Grandjean, P., S.S. Brown, P. Reavey, and D.S. Young. (1994) Biomarkers of chemical exposure: State of the art.
Clin. Chem. 40:1360-1362.
Granick, J.L., S. Sassa, R.D. Granick, R.D. Levere, and A. Kappas. (1973) Studies in lead poisoning II: Correlation
between the ration of activated to inactivated d-aminolevulinic acid dehydrates of whole blood and the blood lead
level. Biochem. Med. 8:149-159.
Graziano, JH; Blum, CB; Lolacano, NJ; et al. (1996) A human in vivo model for the determination of lead
bioavailability using stable isotope dilution. Environ Health Perspect 104:176-179.
Greszta, J. (1982) Correlation between the content of copper, zinc, lead and cadmium in the soil and the content of
these metals in the seedlings of selected forest tree species. Frag Flor Geobot 28:(l):29-52.
Griffin, S; Goodrum, PE; Diamond, GL; et al. (1999) Application of a probabilistic risk assessment methodology to
a lead smelter site. Hum Ecol Risk Assess 5:845-868.
Gulson, BL; Pounds, JG; Mushak, P; et al. (1999a) Estimation of cumulative lead releases (lead flux) from the
maternal skeleton during pregnancy and lactation. J Lab Clin Med 13(6):631-640
Gulson, BL; Mahaffey, KR; Jameson, C; et al. (1999b) Impact of diet on lead in blood and urine in female adults and
relevance to mobilization of lead from bone stores. Environ Health Perspect 107(44):257-263.
Gulson, BL; Mahaffey, KR; Mizon, KJ; Korsch, M; Cameron, MA, Vimpani, G. (1995) Contribution of tissue lead
to blood lead in adult female subjects based on stable lead isotope methods. J. Lab. Clin. Med. 12:703-712.
Gunderson, BL (1995) FDA total diet study, July 1986-April 1991, dietary intakes of pesticides, selected elements,
and other chemicals. Journal of AOAC International, 1353-1363.
Gundersen, JK; Jorgensen, EL; Larsen, E; et al. (1992) Mats of giant sulfur bacteria on deep-sea sediments due to
fluctuating hydrothermal flow. Nature 360:454-455.
Guo, Z. (2002) Review of indoor emission source models. Overview Environ. Pollut 120:533-549.
Gupta, SK; Chen, KY. (1975) Partitioning of trace elements in selective chemical fractions of nearshore sediments.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-15
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
Environ Lett 10:129-158.
Gustafson, JP (2003) Visual MINTEQ, version 2.14; a Windows version of MNTEQA2, version 4.0.
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
(SUNY) at Stony Brook.
Hostynek, JJ; Hinz,RS; Lorence, CR; Price, M; Guy, RH. (1993). Metals and the skin. Grit. Rev. Toxicol. 23:171-
235.
Hostynek, JJ; Hinz, RS, Lorence, CR, Guy, RH. (1998) Human skin penetration by metal compounds. In Roberts,
MS; Walters, KA, eds. Dermal Absorption and Toxicity Assessment, Marcell Dekker Inc., New York.
Hook, S.E., and N.S. Fisher. (2001) Reproductive toxicity of metals in calanoid copepods. Mar. Biol.
138:1131-1140.
Hook, SE; Fisher, NS. (2002) Relating the reproductive toxicity of five ingested metals in calanoid copepods with
sulfur affinity. Marine Environ Res 53:161-174.
Hopkin, SP. (1989) Ecophysiology of metals in terrestrial invertebrates. London: Elsevier Applied Science.
Hostynek, J.J., Hinz, R.S, Lorence, C.R, and Guy, R.H. (1998) Human Skin Penetration by Metal Compounds. In
M.S. Roberts and K.A. Walters, eds. Dermal Absorption and Toxicity Assessment, Marcel Dekker, Inc., NewYork.
Hudson, RJM; Gherini, SA; Watras, CJ; et al. (1994) A mechanistic model for the biogeochemical cycle of mercury
in lakes. In: Watras, CJ; Huckabee, JW, eds. Mercury as a global pollutant. Chelsea, MI: Lewis Publishers.
Huertadiaz, MA; Carignan, R; Tessier, A. (1993). Measurement of trace-metals associated with acid volatile sulfides
and pyrite in organic fresh-water sediment. Environ Sci Technol 27:2367-2372.
Huheey, JE; Keiter EA; Keiter, RI. (1993) Inorganic chemistry—principles of structure and reactivity, 4m ed. New
York: Harper Collins College Publishers, 964 pp.
Hunt, A; Johnson, DL; Watt, JM; et al. (1992) Characterizing the sources of particulate lead in house dust by
automated scanning electron microscopy. Environ Sci Technol 26:1513-1523.
Hursh, JB; Clarkson, TW, Miles, EF, Goldsmith, LA. (1989) Percutaneous absorption of mercury vapor by man.
Arch. Environ.' Health 44:120-127.
HydroQual (1982) Application guide for CMA-HydroQual chemical fate models. Report No. Cl 13 for the
Chemical Manufacturers Association Aquatic Research Task Group, Washington, DC. 201pp. Mahway, NJ.
HydroQual (1997) Documentation for the probabilistic analysis program MONTE. Mahway NJ.
HydroQual (1998) A primer for ECOMSED, Version 1.3. Mahway NJ.
LTI. (1994) Dynamic toxics wasteload allocation model (DYNTOX), version 2.0 user's manual. Ann Arbor, MI..
IARC (International Agency for Research on Cancer) (2004) Inorganic and organic lead compounds. Volume 87.
Available at: http://www-cie.iarc.fr/htdocs/announceinents/vol87.htm
Idaho Department of Health and Welfare (2000) Coeur d'Alene River Basin environmental health assessment.
Agency for Toxic Substances and Disease Registry, Atlanta, GA.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-17
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
llyin, LA; Ivannikov, AT; Parfenov, YD; Stolyarov, VP. (1975) Strontium absorption through damaged and
undamaged human skin. Health Phys 29:75-80.
Ingersoll, CO; Haverland, PS; Brunson, EL; et al. (1996) Calculation and evaluation of sediment effect
concentrations for the amphipod Hyalella azteca and the midge Chironamus riparius. J Great Lakes Res
22:602-623.
Isaure, M.-P; Laboudique, A; Manceau, A; et al. (2002) Quantitative Zn speciation in a contaminated dredged
sediment by nPixe, ji-SXRF, EXAFS spectroscopy and principal component analysis. Geochim Cosmochim Acta
66(9):1549-1567.
Jaffe, D., M. S. Landis, P. Weisspenzias, and E. Prestbo.. (2003) Comments on "Mercury Concentrations in Costal
California Precipitation: Evidence of Local and Trans-Pacific Fluxes of Mercury to North America" By Steding and
Flegal (2002). J Geophys Res-Atmos 108(D17):ACH 11-1 to ACH 11-2.
James, BR (1994) Hexavalent chromium solubility and reduction in alkaline soils enriched with chromite ore
processing residue. J. Environ. Qual. 23:227-233.
James, AC; Stahlhofen, W; Rudolf, GR, et al. (1994) Deposition of inhaled particles. Annals of the ICRP (Pub. No.
66) 24(l-3):231-299.
i
Janes, N; Playle, RC. (1995) Modeling silver binding to gills of rainbow trout (Oncorhynchus mykiss). Environ
Toxicol Chem 14:1847-1858.
Janssen, C; Muyssen, B. (2001) Essentiality of metals: consequences for environmental risk assessments. Fact sheet
#5 on environmental risk assessment. International Council on Metals and the Environment (ICME), Ottawa,
Ontario, Canada.
Janssen, CR; Heijerick, DG. (2003). Algal toxicity tests for environmental risk assessments of metals. IN: Ware,
George W. [Editor,] Reviews of Environmental Contamination and Toxicology. 178 :23-52.
Jarup, L.; Berglund, M.; Blander, C.; Nordberg.G.; Vahter, M. (1998) Health effects of cadmium exposure—a
review of the literature and a risk estimate. Scand. J. Work Environ. Health 24(Suppl l):l-52.
Jeffree, RA. (1988) Patterns of accumulation of alkaline-earth metals in the tissue of the freshwater mussel Velesunio
angasi. Archiv fuer Hydrobiologiell2: 67-90.
Jing, J; Logan, T (1992) Effects of sewage sludge cadmium concentration on chemical extractability and plant
uptake. J. Environ. Qual. 21:73-81.
Jones, MM; Vaughn, WK. (1978) HSAB theory and acute metal ion toxicity.and detoxification processes. J Inorg
Nucl Chem 40:2081-2088.
Kabata-Pendias, A. and H. Pendias (2001). Trace elements in soils and plants. Boca Raton, Fla., CRC Press.
Kagi, J.H.R., and Y. Kogima, eds. 1987. Chemistry and biochemistry of metallothionein. Boston: Birkhauser, pp.
25-61.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-18
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
Kahle, J; Zauke, G-P. (2003) Bioaccumulation of.trace metals in the Antarctic amphipod Orchomene plebs:
evaluation of toxicokinetic models. Marine Environmental Research 55:359-384.
Kamunde, C; Grosell, MH; Wood, CM. (2002) Copper metabolism in actively growing rainbow trout
(Oncorhynchus mykiss): interactions between dietary and waterbome copper uptake. J Exp Biol 205(Pt 2)279-290.
Kapustka, LA; Clements, WH; Ziccardi, L; Paquin, PR; Sprenger, M; Wall, D. (2004) Issue paper on the ecological
effects of metals. September 2004. .
Karunaratne, RS; Baker, JH; Barker, AV. (1986) Phosphorus uptake by mycorrhizal and nonmycorrhizal roots of
soybean Glycine max. J Plant Nut 9:103-1314.
Kemp, PF; Swartz, RC. (1986) Acute toxicity of interstitial and particle-bound cadmium to a marine infauna]
amphipod. Mar Environ Res 26:135-153.
Kheboian, C; Bauer, CF, (1987) Accuracy of selective extraction procedures for metal speciation in model aquatic
sediments. Anal Chem 59:1417-1423.
Khoury, GA; Diamond, GL. (2003) Risks to children from remedial or removal activities at superfund sites: a case
study of the rsr lead smelter Superfund site. J Exp Anal Environ Epidemiol. In press.
Kiffney, PM; Clements, WH. (1996) Effects of metals on stream macroinvertebrate assemblages from different
altitudes. Ecol Appl 6:472-481.
Kimber I, Basketter DA. 1996. Contact Hypersensitivity to Metals. In: Chang LW Toxicology of Metals. Boca
Raton: CRC Press, pp 827-834.
Kjellstrom, T., and G.F. Nordberg. 1978. A kinetic model of cadmium metabolism in the human being. Environ,
Res. 16:248-269.
Klerks, PL; Levinton, JS. (1993) Evolution of resistance and changes in community composition in metal-polluted
environments: a case study on foundry cove. In: Dallinger, R; Rainbow, PS, eds Ecotoxicology of metals in
invertebrates. SET AC Special Publication Series. Ann Arbor, MI: Lewis Publishers; p. 223-241.
Kostial, K., D. Kello, S. Jugo, I. Rabar, and T. Maljkovic. 1978. Influence of age on metal metabolism and toxicity.
Environ. Health Perspect 25:81-86.
Kraak, MHS; Lavy, D; Schoon, H; et al. (1994) Ecotoxicity of mixtures of metals to the zebra mussel Dreissena
polymorpha. Environ Toxicol Chem 13:109-114.
Krantzbeig, C; Stokes, PM. (1988) The importance of surface adsorption and pH in metal accumulation by
chironomids. Environ Toxicol Chem 7:653-670.
Krishnan, K., and M.E. Andersen. 1994. Physiologically based pharmacokinetic modeling in toxicology In: Hayes,
A.W., ed. Principles and methods in toxicology, 3rd ed. New York: Raven Press, Ltd., pp. 149-188.
Landis, M. S., and R. K. Stevens. (2003) Comments on "Measurements of Atmospheric Mercury Species at a Costal
Site in the Antarctic and Over the South Atlantic Ocean During Polar Summer"; Environ Sci Technol
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-19
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
37(i4):3239-3240.
Langmuir, D. (1997) Aqueous environmental chemistry. Upper Saddle River, NJ: Prentice-Hall.
Langmuir, D; Chrostowsk, P; Vigneault, B Chancy, R. (2004) Issue paper on the environmental chemistry of metals.
September 2004, .
Langston, WJ. (1980) Arsenic in U.K. estuarine sediments and its availability to benthic organisms. J Mar Biol Assn
UK 60:869-881.
Lanno, RP; LeBlanc, S; Knight, B, et al. (1998) Application of body residues as a tool in the assessment of soil
toxicity. In: Sheppard, S; Bembridge, J; Holmstrup, M; et al., eds. Advances in earthworm ecotoxicology. Pensacola,
FL: SETAC Press; pp. 41-53.
Laperche, V; Logan, TI; Gaddam, P; et al. (1997). Effect of apatite amendments on plant uptake of lead from
contaminated soil. Environ Sci Technol 31:2745-2753.
Larison, JR; Likens, GE; Fitzpatrick, JW; Crock, JG. (2000). Cadmium toxicity among wildlife in the Colorado
Rocky Mountains. Nature. 406:181-183.
Larsson, P. (1985) Contaminated sediments of lakes and oceans act as sources of chlorinated hydrocarbons for .
release to water and atmosphere. Nature 317:347-349.
Laurie, SH; Manthey, JA. (1994) The Chemistry and Role of Metal Ion Chelation in Plant Uptake Processes. IN:
Manthey, J.A., D.E. Crowley, and D.G. (eds). Biochemistry of metal micronutrients in the rhizosphere Lewis
Publishers, Boca Raton, FL pp. 165-182.
LaVelle, JM; Poppenga, RH; Thacker, BJ; et al. (1991) Bioavailability of lead in mining waste: an oral intubation
study in young swine. In: Proceedings of the International Symposium on the Bioavailability and Dietary Uptake of
Lead. Sci Technol Lett 3:105-111.
Lazo, JS; Kuo, SM; Woo, ES; Pitt, BR. (1998) The protein thiol metallothionein as an antioxidant and protectant
against antineoplastic drugs. Chem Biol Interact. Apr 24,111 -112:255-262.
Lee, B-G; Griscom, SB; Lee, JS; et al. (2000) Influences of dietary uptake and reactive sulfides on metal
bioavailability from aquatic sediments. Science 287:282-284.
Leggett, RW. (1993) An age-specific kinetic model of lead metabolism in humans. Environ Health Perspect
101:598-616.
LePoire, D; Amish, J; Gnanapragasam, ES; et al. (2000) Probabilistic modules for the RESRAD and RESRAD-build
computer codes. NUREG/CR-XX ANL/EAD/TM-91. Argonne National Laboratory, Argonne, IL.
Liess, M; Champeau, O; Riddle, M; Schulz, R; Duquesne, S.(2001) Combined effects of ultraviolet-B radiation and
food shortage on the sensitivity of the Antarctic amphipod Paramoera walkeri to copper. Environ Toxicol Chem 20:
2088-2092.
LePoire, D; Arnish, J; Gnanapragasam, ES; et al. ( 2000) Probabilistic modules for the RESRAD and RESRAD-
11/24/2004 Peer Review Draft <
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-20
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
build computer codes. NUREG/CR-XX ANL/EAD/TM-91. Argonne National Laboratory, Argonne, IL.
Llobet, JM; -Domingo, JL; Gomez, M; et al (1987) Acute toxicity studies of aluminum compounds: antidotal efficacy
of several chelating agents. Ann Rev Pharmacol Toxicol 60:80-3.
Lindqvist, O (1994) Atmospheric cycling of mercury: An overview. In: Watras, CJ and Huckabee, JW, eds. Mercury
pollution: Integration and synthesis. Boca Raton, FL: Lewis Publishers.
Lofts, S; Tipping, E. (1998) An assemblage model for cation binding by natural particulate matter. Geochem
Cosmochim Acta 62(15):2609-2625.
Logan, TJ; Traina, SJ. (1993) Trace metals in agricultural soils. Chapter 10 in: Metals in Groundwater. Eds. Allen,
H.E., E.M. Perdue, and D.S. Brown. New York: Lewis Publishers, pp. 309-347.
Long, ER; Morgan, LG. (1990) The potential for biological effects of sediment sorbed contaminants tested in the
National Status and Trends Program. NOS OMA 52. Technical Memorandum. National Oceanic and Atmospheric
Administration, Office of Oceanography and Marine Assessment, Seattle, WA.
Long, ER; Morgan, LG. (1991). The potential for biological effects of sediment-sorbed contaminants tested in the
National Status and Trends Program. NOAA Tech. Memo. NOS OMA 52. Seattle, WA
Loux, NT; Brown, DS; Chafin, CR; Allison, JD; Hassan, SM. (1990) Modeling geochemical processes attenuating
inorganic contaminant transport in the subsurface region: Adsorption on amorphous iron oxide. In: Murarka, I. and
S. Cordle, eds. Proc. Envir. Res. Conf. on Groundwater Qual. and Waste Disposal. Report EPRIEN-6749. Palo
Alto, CA: Electrical Power Research Inst.
Lu, Y; Allen, HE. (2002) Characterization of copper complexation with natural dissolved organic matter (DOM):
link to acidic moieties of DOM and competition by Ca and Mg. Water Res 36:5083-5101.
Lumsdon, DG; Evans, LJ. (1995) Predicting chemical speciation and computer simulation. In: Ure, AM; Davidson,
CM, eds. Chemical speciation in the environment. London: Blackie; pp. 86-134.
Luoma, SN. (1989) Can we determine the biological availability of sediment-bound trace elements? Hydrobiologia
176/177:379-396.
Luoma, SN; Bryan, GW. (1978) Factors controlling the availability of sediment-bound lead to the estuarine bivalve
Scrobicularia plana. J Mar Biol Assn UK 58:793-802.
Ma, YB; Uren, NC. (1995) Application of new fractionation scheme for heavy metals in soils. Commun Soil Sc
Plant Ana 26:3291-3303.
MacDonald, DD; Carr, RS; Calder, FD; et al. (1996) Development and evaluation of sediment quality guidelines for
Florida coastal waters. Ecotoxicol 5:253-278.
MacDonald, DD; Ingersoll, CG; Berger, TA. (2000) Development and Evaluation of Consensus-Based Sediment
Quality Guidelines for Freshwater Ecosystems. Arch. Environ. Contam. Toxicol. 39:20-31.
Mackay, D. (1979) Finding fugacity feasible. Environ Sci Tech 13(10):1218.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-21
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
Mackay, D. (1991) Multimedia Environmental Models. Chelsea, MI: Lewis Publishers.
Mackay, D; Paterson, S; et al, (1992). Generic models for evaluating the regional fate of chemicals. Chemosphere
24:695-717.
Maddaloni, M; et al. (1998) Meal-weighted bioavailability of soil-bourne lead in adults: use of empirical data
analyzed by stable isotope"dilution, and probabilistic worker exposure scenarios, lexicologist 124:25.
Manceau, A; Lanson, B; Schlegel, ML; et al. (2000) Quantitative Zn speciation in smelter-contaminated soils by
EXAFS spectroscopy. Fr Amer J Sci 300(4):289-343.
Manceau, A; Boisset, M-C; Sarret, G; et al. (1996) Direct determination of lead speciation in contaminated soils by
EXAFS spectroscopy. Environ Sci Technol 30:1540-1552.
Markich, SJ; Brown, PL; Jeffree, RA; et al. (2000) Valve movement responses of Velesunio angasi (Bivalvia:
Hyriidae) to manganese and uranium: an exception to the free ion activity model. Aquat Toxicol 51:155-175.
Markich, SJ; Brown, PL; and Jeffree, RA. (2001) Divalent metal accumulation in freshwater bivalves: an inverse
relationship with metal phosphate solubility. The Science of the Total Environment 275: 27-41.
Marschner, H. (1995) Mineral nutrition of higher plants, ed. San Diego: Academic Press. 2nd Edition.
Marschner, H. (1998) Soil-root interface: biological and biochemical processes. In: Huang, PM, ed. Soil chemistry
and ecosystem health. Soil Science Society of America, Madison, WI; pp. 191-232.
Martin, JL; Medina, AJ. (1998) Metal exposure and transformation assessment models: model documentation for
version 3. Medine Environmental Engineering. Boulder, CO.
Mason, AZ; Jenkins, KD. (1995) Metal detoxification in aquatic organisms. In: Tessier, A; Turner, DR, eds. Metal
speciation and bioavailability in aquatic systems. Chichester, UK: John Wiley & Sons; pp. 479-608.
Masters, BA; Kelly, EJ; Quaife, CJ; Brinster, RL; and Palmiter, RD. (1994) Targeted disruption of metallothionein I
and II genes increases sensitivity to cadmium. Proc. Nat. Acad. Sci. 91:584-588.
Mayer, LM; Weston, DP; Bock, MJ. (2001) Benzo[a]pyrene and zinc solubilization by digestive fluids of benthic
invertebrates: a cross-phyletic study. Environ Toxicol Chem 20:1890-1900.
McCarthy, LS; Mackay, D. (1993) Enhancing ecotoxicological modeling and assessment. Environ Sci Technol
27:1719-1728.
McCarty, K.M., D.B. Senn, M.L. Kile, Q. Quamruzzaman, M. Rahman, G. Mahiuddin, and D.C. Christian. 2004.
Antimony: An unlikely confounder in the relationship between well water arsenic and health outcome in
Bangladesh. Environ Health Perspect. 112:809-811.
McDowell, LR. (2003) Minerals in animal and human nutrition, 21"1 ed. Amsterdam: Elsevier.
McGeer, JC; Playle, RC; Wood.CM; et al. (2000) A physiologically based biotic ligand model for predicting the
acute toxicity of waterborne silver to rainbowtrout in freshwaters. Environ Sci Technol 34:4199-4207.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-22
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
McGeer, JC; Brix, KV; Skeaff, JM; et al. (2003) Inverse relationship between bioconcentration factor and exposure
concentration for metals: implications for hazard assessment of metals in the aquatic environment. Environ Toxicol
Chem 22(5)1017-1037.
McGeer, J, Henningsen, G; Lanno, R; Fisher. N; Sappington, K; Drexler, JW. (2004). Issue Paper on the
Bioavailability and Bioaccumulation of Metals. September 2004.
.
McGrath, SP. (1995) Chromium and nickel. In: Alloway, BJ, ed. Heavy metals in soils. London: Blackie Academic
and Professional; pp. 152-178.
McKinney JD; Richard, A; Waller, C; Newman, MC; Gerberick, F. (2000) The practice of structure activity
relationships (SAR) in toxicity. Toxicol Sci 56: 8-17.
McKone, T.E. 1987. Human exposure to volatile organic compounds in household tap water: The indoor inhalation
pathway. Environ. Sci. Technol. 21:1194-1201.
McLaughlin, MJ; Smolders, E. (2001) A rationale for the metalloregion approach to risk assessments for metals in
soil: a case study for effects of zinc on microbial functions in soil. Environ Toxicol Chem 20:2639-2643.
McLaughlin, MJ; Tiller, KG; Beech, TA; et al. (1994) Soil salinity causes elevated cadmium concentrations in
field-grown tubers. Aust J Soil Res 35:1-17.
McLaughlin, MJ; Hamon, RE; Parker, DR; Pierzynski, GM; Smolders, E; Thornton, I; and Welp, G. (2002). Soil
chemistry. In: Fairbrother, A., P.W. Glazebrook, J.V. Terazona, and N.M. van Straalan, eds. Test methods to
determine hazards of sparingly soluble metal compounds in soils. Pensacola, FL: SETAC Press, pp. 5-16.
Medlin, EA. (1997) Master's thesis, Department of Geological Sciences, University of Colorado. 111 pp.
Menzie-Cura & Associates, Inc. and TN&A, Inc. (2000). Development of a framework to address bioavailability
issues in wildlife. Submitted to Randall Wentsel, U.S. EPA, National Center for Environmental Assessment,
Washington, D.C.
Mertz, W. (1981) The essential elements. Science 213:1332-1338.
Meyer, J. (2002) The utility of the terms "bioavailability" and "bioavailable fraction" for metals. Mar Environ Res
53:41-423.
Meyer, JS; Santore, RC; Bobbitt, JP; DeBrey, LD; Boese, CJ; Paquin, PR; Allen, HE; Bergman, HL; DiToro, DM.
(1999) Binding of Nickel and Copper to Fish Gills Predicts Toxicity When Water Hardness Varies, But Free-ion
Activity Does Not. Environ. Sci. and Technol 33: 913-916.
Meyer, JS; Boese, C J; Collyard, SA. (2002) Whole-body accumulation of copper predicts acute toxicity to an
aquatic oligochaete (Ltanbriculus variegatus) as pH and calcium are varied. Comp. Biochem. Physiol. Part C
133:99-109.
Mills, WB; Porcella, DV; Ungs, MJ; et al. (1985) Water quality assessment: a screening procedure for toxic and
conventional pollutants in surface and ground water. Part 1. EPA-600-6-85-002a.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-23
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
Mitrofanov, D. (1993) Ability of non-woody forest plants to accumulate metals. Aquilo Ser Bot 31:41-45.
Morel, F. (1983) Complexation: trace metal and microorganisms. In: Principles of aquatic chemistry. New York:
John Wiley & Sons; pp. 405-418.
Morgan, JE; Morgan, AJ. (1988) Calcium-lead interactions involving earthworms: part 2. The effects of accumulated
lead on endogenous calcium in Lumbricits rubellus. Environ Pollut 55:41-54.
Morrison, HA; Whittle, DM; Haffner, GD. (2000) The relative importance of species invasions and sediment
disturbance in regulating chemical dynamics in western Lake Erie. Ecol Modeling 125:279-294.
Mowat, FS; Bundy, KJ. (2002) Experimental and mathematical/computational assessment of the acute toxicity of
chemical mixtures from the Microtox assay. Adv Environ Res 6:547-558.
Moya, J., C. Howard-Reed, and R.L. Corsi. 1999. Volatization of chemicals from tap water to indoor air from
contaminated water used for showering. Environ. Sci. Technol. 33:2321-2327.
Mulvey, M; Diamond, SA. (1991). Genetic factors and tolerance acquisition in populations exposed to metals and
metalloids. Pages 301-321 in Newman, M.C., and AW. Mclntosh, editors. Metal ecotoxicology: concepts and
applications. Lewis Publishers, Inc., Chelsea, Michigan.
Munthe, J (1994) The atmospheric chemistry of mercury: Kinetic studies of redox reactions. In: Mercury Pollution:
Integration and Synthesis. Watras, CJ and Huckabee JW, eds. Boca Raton, FL: Lewis Publishers.
Mudroch, A; Azcue, JM; Mudroch, P. (1997) Manual of Physico-Chemical Analysis of Aquatic Sediments. Lewis
Publishers, Boca Raton, Florida.
Mudroch, A; Azcue, JM; Mudroch, P. (1999) Manual of Bioassessment of Aquatic Sediment Quality. Lewis
Publishers, Boca Raton, Florida.
NAS/NRC (National Academy of Sciences/National Research Council) (1980) Mineral tolerance of domestic
animals. Washington, DC: NRC Press.
NAS/NRC (1993) Measuring lead exposure in infants, children, and susceptible populations. Washington, DC:
National Academy of Sciences Press.
NAS/NRC (1994) Nutritional requirements of poultry, 9th rev. ed. National Academy of Sciences, Washington, DC:
National Academy of Sciences Press.
NAS/NRC (1999) Arsenic in drinking water. Washington, DC: National Academy of Sciences Press.
NAS/NRC (2000) Toxicological Effects of methyl mercury. Washington, DC: National Academy of Sciences Press.
NAS/NRC (2001) Arsenic in drinking water: 2001 update. Washington, DC: National Academy of Sciences Press.
NAS/NRC (2002) Bioavailability of contaminants in soils and sediments: Processes, tools, and applications.
Washington, DC: National Academy of Sciences Press.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-24
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
NAS/IOM (National Academy of Sciences/Institute of Medicine) (2000) Dietary reference intakes for vitamin A,
vitamin K, arsenic, boron, chromium, copper, iodine, iron, manganese, molybdenum, nickel, silicon, vanadium, and
zinc. Food and Nutrition Board, Institute of Medicine, Washington, DC. ISBN 0-309-7279-4. Available at:
http ://www.nap.edu/catalog/l 0026.html.
NAS/IOM (2001) Dietary reference intakes for Vitamin A, Vitamin K, arsenic, boron, chromium, copper, iodine,
iron, manganese, molybdenum, nickel, silicon, vanadium, and zinc. Standing Committee on the Scientific
Evaluation of Dietary Reference Intakes, Panel on Micronutrients, Food and Nutrition Board, Institute. Available at:
http://www.nap.edU/catalog/10026.html
NAS/IOM (2003) Dietary Reference Intakes: Guiding Principles for Nutrition Labeling and Fortification. Food and
Nutrition Board, Institute of Medicine, Washington, DC.
NFESC (Naval Facilities Engineering Command) (2000a) Guide for Incorporating Bioavailability Adjustments into
Human Health and Ecological Risk Assessments at U.S. Navy and Marine Corps Facilities. Part 1: Overview of
Metals Bioavailability." Prepared by Battelle and Exponent.
»
NFESC (2000b) Guide for Incorporating Bioavailability Adjustments into Human Health and Ecological Risk
Assessments at U.S. Navy and Marine Corps Facilities. Part 1: Technical Background Document for Assessing
Metals Bioavailability. Prepared by Battelle and Exponent.
Newman, MC (1991) A statistical bias in the derivation of hardness-dependant metals criteria. Environ. Toxicol.
Chem. 10:1295-1297.
,—i
Newman, MC; McCloskey, JT. (1996) Predicting relative metal toxicity and interactions of divalent metal ions:
microtox bioluminescence assay. Environ Toxicol Chem 15:275-281.
Newman, MC; McCloskey, JT; Tatara, CP. (1998) Using metal-ligand binding characteristics to predict metal
toxicity: quantitative ion character-activity relationships (QICARs). Environ Health Perspect 106(Suppl.
6):1262-1270.
Newman, MC; Diamond, GL; Menzie, C; Moya, J; Nriagu, J. (2004). Issue paper on metal exposure assessment.
September 2004. .
Ng, JC; Kratzmann, SM; Qi, L; et al. (1998) Speciation and absolute bioavailability: risk assessment of arsenic
contaminated sites in a residential suburb in Canberra. Analyst 123:889-892.
Nieboer, E; Richardson, DHS. (1980) The replacement of the nondescript term "heavy metals" by a biologically and
chemically significant classification of metal ions. Environ Pollut (Series B) 1:3-26.
Noorishad, JC, Camahan, L.; Benson, LV. (1987) CHMTRNS: A temperature-dependent non-equilibrium reactive
transport code. Earth Sciences Div., Lawrence Berkeley Laboratory.
Ndrwood, WP; Borgmann, U; Dixon, DG; et al. (2003) Effects of metal mixtures on aquatic biota: a review of
observations and methods. Hum Ecol Risk Assess 9: 795-811.
Nott, JA; Nicolaidou, A. (1990). Mediterranean pollution from a ferro-nickel smelter: differential uptake of metals
by some gastropods. Mar. Pollut. Bull. 21, pp. 137-143.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-25
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
Nott, JA; Nicolaidou, A. (1994) Variable transfer of detoxified metals from snails to hermit crabs in marine food
chains. Mar. Biol. 120:369-377,
Nriagu, JO; Pacyna JM. (1988) Quantitative assessment of worldwide contamination of air, water and soils by trace
metals. Nature 333(6169): 134-9.
NTP (National Toxicology Program). 2002.10th Report on carcinogens. U.S. Department of Health and Human
Services, Public Health Service, Washington, DC.
Nyffeler, UP; Santschi, PH; Li,Y-H. (1986) The relevance of scavenging kinetics to modeling of sediment water
interactions in natural waters, Limnol. Oceanogr. 31: 277-292.
Oberdorster, O. (1992) Pulmonary deposition, clearance and effects of inhaled soluble and insoluble cadmium
compounds. In: Nordberg, GF; Herber, RFM; Alessio, L, eds.
Cadmium in the human environment: toxicity and carcinogenicity. Lyon, FR:IARC; pp. 189-204.
Oberdorster, G. (1992) Pulmonary deposition, clearance and effects of inhaled soluble and insoluble cadmium
compounds. In: Nordberg, GF; Herber, RFM; Alessio, L, eds. Cadmium in the human environment: toxicity and
carcinogenicity. Lyon, FR:IARC; pp. 189-204.
O'Connor, DJ. (1988) Models of sorptive toxic substances in freshwater systems: III. streams and rivers. J Environ
Engrll4(3):552-574..
O'Flaherty, EJ. (1998) Physiologically based models of metal kinetics. Crit Rev Toxicol 28:271-317.
O'Flaherty, EJ. 1995. Physiologically based models for bone-seeking elements. V: Lead absorption and disposition
in childhood. Toxicol. Appl. Pharmacol. 131:297-308.
O'Flaherty, E.J. 1993. Physiologically based models for bone-seeking elements. IV: Kinetics of lead disposition in
humans. Toxicol. Appl. Pharmacol. 118:16-29.
OECD (Organization for Economic Cooperation and Development). (2001) Classification of metals and metal
compounds. In: Guidance document on the use of the harmonized system for the classification of chemicals which
are hazardous for the aquatic environment. (OECD Series on testing and assessment, Number 27). pp. 97-115.
Document ENV/JM/MONO(2001)8. Paris, France.
Omernik, JM. (1986) Aquatic ecoregions of the conterminous United States. Supllement to the Annals of the Assoc
of Amer Geographers 77:(1).
O'Rourke, M.K., P.K. van de Water, S. Jin, S.P. Rogan, A.D. Weiss, S.M. Gordon, D.M. Moschandreas, and M.D.
Lebowitz. 1999. Evaluations of primary metals from NHEXAS Arizona: distributions and preliminary exposures. J.
Exposure Anal. Environ. Epidemiol. 9:435-445.
Ott, WR. (1985) Total human exposure. Environ Sci Technol 10:880-886.
Ownby, DR; Newman, MC. (2003) Advances in quantitative ion character-activity relationships (QICARs): using
metal-ligand binding characteristics to predict metal toxicity. Quant Struct Act Relat (QSAR) Comb Sci 22:241-246.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-26
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
Pabalan, RT; Turner, DR, Bertetti, FP; Prikryl, JD (1998) Uranium VI sorption onto selected mineral surfaces: Key
geochemical parameters. Chap. 3. In Adsorption of metals by Geomedia. Ed. E.A. Jenne. Academic Press, New
York: 99-130.
Pagenkopf, GK. (1983) Gill surface interaction model for trace-metal toxicity to fishes: Role of complexation, pH ^
and water hardness. Environ Sci.Technol 17: 342-347
Paquin, PR; Santore, RC; Farley, KJ; Wu, KB; Mooney, K; DiToro, DM (2003) Review of metal fate,
bioaccumulation and toxicity models for metals. Pensacola, FL: SET AC Press. In press.
Paquin, PR; Gorsuch, JW; Apte, S; Batley, GE; Bowles, KC; Campbell, PGC; Delos, CG; Di Toro, DM; Dwyer,
RL; Galvez, F; Gensemer, RW, Goss, GG; Hogstrand, C; Janssen, CR; McGeer, JC; Naddy, RB; Playle, RC;
Santore, RC; Schneider, U; Stubblefield, WA; Wood, CM; Wu, KB (2002a) The biotic ligand model: A historical
overview. Special Issue: The biotic ligand model for metals-Current research, future directions, regulatory
implications. Comp. Biochem. Physiol. C 133:3-35.
Paquin, PR; Zoltay, V; Winfield, RP; et al. (2002b) Extension of the biotic ligand model of acute toxicity to a
physiologically-based model of the survival time of rainbow trout (Oncorhynchus myldss) exposed to silver. Comp
Biochem Physiol (Part C) 133:305-343.
Paquin; PR, DiToro, DM; Santore, RC; Trivedi, D; Wu, KB (1999) A biotic ligand model of the acute toxicity of
metals. Ill: Application to fish and daphnia exposure to silver. Office of Water, Office of Research and
Development.
Parent, L; Campbell, PGC. (1994) Aluminum bioavailability to the green alga Chlorella pyrenoidosa in acidified
synthetic soft water. Environ Toxicol Chem 13:587-598.
Parker, DR; Pedler, JF. (1997) Reevaluating the free ion activity model of trace metal availability to higher plants.
In: Ando, T; Fujita, K; Mae, T; et al., eds. Plant nutrition for sustainable food production and environment.
Dordrecht, The Netherlands: Kluwer; pp. 107-112.
Parkhurst, DL (2002) PHAST:3D reactive transport model. U.S. Geological Survey.
Parkhurst, DL, Appelo, CAJ (1999) User's Guide to PHREEQC (version 2)-A computer program for speciation,
batch-reaction, one-dimensional transport, and inverse geochemical calculations. U.S. Geological Survey, Water
Resources Inv. Report 99-4259.
Parkhurst, D.L., Thorstenson, D.C., and Plummer, L.N., 1980. PHREEQE- a computer program for geochemical
calculations. U.S. Geological Survey Water-Resources Investigations Report 80-96, 210p.
Parkinson, A. 2001 Biotransformation of zenobiotics. In: Klaassen, C.D., ed. Casarett & Doull's toxicology. New
York: McGraw-Hill, pp. 133-224.
Paulson, AJ; Balistrieri, LS (1999) Modeling removal of Cd, Cu, Pb, and Zn in acidic groundwater during
neutralization by ambient surface waters and groundwaters. Environ. Sci. Technol. 33(21 ):3850-3856.
Paustenbach, DJ. (2000) The practice of exposure assessment: a state-of-the-art review. J Toxicol Environ Health
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-27
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
(Part 6)3:179-291.
Pearson, RG. (1973) Hard and soft acids and bases. Stroudsburg, PA: Dowden, Hutchinson & Ross.
Peijnenburg, WJGM; Posthuma, L; Zweers, PGPC; et al. (1999b) Prediction of metal bioavailability in Dutch field
soils for the oligochaete Enchytraeus crypticus. Ecotoxicol Environ Saf 44:170-186.
Peijnenburg, WJGM; Baerselman, R; De Groot, AC; et al. (1999a) Relating environmental availability to
bioavailability: soil-type dependent metal accumulation in the oligochaete Eisenai andrei. Ecotoxicol Environ Saf
44:294-310.
Persaud, D; Jacqumagi, R; Hayton, A. (1989) Development of provincial sediment quality guidelines. Draft Report.
Water Resources Branch, Ontario Ministry of the Environment. Toronto, ON, Canada.
Phalen, RF; Oldham, MJ; Beaucage, CB; et al. (1985) Postnatal enlargement of human tracheobronchial airways and
implications for particle deposition. Anat Rec 212:368-380.
Picciano, M.F. 1996. Pregnancy and lactation. In: Ziegler, E.E. and L.J. Filer, Jr., eds. Present knowledge in
nutrition, 7th ed. Washington, DC: ILSI Press, pp. 384-395.
Pickering, WF. (1986) Metal Ion Speciation- Soils and Sediments (A Review). Ore Geology Reviews 1:8 3.-
Fillet, S., M. Fournier, L.N. Measures, J. Bousqsuegneau, and D.G. Cyr. 2002. Presence and regulation of
metallothioneins in peripheral blood leukocytes of grey seals. Toxicol. Appl. Pharmacol 185:207-217,
Pitt, BR; Schwarz, M; Woo, ES; Yee, E; Wasserloos, K; Iran, S; Weng, W; Mannix, RJ; Watkins, SA; Tyurina,
YY; Tyurin, VA; Kagan, VE; Lazo, JS. (1997) Overexpression of metallothionein decreases sensitivity of
pulmonary endothelial cells to oxidant injury. Am. J. Physiol. 273:856-865.
Playle, RC; Wood, CM. (1990) Is precipitation of aluminum fast enough to explain aluminum deposition on fish
gills? Can J Fish Aquat Sci 47:1558-1561.
Playle, RC; Dixon, DG; Bumison., K. (1993a) Copper and cadmium binding to fish gills: modification by dissolved
organic carbon and synthetic ligands. Can J Fish Aquat Sci 50 2667-2677.
Playle, RC; Dixon, DG; Bumison, K. (1993b) Copper and cadmium binding to fish gills: estimates of metal-gill
stability constants and modelling of metal accumulation. Can J Fish Aquat Sci 50:2678-2687.
Posthuma, L; Janssen, GM. (1995) Genetic variation for life-history characteristics in reference populations of O.
cincta in relation to adaptation to metals in soils. Acta Zoologica Fennica 196:301-306.
Posthuma, L; Van Straalen, NM. (1993) Heavy-metal adaptation in terrestrial invertebrates: A review of occurrence,
genetics, physiology and ecological consequences. Comparative Biochemistry and Physiology. 106C(1):11-38.
Posthuma, L; Baerselman, R; Van Veen, RPM; et al. (1997) Single and joint toxic effects of copper and zinc on
reproduction of Enchytraeus crypticus in relation to sorption of metals in soils. Ecotoxicol Environ Saf 38:108-121.
Posthuma, L; Suter, GW; Traas, TP. eds. (2001). Species Sensitivity Distributions in Ecotoxicology, Lewis CRC-
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-28
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
Presspublishers, Boca Raton, US.
Prosser, CL. (1986). Adaptational Biology; Molecules to Organisms. John Wiley & Sons, New York.
Rainbow, P.S. (2002) Trace metal concentrations in aquatic invertebrates: why and so what? Environmental
Pollution 120, 497-507.
Raskin, I.; Kumar, PBAN; Dushenkov, S; et al. (1994) Bioconcentration of heavy metals by plants. Curr Opin
Biotechnol 5:285-290.
Regoli, F; Orlando, E. (1994) Bioavailability of biologically detoxified lead: Risks arising from consumption of
polluted mussels. Environ. Health Perspectives, Supplements, 102: 335-338.
Reid, SD; McDonald, DG (1988) Effects of cadmium, copper and low pH on ion fluxes in the rainbow trout, Salmo
gairdneri. Can J Fish Aquat Sci 45:244-253.
Reid, SD; McDonald, DG (1991) Metal binding activity of the gills of rainbow trout (Oncorhynchus myidss). Can. J.
Fish. Aquat Sci. 48:1061-1068.
Reinfelder, JR; Fisher, NS; Luoma, SN; et al. (1998) Trace element trophic transfer in aquatic organisms: a critique
of the kinetic model approach. Sci Total Environ 219:117-135.
Reinfelder, JR; Fisher, NS. (1991). The assimilation of elements ingested by marine copepods. Science
251:794-796.
Reuther, R. (1999) Trace metal speciation in aquatic sediments: methods, benefits, and limitations. In: Manual of
Bioassessment of Aquatic Sediment Quality. Lewis Publishers, Boca Raton, Florida. Pp. 1-54.
Richardson, GM (2002) Natural sources and emissions of elements to the atmosphere. International Council on
Mining and Metals. (February).
Robinson, B., K.R. Heid, T.L. Aldridge, and R.D. Glenn. 1983.1976 Hanford americium exposure incident: Organ
burden and radiation dose estimates. Health Phys. 45(4):91 Ir921.
Rodriguez, RR; Baste, NT; Casteel, SW; et al. (1999) An in vitro gastrointestinal method to estimate bioavailable
arsenic in contaminated soils and soil media. Environ Sci Techno! 33:642-649.
Roesijadi, G. (1992) Metallothioneins in metal regulation and toxicity in aquatic animals. Aquat Toxicol 22:81-114.
Roy, I; Hare, L. (1999) Relative importance of water and food as Cd sources to the predatory insect Sialis vlata
(Magaloptera). Can J Fish Aquat Sci 56(7): 1143-1149.
Rozman, K.K. and C.D. Klaassen. 2001. Biotransformation of zenobiotics. In: Klaassen, C.D., ed. Casserett and
Doull's toxicology. New York: McGraw-Hill, pp.107-132.
Ruby, MV; Davis, A; Link, TE; et al. (1993) Development of an in vitro screening test to evaluate the in vivo
bioaccessibility of ingested mine-waste lead. Environ. Sci Technol 27:2870-2877.
Ruby, MV; Davis, A; Schoof, R; et al. (1996) Estimation of lead and arsenic bioavailability using a physiologically
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-29
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
based extraction test. Environ Sci Technol 30:422—430.
Ruby, MV; Schoof, R; Brattin, W; et al. (1999) Advances in evaluating the oral bioavailability of inorganics in soil
for use in human health hsk assessment. Environ. Sci. Technol. 33:3697-3705.
Runkel, RL; Kimball, BA; McKnight, DM; Bencala, KE (1999) Reactive solute transport in streams: A surface
complexation approach for trace metal sorption. Water Resour. Res. 35(12):3829-3840.
Rusk, J.A., Hatnon, R.E., Stevens, D.P., and M.J. McLaughlin, 2004. Adaptation of soil biological nitrification to
heavy metals. Environ. Sci. Technol. 38:3092-97.
Ryan, P.B., K.A. Scanlon, and D.K. Macintosh. 2001. Analysis of dietary intake of selected metals in the NHEXAS-
Maryland investigation. Environ. Health Perspect. 109:121-129.
Ryan, PB. (1998) Historical perspective on the role of exposure assessment in human risk assessment. In: Newman
MC; Strojan, CL, eds. Risk assessment: logic and measurement. Boca Raton, FL: CRC/Lewis Press; pp. 23-43.
Sakai, T. 2000. Biomarkers of lead exposure. Ind. Health 37:127-142.
Salocks, C; Hathaway, T; Ziarkowski, C; et al. (1996) Physical characterization, solubility and potential
bioavailability of arsenic in tailings from a former gold mine. Toxicologist 16:48.
Salomons, W; De Rooij, MM; Kerdijk, H; et al. (1987) Sediments as sources of contaminants? Hydrobiol
149:13-30.
Sample, BE; Aplin, MS; Efroymson, RE, Suter, GW; Welsh, CE. (1997). Methods and tools for estimation of the
exposure of terrestrial wildlife to contaminants. Oak Ridge National Laboratory. ORNL/TM-13391.
Sample, BE; Arenal, CA. (1999) Allometric models for interspecies extrapolation of wildlife toxicity data. Bulletin
of Environmental Contamination & Toxicology. 62(6). 653-663.
Sample, BE; Beauchamp, JJ; Efroymson, RA; et al. (1998a) Development and validation of bioaccumulation
models for earthworms. ES/ER/TM-220. Prepared for the U. S. Department of Energy under Contract DE-AC05-
84OR21400. Code EW 20. Oak Ridge National Laboratory, Oak Ridge, TN.
Sample, BE; Beauchamp, JJ; Efroymson, R; et al. (1998b) Development and validation of bioaccumulation models
for small mammals. ES/ER/TM-219. Oak Ridge National Laboratory, Lockheed Martin Energy Systems
Environmental Restoration Program, Oak Ridge, TN.
Sample, BE; Beauchamp, JJ; Efroymson, R; Suter, GW. (1999). Literature-derived bioaccumulation models for
earthworms: development and validation. Environmental Toxicology and Chemistry. 18: 2110-2120.
Sample, BE; Arenal, CA. (2001). Development of Literature-based Bioaccumulation Models for Terrestrial
Arthropods. Platform presentation at 2001 SETAC Meeting in Baltimore, MD. November, 2001.
Santore, RC; Driscoll, CT. (1995) The CHESS model for calculating chemical equilibria in soils and solutions.
Chemical equilibrium and reaction models. SSSA Special Publication 42. Soil Science Society of America,
American Society of Agronomy, Madison, WI.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any •
Agency determination or policy,
7-30
-------
t
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
Santore, RC; Di Toro, DM; Paquin, PR. (2000) A biotic ligand model of the acute toxicity of metals: II. application
to acute copper toxicity in freshwater fish and Daphnia. Environ Toxicol Chem 20(10):2397-2402
Santore, RC; Di Toro, DM; Paquin, PR; et al. (2001) Biotic ligand model of the acute toxicity of metals: 2.
application to acute copper toxicity in freshwater fish and Daphnia. Environ Toxicol Chem 20:2397-2402.
Santore, RC; Mathew, R; Paquin, PR; et al. (2002) Application of the biotic ligand model to predicting zinc toxicity
to rainbow trout, fathead minnow, and Daphnia magna, Comp Biochem Physiol 113C(l-2):271-286.
Santschi, PH; Nyffeler, UP; Li, Y-H. Li and P. O'Hara. (1986) Radionuclide cycling in natural waters: relevance of
scavenging kinetics. In: Sediments and Water Interactions, P.O. Sly, ed.s Springer Verlag, 183-191.
Santschi, PH; Amdurer, M; et al. (1987) Relative mobility of radioactive trace elements across Ihe sediment-water
interface in the MERL model ecosystems of Narragansett Bay. J Mar Res 45:1007-1048.
Sauve, S; Dumestre, A; McBride, M; et al; (1998) Derivation of soil quality criteria using predicted chemical
speciation of Pb2* and Cu2+. Environ Toxicol Chem 17:1481-1489.
Saxe, JK; Impellitteri, CA; Peijnenburg, WJGM; et al. (2001) Novel model describing trace metal concentrations in
the earthworm, Eisenia andrei. Environ Sci Technol 35:4522-4529.
Saxena, DK, R.C., Murthy, S. V. Chandra. 1989. Zinc protects testicular injury induced by concurrent exposure to
cadmium and lead in rats. Res. Commun. Chem. Pathol. Pharmacol. 64:317-329.
Schecher, WD; McAvoy, DC (1994) MINEQL+: A chemical equilibrium program for personal computers (version
3.01). Environmental Research Software, Hallowell, ME.
Schecher WD; McAvoy DC. (2001) MINEQL+ - A Chemical Equilibrium Modelling System, version 4.5 for
windows User1 s Manual. Environmental Research Software, Hallowell, ME, USA.
Schlekat, CE; Lee, BY; Luoma, SN. (2001) Dietary metals exposure and toxicity to aquatic organisms: implications
for ecological risk assessment. In: Newman,' M, ed. Coastal and estuarine risk assessment. Boca Raton, FL: CRC
Press.
Schnoor, JL; Sata, C; McKechnie, D; etal. (1987) Processes, coefficients, and models for simulating toxic organics
and heavy metals in surface waters. EPA-600-3-87-015.
Schoof, R.A., J. Eickhoff, L.J. Yost, E.A. Crecelius, D.W. Cragin, D.M. Meacher, D.B. Menzel. 1999. Dietary
exposure to inorganic arsenic, pp. 81-88. In: Proc. Third International Conference on Arsenic Exposure and Health
Effects. W.R. Chappell, C.O. Abemathy, and R.L. Calderon (eds). Elsevier Science Ltd.
Schoof, R.A., L.J. Yost, J. Eickhoff, E.A. Crecelius, D.W. Cragin, D.M. Meacher, and D.B. Menzel. 1999. A
market basket survey of inorganic arsenic in food. Food Chem. Toxicol. 37:839-846.
Shroeder, W.H. and Munthe, J. (1998) Atmospheric mercury - An overview. Atmospheric Environment, 32(5): 809-
822.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-31
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
Schulte, P.A. and G. Talaska. 1995. Validity criteria for the use of biological markers of exposure to chemical agents
in environmental epidemiology. Toxicology 101:73-88.
Science and Research Working Group (SRWG) of the Non-Ferrous Metals Consultative Forum on Sustainable
Development (RWG) (2002) Risk assesment of non-ferrous metals and sustainable development.
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
of zinc for the springtail Folsomia Candida. Environmental Toxicology and Chemistry 17, No. 6, pp. 1 132-1 141.
Smith, CM; Wang, X; Hu, H; et al. (1995) A polymorphism in the d-aminolevulinic acid dehydratase gene can
modify the pharmacodynamics and toxicity of lead. Environ Health Perspect 103:248-253.
Smolders, E., S.P. McGrath, E. LombL, M.J. McLaughlin, M Rutgers, A. Fairbrother, L. Van der Vliet, and B. Hale.
(in prep). Hazard assessment of metals and metal compounds in terrestrial systems based on persistency,
bioaccumulation, and toxicity IN: Hazard Identification Approach For Metals And Inorganic Metal Substances,
Adams, W. J. et al. (eds.). SETAC Press, Pensacola, FL.
Spehar, RL; Fiandt, JT. (1 986) Acute and chronic effects of water quality criteria-based metal mixtures on three
aquatic species. Environ Toxicol Chem 5:917-931 .
Sposito, G. (1989) The chemistry of soils. New York: Oxford University Press.
Stauber, JL; Florence, TM; Gulson, BL; Dale, LS. (1 994) Percutaneous absorption of inorganic lead compounds.
Sci. Total Environ. 145:55-70.
Stephen, CE; Mount, DI; et al. (1 985) Guidelines for deriving numerical national water quality criteria for the
protection of aquatic organisms and their uses. National Technical Information Service, Springfield VA, PB 85-22-
7049.
Stem, AH. (1 994) Derivation of a target concentration of Pb in soil based on elevation of adult blood pressure. Risk
Anal 16:201-210.
Stem, AH. (1996) Derivation of a target concentration of Pb in soil based on elevation of adult blood pressure. Risk
Anal 16:201-210.
Storm, G.L., G.J. Fosmire, and E.D. Bellis. 1994. Heavy metals in the environment: Persistence of metals in soil and
selected vertebrates in the vicinity of Palmerton zinc smelters J. Environ. Qual. 23 :5Q8-51 5.
Stumm, W. (1992) Chemistry of the solid-water interface: Processes at the mineral-water and particle-water
interface. New York: Wiley InterScience.
Stumm, W; Morgan, JJ. (1996) Aquatic chemistry, 3rd ed. New York: John Wiley & Sons.
Stumm, W; Morgan, JJ (1970) Aquatic Chemistry. An Introduction Emphasizing Chemical Equilibria in Natural
Waters. New York: John Wiley & Sons.
Suedel, BC; Boracxek, JA; Peddicord, RK; et al. (1 994) Trophic transfer and biomagnification potential of
contaminants in aquatic ecosystems. Environ Toxicol Chem 136:21-89.
Sullivan, J; Ball, J; Brick, E; et al. (1985) Report of the technical subcommittee on determination of dredge material
suitability form-water disposal. Wisconsin Department of Natural Resources, Madison, WI.
Sunderman, FW, Jr. (1978) Carcinogenic effects of metals. Fed Proc 37:40^6.
Suter, GW, II. (1993) Ecological risk assessment. Boca Raton, FL: Lewis Publishers.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-33
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
Suttle, NF; Mils, CF. (1966) Studies of the toxicity of copper to pigs: 1. Effects of oral supplements zinc and iron
salts on the development of copper toxicosis, BrJNutr 20:135-149.
Swartz, RC; Dittsworth, OR; Schults, DW; et al. (1985) Sediment toxicity to a marine infaunal amphipod: cadmium
and its interaction with sewage sludge. Mar Environ Res 18:133-153.
Szebedinszky, C; McGeer, JC; McDonald, DG; et al. (2001) Effects of chronic Cd exposure via the diet or water on
internal organ-specific distribution and subsequent gill Cd uptake kinetics in juvenile rainbow trout (Oncorhynchus
mykiss). Environ Toxicol Chem 20:597-660
Szulczewski, MD; Helmke, PA; Bleam, WF. (1997) Comparison of XANES analyses and extractions to determine
chromium speciation in contaminated soils. Environ Sci Technol 31:2954-2959.
Tatara, CP; Newman, MC; Williams, P; McCloskey, JT. (1997) Predicting relative toxicity with divalent ion
characteristics: Caenorhabditis elegans. Aquat Toxicol 39: 279-290.
Tatara, CP; Newman, MC; McCloskey, JT; Williams, PL. (1998) Use of ion characteristics to predict relative
toxicity of mono-, di-, and trivalent metal ions: Caenorhabditis elegans LC50. Aquat Toxicol 42: 255-269.
Templeton, DM; Ariese, F; Cornells, R.; Danielsoa, LG; Muntau, H; Van Leeuwen, HP; Lobinski, R. (2000)
Guidelines for terms related to chemical speciation and fractionation of elements. Definitions, structural aspects, and
methodological approaches. Pure Appl. Chem. 72(8): 1453-70.
Tessier, A. 1992. Sorption of trace elements on natural particles in oxic environments. In Environmental Particles.
Edited by J. Buffle and H.P. Van Leeuwen. Lewis Publishers, Boca Raton, FL, USA. pp. 425-453.
Tessier, A; Campbell, PGC. (1987) Partitioning of trace metals in sediments: Relationships with bioavailability.
Hydrobiologia 149:43-52.
Tessier, A; Campbell, PGC. (1988) Partitioning of trace metals in sediments. In: Kramer, JR; Allen, HE, eds. Metal
speciation: theory, analysis and application. Chelsea, MI: Lewis Publishers; pp. 183-199.
Tessier, A; Campbell, PGC; Bissom, M. (1979) Sequential extraction procedure for the speciation of particulate
trace metals. Anal Chem 51:844-850.
Tessier, A; Campbell, PGC; Auclair, JC; et al. (1984) Relationships between the partitioning of trace metals in
sediments and their accumulation in the tissues of the freshwater mollusc Elliptic complanata in a mining area. Can
JFish Aquat Sci 41:1463-1472.
Tessier, A.; Rapin, F; Cariganan, R. (1985) Trace metals in oxic lake sediments: possible adsorption onto Iron
oxyhydroxides. Geochim. Cosmochim. Ac. 49:.183-194.
Tessier, A; Carignan, R; Dubreuil, B; Rapin, F. (1989) Partitioning of zinc between the water column and the oxic
sediments in lakes. Geochim Cosmochim Acta 53:1511-1522.
Tessier, A; Couillard.Y; Campbell, PGC; Auclair, JC. (1993) Modeling Cd partitioning in oxic lake sediments and
Cd concentrations in the freshwater bivalve Anodonta grandis. Limnology and Oceanography, 38(1): 1-17.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-34
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
Thawley, DG; Pratt, SE; Selby, LA, (1977) Antagonistic effect of zinc on increased urinary delta-aminolevulinic
acid excretion in lead intoxicated rats. Environ Resl 4:463-475.
Thomas, KW; Pellizzari, ED; Berry, MR. (1999) Population-based dietary intakes and tap water concentrations for
selected elements in the EPA Region V National Exposure Assessment Survey (NHEXAS). J Exp Anal Environ
Epidem 9:402^13.
Tipping, E (1994) WHAM: a chemical equilibrium model and computer code for waters, sediments, and soils
incorporating a discrete site/electrostatic model of ion-binding by humic substances. Computers & Geosciences
20:973-1023.
Tipping, E (1998) Humic ion binding model VI: An improved description of the interactions of protons and metal
ions with humic substances. Aqua. Geochem. 4(l):3-48.
Tolstykh, El; Degteva, MO; Kozheurov, VP; et al. (1998) Strontium transfer from maternal skeleton to the fetus
estimated on the basis of the Techa River data. Radiat Prot Dosim 79(1-4):307-310.
Tomerlin, JR; Berry, MR; Tran, NL; et al. (1997) Development of a dietary exposure potential model for evaluating
dietary exposure to chemical residues in food. J Exp Anal Environ Epidem 7:81-101.
Tsuda T., Inoue T., Kojima M. and Shigeru A. (1995) Market basket and duplicate portion estimation of dietary
intakes of cadmium, mercury, arsenic, copper, manganese, and zinc by Japanese adults. Journal of AOAC
International 78,1363-1368.
Tye, AM; Young, SD; Grout, NMJ; Zhang, H; Preston, S; Barbosa-Jefferson, VL; Davison, W;
McGrath, SP; Paton, GI; Kilham, K; Resende, L. (2003) Predicting the activity of Cd2+ andZn2+ in soil pore water
from the radio-labile metal fraction. Geochim. et Cosmochim. Acta 67(3): 375-85.
Tyler, G; Olsson, T (2001 a) Concentrations of 60 elements in the soil solution as related to the soil acidity. Eur. J.
Soil Sci. 52:151-165.
Tyler, G; Olsson, T (2001b) Plant uptake of major and minor mineral elements as influenced by soil acidity and
liming. Plant Soil 230:307-321.
Ullrich, S.M., Tanton, T.W., & Abdrashitova, S.A., "Mercury in the Aquatic Environment: A Review of Factors
Affecting Methylation" 2001, Critical Rev. Environ. Sci. Tech. 31,3, 241-293.
U.S. Army Center for Health Promotion and Preventative Medicine (Army). (2001). Development of Terrestrial
Exposure and Bioaccumulation Information for the Army Risk Assessment Modeling System (ARAMS). Contract
No. DAAD050-00-P-8365. March, 2001.
U.S. DHHS. (Department of Health and Human Services). (1996) National Center for Health Statistics. Third
National Health and Nutrition Examination Survey, 1988-1994, NHANES III Individual Foods Data File form the
Dietary Recall (CD-ROM). Public use data file documentation number 76700. Hyattsville, MD:.
U.S. DOE (U.S. Department of Energy). 2001. User's manual for RESRAD, version 6. ANL/EAD-4. Argonne
National Laboratory, Argonne, IL.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information.
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-35
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
U.S. DOE (Department of Energy). (1994) Estimating exposure of terrestrial wildlife to contaminants. ES/ER/TM-
125.
U.S. FDA (U.S. Food and Drug Administration). 2001. Total diet survey. Center for Food Safety and Applied
Nutrition, Office of Plant and Dairy Foods and Beverages. October. (Data summaries from
http ://vm. cfsan. fda. gov/~comm/tds-toc .html.)
Van Straalen, MM. (1996) Critical body concentrations: their use in bioindication. In: Van Straalen, NM;
Krivolutsky, DA, eds. Bioindicator systems for soil pollution. Dordrecht, The Netherlands: Kluwer; pp. 5-16.
Van Straalen, NM. (2001) Distribution-based extrapolation approaches in the risk assessment of metals in the
environment Available at: http://www.icme.coni/uploads/l~Fact3.pdf.
Van Wensem, J; Vegter, JJ; Van Straalen, NM. (1994) Soil quality criteria derived from critical body concentrations
of metals in soil invertebrates. Appl Soil Ecol 1:185-191.
Van Assche, F; van Tilborg, W; Waeterschoot, H. (1997) Environmental risk assessment for essential elements.
Case study: Zinc. International Zinc Association, Brussels, Belgium. Available at:
http://\vww.iza.com/zwo_org/Environment/040106.htm.
von Lindern, I; Spalinger, S; Petroysan, V; et al. (2003) Assessing remedial effectiveness through the blood lead:
Soil/dust lead relationship at the Bunker Hill Superfund site in the Silver Valley of Idaho. Sci Total Environ
303:39-170.
Waalkes, M. (1995) Metal carcinogenesis. In: Goyer, RA; Klaassen, CD; eds. Metal toxicology. New York:
Academic Press; pp. 47-67.
Wade, M.J., David, B.K., Carlisle, J. S. Klein, A. K. & Valoppi, L.M. "Environmental Transformations of Toxic
Metals" 1993, Occup. Med. 8,3, 575-601.
Wallace, B; Srb, A. (1961) Adaptation. Edgewood Cliffs, NJ: Prentice Hall.
Wallace, WG; Lopez, GR. (1996) Relationship between subcellular cadmium distribution in prey and cadmium
trophic transfer to a predator. Estuaries 19, 923-930.
Wallace, WG; Lopez, GR. (1997) Bioavailability of biologically sequestered cadmium and the implications of metal
detoxification. Mar. Ecol. Prog. Ser. 147, pp. 149-157.
Wallace, WG; Lopez, GR; Levinton, JS. (1998) Cadmium resistance in an oligochaete and its effect on cadmium
trophic transfer to an omnivorous shrimp. Mar. Ecol. Prog. Ser. 172:225-237.
Wallace, WG; Lee, BG; Luoma, SN. (2003) Subcellular compartmentalization of Cd and Zn in two bivalves. I.
Significance of metal-sensitive fractions (MSF) and biologically detoxified metal (BDM). Marine Ecology-Progress
Series 249,183-197.
Wallace, WG; Luoma, SN. (2003) Subcellular compartmentalization of Cd and Zn in two bivalves. II. Significance
of trophically available metal (TAM). Marine Ecology-Progress Series 257,125-137.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-36
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
Wang, XL. (2002) Interactions of trace metals and different marine food chains. Mar Ecol Prog Ser 243:295-309.
"Wang, XL; Fisher, NS. (1999) Assimilation efficiencies of chemical contaminants in aquatic invertebrates: a
synthesis. Environ Toxicol Chem 18:2034-2045.
Wang, XL; Zauke, G-P. (2004) Size-dependent bioaccumulation of metals in the amphipod Gammarus zaddachi
(Sexton 1912) from the River Hunte (Germany) and its relationship to the permeable body surface area.
Hydrobiologia 515:11-28.
Washington; Washington; Wilson; eds. (2001) Physiological pharmaceutics: barriers to drug absorption, 2nd ed.
New York: Taylor and Francis.
Wehrli, B, Dinkel, C; Muller, B. (1994) Measurement of benthic gradients in deep lakes with ion selective
electrodes and video endoscopy. Mineral Mag 58A:961-962.
Weis, CP; LaVelle, JM. (1991) Characteristics to consider when choosing an animal model for the study of lead
bioavailability. In: The proceedings of the international symposium onthebioavailability and dietary uptake of lead.
Sci Technol Lett 3:113-119.
Weis, C; Hemlein, K; Drexler, J. (1992) Collection and characterization of mine waste for the purpose of
determining lead bioavailability. Toxicologist 12:211.
Weis, C; Henningsen, G; Poppenga, R; et al. (1993a) Pharmacokinetics of lead in blood of immature swine
following acute oral and iv exposures. Toxicologist 13:175.
Weis CP; Poppenga, R; Thacker, BJ; et al, (1993b) Systemic availability of lead (Pb2+) from mineral extraction
process waste to young swine: a model for human children. Abstract. Presented at Annual ASTM Meeting, Boulder,
CO, Jul 26.
Welter, E; Calmano, W; Mangold, S; et al. (1999) Chemical speciation of heavy metals in soils by use of XAFS
spectroscopy and electron microscopical techniques. Fr J Anal Chem 364:238-244.
i
Weltje, L. (1998) Mixture toxicity and tissue interactions of Cd, Cu, Pb and Zn in earthworms (Oligochaeta) in
laboratory and field studies: a critical evaluation of the data. Chemosphere 36:264-2660.
Weltje, L. (2002) Bioavailability of Lanthanides to freshwater organisms. Speciation, accumulation and toxicity.
Delft University of Technology, Delft university Press, Delft, The Netherlands.
WestalL, JC, Zachary, JL; Morel, FM. (1976) A Computer Program for the Calculation of Chemical Equilibrium
Composition of Aqueous Systems, Technical Note 18. Department of Civil Engineering, Massachusetts Institute of
Technology, Cambridge, MA.
Wester, RC; Maibach, HI; Sedik, L; Melendres, J; Dizio, S; Wade, M. (1992) In vitro percutaneous absorption of
cadmium from water and soil into human skin. Fundam. Appl. Toxicol. 19:1-5.
t
Wester RC, Hui X, Barbadillo S, Maibach HI, Lowney YW, Schoof RA, Holm SE, Ruby MV. In vivo percutaneous
11/24/2004 Peer Review Draft
DISC L AIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-37
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
absorption of arsenic from water and CCA-treated wood residue. Toxicol Sci. 2004 Jun;79(2):287-95.
Weston, DP; Maruya, KA. (2002) Predicting bioavailability and bioaccumulation with in vitro digestive fluid
extraction. Environ Toxicol Chem 21:962-971.
Whits, PD; VanLeeuwen, P; Davis, BD; et al. (1998) The conceptual structure of the integrated exposure uptake
biokinetic modeller lead in children. Environ Health Perspect 106(Suppl. 6): 1513-30.
Whitehead, D.C. 1984. The distribution and transformations of iodine in the environment. Environ. Int. 10:321-339.
WHO (World Health Organization) (1996a) Trace elements in human health and nutrition. Chapter 3: Trace element
bioavailability and interactions. Geneva, pp. 23-41
WHO (1996b) Trace elements in human health and nutrition. Chapter 10: Manganese. Geneva, pp. 163-167.
WHO (1996c) Trace elements in human health and nutrition. Chapter 17: Arsenic. Geneva, Switzerland.
WHO/IPCS (World Health Organization/International Programme on Chemical Safety) (1983) Arsenic.
Environmental Health Criteria Document No. 18. Geneva, Switzerland.
WHO/IPCS (1990a) Methylmercury. Environmental Health Criteria Document No. 101. Geneva, Switzerland.
IPCS(1990b) Beryllium. Environmental Health Criteria Document No. 106. Geneva, Switzerland.
WHO/IPCS (1991) Inorganic mercury. Environmental Health Criteria Document No. 118. Geneva, Switzerland.
WHO/IPCS (1993) Biomarkers and risk assessment: Concepts and principles. Environmental Health Criteria
Document No. 155. Geneva, Switzerland.
WHO/IPCS (1995) Inorganic lead. Environmental Health Criteria Document No. 165. Geneva, Switzerland.
WHO/IPCS (1998) Copper. Environmental Health Criteria Document No. 200. Geneva, Switzerland.
WHO/IPCS (2002) Principles and methods for the assessment of risk from essential trace elements. Environmental
Health Criteria Document No. 228. Geneva, Switzerland
Wilkes, CR. (1998) Case study. In: Olin, S, ed. Lead. Boca Raton, FL: CRC Press LLC., pp. 183-224.
Williams, MW; Turner, JE (1981) Comments on softness parameters and metal ion toxicity. J inorg nucl Chem 43:
1689-1691.
Willemson, A; Vaal, MA; De Zwart, D. (1995). Microtests as Tools for Environmental Monitoring. NIPH and Env.
Plan. (RIVM) The Netherlands, Report #9,607042005, 39 pp.
Wilson, JB. (1988) The cost of heavy-metal tolerance: an example. Evolution 42:408-413.
Winge, D; Krasno, J; Colucci, AV. (1974) Cadmium accumulation in rat liver: correlation between bound metal and
pathology. In: Hoekstra, WG; Suttie, JW; Ganther, HE; et al., eds. Trace element metabolism in animals, Baltimore:
University Park Press; pp. 500-502.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-38
t
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
Woo, ES; Monks, A; Watkins, SC; Wang, AS; Lazo, JS. (1997) Diversity of metallothionein content and subcellular
localization in the National Cancer Institute tumor panel. Cancer Chemother. Pharmacol, 41:61-68.
Wu, L; Chen, J; Tanji, KK; et al. (1995). Distribution and biomagnification of selenium in a restored upland
grassland contaminated by selenium from agricultural drain water. Environ Toxicol Chem 4:733-742.
Xu, GB; Yu, CP. (1986) Effects of age deposition of inhaled aerosols in the human lung. Aerosol Sci Technol
5:349-357.
Xu, J; Thornton, I. (1985) Arsenic in garden soils and vegetable crops in Cornwall, England: implications for human
health. Environ Geochem Health 7:131-133.
Xue, H-B; Stumm, W; Sigg, L. (1988) The binding of heavy metals to algal surfaces. Water Res 22: 917-926.
Yeh, GT; Tripathi, VS (1991) A model for simulating transport of reactive multispecies components: Model
development and demonstration. Water Resour. Res. 27(12):3075-3094.
Yokel, RA; McNamara, PJ. (2001) Aluminum toxicokinetics: an updated minireview. Annu Rev Pharmacol Toxicol
88:159-67.
Yokel, RA; McNamara, PJ. (2001) Aluminum toxicokinetics: an updated minireview. Annu Rev Pharmacol Toxicol
88:159-67.
Young, LB; Harvey, HH. (1991) Metal concentrations in chironomids in relation to the geochemical characteristics
of surficial sediments. Arch Environ Contain Toxicol 21:202-211.
Zartarian VG, Ozkaynak H, Burke JM, Zufall MJ, Rigas ML, Furtaw EJ Jr. A modeling framework for estimating
children's residential exposure and dose to chlorpyrifos via dermal residue contact and nondietary ingestion. Environ
Health Perspect. 2000 Jun;108(6):505-14. .
Zhang, H; Davidson, W; Miller, S; et al. (1995) In situ high resolution measurements of fluxes of Ni, Cu, Fe and Mn
and concentrations of Zn and Cd in pore waters by DOT. Geochim Cosmochim Acta 59:4181—4192.
U.S. Environmental Protection Agency:
U.S. EPA. (1986a) Guidelines for the health risk assessment of chemical mixtures. Fed. Reg. 51(185):34014-14025.
U.S. EPA. (1986b) Air Quality Criteria for Lead Volumes I - IV. EPA 600/8-83-028 a-d.U.S. EPA 1992. Ground
Water Issue. Behavior of Metals in Soils. EPA/540/S-92/018 October 1992
U.S. EPA, (1987) Selection criteria for mathematical models used in exposure assessments: Surface water models.
US EPA Office of Health and Environmental Assessment, Washington, DC. EPA-600-8-87-042.
U.S. EPA. (1989a) Exposure factors handbook. EPA/600/8-89/043. Office of Health and Environmental Assessment.
U.S. EPA. (1989b) Risk assessment guidance for superfund (RAGS); Volume 1: human health evaluation manual
(HHEM), (Part A); Interim Final. EPA 540-1-89-002, OSWER 9285.70-02B. Office of Emergency and Remedial
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-39
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
Response, Washington, DC.
U.S. EPA. (1989c) Rapid bioassessment protocols for use in streams and rivers: Benthic macroinvertebrates and fish.
EPA 440/4-89-001. Washington, DC.
U.S. EPA. (1989d) Sediment classification methods compendium. PB92-231679. NTIS, Springfield, VA.
U.S. EPA (1990) Methods for derivation of Inhalation Reference Concentration and Application of Inhalation
Dosimetry. EPA/600/8-90/066F, Office of Research and Development, Washington D.C.
U.S. EPA (1991a) Guidelines for Developmental Toxicity Risk Assessment. U.S. Environmental Protection Agency.
Risk Assessment Forum, Washington, DC, EPA/600/FR-91/001,1991.
U.S. EPA (1991b) MINTEQA2, A geochemical assessment data base and test cases for environmental systems: Ver.
3.0 user's manual. EPA/600/3-91/-21. Athens, GA. Allison, JD; Brown, DS; Novo-Gradac, KJ.
U.S. EPA. (1991c) Statement of policy and guidance for petitions under Section 313 of the Emergency Planning and
Community Right-to-Know Act of 1986. Federal Register, Volume 56(100), May 23, p. 23703.
U.S. EPA. (1992a) Ground water issue: Behavior of metals in soils. EPA 540-S-92-018. October.
U.S. EPA. (1992b) Guidelines for exposure assessment. EPAASOO/Z-92/001. Fed. Reg. 57(104):22888-22938.
U.S. EPA. (1992c) Dermal exposure assessment: Principles and application. EPA/600/8-91/01 IB.
U.S. EPA (1994a) EPA health effects notebook for hazardous air pollutants—Draft. EPA/452/D-95-003.
U.S. EPA. (1994b) Guidance Manual for the Integrated Exposure Uptake Biokinetic Model for Lead in Children.
EPA/540/R-93/081, PB93-963510.
U.S. EPA. (1994c) Interim guidance on determination and use of water-effect ratios for metals. EPA-823-B-94-001.
Office of Water, Office of Science and Technology, Washington, DC.
U.S. EPA. (1994d) Water quality standards handbook, 2nd ed. Appendix L: Interim guidance on determination and
use of water effect ratios for metals. EPA-823-B-94-005a. Office of Water.
U.S. EPA. (1994b) Water quality standards handbook, 2nd ed., update #1. EPA-823-B-94-006. Office of Water.
U.S. EPA. (1995a) A guide to the biosolids risk assessments for the EPA Part 503 Rule. EPA 832-B-93-005. Office
of Water, Office of Wastewater Management, Washington, DC. September.
U.S. EPA. (1995b) Final water quality guidance for the Great Lakes system; final rule. Fed. Reg. 60(56):15366-
15425. March 23.
U.S. EPA. (1996a) Water quality criteria for document forthe protection of aquatic life in ambient water; 1995
updates. Washington, DC, Environmental Protection Agency. EPA 820-B-96-001
U.S. EPA. (1996b) Review of the National Ambient Air Quality Standards for Particulate matter, policy assessment
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-40
-------
t
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
of scientific and technical information. OAQPS Staff Paper. EPA-452/R-96-013.
U.S. EPA. (1996c) Recommendations of the technical review workgroup for lead for an approach to assessing risks
associated with adult exposures to lead in soil. EPA-540-R-03-001.
U.S. EPA. (1996d) Bioavailability of arsenic and lead in environmental substrates 1. Results of an oral dosing study
of immature swine. EPA 910/R-96-002 and http://www.epa.gov/rlOeartWoffices/oea/risk/bioavail.pdf
U.S. EPA (1997d) Progress report of the ecological committee onFIFRA risk assessment methods: VII. Terrestrial
risk assessment. Available at: http:/www.epa.gov/oppefedl/ecorisk/risk.htm.
U.S. EPA. (1997a) Compendium of tools for watershed assessment and TMDL development US EPA Office of
Wetlands, Oceans, and Watersheds, Washington, DC. EPA-841 -8-97-006.
U.S. EPA. (1997b). Mercury study report to Congress. Office of Air and Radiation. Available at:
http://www.epa.gov/oar/mercury.html.
U.S. EPA. (1997c). Air Quality Criteria for Particulate Matter. Volume I of III. EPA/600/P-95/001aF. Office of
Research and Development. Washington, DC.
U.S. EPA. (1997d) The incidence and severity of sediment contamination in surface waters of the United States, Vol
1: National sediment quality survey. EPA 823-R-97-006. Office of Water, Washington, DC.
U.S. EPA. (1997e)The incidence and severity of sediment contamination in surface waters of the United States, Vol
2: Data summaries for areas of probable concern. EPA 823-R-97-007. Office of Water, Washington, DC.
U.S. EPA. (1997f) The incidence and severity of sediment contamination in surface waters of the United States, Vol
3: National sediment contaminant point source inventory. EPA 823-R-97-008. Office of Water, Washington, DC.
U.S. EPA. (1997g) Ecological risk assessment guidance for superfund: process for designing and conducting
ecological risk assessments. EPA/540-R-97-006. Office of Solid Waste and Emergency Response, Washington, DC.
June.
U.S. EPA. (1997h) Exposure factors handbook. EPA/600/P-95/002Fc. Office of Research and Development, U.S.
EPA, Washington, DC. http://www.epa.gov/ncea/exposfac.htm
U.S. EPA. (1998a) Guidelines for ecological risk assessment. EPA 630-R-95-002F. Office of Research and
Development, National Center for Environmental Assessment, Risk Assessment Forum, Washington, DC.
U.S. EPA.(1998b) Human health risk assessment protocol for hazardous waste combustion facilities. EPA 530-D-
98:001.
U.S. EPA (1999a) Understanding variation in partition coefficient, Kd, values. Office of Air and Radiation. EPA
402-R-99-004A&B. \,
U.S. EPA (1999b) Compendium of methods-tor Ihe determination of inorganic compounds in ambient air. Office of
Research and Development. EPA/625/R-96/010.
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-41
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
U.S. EPA. (1999c) April 6-7, 1999. Integrated Approach to Assessing the Bioavailability and Toxicity of Metals in
Surface Waters and Sediments, a report to the EPA Science Advisory Board, Office of Water, Office of Research
and Development, Washington, DC.'EPA-822-E-99-001.
U.S. EPA. (1999d) IEUBK model bioavailability variable. EPA/540-F-00-006. Technical Review Workgroup for
Lead, Office of Emergency and Remedial Response, Washington DC. October.
http://www.epa.gov/superfund/programs/lead/products/sspbbioc.pdf
U.S. EPA. (1999e) Notice of Intent to revise aquatic life criteria for copper, silver, lead, cadmium, iron and
selenium. Fed. Reg. 64(209):58409-58410. October 28.
U.S. EPA. (1999f) Sociodemographic Data Used for Identifying Potentially Highly Exposed Populations,
EPA/600/R-99/060, Office of Research and Development, Washington, DC.
U.S. EPA. (2000a) Region 10 guidance for Superfund human health risk assessment: interim. EPA Region 10,
Seattle, WA. September.
U.S. EPA. (2000b) Supplementary guidance for conducting health risk assessment of chemical mixtures.
EPA/630/R-00/002.
U.S. EPA. (2000c) Ecological soil screening level guidance; draft. Office of Emergency and Remedial Response,
Washington, DC. July 10.
U.S. EPA. (2000d) Estimated Per Capita Water Ingestion in the United States. Office of Water. EPA-822-00-008.
April 2000.
U.S. EPA. (2000e) "An SAB Report: Review of the Biotic Ligand Model of the Acute Toxicity of Metals," prepared
by the Ecological Processes and Effects Committee of the Science Advisory Board, EPA-SAB-EPEC-00-0006.
U.S. EPA. (2001 a) Description and Evaluation of Atmospheric Mercury Simulation Using the CMAQ Model;
EPA/600/J-03/227.
U.S. EPA (2001b). ECO update; The role of screening-level risk assessments and refining contaminants of concern
in baseline ecological risk assessments. Office of Solid Waste and Emergency Response. OSWER 9345.0-14.
U.S. EPA. (2001c) Lead and lead compounds: Lowering of reporting thresholds; Community right-to-fcnow toxic
chemical release reporting; final rule. Fed. Reg. 66(11):4500-4547, January 17.
U.S. EPA. (2001d) User's guide for the integrated exposure uptake biokinetic model for lead in children
(IEUBKwin). Office of Emergency and Remedial Response. Washington, DC.
U.S. EPA. (200le) Integrated exposure uptake biokinetic model for lead in children (lEUBKwin vl.O). Office of
Emergency and Remedial Response, Washington, DC.
U.S. EPA. (2001f) Risk Assessment Guidance for Superfund (RAGS), Volume I: Human Health Evaluation Manual
(Part E, Supplemental Guidance for Dermal Risk Assessment) Interim. Office of Solid Waste and Emergency
Response, Washington, D.C.. EPA/540/R-99/005. September 2001. Available at
http://www.epa.gov/superfund/programs/risk/ragse/index.htm
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-42
%s^
f
-------
t
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
U.S. EPA (2001g) Guidance for characterizing background chemicals in soil at Superfund sites; External review
draft. OSWER. 9285.7-41. Office of Emergency and Remedial Response. [Replaced by Guidance for Comparing
Background and Chemical Concentrations in Soil for CERCLA Sites, EPA 540-R-01-003, September 2002.]
U.S. EPA. (2002a) Development of a framework for metals assessment and guidance for characterizing and ranking
metals; draft action plan. EPA/630/P-02/003A.
U.S. EPA. (2002b) Guidance for comparing background and chemical concentrations in soil for CERCLA sites.
OSWER 9285.7-41; EPA 540-R-01-003. Office of Emergency and Remedial Response, Washington, DC.
U.S. EPA. (2002c) Policy considerations for the application of background data in risk assessment and remedy
selection: Role of background in the CERCLA cleanup program. OSWER 9285.6-07P. Office of Emergency and
Remedial Response, Washington, DC.
U.S. EPA. (2002d) Draft action plan: Development of a framework for metals assessment and guidance for
characterizing metals. EPA/630/P-02/003A. Washington, DC.
U.S. EPA. (2002e) Child-Specific Exposure Factors Handbook. Interim Final. EPA/600/P-00/002B/01
U.S. EPA. (2002f) Procedures for Deriving Equilibrium Partitioning Sediment Benchmarks (ESBs) for the
Protection of Benthic Organisms: Metal Mixtures (Cadmium, Copper, Lead, Nickel, Silver, and Zinc). EPA-600-R-
02-011. Office of Research and Development. Washington, DC 20460.
U.S. EPA. (2002g) Equilibrium partitioning sediment guidelines (ESGs) for the protection of benthic organisms:
Metal mixtures (cadmium, copper, lead, nickel, silver and zinc). EPA-822-R-00-005.
U.S. EPA (2002h) TRIM.FaTE technical support document. Volume II: Description of chemical transport and
transformation algorithms. EPA/R-02-01 Ib. U.S. Environmental Protection Agency, Office of Air Quality Planning
and Standards, Research Triangle Park, NC.
U.S. EPA (2003a) Human Health Research Strategy. Office of Research and Development, Washington, DC.
September 2003. Available at: http://www.epa.gov/nheerl/humamiealth/HHRS_final_web.pdf.
U.S. EPA. (2003b) Draft final guidelines for carcinogen risk assessment. (External review draft, February 2003).
EPA/630/P-03/001A, NCEA-F-0644A. Risk Assessment Forum, U.S. EPA, Washington, DC. pp. 120. Available at:
http ://www.epa.gov/ncea/raf/cancer2003 .htm
U.S. EPA. (2003c) Guidance for the development of ecological soil screening levels. Office of Solid Waste and
Emergency Response, Washington, DC. March. OSWER Directive 92857-55.
U.S. EPA. (2003d) Literature review and evaluation of the atmospheric persistence of air toxic compounds; Internal
EPA Report, EPA/600/X-03/011, Research Triangle Park, NC. http://www.epa.gov/nerl/research/2003/gl-2.html.
U.S. EPA. (2003f) Reference dose for chronic oral exposure: cadmium. IRIS. Available at: http://www.epa.gov/iris/.
U.S. EPA. (2003g) Reference dose for chronic oral exposure: manganese. IRIS. Available at:
http ://www. epa/gov/irisA
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-43
-------
Note: The metals framework reference section is currently undergoing review, and revisions are anticipated.
U.S. EPA. (2003h). Estimation of Water Chemistry Parameters for Acute Copper Toxicity Tests. Office of Water,
Office of Science and Technology, Health and Ecological Criteria Division, Washington, D.C.
U.S. EPA. (2003i) Saltwater Conversion Factors for Dissolved Values. Office of Water, Office of Science and
Technology, Health and Ecological Criteria Division, Washington, D.C.
U.S. EPA. (2004a) Estimation of relative bioavailability of lead in soil and soil-like materials using in vivo and in
vitro methods. OSWER 9285.7-77, Office of Solid Waste and Emergency Response, Washington, DC.
U.S. EPA (2004b) Air Quality Criteria for Paniculate Matter (External Review Draft). Office of Research and
Development, National Center for Environmental Assessment. Available at:
http://crpub2.epa.gov/ncea/cfm/partmatt.cfm.
U.S. EPA (2004c) Air Toxics Risk Assessment Reference Library, Volume I. Technical Resource Manual. EPA-
453-K-04-001 A. OAQPS, RTP. Available at: http://www.epa.gov/ttn/fera/risk_atra_main.html.
U.S. EPA (2004d) A Discussion with the FIFRA Scientific Advisory Panel Regarding the Terrestrial and Aquatic
Level II Refined Risk Assessment Models (Version 2.0). U.S. Environmental Protection Agency, Office of
Pesticide Programs, Washington, DC. March 4,2004).
11/24/2004 Peer Review Draft
DISCLAIMER: This information is distributed solely for the purpose of peer review under applicable information
quality guidelines. It has not been formally disseminated by the EPA and should not be construed to represent any
Agency determination or policy.
7-44
-------
s
I
s
CA
§
u
at
o
z
g
At MJ
ft. **
5
t/i
t
s
s
01
2
O)
<
05
•a
i
EP
g
|
w
I
o-
<
's
I
u
1
as
^
* 3
•8
S
(L>
Silal f
43 £ &J3 « 13
f! !i 5
=3 CJ
1
ll
lit
ills
^ 1 .a -S 3
™ -. O
8?J5-a
cult to include
toxicily of in
tl
£
53 "o
l«s
S£
« .2
o o
O"
S|-a
isi
l
o
o
If tlfP
1S ^t3 o ^ ^ 'S3 w *^
w Si S ^ ** m ca S
^C^H,S«C£3^S
gSS'S'g^u'B
ca 1.5 8 I 5V S .a
&ll
f^ § oa
=3 §
1'^S
«I §
6^2S
"Sa w
«8P|
i|^
fc § "S'
^ B 3
§ < S
t* *1 -3
?tvti
S2 1
S a =3
-------
I 111
° 1§ & <3
g B T3 >
ff
1
^
"S
o
f
~ o
-------
t
ft
^ Hi O
||f,
*" ^ 'i
iPf
M «j ^ ^
"2 ~ « a
I SI § *
m y *»••< fl%
•^H v >pH O
« Kl =Q
a a-is
T8 g|
•g, So
5 S £
*•* M *3
O
••1?
^
.§ s 1
a"£3 ^
03 ^
o
,3
-3
&
a -a
O X
•a g $
.one
•'a
I
1
00
1
en
\
-------
Ill s1 i
>, I l"s 3
O 1)
i»h
> >? o. .
^ 4> ft O
tn *O ,. .S
jj-f3-
nil
»-0 § s
its!
!«;&,
8-»f &S
5 1 "i ,2 3
•2 « a °"«
1 II1
w «s a I •§
c o
2'S
The
tool
sedi
5 I Ji 3 t
.8
1
t
-------
|