EPA910/R-98-001
4>EPA
Alaska
United Stales Region 10 Idaho
Environmental Protection 1200 Sixth Avenue Oregon
Agency Seattle WA9B101 Washington
Office of Waste and Chemicals Management January 1998
Interim Final Guidance:
Developing Risk-Based Cleanup Levels
At Resource Conservation and Recovery
Act Sites in Region 10
-------
INTERIM FINAL GUIDANCE: DEVELOPING
RISK-BASED CLEANUP LEVELS AT
RESOURCE CONSERVATION AND RECOVERY ACT SITES IN REGION 10
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Waste and Chemicals Management
Seattle, WA 98101
Work Assignment No.
EPA Region
Date Prepared
Contract No.
Prepared by
Tetra Tech EM Inc. Project Manager
Telephone
EPA Work Assignment Manager
Telephone
R10018
10
January 5,1998
068-W4-0004
Tetra Tech EM Inc.
Paul Racette
(206) 587-4646
Dr. Marcia Bailey
(206) 553-0684
5 mfP*»ii«iW?ASKr*VtSiJWM*UMAST» WMSI-ftmimmWInM !1mi
-------
DISCLAIMER
This guidance document sets forth recommended approaches to conduct risk assessment and
other activities which are integral to the process of developing risk-based cleanup levels at RCRA
corrective action facilities in Region 10 of the United States Environmental Protection Agency.
Alternative approaches may be more appropriate at specific sites. All approaches used should be
described in full in documents generated as part of the cleanup level decision-making process.
This guidance is intended to be updated as scientific developments occur and U.S. EPA and
state rules and policies change. The user is encouraged to use the latest and best information
available for developing media cleanup levels. Users are also encouraged to submit suggestions
for updates to the guidance, or to report any errors noted, to Marcia Bailey, U.S. EPA Region 10,
at (206) 553-0684, or Bailey.marcia@epamail.epa.gov.
This guidance is intended as guidance to U.S. EPA Region 10 personnel and to RCRA-regulated
facilities undergoing corrective action or clean closures in Region 10. It does not constitute final
U.S. EPA action and does not constitute rulemaking. It is not intended, nor can it be relied upon,
to create any rights enforceable by any party in litigation with the United States government. U.S.
EPA officials may decide that the guidance provided in this document should be follwed, or may
decide to act at variance with the guidance, based on an analysis of specific site circumstances.
U.S. EPA reserves the right to change the guidance at any time without public notice.
-------
Chapter Ease
ACRONYMS AND ABBREVIATIONS viii
GLOSSARY xi
EXECUTIVE SUMMARY ES-I
1 INTRODUCTION I-l
1.1 Statutory and Regulatory Authorities Overview 1-2
1.2 Human Health and Ecological Risk Assessments:
An Overview ........ 1 -2
1.3 Risk Characterization Principles 1-9
2 STATE PROGRAMS 2-1
2.1 Alaska Department of Environmental Conservation ........................... 2-2
2.2 Idaho Division of Environmental Quality 2-2
2.3 Oregon Department of Environmental Quality 2-2
2.4 Washington Department of Ecology 2-3
3 DATA COLLECTION TO CHARACTERIZE FACILITY AND DETERMINE HAZARDOUS
CONSTITUENTS OF POTENTIAL CONCERN 3-1
3.1 Data Quality Objectives Process 3-1
3.1.1 Step 1: State the Problem 3-3
3.1.2 Step 2: Identify the Decision 3-3
3.1.3 Step 3: Identify the Inputs to the Decision 3-4
3.1.4 Step 4: Define Boundaries of the Study .............................. 3-4
3.1.5 Step 5: Develop a Decision Rule 3-5
3.1.6 Step 6: Specify Tolerable Limits on Decision Errors .................... 3-6
3.1.7 Step 7: Optimize the Design for Obtaining Data ................. 3-8
3.2 Data Useability 3-9
3.2.1 Data Sources 3-9
3.2.2 Documentation 3-10
3.2.3 Analytical Methods, Detection Limits, and Quantitation Limits 3-11
3.2.4 Data Quality Indicators 3-11
3.2.5 Data Review 3-12
3.2.6 Reports from Sampling and Analysis 3-13
3.3 Data Quality Assessment 3-13
S REMJUi«MrTASKrMVlSEJWNM,\MASTE* WTTKU/VI/YW
-------
CONTENTS (Continued)
Chapter Page
3,4 Identification of Hazardous Constituents of
Potential Concern 3-14
4 HUMAN HEALTH RISK ASSESSMENT PROCEDURES 4-1
4.1 Exposure Assessment: Identification of Exposure Pathways 4-2
4.2 Classification of Land Use 4-5
4.2.1 U.S. Environmental Protection Agency Land Use Policy and
Soil Cleanup Levels 4-6
4.2.2 Region 10 State Land Use Policies and Soil Cleanup Levels .............. 4-7
4.2.3 Land Use Policies and Other Media ..................... ........ 4-9
4.3 Identification of Promulgated Standards and Criteria 4-13
4.3.1 Federal Standards and Criteria 4-13
4.3.2 State Standards and Criteria . 4-13
4.4 Risk-based Screening 4-15
4.5 Exposure Assumptions ................................................ 4-11
4.5.1 Dermal Absorption Factors 4-23
4.5.2 Fate and Transport Models 4-25
4.6 Toxicity Assessment 4-39
4.6.1 Dose-Response Information ...................................... 4-39
4.6.2 U.S. Environmental Protection Agency Toxicity Factors ................. 4-43
4.63 Chemicals or Exposure Pathways With No U.S. Environmental Protection
Agency Toxicity Values 4-45
4.7 Calculation of Risk-Based Concentrations 4-57
4.7.1 Selection of Target Risks and Hazard Quotients ....................... 4-58
4.7.2 Risk-Based Concentration Calculation Equations 4-59
4.7.3 Examples of Risk-Based Concentration Calculations ................... 4-66
4.7.4 Adjustment of Risk-Based Concentrations for
Multiple Hazardous Constituents 4-69
4.7.5 Risk-Based Concentrations Based on Hazardous
Constituent Migration 4-72
4.7.6 Cleanup Levels for Lead 4-73
SmEFA*»in»iOTA5K«*YlSf luw
-------
CONTENTS (Continued)
Chapter Page
4.8 Uncertainly Analysis 4-75
4 J. 1 Data Uncertainly 4-76
4.8.2 Exposure Assessment Uncertainty 4-77
4.8.3 Toxicity Assessment Uncertainty 4-79
4.8.4 Cleanup Level Uncertainty 4-81
5 ECOLOGICAL RISK ASSESSMENT PROCESS 5-1
5.1 Screening-Level Ecological Risk Assessment 5-4
5.1J Preliminary Problem Formulation 5-6
5.1.2 Analysis 5-15
5.1.3 Preliminary Risk Characterization 5-22
5.1.4 Ecological Adversity 5-26
5.1.5 Uncertainty Analysis 5-27
5.1.6 Risk-Based Corrective Action 5-27
6 PROBABILISTIC RISK ASSESSMENT 6-1
6.1 Populations, Parameters, and Statistics 6-2
6.2 Probabilistic Analysis 6-5
6.2.1 Uncertainty and Variability 6-6
6.2.2 Other Methods of Estimating Uncertainty in Human Health
Risk Assessment and Ecological Risk Assessment 6-8
6.2.3 Monte Carlo Simulation 6-9
6.3 How to Conduct a Monte Carlo Simulation 6-10
6.3.1 Selection of an Appropriate Exposure Models 6-11
6.3.2 Selection of Probability Distributions 6-12
6.3.3 Running the Monte Carlo Simulation 6-20
6.4 Probabilistic Risk Assessment Case Study 6-21
6.4.1 Monte Carlo Simulation 6-21
6.4.2 Deterministic Risk Assessment 6-29
6.4.3 Summary of Results 6-29
6.5 Work Plans, Reports, and Presentations 6-31
S MFAJUtwMTASKWtlVSEXBWMWASTIa WPfMJ/lWS;? Mantac
-------
CONTENTS (Continued) ^
Chapter Page { /
? DETERMINATION OF COMPLIANCE WITH TARGET CLEANUP LEVELS 7-1
7.1 Data Quality Objective Steps 1 Through 3 . .. 7- i
7.2 Step 4: Study Boundaries .... 7-1
7.3 Step S; Develop Decision Rule 47-3
7.4 Step 6: Specify Tolerable Limits on Decision Errors 7-4
7.5 Step 7: Optimize the Design for Obtaining Data 7-5
7.6 Data Useability and Data Quality Assessment 7-7
7.7 Detection Limits 7-8
REFERENCES R-l
INDEX 1-1
ATTACHMENTS
A U.S. Environmental Protection Agency Carol Browner Memorandum on Risk Assessment
B RCRA/CERCLA Interface-Interim Final Guidance, EPA Region 10 Memorandum
C Examples of Data Quality Assessment Applications
D Data Useability Worksheets
E U.S. Environmental Protection Agency Region 9 Preliminary Remediation Goals
F U.S. Environmental Protection Agency Region 8 Superfund Technical Guidance Number
RA-03: Evaluating and Identifying Contaminants of Concern for Human Health
G U.S. Environmental Protection Agency Land Use Memorandum
H Soil Screening Guidance Chemical Properties Table C
1 Region 10 EPA Memorandum on Inorganic Arsenic
J World Health Organization Abstract on Dioxin Toxic Equivalency Factors
K EPA Region 1 Risk Updates, Revised Manganese Reference Dose
L An Approach for Determining Toxicity Values for Dermal Exposure, Oak Ridge National
Laboratory Internal Paper
s »mTMKrMvHtr*mAL'.MA*TWwromnniCTWii'n
-------
CONTENTS (Continued)
ATTACHMENTS
M EPA Guidance for Calculating Exposure Point Concentrations
N A Bibliography Related to Ecological Risk Assessment
O EPA Guidance and Policy for Probabilistic Risk Assessment
P Dealing with Data Below Detection Limits, Quality Assurance Course Module 492, EPA
National Center for Environmental Research and Qualify Assurance
EXHIBITS
Exhibit fee
1-1 Selection Process for Hazardous Constituents of Potential Concern ...... 1-5
1 -2 Human Health Risk Assessment Resource Conservation and Recovery Act
Cleanup Level Development 1-?
1-3 Ecological Risk Assessment Risk-Based Corrective Action Flowchart 1-8
4-1 Reduced Risk-Based Concentration Equations 4-62
4-2 Reduced Risk and Hazard Index Equations .... 4-64
4-3 Adjusting Cleanup Levels for Carcinogenic Compounds 4-71
4-4 Adjusting Cleanup Levels for Noncarcinogenic Health Effects 4-72
4-5 Risk Assessment and Cleanup Level Determination Process ..... 4-75
5-1 Sources for Ecological Risk Assessment Methods 5-1
5-2 Ecological Risk Assessment Steps and Decision Points 5-2
5-3 Ecological Receptor Characterization 5-7
5-4 Exposure Pathway Examples 5-9
5-5 Sample Endpoints and Relationships 5-13
¦ i*«i i« t»«;« juvi:• HN/U.XA5TER WTOi : sunn* V
-------
CONTENTS (Continued)
EXHIBITS
Exhibit Page
5-6 Suggested Measurement Endpoints 5-14
5-7 Sample Intake Calculation for The Deer Mouse 5-19
5-8 Evaluation Criteria for Current and Future Ecological Adversity 5-26
5-9 Possible Uncertainties In Screening-level Ecological Risk Assessment 5-28
6-1 Common Probability Distributions 6-14
6-2 Probability Density Function Intervals-Latin Hypercube Sampling 6-21
6-3 Exposure Parameter Probability Density Functions - Adult 6-23
6-4 Exposure Parameter Probability Density Functions - Child 6-24
6-5 Case Study Probability Density Function for Benzo(a)pyrene Risk-Based
Soil Concentrations Region 10 RCRA Risk-Based Cleanup Level Guidelines 6-25
6-6 Case Study Sensitivity Chart for Exposure Parameters 6-27
S ftEPA^i«HrTASKR\REV)SE2\flNAL\MASTEIl WPtXilrt|W7/» JUmYac
-------
CONTENTS (Continued)
TABLES
lahis Ease
4-1 Potential Exposure Pathways for Human Receptors 4-3
4-2 Standard Default Exposure Factors 4-18
4-3 U.S. Environmental Protection Agency Region 3 and Region 9
Soil, Water, and Air Exposure Pathways .,. 4-20
4-4 Washington State Model Toxics Control Act Cleanup Level Exposure Factors 4-21
4-5 Recommended Dermal Absorption Factors for Soil 4-24
4-6 Recommended Defaults for Dermal Exposure Factors .............................. 4-26
4-7 Derivation of The Volatilization Factor 4-28
4-8 Derivation of The Particulate Emission Factor 4-29
4-9 Dispersion Factor Values by Source Area, City, and Climatic Zone ................... 4-31
4-10 Derivation of The Soil Saturation Limit 4-34
4-11 Ranges of Human Potency and Cancer Potency Factors for Environmental
Mixtures of PCBs 4-53
6-1 Descriptive Statistics for Benzo(a)pyrene Risk-Based Criteria Probability
Density Function .................... 6-28
6-2 Population Percentiles for Benzo(a)pyrene Risk-Based Criteria Probability
Density Functions [[[ 6-29
-------
(THIS PAGE INTENTIONALLY LEFT BLANK)
i ir,TAJi;»JUEVHl}*»M
-------
ACRONYMS AM) ABBREVIATIONS
2,4-D
2,4-Dichlorophenoxyacetic Acid
40 CFR
Title 40 of the Code of Federal Regulations
95th UCL
95th upper confidence limit on the arithmetic mean
AAC
Alaska Administrative Code
ABS
Dermal absorption factor
ADD
Average daily dose
ADEC
Alaska Department of Environmental Conservation
Arl254
Aroclor 1254
ASTM
American Society for Testing and Materials
ATSDR
Agency for Toxic Substances and Disease Registry
AUF
Area use factor
AWQC
Ambient water quality criteria
BDL
Below detection limit
Cal/EPA
California Environmental Protection Agency
CDD
Chlorinated dibenzo-p-dioxins
CDF
Chlorinated dibenzofurans
GDI
Chronic daily intake
CERCLA
Comprehensive Environmental Response, Compensation, and Liability Act of 1980
CMS
Corrective measures study
COPC
Constituent of potential concern
CPF
Cancer potency factor
CSGWPP
Comprehensive state groundwater protection plan
CSM
Conceptual site model
DDT
Dichlorodiphenyltrichloroethane
DEFT
Decision error feasibility trials
DOD
U.S. Department of Defense
DOE
U.S. Department of Energy
DQA
Data quality assessment
DQ1
Data quality indicators
DQO
Data quality objectives
Ecology
Washington Department of Ecology
EDQL
Ecological data quality levels
EPA
U.S. Environmental Protection Agency
FR .
Federal Register
HI
Hazard index
HEAST
Health Effects Assessment Summary Tables
HHRA
Human health risk assessment
HQ
Hazard quotient
HSWA
Hazardous and Solid Waste Amendments of 1984
1AC
Idaho Administrative Code
1DEQ
Idaho Division of Environmental Quality
1EUBK
Integrated exposure uptake biokioetic model
IRIS
Integrated Risk Information System
LADD
Lifetime average daily dose
LDS#
Lethal dose 50
S A£P»*l'>->H«y.rJl£V1S£J*l>
-------
ACRONYMS AND ABBREVIATIONS (Continued)
L/day
Liters per day
LOAEL
Lowest-observable-adverse-effect-level
jUg/dL
Micrograms per deciliter
^g/L
Micrograms per liter
m3/hr
Cubic meter per hour
MCL
Maximum contaminant level
MCLG
Maximum contaminant level goal
MDEP
Massachusetts Department of Environmental Protection
mg/kg
Milligrams per kilogram
mg/kg-day
Milligrams per kilogram per day
MRL
Minimal risk levels
MTCA
Model Toxics Control Act
NCEA
National Center for Environmental Assessment
NOAA
National Oceanic and Atmospheric Administration
NOAEL
No-observable-adverse-effect-level
OAR
Oregon Administrative Rules
ODEQ
Oregon Department of Environmental Quality
PAH
Polynuclear aromatic hydrocarbon
PCB
Polychlorinated biphenyl
PDF
Probability density function
PEF
Particulate emission factor
PRA
Probabilistic risk assessment
PRG
Preliminary remediation goal
Q/C
Dispersion factor
OA
Quality assurance
QAPP
Quality assurance project plan
QC
Quality control
QMP
Quality management plan
RAGS
Risk Assessment Guidance for Superfund
RBC
Risk-based concentration
RCRA
Resource Conservation and Recovery Act
RfC
Reference concentrations
RfD
Reference dose
RFI
RCRA facility investigation
RI/FS
Remedial investigation and feasibility study
RME
Reasonable maximum exposure
SAP
Sampling and analysis plan
SOP
Standard operating procedures
SWMU
Solid waste management unit
svoc
Semivolatile organic compounds
TCDD
2,3»7»8-Tetrachlorodibenz0-p-dioxin
TEF
Toxicity equivalency factors
TEQ
Toxic equivalent
TPH
Total petroleum hydrocarbons
TR
Target risk
i •¦»*n»n«'.f*iK»«wse«MAMMASrB» vtmtw irrnt imi
-------
ACRONYMS AN® ABBREVIATIONS (Continued)
TRV
Toxicity reference value
TWA
Time-weighted average
VF
Volatilization factor
VFS
Soil-to-air volatilization factor
VP.
Groundwater-to-indoor air volatilization factor
voc
Volatile organic compounds
WAC
Washington Administrative Code
I ItffWJlSwi* TASXtMVtSEJtflNAUMASTER WfDUJnirtW I8*mvauur
x
-------
(THIS PAGE INTENTIONALLY LIFT BLANK)
s REP»i»iw>i*T«i;»«viSB'WAUMMl» wn»mji/»jm
-------
GLOSSARY
GENERAL RISK ASSESSMENT TERMS
absorbed dose: The amount of a substance penetrating the exchange boundaries of an organism after
contact. Absorbed dose is calculated from the intake and the absorption efficiency. It usually is
expressed as mass of a substance absorbed into the body per unit body weight per unit time
(e.g., mg/kg-day).
acute effects: Adverse human or ecological impacts caused by very short-term exposure to hazardous
constituents.
administered dose: The mass of a substance given to an organism and in contact with an exchange
boundary (e.g., gastrointestinal tract) per unit body weight per unit time (e.g., mg/kg-day).
carcinogenic risks: Incremental probability that an individual will develop cancer over a lifetime as a
result of exposure to a carcinogen.
chronic effects: Adverse human or ecological impacts caused by long-term exposure to hazardous
constituents.
cleanup levels: The hazardous constituent concentrations to which a contaminated environmental
medium (e.g., soil, groundwater, surface water, sediment) must be remediated, EPA establishes cleanup
levels on a facility-by-facility basis during the remedy selection process. Determination of target
cleanup levels is a risk management decision.
conceptual site moid: Schematic and/or narrative presentation of information about a facility
conditions including known and potential sources of releases of hazardous constituents, exposure
pathways, receptors, and all available information about constituents of potential concern at the facility.
data quality objectives (DQOs): Qualitative and quantitative statements relevant to facility-specific
circumstances which are to ensure that sampling and analysis data of known, documented and adequate
quality are obtained to support a risk assessment
dose-response evaluations: The process of quantitatively evaluating toxicity information and
characterizing the relationship between the dose of a hazardous constituent received and the incidence of
adverse health effects in tie exposed population.
exposure pathways: The various ways a hazardous constituent in a given medium can come into
contact with a receptor. For example, possible exposure pathways for contaminated soil include
ingestion of the soil, inhalation of the soil as dust, inhalation of volatile organics emanating from the soil,
and dermal contact with the soil.
exposure route: The way an environmental hazardous constituent can enter an organism. The three
primary routes are ingestion, inhalation, and dermal contact.
S REMASTER WPCnilrtlftT/*} l*aroi«e
xi
-------
RCRA hazardous constituent: Substances that have been shown in scientific studies to have toxic,
carcinogenic, mutagenic, or teratogenic effects on humans or oiler life forms. RCRA hazardous
constituents used in 40 CFR 261, Appendix VIII.
CERCLA hazardous sabstance: Elements, compounds, mixtures, solutions, and substances, which,
when released into the environment may present substantial danger to the public health or welfare or the
environment. The terms means any substances designated under the federal water pollution control act,
CERCLA, RCRA, the clean air act, and the toxic substances control act, CERCLA hazardous substance
are listed in 40 CFR 3024.
hazard index: An estimate of the risk associated with a specified exposure to a noncarcinogenic
hazardous constituent, expressed as the ratio of a substance exposure level over a specified time period to
a reference dose for that substance derived from a similar exposure.
lifetime average daily intake: Exposure expressed as mass of a substance contacted per unit body
weight per unit time, averaged over a lifetime.
linearized multistage model: One of a number of mathematical models and procedures used to
extrapolate from carcinogenic responses observed at high doses to responses expected at low doses.
95 percent upper confidence limit (95 UCL) on the arithmetic mean: Value that, when calculated
repeatedly for randomly drawn subsets of facility data, equals or exceeds the true mean 95 percent of the
time. Provides a conservative estimate of the average concentration.
quality assurance project plan (QAPP)t Describes the policy, organization, functional activities, and
quality assurance and quality control protocols necessary to achieve DQOs dictated by the intended use
of the data.
receptor: An organism (human, plant, or animal) that is potentially exposed to chemical contamination
from a facility.
reference dose: An estimate (with uncertainty spanning an order of magnitude or greater) of a daily
exposure level for the human population, including sensitive subpopulations, that is likely to carry no
appreciable risk of deleterious effects during a lifetime.
cancer potency factor: A plausible upper-bound estimate of the probability of an individual developing
cancer as a result of a lifetime of exposure to a particular level of potential carcinogen.
toxicity value: A numerical expression of a dose-response relationship for a particular substance. The
most common values used in EPA risk assessments are reference doses (for noncarcinogenic effects) and
cancer potency factors (for carcinogenic effects).
weight-of-evidence classification: An EPA classification system for characterizing the extent to which
the available data indicate that an agent is a human carcinogen. Recently, EPA has developed weight-of-
evidence classification systems for some other kinds of toxic effects, such as development effects.
f»AMt?t«)!l TASK»\R£VTSEJ\nNAl,\MASTER WFOM2/3
xir
-------
weigbt-of-evidence: Classification of evidence from human and animal studies into categories of
sufficient, limited, inadequate, no data, or no evidence of cancer effects.
ECOLOGICAL TEEMS USED BY THE U.S. ENVIRONMENTAL PROTECTION AGENCY
(1992f and 19941)
area use factor; The fraction of an organism's home range, breeding range, or foraging range to the
area of contamination or the facility area under investigation.
assessment endpoiat: A clearly defined statement of the environmental value that is to be protected.
bioaccrnnulation: General term describing a process by which chemicals are taken up by an organism
either directly from exposure to a contaminated medium or by consumption of food containing the
chemical, EPA's 19941 (and new 1997) Ecological Risk Assessment Guidance for Superfund: Process
for Designing and Conducting Ecological Risk Assessments.
bioavailability: The degree to which a material in environmental media is assimilated by an organism.
constituents of potential concern: Chemicals detected at a facility which have the potential to
adversely affect ecological receptors because of their concentration, distribution, and mode of toxicity.
complete exposure pathway: Includes a source or release from a source, an exposure route {that is,
soil), and an exposure point (that is, dermal contact). If the exposure point differs from the source,
transport and exposure media are also included in the exposure pathway.
baseline ecological risk assessment: A comprehensive ecological risk assessment where uncertainties
of the screening-level assessment are reduced, and nonfaci 1 ity-specific TRVs are refined by
incorporating data on facility-specific results from fate and transport modeling as well as exposure and
ecological effects analyses.
conceptual site model: The conceptual site model describes a series of working hypotheses of how a
stressor might affect ecological components. It also describes the ecosystem potentially at risk, the
relationship between assessment endpoint and measurements and exposure scenarios.
ecological effect: An effect where the stressor acts directly on the ecological component of interest
(direct effect). Also, an effect where the stressor acts on supporting ecological components of the
ecosystem, which in turn have an indirect effect on the ecological components of interest
ecological niche: The functional position of an organism in its environment, comprising the habitat in
which the organism lives, the periods of time during which it occurs and is active there, and the resources
it obtains there.
ecological receptor: The biotic component (for example, organism, population, community) exposed to
a stressor.
5 ¦>!« TASKrAEVtSErflKALMASTOt tiaoue
siii
-------
ecological relevance: This term is typically used In the context of Identifying assessment endpoints.
Ecologically relevant assessment endpoints reflect important ecosystem components that are functionally
related to other ecosystem components and assessment endpoints
ecological risk assessment: "The process that evaluates the likelihood that adverse ecological effects
may occur or are occurring as a result of exposure to one or more stressors,
ecosystem: The biotic community and the abiotic environment within a specified location in space and
time. The abiotic environment includes non-living environmental media (for example, water, soil,
sediment) and associated physical and chemical influences (far example, light, temperature, pH,
humidity).
ecotone: A narrow and fairly sharply defined transition zone between two or more different biotic
communities. These "edge" communities are typically species-rich, Reference: Allaby, M., editor.
1994. The Concise Oxford Dictionary of Ecology. Oxford University Press.
exposure: The contact or co-occurrence of a stressor with an ecological a receptor.
exposure area: A contaminated habitat where ecological receptors may be exposed to hazardous
constituents that may cause adverse ecological effects.
exposure point concentration: The concentration of a constituent that an ecological receptor is exposed
to through exposure routes such as ingestion, dermal contact, inhalation.
exposure profile: The product of the exposure analysis step in the ecological risk assessment. The
exposure profile summarizes the magnitude and spatial and temporal patterns of exposure for the
exposure scenarios described in the conceptual site model.
exposure scenario: A set of assumptions concerning how an exposure may occur, including
assumptions about the exposure setting, stressor characteristics, and activities that may lead to exposure.
guild: A group of species that share common ecological characteristics (for example, feeding behavior).
Guilds are defined by guild descriptors (for example, feeding guild) that may be general or specific.
Guilds may contain many or few species in response to the number of guild descriptors.
hazard index: A sum of hazard quotients for hazardous constituents of ecological concern with the
same ecological effect endpoint and/or the same mechanism of toxic effect.
hazard quotient: The ratio of a single exposure concentration or dose to a toxicity value selected for the
risk assessment (for example, lowest observed adverse effect level or no observed adverse effect level).
keystone species: A species, the presence or abundance of which can be used to assess the extent to
which ecological components of an ecosystem are impacted.
lowest observed adverse effect level: The lowest level of a stressor evaluated in a test that causes
statistically significant differences from the controls.
S REPA\R^w|*-TASKrJlEV1SEjyftNALWASTER M
-------
measurement endpoint: A measurable ecological characteristic that is indirectly related to the
assessment endpoint.
no observed adverse effect level: The highest level of a stressor evaluated in a lest that does not cause
statistically significant differences from lie controls.
population: A group of organisms of the same species, occupying a given area, and capable of
interbreeding.
risk characterization: A phase of ecological risk assessment that integrates the results of the exposure
and ecological effects analysis to evaluate the likelihood of adverse ecological effects associated with
exposure to a stressor. Hie 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.
screening-level ecological risk assessment: Simplified assessments that can be conducted with limited
data by assuming values for parameters for which data are lacking. Where data are lacking, assumed
values are biased in the direction of overestimating risk so the assessment can provide a defensible
conclusion of no unacceptable ecological risk.
stressor: Any physical, chemical, or biological entity that can induce an adverse response.
subpopulation: A portion of the population known or likely to be exposed to hazardous constituents at
or from the facility.
lexicological test: Tests used to evaluate relative potency of a chemical by comparing its effect on
living organisms with the effect of a standard preparation on the same type of organism.
trophic level: A functional classification of taxa within a community that is based on feeding
relationships (for example, aquatic and terrestrial plants make up the first trophic level, and herbivores
make up the second).
% .R£F*'Jm»»TA5K«m»BNAOM«TEII WPWI»i«»i thorax
-------
EXECUTIVE SUMMARY
This guidance document provides procedures for developing human and ecological risk-based cleanup
levels for facilities undergoing corrective action and clean closure under the Resource Conservation and
Recovery Act (RCRA). The procedures are intended for use by U.S. Environmental Protection Agency
(EPA) permit writers and regulatory compliance officials as well as RCRA-regulated facilities.
This guidance document references EPA Region 10 state RCRA corrective action programs and relevant
laws and regulations, EPA guidance on determining data quality objectives and performing a data quality
assessment is summarized. The major risk assessment steps, including data evaluation, exposure
assessment, toxicity assessment, and risk characterization, are described. Methods for determining human
and ecological risk-based cleanup levels using deterministic and probabilistic techniques are presented.
Screening-level ecological risk assessment methods are described. Procedures to follow when determining
compliance with cleanup levels are also described. Federal, stale, and general literature references that
provide further details on the risk calculation processes are identified throughout the document.
Consultation with Region 10 human health scientists and ecologists is recommended if complex aspects of
the risk assessment process are encountered.
5 REPA* J«*J}*vrASKrAEVIser/wALXMAHlft WKKIJ1 4UUMZ/51
ES-1
-------
-------
CHAPTER I
INTRODUCTION
This guidance document provides procedures for developing human and ecological risk-based cleanup
levels for contaminated facilities undergoing corrective action and clean closure under lie federal Resource
Conservation and Recovery Act (RCRA). The procedures are intended for use by permit writers and
enforcement officials as well as by RCRA-regulated facilities. The guidance is intended to enable RCRA
project managers to recommend cleanup level determinations based on risks posed to human health and the
environment by releases from the facility. The document also describes situations likely to require expert
technical assistance. A risk assessor or lexicologist should be involved at the beginning of the RCRA
facility investigation or corrective action order negotiation process.
This guidance document updates and supersedes the previous U.S. Environmental Protection Agency
(EPA) Interim Final Guidelines for Developing Risk-Based Cleanup Levels at RCRA Sites in Region 10
document (1992a). The document also complements RCRA Facility Investigation (RFI) Guidance
(EPA 1989a). The approach in this document is intended to be consistent with the Comprehensive
Environmental Response, Compensation, and Liability Act of 1980 (CERCLA), known as Superfund.
In circumstances where there are no RCRA-specific guidelines or rules, Superfund guidance should be
used. It is EPA Region 10's objective that cleanup activities conducted under the auspices of either
Superfund or RCRA are comparably protective of human health and the environment (EPA 1994a).
Sections 1,1 through 1.3 provide overviews of (1) EPA's statutory and regulatory authorities for
requiring corrective action, (2) processes for setting the human health and ecological cleanup levels, and
(3) risk characterization principles, respectively.
Additional sections of this guidance document summarize EPA Region 10 state programs (Chapter 2),
data collection and useability issues (Chapter 3), human health risk-based methods for calculating
cleanup levels (Chapter 4), ecological screening-level risk assessment and cleanup levels (Chapter 5),
probabilistic risk assessment methods and applications (Chapter 6), and determination of compliance
with cleanup levels (Chapter 7).
* fcEf>A\ftHH(KTASMr$EV15E2#fN&lyMlASTEIt lUumVUc
-------
1.1 STATUTORY AND REGULATORY AUTHORITIES OVERVIEW
Application of the procedures described in this guidance is intended for RCRA facilities where releases
of hazardous constituents require corrective action or where corrective action is necessary so that a
RCRA-regulated unit may be clean-closed. EPA derives its authority for compelling corrective action at
facilities regulated under RCRA Subtitle C by a variety of statutory provisions. Before the Hazardous
and Solid Waste Amendments of 1984 (HSWA) were passed, the RCRA corrective action authorities
were limited to Section 7003, which provides authority to compel action where solid or hazardous waste
may present an imminent and substantial endangerment to human health or the environment, and
Section 3013, which provides authority for requiring investigations where the presence of hazardous
waste or releases of hazardous waste may present a substantial hazard to human health or the
environment. HSWA substantially expanded corrective action authorities for both permitted RCRA
facilities and facilities operating under interim status. Section 3004(u) of HSWA requires that any
RCRA permit issued after November 8, 1984, address corrective action for releases of hazardous wastes
or hazardous constituents from any solid waste management unit. Section 3004(v) authorizes EPA to
require corrective action by permitted facilities beyond the facility boundary where appropriate.
Section 3008(h) provides the authority to require corrective action when there has been a release of
hazardous waste or hazardous constituents from a RCRA facility operating under interim status. EPA
authority for setting cleanup levels at closing units stems from RCRA Section 3004 with regulations
promulgated in Title 40 of the Code of Federal Regulations parts 264 and 265, subpart G» which require
that, among other things, the facility must be closed in a manner that "controls, minimizes or eliminates,
to the extent necessary to protect human health and the environment, post-closure escape of hazardous
waste, hazardous constituents, leachate, contaminated run-off, or hazardous waste decomposition
products to the ground or surface waters or to the atmosphere."
1.2 HUMAN HEALTH AND ECOLOGICAL RISK ASSESSMENTS: AN
OVERVIEW
Human health risk assessment (HHRA) and ecological risk assessment methods can be used to either
(1) calculate the risk associated with exposure to a hazardous constituent or (2) calculate a risk-based
concentration (RBC) that represents a level of exposure to a hazardous constituent that is not expected to
result in unacceptable risks to human health or the environment health. RBCs may then be used as a
5. imlrTASKnUVISElWNWMASRK Wmtte 1 "2
-------
basis for determining risk-based cleanup levels. As described in Chapter 5, screening-level ecological
risk assessments may be performed to determine the need for settings ecological RBCs. The human
health and ecological risk assessment processes are similar in that they involve the identification of
potential exposure pathways, the assessment of constituent toxicity, and the characterization of risk
based on exposure and toxicity Information. The output of a risk assessment is typically an estimate of
the risk of getting cancer over a lifetime (for humans) or the likelihood of other toxic effects (referred to
as hazards) occurring in humans or ecological receptors. The direct calculation of cancer risks or hazards
can incorporate cumulative exposure occurring from more than one medium (for example, soil and
groundwater exposures). Risk assessments require that data of sufficient quantity and quality be
collected to determine the nature and magnitude of contamination released from a facility and the
resulting level of potential exposures to human and ecological receptors. Uncertainty associated with the
various risk assessment steps must be described, and in some cases, it may be quantified. When relevant,
both human and ecological procedures should be applied at each facility, and the processes can be
conducted either simultaneously or sequentially. For facilities where it has been decided that both
human and ecological receptors should be protected, the protective levels for each should be compared,
and the more stringent of the two should be proposed as the cleanup level.
When risk assessment methods are used to calculate RBCs, the output is the concentration of a specific
hazardous constituent in a specific medium (for example, soil) that will not cause unacceptable cancer
risks, systemic hazards, or ecological effects. For some hazardous constituents, federal standards or
criteria have been promulgated, such as maximum contaminant levels (MCL) under the federal Safe
Drinking Water Act. Criteria and standards consider exposure and toxicity information but may also
incorporate other factors such as cost, treatment technology, and available analytical methods. Criteria
and standards promulgated by both federal and state agencies should be considered when making
cleanup level decisions, but they may not be deemed sufficiently protective on a site-specific basis.
Chapter 2 summarizes state programs, while Chapters 4 and 5 provide additional details on federal and
state agency programs related to cleanup level determination methods.
Where promulgated criteria and standards are not available or are determined to be insufficiently
protective of human health or ecological components, RBCs should be calculated using risk assessment
methods. Hazardous constituents of potential concern (COPC), contaminated media, and important
<; m»»rTASKKftEVlSEraHALMASTO
-------
exposure pathway information are first identified. Exposure assumptions and toxicity values are then
incorporated into risk assessment equations to derive RBCs for specific environmental media that do not
pose unacceptable cancer risks, hazards, or ecological risks. Tie RBCs can be used in the risk
management process to support the setting of cleanup levels.
A quantitative approach to deriving human health-based RBCs is presented in this guidance, while the
development of ecological RBCs usually involves a tiered approach including a screening-level
assessment (a qualitative assessment that is presetted in this document) and a subsequent comprehensive
assessment For a HHRA, exposure and toxicity information is used to calculate specific constituent
RBCs for each environmental medium. As explained In Section 4.4, constituent screening can also be
performed using available human health risk-based concentrations that are based on significant exposure
pathways. The HHRA is concerned with just one type of receptor: the human populations potentially
affected by the facility. In an ecological risk assessment, there may be many potential ecological
receptors, including both aquatic and terrestrial organisms. A complex set of exposure pathways may be
associated with these receptors, based on how constituents may migrate through soil, water, sediment, air,
and the food chain. Each facility will have a specific set of conditions based on the types of habitat and
ecological receptors present in exposure areas. Ecological effects that may result from the complex
interrelationships among chemicals taken up by the ecological receptors must be weighed and assessed in
a series of judgements on the relative risks. In this way, one or more constituent-pathway-receptor
combinations are identified as the greatest threats to ecological health at a facility.
Flowcharts presented in Exhibits 1-1 through 1-3 summarize the dual process of developing cleanup levels
for a RCRA facility. Only environmental data of sufficient
quality and quantity are used to identify COPCs. Data needs Exhibit l-l demonstrates how COPCs
are Identified by considering data quality,
specific to the future determination of human health and background chemical concentrations, and
ecological cleanup levels should be determined and risk-based screening,
incorporated into the RCRA facility investigation (RFI)
work plan.
S HEMMama
-------
OATE OB-ZBif ... DM FILENAME RVCAQ\251 RI001B0700\SE(.CTNDS4
YES
f
ARE AVAILABLE HUMAN HEALTH
PROMULGATED STANDARDS AND CRITERIA\
SUFFICIENTLY PROTECTIVE? IF YES,
USE AS CLEANUP LEVELS, If NOT
AVAILABLE OR NOT PROTECTIVE,
^ PROCEED TO RISK-BASED SCREENING
SUSPECTED CONSTITUENTS
DATA USABILITY REVIEW
- ANALYTICAL METHODS
- DETECTION LIMITS
- DETECTION FREQUENCY
- QA/QC
CONTAMINANT DETECTED
YES
f HUMAN HEALTH RISK-BASED
( SCREENING
I SITE CONCENTRATION IS GREATER
\ THAN RESIDENTIAL RISK-BASED
Vw CONCENTRATION
REMOVED
FROM FURTHER
CONSIDERATION
VES
NO
REMOVED
FROM FURTHER
CONSIDERATION
i >
ECOLOGICAL PROMULGATED
CRITERIA AND STANDARDS SCREENING
COMPARE SITE CONCENTRATIONS TO
NONSITE-SPECIF1C SCREENING LEVELS
- EPA AMBIENT WATER QUALITY
CRITERIA
- SEDIMENT QUALITY GUIDELINES
SITE CONCENTRATION IS GREATER
THAN CRITERIA 01 STANDARD
KETI
NO
REMOVED
FROM FURTHER
CONSIDERATION
CONSTITUENTS Of POTENTIAL \
CONCERN, DEVELOP HUMAN
HEALTH RISK-BASED
CLEANUP LEVEL
LEGEND
I
CONSTITUENTS OF POTENTIAL
ECOLOGICAL CONCERN
EVALUATE SIGNIFICANCE
OF FACILITY CONCENTRATIONS
AND NATURALLY OCCURRING
OR ANTHROPOGENIC
CONCENTRATIONS
O
ANALYSIS PHASE OF THE
PHASE I SCREENING-LEVEL
ECOLOGICAL RISK-ASSESMENT
EXHIBIT 1-1
HUMAN HEALTH RISK
ECOLOGICAL RISK
MODIFIED FROM U.S. DEPARTMENT OF ENERGY (1995) AND FEDERAL REGISTER 19439 (May 1,1996)
SELECTION PROCESS FOR HAZARDOUS
CONSTITUENTS OF POTENTIAL CONCERN
"H; TETRA TECH EM INC.
-------
(THIS PAGE INTENTIONALLY LEFT BLANK)
I vREP4Aii»irtASK«\R£VISE21JnNAI\MAS,ny( WFO!S!JttimiW*Kivl2/ll*7/H
-------
Sampling and analysis activities undertaken during the RFl should provide adequate data lo evaluate all
appropriate exposure pathways and chosen ecological endpoints. Chapter 3 provides guidance on data
collection, data useability, and data evaluation issues. Steps to identify COPCs are also discussed in
Chapter 3. Risk-based screening can be performed lo focus cleanup level determinations on hazardous
constituents that represent significant health concerns. Risk-based screening should be performed after
facility concentrations have been compared with promulgated standards and criteria.
as cleanup levels. If no promulgated standards exist for a specific COPC or pathway, appropriate exposure
assumptions should be made and combined with toxicity criteria lo calculate RBCs. If no toxicity criteria
exist, an experienced risk assessor should be consulted. If numerous COPCs are present on a facility,
cleanup levels may require adjustment to assure that the total facility risk or hazard remaining after
cleanup is acceptable.
Problem formulation, the first step in the ecological assessment 7 ! L
pathways of ecological concern. Ecological COPCs are also
identified during problem formulation. Incomplete exposure
pathways are removed from the ecological risk assessment process but should be reevaluated if exposure
pathways may be created based on future land use plans. If exposure pathways are identified for specific
receptors, it should be determined whether promulgated standards or criteria exist for concentrations of
COPCs for the appropriate medium. If applicable RBCs are identified, these RBCs should be considered
cleanup levels. If not, it is likely that a comprehensive ecological risk assessment will be necessary to
identify facility-specific cleanup levels. Interim corrective action may be determined appropriate if acute
health effects (for example, fish kills) are occurring. Long-term risks require evaluation following interim
Exhibit 1-2 demonstrates how
human health-based cleanup
levels are determined.
Once facility hazardous constituents are identified, the first step in
the HHRA is identification of land use and exposure pathways. If
promulgated standards and criteria are available for an identified
exposure pathway and hazardous COPCs, these criteria can be used
process, includes a facility reconnaissance to identify ecological
components (habitats and biota) and potentially complete exposure
Exhibit 1 -3 shows the steps in
developing cleanup levels
based on ecological
assessment
S S !ss?§« TASls* R£VJSEJ>FTNA!\MASTER WPCM Ji Itasua*
-------
(THIS PAGE INTENTIONALLY LEFT BLANK)
«. *EMftlwirTASU'MVnuiFnuaiMASIE* wmiSMMnilBffl»llfflWf»li
-------
YES
NO
NO
YES
CONTACT RISK
ASSESSMENT
SPECIALIST
CLEAN UP TO
SPECIFIC LEVEL
UNLESS DETERMINED
TO BE INADEQUATE
IDENTIFICATION OF EXPOSURE
PATHWAYS AND LAND USE
USE RISK-BASED SCREENING
TO IDENTIFY CONSTITUENTS
CONTRIBUTING GREATEST RISK
IDENTIFY EXPOSURE ASSUMPTIONS
FOR SELECTED PATHWAYS
AND LAND USE
ARE PROMULGATED
CRITERIA OR STANDARDS
AVAILABLE?
DO TOXICITY VALUES EXIST?
(REFERENCE DOSES, SLOPE FACTORS)
IDENTIFY CONSTITUENTS OF POTENTIAL
CONCERN BASED ON DATA REVIEW
SCREEN OUT CONSTITUENT
IF FACILITY CONCENTRATION
IS LESS THAN BACKGROUND
CONCENTRATION
CALCULATE CLEANUP LEVELS BASED ON
EXPOSURE ASSUMPTIONS, TOXICITY VALUES.
TARGET RISKS AND HAZARDS, AND OTHER
CHEMICAL-SPECIFIC PARAMETERS.
CONSIDER CUMULATIVE EFFECTS OF
MULTIPLE CONSTITUENTS.
EXHIBIT 1-2
HUMAN HEALTH RISK ASSESSMENT
RCRA CLEANUP LEVEL DEVELOPMENT
"It TETRA TECH EM INC.
I-7
-------
(THIS PAGE INTENTIONALLY LEFT BLANK)
S AEMAM«U« TA$K*tt£VTSEl\nNM.\MASTW WI»D\l5|.*IWI*mmitm«7/U
-------
PHASE SCREENING-LEVEL
INTERIM CORRECTIVE ACTION
TO REDUCE ACUTE RISKS
REEVALUATE RISKS FOLLOWING
ACTION
NO
YES
COPCsEXCEED
NON-SITE-SPECIFIC RBSLs?
J YES
CORRECTIVE ACTION TO
NON-SITE-SPECIFIC
RBSLs PRACTICABLE?
i/W
INTERIM CORRECTIVE
ACTION APPROPRIATE
TO REDUCE ACUTE RISKS
IB
PHASE II COMPREHENSIVE
ECOLOGICAL RISK ASSESSMENT
NO
NO
_j», CORRECTIVE ACTION PROGRAM
INTERIM CORRECTIVE
ACTION APPROPRIATE?
m
—I
NO
COCs EXCEED
SITE-SPECIFIC RBSLs?
YES
CORRECTIVE ACTION TO
SITE-SPECIFIC RBSLs
PRACTICABLE?
CONTINUED MONITORING
REQUIRED?
________
COMPLIANCE MONITORING
NO
YES
NO
NO FURTHER ACTION
FIGURE 1-3
MODIFIED FROM AMERICAN SOCIETY TESTING
AND MATERIALS (1995b)
ECOLOGICAL RISK ASSESSMENT
RISK-BASED CORRECTIVE
ACTION FLOWCHART
mc ENVIRONMENTAL MANAGEMENT, INC.
-------
(THIS PAGE INTENTIONALLY LEFT BLANK)
* MMW r*.T A.wr«vts£?*!N *L«*rr» «mn I >t •» Wimw i in i mm Ihm
-------
action; the focus of this guidance document is on long-term, chronic risks. Once all the necessary data are
collected and evaluated, cleanup levels can be determined,
IJ RISK CHARACTERIZATION PRINCIPLES
Data collected at a facility will typically be evaluated to determine whether corrective action is necessary
and appropriate. Where risks are deemed to be significant enough to trigger remediation, cleanup levels
must be determined. Risk characterization principles, which are summarized below, are an important part
of the risk assessment process including the determination and communication of corrective action
decisions.
Risk characterization is an important and requisite section of every risk assessment. Although the
principles of risk characterization should be evident throughout, a separate section must summarize risk
characterization. Risk characterization integrates information from the preceding components of the risk
assessment (primarily exposure and toxicity assessment components) and synthesizes information in a
manner that is complete, informative, and useful for risk managers, stakeholders, and the public. To
support quantitative and qualitative estimates of risk, it is critical to provide information to explain and
justify assumptions, methodologies used, and conclusions drawn. Risk characterizations should state and
explain why any potential COPCs or exposure pathways were eliminated from the risk assessment at any
time during the process. Risk characterizations should also discuss relative confidence in the
methodologies used, the potential impact of alternative choices, and the limitations of the analysis.
Particularly critical to complete risk characterization is a clear and complete discussion of the uncertainties
and variabilities associated with each of the components of the risk assessment. Uncertainty can be
defined as a qualitative or quantitative lack of precise knowledge about the truth. Uncertainty is typically
reducible through further measurement or study. Variability refers to the heterogeneity in a population
and is usually not reducible through further measurement or study. Uncertainly discussions help to
identify where additional information couid contribute significantly to reducing uncertainties in risk
assessment and aid decision-makers in deciding whether reduced uncertainty would add value to the
overall objectives of the project. Sections 4 J and 5.1.5 of this document identify specific risk assessment
uncertainty issues for human and ecological receptors, respectively.
S UFA Pf«»»rTA5kr*£VIStr/I«Al.NMA5mi WHM):-«»»!«!*>«IHow 1"*?
-------
In a 1995 memorandum and associated policy statement concerning the EPA Risk Characterization
Program (EPA 1995a and Attachment A), EPA Administrator Carol Browner staled that all risk
assessments and in particular, risk characterizations, must embrace the following fundamental values:
Transparency in decision making process
Clarity in communication
Consistency between EPA programs
Reasonableness of assumptions and policies
"Transparency" refers to the decision-making process; risks must be characterized fully and openly. Risk
characterizations should disclose the scientific analyses, uncertainties, and assumptions (both science- and
policy-based) that underlie all decisions. "Clarity" refers to communication; the risk assessment process
should help the public better understand the relative significance of environmental risks, It is important to
note thai risk characterization is a key component of risk communication, an interactive process involving
exchange of information and expert opinion among individuals, groups, and institutions. "Consistency"
and "reasonableness" refer to lie core assumptions and scientific policies that are part of the risk
assessment process. Consistency among EPA risk assessments is an important goal of EPA's risk
characterization policy. For example, CERCLA risk assessment guidance is considered appropriate for the
RCRA program, particularly because more detailed CERCLA guidance is often available. For RCRA
corrective actions and Superfund remedial actions, the actual environmental results achieved through
cleanup are expected to be environmentally equivalent. Further information on how the RCRA and
CERCLA programs overlap is documented in the EPA Region 10 (1994a) RCRA/CERCLA Interface-
Interim Final Guidance, which is included as Attachment B.
it is important that risk characterizations, like risk assessments themselves, be separate from risk
management decisions; scientific information should be selected, evaluated, and presented without
considering issues such as cost, feasibility, or how the scientific analysis might influence regulatory or
facility-specific decisions. In addition, the risk assessment process does not include decisions on the
public acceptability of risk levels, the value of reducing uncertainty by conducting further studies, and the
appropriate procedures for reducing facility-specific risks. The risk assessment process should delineate
both current and future risks because the time variable can impact site risks (for example, future risks may
MO
-------
be greater than current risks because of remedial activities and/or potential land usage at the site). Current
or future site risks for decision making should be selected during the risk management proceedings rather
than the risk characterization. Hie EPA (1997g) Rules of Thumb for Superfund Remedy Selection
identifies principles that should be consulted for risk assessments and risk management decisions. These
principles can be applied to RCRA corrective action programs.
5 tow
-------
3
>
$
3
§
-------
CHAPTER 2
STATE PROGRAMS
U.S. Environmental Protection Agency (EPA) may grant stales the authority to administer Resource
Conservation and Recovery Act (RCRA) corrective action as part of their authorized RCRA permit
programs. States may promulgate their own regulations. For authorized programs, these regulations must
be at least as stringent as federal regulations, and EPA retains corrective action authority through statutory
enforcement orders in all states regardless of authorization status.
Facilities in states without authorized RCRA permit program regulations must comply with federal
regulations; however, according to the Federal Register 19457 (May 1, 1996), EPA recognizes that many
states have developed independent Superfund-like authorities and cleanup programs. Consequently, when
developing cleanup levels for a facility, the project manager should consider other promulgated state
standards or criteria, including those regarding land use classifications that may influence cleanup level
selection. Whether the corrective action is state-lead or EPA-lead, cleanup levels typically should be at
least as stringent as state standards or criteria to avoid the need for the state to revisit the corrective
measure taken at a facility. The burden is on the facility to ensure that both state and federal requirements
are met. Unlike the Comprehensive Environmental Response, Compensation, and Liability Act of 1980
(CERCLA), RCRA statutory language does not include requirements to follow state standards as
applicable or relevant and appropriate requirements. RCRA only requires that cleanups are "protective."
The following sections summarize the authorization status of each Region 10 state RCRA program. Only
Washington State has specific regulations for an authorized RCRA corrective action program; however,
other non-RCRA state cleanup programs are also identified, and the state agency and phone number are
provided for obtaining further information. More specific information on state programs is included where
appropriate under the human health risk assessment (HHRA) and ecological risk assessment procedure
sections of this guidance. For example, state-promulgated human health standards and criteria are
described in more detail under Section 4.3, Identification of Promulgated Standards and Criteria.
S WWvlSI-Ritaw
2-1
-------
2.1
ALASKA DEPARTMENT OF ENVIRONMENTAL CONSERVATION
The Alaska Department of Environmental Conservation (ADEC) las not applied for authorization to
manage any of the RCRA programs in lieu of EPA. ADEC is currently preparing cleanup standard
regulations for contaminated sites and associated guidance documents on HHRA, petroleum risk
evaluation methodology, and background calculation methodology. These cleanup regulations were
proposed for public comment on December 18, 1996, Current and proposed regulations are available for
download (http://www.state.ak.us/dec/dec-cal.htm#Regulation). The Contaminated Sites Remediation
Program may be contacted for further information at (907) 465-5390.
2.2 IDAHO DIVISION OF ENVIRONMENTAL QUALITY
The Idaho Division of Environmental Quality (IDEQ) is authorized to operate a RCRA hazardous waste
program; however, the state only has the authority to compel corrective action at RCRA facilities with
RCRA permits. Idaho has not promulgated specific rules for setting cleanup levels at RCRA facilities and
has followed EPA guidance. Idaho statutory language (Section 39-4404 of Idaho Code, Hazardous Waste
Management Act of 1983) prohibits IDEQ from promulgating rules more stringent than existing EPA
RCRA regulations. The IDEQ can be contacted at (208) 373-0502.
2.3 OREGON DEPARTMENT OF ENVIRONMENTAL QUALITY
The Oregon Department of Environmental Quality (ODEQ) is authorized to operate a RCRA hazardous
waste program, including the corrective action program. According to its corrective action authorization
application, ODEQ intends to rely on EPA risk assessment guidance documents and lexicological
databases to determine the appropriate cleanup levels for a given facility. Oregon has not promulgated
specific rules for setting cleanup levels at RCRA facilities. Facilities may calculate site-specific cleanup
levels that must be approved in advance by ODEQ.
In 1995, Oregon amended its statutory authority for environmental cleanup rules (Oregon Revised
Statutes 465.315 and 465.325), requiring that new rules be adopted for conducting risk assessments and
defining hot spots. The rules, adopted on January 10, 1997, establish protocols for HHRA and ecological
risk assessment that include deterministic and probabilistic methods. The rules apply to facilities subject
V wmwio TASKIJUVISE7#m*I.MAST» IWW 1-2.
-------
to the stale's Superfund-like program. These rules are not currently a part of the state-authorized RCRA
program. ODEQ's Waste Management and Cleanup Division can be contacted at (503) 229-5913 or on
the Interne! at http://www.deq.state.or.us/ for further information on the slate's RCRA program.
2.4 WASHINGTON DEPARTMENT OF ECOLOGY
The Washington Department of Ecology (Ecology) is authorized to operate a RCRA hazardous waste
program, including the corrective action program, Washington's Model Toxics Control Act (MTCA)
Cleanup Regulation, amended in January 1996 under Chapter 173-340 of the Washington Administrative
Code, establishes methods for calculating cleanup levels. Under the alternative authorities initiative of
Ecology's corrective action authorization application, Ecology was authorized for a RCRA corrective
action program that allows for the option of incorporating a MTCA order into RCRA permits to fulfill the
RCRA Section 3004{u) and (v) requirements that all RCRA permits must include corrective action permit
conditions. As previously noted, however, EPA retains corrective action authority through statutory
enforcement orders regardless of Ecology's authorization status. In February 1996, Ecology published a
MTCA Cleanup Levels and Risk Calculations (CLARCII) Update reference document (Ecology 1996).
The MTCA cleanup regulation is described further in Chapter 4. Ecology has a comprehensive Internet
site where dozens of guidances and regulations are available for download, including those pertinent to the
Toxics Cleanup Program (http://www.wa.gov/ecology/tcp/cleanup.html). The Toxics Cleanup Program
can also be contacted toll-free at (800) 826-7716.
The Guidance for Clean Closure of Dangerous Waste Facilities (Ecology 1994) provides closure guidance
for interim and final status treatment, storage, and disposal facilities. The document provides direction for
demonstrating compliance with the clean closure performance standards and recommends the use of
MTCA residential cleanup standards.
S *EPAWmU\TASKl\REVlS&\nNAL\MACT^ 1lani\uc ~
-------
r
pi
3
O
z
-------
CHAPTER 3
DATA COLLECTION TO CHARACTERIZE FACILITY AND DETERMINE HAZARDOUS
CONSTITUENTS OF POTENTIAL CONCERN
Site-specific data of sufficient quality must be collected to determine facility conditions and the extent of any
necessary cleanup. Guidance and reference documents that describe data collection and data review
methods are summarized in the following subsections. Many of the EPA documents cited are available on
the EPA Region 10 web site (http://www.epa.gov/r 1 Oearth/office/oea/r 1 Oqahome.htm). Two primary
issues are addressed in Chapter 3. The first issue is identifying what data must be collected to characterize a
facility. This issue can be addressed by following the data quality objectives (DQO) process defined by U.S.
Environmental Protection Agency (EPA). Section 3.1 summarizes the DQO process. The second issue is
identifying how the samples should be collected and analyzed to assure that the data meet useability
requirements. Issues that should be considered to address data useability are described in Section 3,2.
Following data generation, a data quality assessment is completed to assure that DQOs have been met.
Following this assessment, the data are then used to identify constituents of potential concern (COPC). The
data quality assessment and COPC identification steps are described in Sections 3.3 and 3.4, respectively.
Existing facility data that have been or can be validated should be considered during the DQO process as
well as used in COPC identification, risk assessment, and compliance determinations.
The DQO process must also be applied to determine whether compliance with cleanup goals has been
achieved following remediation. Determining compliance with cleanup goals is described in Chapter 7 of
this guidance.
3.1 DATA QUALITY OBJECTIVES PROCESS
EPA's DQO guidance applies to all EPA
programs and can be used for Resource
Conservation and Recovery Act (RCRA)
corrective action situations, where the facility
is typically responsible for proposing
sampling and analysis activities through draft and final RCRA facility investigation (RFI) work plans. The
DQO process provides a procedure for defining criteria that a data collection design should satisfy,
EPA's Guidance for the Data Quality Objectives
Process (1994b) outlines a systematic planning
process for ensuring that data of sufficient quantity
and quality are collected to support defensible
decision making.
S RJEPA'AhwiirJASKrJlEVISEJvFtNAL^MASTER ISI 'ft I MM 12/) I 1t»mu>e
3-I
-------
including when, where, and how many samples to coiled. The DQO guidance recommends the use of
statistical methods for identifying tolerable levels of decision errors (that is, type 1 and type 2 errors) and
the number of samples required to meet these decision error levels. DQOs are defined during the first six
steps of the process. The data collection design is then "developed in a seventh step, based on the DQOs,
All of the DQO steps should be specified in work plans submitted to the agency when sampling and
analyses are being proposed.
EPA*8 DQO process is currently being developed and may change as guidance documents are developed
and updated. The seven DQO steps are highlighted as follows aid are briefly described in Sections 3.1,1
through 3 J .7, Current EPA guidance documents that provide more detail on the DQO process are also
identified.
The DQO process combines elements of both planning and problem formulation in its seven-step
format.
Step I: State tie problem. Review existing information to concisely describe the problem to be
studied.
Step 2: Identify the decision. Determine what questions the study will try to resolve and what
actions may result.
Step 3: Identify tie inputs to the decision. Identify information and measures needed to resolve
the decision statement
Step 4: Define boundaries of the study. Specify time and spatial parameters as well as where and
when data should be collected.
Step 5: Develop a decision role. Define statistical parameter, action level, and logical basis for
choosing alternatives.
Step 6: Specify tolerable limits on decision errors. Define limits based on the consequences of an
incorrect decision.
Step 7: Optimize the design for obtaining data. Generate alternative data collection designs and
choose most resource-effective design that meets all DQOs.
% REmti<»»nTASK«TCV1SE2vnNALWASTE* WPD\ I51 R I WW* S 2/* i > IujjUk
3-2
-------
3.1.1
Step 1: State the Problem
Step I requires that the problem be defined. This step will include summarizing existing facility
information, such as historical waste management activities and environmental data {including but not
limited to information in the RCRA facility assessment). In the RFI stage, the problem is typically
determining what additional data are required to characterize the type and concentration of hazardous
constituents associated with releases from solid waste management units (SWMU). The problem should
be defined as concisely as possible, focusing on such issues as the media of concern, land use, location of
human and ecological receptors, and magnitude of contamination. A more specific problem may be what
data are needed to determine whether hazardous constituents are present at concentrations greater than
either preliminary background or risk-based screening concentrations.
A conceptual site model (CSM) is a useful tool for defining facility conditions and the types of data
collection that may be required, EPA Superfund Guidance for Conducting Remedial Investigations and
Feasibility Studies Under CERCLA, Interim Final (EPA 1988) provides information on how a CSM can be
developed. As stated in flat document, the CSM should include known and suspected sources of
contamination, types of constituents and affected media, known and potential routes of migration, and
known or potential human and environmental receptors. The American Society for Testing and Materials
(ASTM) Standard Guide for Developing Conceptual Site Models for Contaminated Sites also provides
guidance for developing a CSM (ASTM 1995a). The CSM can be used to help identify locations where
sampling is necessary. The CSM can also be used to identify tie types of exposures that may result from
facility contamination and the cleanup levels required to address these potential exposures. Human health
and ecological exposure issues are further discussed in Chapters 4 and 5 of this cleanup level guidance.
3.1.2 Step 2: Identify the Decision
Step 2 requires that the principle study question be identified and a decision statement defined to link the
study question to possible alternative actions. The principle study question is defined by reviewing the
Step I problem. In the RFI, the study question is likely to be "are hazardous constituents present on a
SWMU at concentrations that exceed screening or preliminary background levels?" or a similar issue.
Possible alternative actions that may be taken are then identified, including the alternative that "no action"
is required. For example, the.action related to the principle study question may be "remediation is
5 *EM *K'i|« TAiKKEVTSES/WML'MArnE* WMfeac 3-3
-------
required if the screening level is exceeded" or alternatively, "no remediation is required if the screening
level is no! exceeded." These alternative actions form the basis for defining decision performance criteria
in Step 6.
3.1.3 Step 3: Identify the Inputs to the Decision
In Step 3, the types of data or information that will be required to resolve the Step 2 decision statements
are identified. This includes determining whether environmental measurements are required and further
defining the types of measurement data. The soirees of this information should be identified (for
example, historical data or new data collection). For the RFl, constituent concentration levels in the media
of concern will likely be required (for example, the average concentration of a constituent in soil). Other
measurements such as soil and hydrogeological parameters may also be required to support site
characterization through fate and transport information.
The basis for setting screening levels should also be defined in this step. Screening levels may be based on
existing standards and criteria (for example, groundwater maximum contaminant levels), risk-based
concentrations (RBC) (for example, soil RBCs based on residential land use), or preliminary site-specific
background data (for example, background metals data).
3.1.4 Step 4: Define Boundaries of tbe Study
Step 4 requires that the spatial and temporal boundaries of the problem be defined. Spatial boundaries
define the physical area to be studied and locations to collect samples. According to the Geostatistical
Sampling and Evaluation Guidance for Soils and Solid Media published by EPA in 1996, temporal
boundaries determine the time frame that the study data will represent and when samples should be
collected (EPA 1996a).
The main purpose of this step is to identify, to the extent possible, a well defined data population that can
be statistically evaluated. In an RFl, sampling should be conducted to characterize the nature and extent of
contamination in areas where releases are suspected to have occurred. These study areas should focus on
waste management activities to define areas with similar contamination (for example, the concentration of
A JUIPA'RIiiKr.TASkKREVlSEWNAUM ASTER WTO\|SI«RHWt«07tliMimm«
-------
a constituent released from a single SWMU). In practice, areas of homogenous contamination may not be
present or readily identifiable. The EPA Soil Screening Guidance: User "s Guide (1996b) recommends
defining study areas by stratifying the site into known, suspected, and unlikely contaminated areas. The
study area can also be defined as an area where contamination may be present as a hot spot, and a
sampling program could provide acceptable probability that hot spots of a specific size will be detected.
In addition to segregation by waste management activities, study boundaries may be defined by the type of
exposure that could occur. For example, if a large area of contaminated soil may be subdivided for future
residential development, it may be necessary to subdivide the SWMU into smaller "residential size"
exposure areas for data collection. Constituent releases that leave the SWMU area may enlarge the
boundary of the study area (for example, a groundwater contaminant plume). In this situation, monitoring
wells located along a plume's center line may define the study boundaries. The average groundwater
concentration over a minimum of four calendar quarters of groundwater monitoring may define the
temporal boundaries of the study (EPA 1993a),
Chapter 3 of Methods for Evaluating the Attainment of Cleanup Standards, Volume I; Soils and Solid
Media (EPA 1989b) provides further information on the identification of discrete study areas to determine
cleanup decisions for soil. Section 2.3 of EPA's Soil Screening Guidance: Users Guide (1996b) also
provides guidance on identifying surface and subsurface soil study areas. EPA's Supplemental Guidance
to RAGS: Estimating Risk from Groundwater Contamination (1993a) discusses approaches for delineating
groundwater exposure areas for Superfund risk assessments. Once study areas are defined, the areas will
be sampled to determine whether the constituent's concentration exceeds a screening level or whether hot
spots are present.
3,1,5 Step 5: Develop a Decision Mile
Step 5 requires that a decision rule be developed to define the conditions that would necessitate the choice
of an alternative action. The decision rule is formed from elements defined during previous DQO tasks,
including (1) the parameters of interest defined in Step 3, such as the average concentration of a
constituent in soil; (2) the screening levels defined in Step 3, such as the soil RBC; (3) the study boundary
defined in Step 4, such as soil located in a SWMU spill area; and (4) the principle study question and
alternative actions defined in Step 2. The decision rule is an "if... then" statement that incorporates the
S BEPA RH«»MTASKrjtEV1SEJ\flNAt\MASTER WPOW)lMZmwW ItasnMag
3-5
-------
previous information. For example. If the parameter of interest (average concentration of constituent
released to soil) within the study area (the SWMU spill area) is greater than the screening level (the soil
RBC), then alternative action A should be taken (for example, remove contaminated soil); otherwise,
alternative action B should be taken (for example, leave soil in place).
Step 3 requires that a general basis for defining facility conditions and screening criteria be defined. For
example, the facility parameter of Interest may be the constituent concentration at a SWMU, while the
screening criteria may be a promulgated standard, a RBC, or a preliminary background level. In Step 5,
the facility parameter and screening criteria must be specifically defined and incorporated into the decision
rule. In the previous example, the average soil concentration of the constituent in a SWMU area is defined
as the specific facility parameter, while the screening criteria is defined as a specific soil RBC (for
example, residential land use RBC based on soil contact).
When determining the presence of hot spots, the decision rale must incorporate the size (radius) of the
potential hot spot that may exist in the study area and the distance between sampling locations within the
study area sampling grid.
More than one constituent may be present at a SWMU. For the purpose of SWMU characterization, the
constituent that will require the most significant data collection to determine whether a release has
occurred above screening levels should be identified and used in the decision rale. This constituent
typically will exhibit the highest variability in concentration or will be detected in a concentration closest
to the screening level.
3,1.6 Step 6: Specify Tolerable Limits on Decision Errors
Step 6 requires that the decisions maker's tolerable limits on decision errors be specified. The true value
of the population parameter being measured (for example, the average constituent concentration) can
never be exactly defined based on sampling design and measurement design errors. An error may be made
during Step 5 since the decision is based on measurement data, A decision error occurs when the data
mislead the decision maker into concluding that the parameter of interest is on one side of a screening
level (for example, greater than the screening level) when it is actually on the other side (that is, less than
5. m?*. RMiirTASKsTOVKEftFtNALVMSTCR immstsm
3-6
-------
the screening level). The possibility of decision errors can never be totally eliminated, but it can be
controlled. For example, a large number of samples may be collected to control sampling design errors.
Chapter 6 of Guidance for the Data Quality Objectives Process {EPA 1994b) explains how the probability
of decisions errors can be controlled by adopting a scientific approach that incorporates hypothesis testing.
The method includes (I) defining the two types of decision errors (that is, false positive [a] and false
negative [P] decision errors), (2) evaluating the consequence of each error, (3) identifying the error with
more severe consequences near the cleanup level, (4) defining a null hypothesis that is equal to the true
state of nature that exists when tie more severe decision error occurs, (5) estimating the range of
parameter values near the cleanup level where the consequences of decision errors are relatively minor
(defined as the grey area), and (6) assigning probability values to points above and below the grey area
that reflect the decision maker's tolerable limits for making an incorrect decision.
For example, the decision maker may want to know whether a hazardous constituent is present in a
SWMU at an average concentration that exceeds a screening level. He decision maker may view the
consequence of deciding that the average concentration is less than the screening level when it is actually
greater than the screening level as the more severe decision error. The null hypothesis would then be that
the average concentration exceeds the screening level (enough data must now be collected to reject the null
hypothesis if it is false). A conclusion that the concentration is less than the screening level when it is
actually greater would be a false positive error, while a conclusion that the concentration is greater than
the screening level when it is actually less would be a false negative error. The decision maker then
establishes a grey area near the action level. The boundaries of the grey area are the action level and the
point below the action level where the consequences of a false negative error begin to become significant.
The actual grey area interval is the concentration range near the action level where the decision maker
determines it is not necessary or feasible to control the probability of a false negative error (for example,
because the consequences of this error are minor, or because the costs of collecting enough samples are
prohibitively high). The decision maker then sets allowable decision error probabilities at points above
(false positive errors) and below (false negative errors) the screening level, starting at the boundaries of
the grey area where the consequences of errors are minor and/or expensive to control. The "tolerable
limits" are the intervals of concentration above or below the grey area where allowable decision errors are
set. Generally, the wider the interval, the lesser the decision error probability that will be accepted.
s «»wnMWT»s»jttwse«*iwA£reii wMxisi-Mmummiinum inaut 3-?
-------
A similar DQO evaluation is performed when determining compliance with a cleanup level (see
Chapter 7). Detailed examples of DQO evaluations have been developed by EPA (Guidance for the Data
Quality Objectives Process [1994b], and Data Quality Objectives Decision Error Feasibility Trials
(DQO/DEFT) Users Guide [1994c]) and are included in Attachment C.
3.1.7 Step 7: Optimize tie Design for Obtaining Data
As staled in the Guidance for the Data Quality Objectives Process (EPA 1994b), DQOs are qualitative and
quantitative statements derived from the outputs of Steps 1 through 6 that help to accomplish the following
tasks:
• Clarify study objectives
• Define appropriate type of data to coiled
• Determine conditions from which to collect the data
• Specify tolerable limits on decision errors to be used as the basis for establishing the
quality and quantity of data needed to support the decision
Step 7 includes identifying a resource-effective data collection (sampling) strategy for generating data that
are expected to satisfy the DQOs, The sampling strategy typically will focus on the sampling design, the
sample size, and the analytical methods required to meet the DQOs. A primary requirement of Step 7 will
be to define a statistical method for testing the Step 6 hypothesis and a sample size formula that
corresponds to the statistical method and the sample design. EPA has published several guidance
documents on the selection of appropriate statistical models for determining sample sizes and sample
designs. These documents are listed in the Chapter 7 discussion of compliance with cleanup levels.
Similar statistical models can be used in both site characterization and compliance determinations.
It is preferable to selected analytical methods that can be used to detect constituents at reasonable
concentrations well below the screening levels, to reduce the potential for false negatives, and to increase
confidence in the quantification of positive hits. A cost function that relates the number of samples to total
cost may also be defined. The cost function may be used to support the proposal of a cost-effective
S IS TASKivREVISEiJFlNAL^vMASttR WPR! 31
-------
sampling strategy that meets the DQOs. Generally, the lower the probability of an error that the decision
maker is willing to accept, the greater the sampling effort and costs required to meet the DQOs.
The output of Step 7 will be a sampling strategy that defines sampling design, sample numbers, and
analytical methods. Section 3,2 farther describes methods that should be followed to assure that the
sample data collected are of acceptable quality,
3.2 DATA USEAJBILITY
Sampling and analysis activities should provide adequate data to evaluate all appropriate exposure
pathways and chosen ecological cndpoints. The sampling plan should be designed with all data uses
(human health, ecological, and others uses) in mind, including attainment of cleanup levels. Hence,
human health and ecological risk assessors and others must be Involved in the designing of sampling
plans. These plans should ensure that the issues in the data useability work sheet (Attachment D) can be
adequately responded to after data have been collected.
The Final Guidance for Data Useability in Risk Assessment Pari A
(EPA 1992b) provides risk assessors and remedial project
managers with nationally consistent procedures to plan aid assess
sampling and analysis of useable environmental data.
Chapters 4 and 5 of EPA's Risk Assessment Guidance for
Superfund, Volume 1, Human Health Evaluation Manual (Pari A)
(EPA 1989c) discuss data collection and data evaluation,
respectively.
3,2,1 Data Sources
The data sources selected (for example, field screening, field analytical, fixed analytical) depend on the
type of data required and their intended use. The data sources must be comparable if data are combined
for quantitative use. For example, field screening and fixed laboratory data should not be combined for a
quantitative analysis. These separate data sources can, however, be used to complement one another.
Field screening data may be used to delineate soil contamination, while fixed laboratory data would be
The Final Guidance for Data
Useability in Risk Assessment
(Part A) (EPA 1992b)
discusses the six data
useability criteria involved in
planning an initial site
investigation or an
investigation to determine
compliance with cleanup
levels: data sources,
documentation, analytical
methods and quantitation
limits, data quality indicators
(DQ1), data review, and
reports to risk assessors.
S RCPA RlnrMit TASK* RE VISE* TIN A4. WASTER WPtM5l«R|4Mt»llwn11/1| W/V tew
3-9
-------
used to quantify the contamination. If field analytical screening results are going to be used quantitatively,
it roust be demonstrated that they are of sufficient quality to meet DQOs.
3.2.2 Documentation
Sample collection and analysis procedures must be fully and accurately documented to substantiate the
reliability of the data "derived from its analysis. The major types of documentation are quality assurance
project plans (QAPP), quality management plans (QMP), standard operating procedures, field and
analytical records, and chain-of custody records.
In addition, data quality indicators (DQI) (see Section 3.2.4) for assessing results against stated
performance objectives should be documented in the QAPP.
EPA policy requires that all environmental data used in decision making be supported by an approved
QAPP. A QAPP is required for each specific project or continuing operation. The QAPP documents how
quality assurance (QA) and quality control (QC) are applied to an environmental data operation to assure
that the results obtained are of the type and quality needed for a specific decision or use. QA is an
integrated system of management activities involving planning, implementation, assessment, reporting,
and quality improvement to ensure that a process, item, or service is of the type and quality needed and
expected. QC is the overall system of technical activities that measures the attributes and performance of
a process, item, or service against defined standards to verify that they meet the stated requirements.
Current EPA requirements for QAPPs are presented in the Interim Guidelines and Specifications for
Preparing Quality Assurance Project Plans (EPA 1980). Current EPA requirements for QMPs are
presented in the Guidelines and Specifications for Preparing Quality Assurance Program Plans
(EPA 1979). EPA is updating the QA/QC requirements. The updated document, EPA Requirements for
Quality Assurance Project Plans for Environmental Data Operations, is currently in the draft interim final
stage (EPA 1994d). A draft interim final version of EPA Guidance on Quality Assurance Project Plans
(QA/G-5) is also available, as is the draft updated document, EPA Requirements for Quality Management
Plans (EPA I994e) which will replace the quality assurance program plan guidelines. EPA Region 10's
quality assurance office has indicated that use of the draft documents is preferable to the use of older
'ASkSJlEVTSE2^NXUMA$TT3l WP0M3I f?Hsj ®r>sr IZflimttttaoAM*
3-10
-------
documents, but ultimately that is the decision of the RCRA project manager. The EPA Region 10 web site
contains news about the finalization of the draft documents (http://www.epa.gov/rlOeartli/offlce/oea/
rIOqahome.htm). In addition, federal Register 19445 (May I, 1996) cites Quality Assurance Project
Plans for RCRA Ground-Water Monitoring and Corrective Action Activities (EPA 1993b) guidance for
information on incorporating DQOs in the decision-making process at RCRA facilities.
The Interim Guidelines and Specifications for Preparing Quality Assurance Project Plans (EPA 1980)
describe 16 elements that must be considered for inclusion in all QAPPs, recommend the format to be
followed, and specify how plans will be reviewed and approved, AH QAPPs must describe procedures that
will be used to document mid report precision, accuracy, and completeness of environmental
measurements. The 16 essential elements described must each be considered and addressed unless it is
documented that a particular element Is not relevant to the project.
3.2.3 Analytical Methods, Detection Limits, and Quantitation Limits
Analytical methods selected should meet the required detection limits for metals and quantitation limits
for nonmetals that are at or below facility-specific screening or cleanup levels. If facility-specific cleanup
levels have yet to be determined, the EPA Region 9 preliminary remediation goals (PRG) described in
Section 4.5 and presented in Attachment E can be used to determine adequate detection and quantitation
limits for protection of human health (using a hazard quotient [HQ] ofO.l; Region 9 PRGs are based on a
HQ of 1.0 ) (EPA 1996c). The term "PRG" has the same meaning as "risk based concentration (RBC)."
Appendix III of the Final Guidance for Data Useability in Risk Assessment (Part A) (EPA 1992b) lists
various analytical methods and associated detection and quantitation limits by chemical for human health.
Ecological data quality levels developed by EPA Region 5 (1995b) can be used to determine adequate
detection and quantitation limits for ecological health. A chemist should be consulted for assistance in
choosing an analytical method when those available have detection or quantitation limits near the cleanup
level or PRG.
3.2.4 Data Quality Indicators
DQls are identified during the development of the DQOs to quantitatively measure the achievement of QA
objectives. The five DQls discussed by EPA (1992b) are completeness, comparability, representativeness,
5 MSP*«tO"l«T«K«WTOEirfNAWMASTC«
3-11
-------
precision, and accuracy, QA objectives for these DQls should be listed in the QAPP. Precision is a
quantitative measure of variability, comparing facility samples to the mean. Results of QC samples (field
aid/or laboratory duplicates) are used to calculate the precision of the analytical or sampling process.
Accuracy is a measure of the closeness of a reported concentration to the true values. This measure is
usually expressed as bias (high or low) and is determined by calculating percent recovery from spiked
samples. Completeness is a measure of the amount of useable data resulting from what was planned for a
data collection effort, Representativeness is the extent to which data define the conditions at a facility.
Comparability refers to the ability to combine or compare results across sampling episodes and time
periods,
3.2.5 Data Review
DQOs dictate the level aid amount of data review required. The level of data review refers to which
evaluation criteria arc selected, ranging from generalized criteria (for example, holding time) to
analyte-specific criteria (for example, recovery of a surrogate spike for organic compounds or analyte
, spike recovery for inorganic compounds). Analytical results, QC sample results, and raw data for
chemicals analyzed to determine compliance with cleanup levels should undergo a full data review. A full
data review is very labor intensive and includes checking the raw laboratory data against a number of data
review criteria and spot checking (recalculation) values reported by the laboratory. A partial data review
may only involve looking at the summary QC information reported by the laboratory. A full data review
minimizes false positives, false negatives, calculation errors, and transcription errors. EPA data review
guidance for Contract Laboratory Program data includes U.S. Environmental Protection Agency Contract
Laboratory Program National Functional Guidelines for Organic Data Review (19940 and U.S.
Environmental Protection Agency Contract Laboratory Program National Functional Guidelines for
Inorganic Data Review (1994g). Data review criteria presented in these guidance documents must be
considered when developing the site-specific QAPP; however, some aspects of these documents may not
be applicable to a specific site or some types of analyses. Generalized and analyte-specific criteria must
be presented in the site-specific QAPP. When large numbers of samples (50 or more) are collected, the
amount of data to be reviewed should be determined based on expense, types of analyses, and historical
knowledge.
S *FI»VJftlMHirTA$krAEV1SE2\nNAUMASTER
3-12
-------
The data review must provide a narrative summary describing specific sampling or analytical problems,
data qualification flags, level of review (full or partial and what was reviewed), detection and quantitation
limit definitions, and interpretation of QC data.
The decision maker must consider the completeness of the data and verify that detection and quantitation
limits were adequate to distinguish constituent concentrations from cleanup levels. All validated, useable
data should then be considered during the data quality assessment (Section 3.3) to determine whether the
sample design and the resulting validated data set are adequate for characterizing the facility and
determining compliance with cleanup levels.
3.2.6 Reports from Sampling ami Analysis
Preliminary reports assist in identifying sampling or analytical problems early enough so that corrections
can be made during data collection and before sampling or analysis resources are exhausted.
3 J DATA QUALITY ASSESSMENT
Data collected and validated in accordance with the QAPP must be assessed to determine whether the
DQOs have been satisfied. EPA Guidance for Data Quality Assessment (1996d) describes data quality
assessment (DQA) methods. The primary objective is to determine whether the collected data meet the
DQO assumptions and whether the data user can then make a decision with the desired confidence. A
preliminary data review should also be performed to evaluate the structure of the data (for example,
common statistical parameters and data distribution type) and assess the accuracy of the sampling design.
Data characteristics should be consistent with statistical assumptions made during the DQO process (for
example, distribution type, nondetection frequency, variance). If the data do not support underlying
statistical assumptions, corrections must be performed to meet the decision maker's needs. This may
include the selection of a different statistical approach or the need for additional data collection and a
revised sampling design. Once an appropriate statistical test has been identified, the DQO decision rule is
tested to reach a conclusion regarding compliance with the screening level (EPA 1996d).
S REPAIR !tBsjmTASKB'vREM5ErftKAI.iWASTER Siasmm
-------
3.4 IDENTIFICATION OF HAZARDOUS CONSTITUENTS OF POTENTIAL
CONCERN
COPCs should be identified so that a risk assessment or cleanup level determination can be focused on
hazardous constituents that pose the primary health threat. To identify COPCs, the data useability review
and DQA discussed in Sections 3,2 and 33 are first performed. Sampling data that have been validated
and determined to meet QAPP requirements should be considered, and all constituents detected should
initially be included as COPCs, All data should be summarized and submitted to regulatory agency
personnel. Deliverables should include the following information for all data including detected and non-
detection sample results: sampling date, sample location map, sample media, sample detection limits,
sample results, and sample qualifiers.
If the purpose of COPC identification is to perform a risk assessment, then risk-based screening is the next
step of the COPC identification process. Hazardous constituents in each medium present at maximum
concentrations which are below certain RBCs can be screened from further consideration. If the purpose
of the COPC identification is to calculate cleanup levels, the risk-based screening step should be
performed after promulgated standards and criteria have been considered for cleanup levels. As further
described in Section 4.3, available criteria and standards should be used as cleanup levels unless
determined on a facility-specific basis to be insufficiently protective of human health and the environment.
Since human health and ecological risk-based screening methods may vary, risk-based screening
approaches are described in Section 4.4 (human health) and Chapter 5 (ecological). Following data
evaluation and relevant risk-based screening, the remaining COPCs should be carried through the human
health risk assessment or considered in the cleanup level determination. Exhibit 1-1 presents the COPC
screening process relative to setting cleanup levels.
A preliminary background evaluation may be performed to determine whether the detected constituents
are related to the facility or potentially associated with background or ambient conditions. The
background evaluation is typically performed only for inorganic compounds that may naturally occur in
soil or water; however, the evaluation of organic constituents may also be performed on a case-by-case
basis if it can conclusively be determined that nonfacility-related organic contamination is present.
Likewise, inorganic compounds associated with nonfacility-related anthropogenic sources may require
evaluation, such as releases of lead from leaded gasoline. The DQO process and associated guidance
S MPMt I l»>rrA5»CrJltVISEi'/W/a«AJTW WnXIJI-KOTtW*" Urtl.W. -Umrtm 3" 1 4
-------
documents discussed in Section 3.1 and Chapter 7 should be followed when collecting data for facility and
background comparisons. The collected data are evaluated to determine thai requirements specified in the
QAPP, including DQOs, have been met. The background data are then compared with site data using the
appropriate statistical tests identified during tie DQO and DQA steps.
Elimination of constituents from the quantification of facility risks, based on claims or assumptions that
the constituents are not related to the facility, should not be allowed unless the following can be
demonstrated: adequate sampling, analyses, and statistical tests conducted following DQO and DQA
procedures should indicate that the constituents truly are not related to the facility. Any constituents that
are eliminated from the quantification of risk based on background conditions must be carried through to
the risk characterization section of the risk assessment, where they should be discussed qualitatively along
with a description of the justifications for the elimination. In the absence of sufficient evidence,
background screening should not be performed before risk assessment or cleanup level determinations.
Additional data may be collected if constituents suspected (but not adequately demonstrated) to be
nonfacility-related have a significant impact on cleanup level determinations (for example, if they are
present at concentrations above cleanup levels or if they have a significant effect on a cleanup level
incorporating risks from multiple constituents).
A generalized approach for comparing facility conditions with background is presented by the
EPA (1994h) in the Region 8 Superfund Technical Guidance, RA-03: Evaluating and Identifying
Contaminants of Concern for Human Health, included as Attachment F. In the guidance, EPA describes
two types of statistical comparisons that can be made between samples collected from background and
contaminated facilities: (1) distributional tests and (2) extreme value tests. EPA describes each
comparison type and recommends specific distributional tests. The distribution of the facility and
background data sets as well as the percent of detections in each data set are considered when selecting
appropriate tests. Attachment F should be consulted for further information on background statistical
testing. In addition, the following updated information should also be considered:
• The Guidance for Data Quality Assessment, Practical Methods for Data Analysis
(EPA 1996d) provides information on summarizing data and performing statistical tests.
EPA will also soon publish the "Data Quality Evaluation Statistical Toolbox
(DATAQUEST)" software, which will include software for running some statistical tests.
3-15
-------
In addition to the Wilcoxon rank sum test and student's t-test cited by EPA Region 8
(1994h) for distributional tests, the quantile test can be used to check for extreme values.
Other non parametric (distribution free) outlier tests that are designed to test groups of data
may be substituted for the quantile test.
At the decision points involving percent detections on Figure 2 in Attachment F (EPA
1994h), if either data set (background or facility) is less than the criterion, use the "less
than" or "yes" branch.
If more than half of the results in either data set are nondetected, use a test of proportions
with a suitable choice of percentile (see Section 3.3.2.1 of EPA 1996d) instead of the
Wilcoxon rank sum or quantile tests,. If the data set are not comparable (thai is, there are
major differences in the quantitation limits for the nondetected results), consult a
statistician.
Additional reference information is presented in a letter regarding background comparison methods for
Rocky Flats Plant prepared by Richard O. Gilbert, Ph.D. (1993). The letter identifies a variety of
statistical tests that can be used for background comparisons, including many of the tests identified in the
EPA documents cited in Chapters 3 and 7. The report describes a series of parametric and nonparametric
tests and also recommends that a "hot measurement" comparison be performed to identify hot spots
(extreme values). Examples of how to perform the statistical tests are presented. More detailed
information on statistical testing can be obtained from Statistical Methods for Environmental Pollution
Monitoring (Gilbert 1987).
The EPA Determination of Background Concentrations of Inorganics in Soils and Sediments at Hazardous
Waste Sites (1995c) report also provides guidance on technical issues that must be considered when
determining whether a site contains elevated levels of inorganic compounds relative to the local
background concentrations. Technical issues discussed include the selection of background sampling
locations, considerations in the selection of sampling procedures, and statistical analyses for determining
whether constituent levels are significantly different on a potential waste site and a background site.
s HEPA'R p>n»-TA5Kr®EViSEim»*WIASre» WHWM-* JJHttMK
3-16
-------
-------
CHAPTER 4
HUMAN HEALTH RISK ASSESSMENT PROCEDURES
Data collected during a Resource Conservation and Recovery (RCRA) facility investigation (RFI).
ongoing monitoring results, and any other available applicable arid useful environmental sampling
information are used to assess risks to human health and the environment so it can be determined whether
there is a need for corrective action.
The data collection and evaluation step is described in
Chapter 3: relevant site data are collected, and the data
are determined to be of acceptable quality for risk
assessment. Exposure assessment, toxicity assessment,
and risk characterization are described in this chapter.
The exposure assessment and toxicity assessment steps may
be performed concurrently. During the exposure assessment step, (I) constituent releases, (2) exposure
pathways into which constituents may then migrate, and (3) potential human exposures to constituents that
may occur are all identified. Constituent concentrations in exposure pathway media and resulting
constituent doses to humans are calculated during the exposure assessment. During the toxicity
assessment, toxicity factors are compiled. Toxicity factors represent the relationship between the
constituent dose received by a person and the resulting adverse response that may occur. The results of the
exposure and toxicity assessments are combined during the risk characterization step to characterize the
potential for adverse health effects to occur. During the risk characterization step, cancer risks and
noncancer hazards are estimated quantitatively where possible.
This chapter on human health procedures focuses on how to develop facility-specific media cleanup
standards for the protection of human health. The type of exposure pathways and land uses that may occur
on the facility must first be determined (Sections 4.1 and 4.2). Available media-specific criteria and
standards are then identified (Section 4.3) since they are frequently used as cleanup levels. When
promulgated standards and criteria do not exist, protective media cleanup standards can be developed
using the human health risk assessment (HHRA) method. Hazardous constituents can first be screened
using preliminary (that is, nonfacility-specific) risk-based concentrations (RBC) to identify constituents of
potential concerns (COPC) (that is, those constituents present at concentrations at or above concentrations
Four primary risk assessment steps:
• Data collection and
evaluation
• Exposure assessment
« Toxicity assessment
• Risk characterization
f *£FAmt»l(ttTASK« JU;VlSP\nNAl\MASTER JUnNttt
-------
that may be associated with significant health concern) (Section 4.4). Facility-specific RBCs are then
calculated for the COPCs using the exposure assessment, toxicity assessment, and risk characterization
steps. Exposure parameters and toxicity values are identified for COPCs {Sections 4,5 and 4.6). The
exposure and toxicity information is then combined to calculate a health-based RBC that correlates with a
target, or acceptable, risk level once that level has been determined (Section 4.7). Facility-specific RBCs
may then be used as the basis for setting risk-based cleanup levels. COPCs that may be nonfacility-
related, such as naturally occurring metals, should be compared with background levels before final
cleanup level determinations are made. This process is discussed in Section 3.4.
4.1 EXPOSURE ASSESSMENT: IDENTIFICATION OF EXPOSURE PATHWAYS
The steps of an exposure assessment include characterization of the exposure setting, identification of
exposure pathways, and quantification of exposure. In the first step, the facility is characterized with
respect to the general physical characteristics (for example, climate, vegetation, groundwater hydrology,
surface water) and the characteristics of the potentially exposed populations on or near the facility. This
section discusses the second step: identification of exposure pathways.
The third step, quantification of exposure, is discussed in Sections 4.5
through 4.7. For additional exposure assessment discussions, see the
U.S. Environmental Protection Agency (EPA) Exposure Assessment
Guidelines (1992c), Risk Assessment Guidance for Superfund,
Volume J, Human Health Evaluation Manual (Part A), (RAGs)
Chapter 6 (1989c), and the Exposure Factors Handbook (1989d and
1996e update).
Table 4-1 summarizes potential exposure pathways (also called
exposure routes) for human receptors at a typical facility. The
identification of complete exposure pathways at a facility is necessary
to ensure that health-based cleanup levels protective of all potential
receptors can be developed. Any person who might be exposed to
facility-related constituents by one or more pathways is considered a
receptor.
Exposure is defined as the
contact of a person with a
chemical or physical agent.
There are three primary routes
by which hazardous constituents
released to the environment can
enter the body: ingestion,
inhalation, and dermal contact
Exposure pathways are the
course a constituent takes from a
source to an exposed organism
(for example, soil ingestion, or
inhalation of volatiles from
groundwater). A complete
exposure pathway consists of
ihe following four elements:
(I) a source and mechanism of
chemical release, (2) a retention
or transport medium, (3) a point
of potential human contact with
the contaminated medium, and
(4) an exposure route (for
example, ingestion) at the
contact point.
5 «?r* * MRU T *Skr*E VrtCWTNAltMA JTW UamUM
-------
TABLE 4-1
POTENTIAL EXPOSURE PATHWAYS FOR HUMAN RECEPTORS
REGION 10 RCRA RISK-BASED CLEANUP LEVEL GUIDELINES
Contaminated
Medium
Exposure Scenario ) Potential Exposure Pathway
Important for
Calculation of
Cleanup Levels
Groundwater
Residential use as
potable water
Ingestion of water
Yes
Inhalation of volatile compounds
Yes, if volatiles
present
Dermal contact with water
Site-specific
determination
Agricultural uses
Transfer to food crops or lives
and subsequent ingestion
Site-specific
determination
Industrial use as potable water
Ingestion of water
Yes
Inhalation of volatile compounds
Site-specific
determination
Dermal contact with water
Site-specific
determination
Surface water
and sediment
Residential or industrial use as
potable water
Ingestion of water
Site-specific
determination
Inhalation of volatile compounds
Site-specific
determination
Dermal contact with water
Site-specific
determination
Agricultural uses
Transfer to food crops or livestock
and subsequent ingestion
Site-specific
determination
Recreational or subsistence
fishing
Consumption of fish and seafood
Site-specific
determination
Recreational use or trespasser
Ingestion of water
Site-specific
determination
Dermal contact with water
Site-specific
detenu iiiation
Ingestion of sediment
Site-specific
determination
Dermal contact with sediment
Site-specific
determination 1
S R£PAA|niMftTASKrJlEVl$E2\nNALVMASTEftWPDtlSl*it»M*Ml»M;irtlW}* Jlunvuc
-------
TABLE 4-1 (Continued)
POTENTIAL EXPOSURE PATHWAYS FOE HUMAN RECEPTORS
REGION 10 RCRA RISK-BASED CLEANUP LEVEL GUIDELINES
Contaminated
Medium
Exposure Scenario
Potential Exposure Pathway
Important for
Calculation of
Cleanup Levels |
soil
Residential uses
Incidental soil ingestion
Yes |
Dermal contact with soil
Yes
Inhalation of particulates/volatile
compounds from soil
Yes
Soil as potential source to
groundwater
Site-specific
determination
Agriculture uses
Consumption of produce, meat,
milk'
Site-specific
determination
Industrial uses
Soil ingestion
Yes
Dermal contact with soil
Yes
Inhalation of particulates/volatile
compounds from soil
Yes
Soil as potential source to
groundwater
Site-specific
determination
Air
Residential uses
Inhalation of particulates/volatile
compounds from stack or other
emissions
Site-specific
determination
Industrial uses
Inhalation of particulates/volatile
compounds from stack or other
emissions
Site-specific
determination
Source; Modified from U.S. Environmental Protection Agency 1991a
S ••RE»\RH>i!»WASK«*IWSeiBN«.*«ST«* WHKIiMMUItlOTRftlinMm MantoK
-------
Exposure pathways are identified within each pertinent exposure scenario. Several exposure scenarios
may be applicable at a given facility. Those most commonly evaluated are tie industrial and residential
exposure scenarios. Other scenarios, including agricultural, recreational, and trespasser, may be important
depending on facility location and identification of the most exposed individual. For example, if the
individual subject to the greatest exposure to facility-related constituents is a recreational user of a surface
water body, a recreational exposure scenario may be sufficiently protective. Classification of land use is
further discussed in Section 4.2.
Cleanup levels should be determined for all exposures to a specific medium, such as soil. The cleanup
level calculated for each medium should take into consideration all exposure pathways and all facility-
related constituents that contribute to risk or hazard. For example, the cleanup levels for soil should be
developed using all possible exposure routes for soil that are appropriate at a facility. It is recommended
that ingestion, dermal contact, and inhalation exposure routes be considered when developing cleanup
levels for all media. Where exposure may occur to a constituent in both soil and water, media-specific
cleanup levels may require further downward adjustment to assure that an acceptable cumulative target
risk level is met.
4.2 CLASSIFICATION OF LAND USE
The evaluation of a facility to determine appropriate cleanup levels is based in part on the appropriate land
use scenario.
A residential scenario results in more conservative (that is, lower)
cleanup levels, because it is assumed that adults and children live
on the site and are exposed to hazardous constituents 24 hours a
day. In the industrial scenario, exposure is assumed only for adults
and only during working hours.
Selection of land use is most relevant when calculating cleanup levels that address direct contact with
soils. The following two subsections discuss EPA and Region 10 state policies or regulations regarding
land use and soil cleanup levels. An additional subsection discusses land use issues associated with setting
Depending on assumptions
regarding future facility uses,
either a residential scenario or an
industrial scenario is typically
chosen.
«{«* !* TASK«.*lEVT$Ej\nNAi\MASTUR WTO\mRM»H*>?OT!2A|/W5» t limbic
4-5
-------
groundwater, surface water, and air cleanup levels. Later discussions In Section 4.3 identify specific
numerical standards and criteria that must be considered when setting cleanup levels.
Of the Region 10 states, only the Washington Model Toxics Control Act (MTCA) regulations have been
authorized for setting RCRA corrective action cleanup levels. Details of MTCA regulations relevant to
land ise are therefore summarized. More general information on other state regulatory programs is also
presented because project managers and facilities are encouraged to take into consideration non-RCRA
state standards or criteria, including those regarding land use classifications that may influence cleanup
level selection. Consideration of stale regulations and land use classifications will avoid the need for the
slate to revisit the corrective measure taken at a facility. Overall, the most stringent land use requirement
that may apply to a facility should be used for risk assessment and setting cleanup levels.
4.2.1 U.S. Environmental Protection Agency Land Use Policy and Soil Cleanup
Levels
EPA's policy is that "current and reasonable expected future land use and corresponding exposure
scenarios should be considered both in the selection and timing of remedial actions" (Federal Register
[PR] 19452, May 1, 1996). EPA's May 25, 1995, directive, Land Use in the CERCLA Remedy Selection
Process addresses land use consideration under the Comprehensive Environmental Response,
Compensation and Liability Act of 1980 (CERCLA) process (directive is attached as Attachment G)
(EPA 1995d). The directive identifies sources and types of information that may aid EPA in determining
the reasonable anticipated future land use at a site. Examples of such information include zoning laws,
community master plans, site location in relation to current land uses and populations, groundwater
protection programs, and environmental justice issues. The directive recommends early discussions
between EPA and local land use planning authorities, local officials, and the public regarding reasonably
anticipated future land uses.
The principles identified in the directive are equally applicable to the RCRA corrective action program
(proposed Title 40 of the Code of Federal Regulations, Part 264 Subpart S Amendment, FR 19439, May 1,
1996). Available information and local input may indicate that nonresidential land use assumptions are
appropriate for corrective action facilities if there is reasonable certainty that the facility will remain
industrial. Factors such as residential properties located adjacent to or on an industrialized facility or child
care areas operated on commercial and industrial facilities should be considered when determining land
S «EP*«1>»i»'.T*SO'*£VISEWT»ljy.tt«ASn* WMM JJ-RHnlsnnnll/ll/'nn tan
-------
use. If the type of future land use cannot be reasonably predicted or if future residential land use (as well
as child care centers and recreational parks) cannot be reasonably ruled out, residential land use should be
assumed.
EPA is committed to ensuring that the public fully participates in all aspects of the RCRA corrective
action. EPA released a detailed guidance manual on public participation in RCRA programs (EPA 1993c)
and followed this guidance with a RCRA Expanded Public Participation Rule (FR 63417, December 11,
1995), EPA regards public participation as an important activity tlat empowers all communities,
including minority and low income communities, to become actively involved in local waste management
activities. This should include public participation in making land use and exposure assumption decisions,
particularly for communities potentially impacted by waste management activities and the risk
management decisions associated with corrective action.
EPA expects that contaminated soils will be cleaned up as necessary to prevent the transfer of
unacceptable concentrations of hazardous constituents from soils, including subsurface soils, to other
media; therefore, the uses of groundwater and surface water potentially impacted by constituent migration
from soil must be considered when cleanup levels are set. Likewise, the location of adjacent or nearby
residents potentially exposed to airborne constituents must also be considered.
4,2.2 Region 1® State Land Use Policies and Soil Cleanup Levels
The Washington State MTCA cleanup regulations specify when cleanup levels may be based on
residential or industrial land uses and were amended in 1996 under Chapter 173-340 of the Washington
Administrative Code (WAC). WAC 173-340-740 requires thai the residential land use scenario be
assumed unless a demonstration of nonapplicability can be made under subsection 740(1 Xa). Industrial
property soil cleanup levels can be established as follows if the site meets the definition of an industrial
property cited under the WAC 173-340-200:
s upa *!••>!« TASK«*iv>»rjw*i.»usTW »n*isi•«i>««»*«*11/>im,-> mm-jx
-------
"Industrial properties" mew properties that are or have been characterized by or are to be committed to traditional
industrial uses such as processing or manufacturing of materials, marine terminal and transportation areas and
facilities, fabrication, assembly, treatment, or distribution of manufactured products, or storage of bulk materials.
One of the following statements is true for industrial properties:
• Zoned for industrial use by a city or county conducting land use planning under Chapter 36.70A
Revised Code of Washington (RCW) (Growth Management Act)
• For counties not planning under Chapter 36.70A RCW (Growth Management Act) and the cities
within them, zoned for industrial use and adjacent to properties currently used or designated for
industrial purposes
WAC 173-340-745 provides additional criteria for determining whether a land use not specified in the
definition meets the "traditional industrial use" requirement or whether a land use zoning category meets
the requirement of being "zoned for industrial use." In addition, WAC 173-340-745 requires an evaluation
of comprehensive plan text or zoning code to verify that only industrial land uses may occur on the site.
WAC 173-340-745 also requires that residential soil cleanup levels be used at industrial properties in close
proximity to (generally, within a few hundred feet) residential areas, schools, or child care facilities, unless
site or constituent inaccessibility and constituent immobility can be demonstrated. Likewise, residential
soil cleanup levels should be used for current or potential future residential areas adjacent to properties
currently used or designated for industrial purposes. State of Washington Guidance for Clean Closure of
Dangerous Waste Facilities (Washington Department of Ecology [Ecology] 1994) specifies that according
1© WAC 173-303-610{2)(b)(I), numeric clean closure levels for soils, groundwater, surface water, and air
must be determined using residential exposure assumptions.
State of Oregon environmental cleanup law (Oregon revised Statutes 465.315 and 465.325), promulgated
on January 10, 1997, requires that current and reasonably anticipated future land uses be considered in risk
assessments and feasibility studies. As previously stated, this law (and the rules promulgated pursuant
thereto) is for state Superfund sites and is technically not part of the Oregon Department of Environmental
Quality authorized corrective action program for RCRA-regulated facilities. Idaho and Alaska have not
developed specific regulations or guidance documents that address land use issues. Idaho accepts the use
of cleanup levels based on residential and industrial land use (Tetra Tech 1996a). Land use is addressed in
proposed Alaska Cleanup Standard Regulations and Risk Assessment Guidance (proposed for public
comment on December 18, 1996) (Tetra Tech 1996b). State-specific rules and regulations, which are
evolving, should be consulted for further information.
-------
4.2.3
Land Use Policies and Other Media
Determination of cleanup levels for groundwater, surface water, and air is the same for all facilities
regardless of the site's land use classification as industrial or residential. EPA Region 10 recommends that
residential exposures be assumed and flat exceptions be made-only for extenuating circumstances.
Corrective actions for soil and groundwater media should assure that discharges from either media do not
exceed surface water, sediment, or air quality standards or risk-based criteria. Sections 4.2.3.2 (Surface
Water), 4.2.3.3 (Air), 4.3 (Standards and Criteria) and Chapter 5 (Ecological Sediment Criteria) provide
information on land use and standards and criteria for these media.
4.2.3.1 Groundwater
EPA expects to return useable groundwaters to their maximum beneficial uses wherever practicable.
When restoration of groundwater is not practicable, EPA expects to prevent or minimize further migration
of the plume, prevent exposure to the contaminated groundwater, and evaluate further risk reduction
(PR 19448, May 1, 1996). State-designated uses should be considered when setting cleanup levels. As
previously noted, however, although RCRA statutory language does not require nonauthorized state
standards to be followed, it does require cleanups to be protective of human health and the environment.
EPA has initiated a comprehensive state groundwater protection program (CSGWPP) to encourage each
state to coordinate its current and planned groundwater protection activities through a CSGWPP. EPA
remediation program personnel should be familiar with and utilize CSGWPPs (EPA 1997h). Washington
State has submitted a draft CSGWPP to EPA and was responding to EPA comments to the proposed plan
at this printing. No other Region 10 states have submitted a CSGWPP to EPA; however, Oregon and
Idaho have initiated CSGWPPs, and the following listed persons can be contacted regarding groundwater
use issues (for example, for information on groundwater classified as a source of drinking water or as
having significant ecological value).
1 TASmHEvKErriNAUMASTE* llama*
4-9
-------
STATE GROUNDWATER PROTECTION PROGRAM CONTACTS
Washington
Groundwater protection
Kurt Cook, Department of Ecology
(360)407-6415
Wellhead protection
David Jennings, Department of Health
(360) 586-9041
Oregon
¦ Groundwater protection
Amy Pattern, ODEQ
(503)229-5878
Wellhead protection
Cheree Stewart, ODEQ
(503) 229-5413
Idaho
Groundwater and
wellhead protection
Donna Rodman
(208) 373-0260
The Washington Slate draft CSGWPP proposes a groundwater protection goal based on the state's anti-
degradation policy contained within Chapter 90.48 of the Revised Code of Washington (RCW) (Water
Pollution Control), Chapter 90.54 RCW (Water Resources Act of 1971), and Chapter 173-200 WAC
(Water Quality Standards for Groundwater). Tie antidegradation policy applies to all state regulatory
programs and requires that existing and future beneficial uses of groundwater be maintained and that
degradation that interferes with these uses be prevented. All groundwater in the state is similarly
protected. Three tiers of groundwater quality standards are specified: (1) numerical Federal Safe
Drinking Water Act maximum contaminant levels (MCL), (2) natural background concentrations, or
(3) site-specific early warning values (WAC 173-200) (Ecology 1995a), When prevention of groundwater
contamination is not possible and where remediation measures are required, Washington State has set
attainment of federal safe drinking water act MCLs as remediation goals (Ecology 1995a). In addition,
Washington State has promulgated cleanup regulations under MTCA (WAC 173-340).
Washington State MTCA regulations require that the highest beneficial use of groundwater (that is,
drinking water and other domestic uses) be assumed when setting cleanup levels unless the following
criteria cited in WAC 173-340-720 can be demonstrated. In general, these criteria include demonstrating
5 •»EPt«BBim1'*SK»JtEWISt5«il*UMASIE* WFDvl JMUinwnmuinim* Wjmm
-------
that the groundwater is not a current or future source of drinking water based on the following: (I) tie
groundwater is present In insufficient quantity to yield water for a domestic well; (2) natural background
concentrations of organic or inorganic constituents are present and make use of the wafer for drinking not
practicable; and (3) the groundwater cannot technically be recovered for drinking water purposes based on
depth or location. It also must also be demonstrated that migration of contamination from an unusable
aquifer to a useable aquifer cannot occur. Information on specific MTCA cleanup goals is presented in
Section 4.3.2.
State of Alaska beneficial use regulations, applicable to both groundwater and surface water, are cited in
Title i 8 Alaska Administrative Code (AAC) Chapter 70, "Water Quality Standards," State of Oregon,
groundwater beneficial use regulations are cited in Oregon Administration Rules (OAR), Chapter 340,
Division 40, "Groundwater Quality Protection," State of Idaho groundwater beneficial use regulations are
cited in the Idaho Administrative Code (1AC), Title 01, Chapter 02, Section 299, "Groundwater Quality
Standards." The state-specific rules and regulations should be consulted for farther information.
4.23.2 Surface Water
The proposed RCRA corrective action regulations recommend that state-designated uses of surface water
be considered when setting cleanup levels (proposed 40 CFR 264 Subpart S Amendment,
FR 30804, July 27, 1990). Promulgated state and federal drinking water standards or risk-based levels
based on water ingestion can be used as cleanup levels for surface water designated for drinking water use.
If surface water has been designated by a state for uses other than drinking water, the EPA may consider
the state-designated use when establishing cleanup levels (FR 30818, July 27, 1990). In any case, federal
Clean Water Act (CWA) prohibitions of releases of hazardous substances and oil to surface waters should
be considered when determining cleanup levels [(CWA 311 (b) (3)].
Washington State MTCA regulations require that the highest beneficial use of surface water be assumed
when setting cleanup levels (WAC 173-340-730). Promulgated standards or risk-based levels based on
water ingestion should be used as cleanup levels for surface waters representing a source or potential
future source of drinking water. Risk-based standards based on human ingestion of fish or shellfish should
be considered when determining cleanup levels for surface waters that support or have the potential to
support fish or shellfish populations. Federal (CWA) (EPA 1986a) and state water quality criteria based
-------
on the protection of aquatic organisms must also be achieved. Washington Slate surface water beneficial
use and water quality standard regulations are cited in WAC 173-201,
As noted in Section 4.2.3 J, State of Alaska surface water beneficial use and water quality standards are
cited in Title 18, AAC, Chapter 7©, Oregon surface water beneficial use and water quality standards are
cited in OAR, Chapter 340, Division 41, "State-Wide Water Quality Maintenance Plan, Beneficial Uses,
Policies." Idaho surface water beneficial use standards are cited in IAC, Title 01, Chapter 02, Section 200,
"General Surface Quality Criteria," The state-specific rules and regulations should be consulted for
further information.
4.233 Air
A residential exposure scenario is assumed when calculating air cleanup levels (FR 30831, July 27, 1990);
however, the location of the most exposed resident must be identified when determining compliance with
the cleanup level. The point where the maximum long-term human exposure to air releases would occur is
typically outside of a facility boundary. Under the corrective action process, the most exposed individual
is identified on a site-specific basis and may be identified as someone living across tie street from the
facility, as a worker living on the site, or wherever else maximum long-term human exposure would occur
based on site characteristics (PR 30831, July 27, 1990). For clean closure, it is assumed that the most
exposed individual is at the unit boundary and that air constituent concentrations must be equal to or below
the health-based level at the unit boundary (EPA 1989e).
Alaska, Idaho, and Oregon do not have specific requirements for setting air cleanup standards at RCRA
facilities. Washington State MTCA regulations require that cleanup levels to protect air quality be based
on estimates of reasonable maximum exposure (RME) (WAC 173-340-750), the highest exposure that can
be reasonably expected to occur at a site under current or potential future uses. The cleanup level,
therefore, should be based on the residential scenario unless the criteria specified in WAC 173-340-750 for
nonresidential site uses can be demonstrated. The WAC 173-340-750 criteria generally require that no
current or future residential use of the site occur and that air emissions from the site not reduce the air
quality of adjacent residential areas. WAC 173-340-750 requires that ambient air cleanup levels for
nonresidential uses be established on a case-by-case basis.
" rcp«s »-lt WIOII JIJUnnarNMifllffM ItMrik
-------
4.3
IDENTIFICATION OF PROMULGATED STANDARDS AND CRITERIA
When developing cleanup levels for a facility, promulgated standards or criteria for environmental media
contaminated by facility activities must be considered. Federal'and slate standards and criteria that should
be considered when developing cleanup levels are summarized in Sections 43,1 (federal) and 4.3,2 (state).
The stale and federal policies and regulations addressing land use discussed in Section 4.2 require
consideration when selecting appropriate standards and criteria.
43.1 Federal Standards and Criteria
Federal standards exist for drinking water supplies and surface water bodies. Under the federal Safe
Drinking Water Act, MCLs, maximum contaminant level goals (MCLG), and secondary MCLs are
established for a number of inorganic and organic chemicals in drinking water supplies. MCLs may be
used as cleanup levels for groundwaters and surface waters that are current or potential drinking water
resources, provided they are deemed to be sufficiently protective given the overall contamination at or
from the facility. In other words, it is discretionary to use risk-based levels rather than MCLs for
determining cleanup levels which may impact drinking water because some MCLs are based on
technological barriers that no longer exist and/or do not consider additivity of risks when other chemicals
are present at levels of concern (see Section 4.7.4 for an example risk calculation where multiple
constituents with MCLs are present in groundwater). Other federal criteria for the protection of human
health include CWA ambient water quality criteria (AWQC), which are established for surface waters.
For the protection of human health, two types of AWQC have been established: (1) for tie ingestion of
water and fish aid (2) for the ingestion of fish only (EPA 1986a). The CWA prohibits the release of oil to
navigable surface waters in any quantity that causes a sheen, an emulsion, or a sludge, regardless of
cleanup standards that may be imposed and/or complied will [CWA 311 (b) (3)]. This prohibition is
federally enforceable in all states.
4.3.2 State Standards and Criteria
Washington State has promulgated cleanup regulations under MTCA (WAC 173-340). As stated in
Chapter 2, Ecology was authorized for a corrective action program that uses MTCA regulations. MTCA
establishes cleanup levels for soil, groundwater, surface water, and air.
S "JIEPAW r>«5l>T*SKrm¥lSESWNM.WASTEIt wnMSI«IMIIII*nliniM» Imt 4-13
-------
MTCA requires that where groundwaters and surface waters are current or potential future sources of
drinking water, cleanup levels be at least as stringent as promulgated federal and state standards. Including
federal MCLs and MCLGs, and Washington state MCLs published under WAC 248-54, In the absence of
promulgated standards, MTCA provides a method for calculating RBCs based on drinking water ingestion
(WAC 173-340-720).
For surface waters that support or have the potential to support fish or shellfish populations, MTCA
requires that human health cleanup levels be at least as stringent as federal AWQC established for the
protection of humans ingesting water and fish (WAC 173-340-730), In the absence of promulgated
standards, MTCA provides a method for calculating risk-based surface water cleanup levels based on fish
ingestion only (WAC 173-340-730).
MTCA requires that soil and air cleanup levels be at least as stringent as applicable state and federal laws.
MTCA provides methods for calculating soil RBCs based on soil ingestion (WAC 173-340-740) and air
RBCs based on inhalation (WAC 173-340-750). MTCA also requires that constituent concentrations in
soil not cause contamination of groundwater to concentrations that exceed groundwater cleanup levels.
MTCA provides three methods (A, B, and C) for determining risk-based cleanup standards. Method A
involves use of a table that identifies conservative, default cleanup levels for a limited number of common
hazardous substances. Method B is applicable when there are more hazardous substances involved.
Method C is applicable when Methods A and B cleanup levels are technically impossible to achieve, lower
than background, or may cause more environmental harm than good. MTCA also defines a second
Method C procedure for determining when soil cleanup levels for industrial land use can be used and how
these cleanup levels should be calculated. Method B and Method C involve calculations of cleanup levels
based on default exposure assumptions for different land-use scenarios. At this printing Ecology is
preparing revised MTCA regulations and should be consulted for the latest regulatory update.
Washington State published a MTCA Cleanup Levels and Risk Calculations (CLARCII) Update
(Ecology 1996) reference document. The document provides guidance on when Methods A, B, and C
should be used and provides tables of chemical-specific Methods B and C cleanup levels for hazardous
constituents in groundwater, surface water, and soils.
w»\T*sa*Evtsinnn«.«ASira
-------
The ODEQ has no specific regulations for setting cleanup levels at RCRA facilities- As noted in
Chapter 2, ODEQ was authorized to include corrective action in RCRA permits using EPA risk assessment
guidance to determine cleanup levels for the corrective action program. Oregon does have rules that
address contamination at state Superfund sites and recently enacted statutory language thai, among other
things, establishes target risk levels for cleanups and also requires development of a protocol for
probabilistic risk assessment (Oregon Revised Statutes 465.315 and 465.325). ODEQ promulgated rules
for these statutes on January 10,1997. The state Superfund cleanup rules are not currently a part of
Oregon's authorized RCRA corrective action program.
As noted in Chapter 2, Idaho and Alaska do not have authorized state-specific cleanup level regulations.
Alaska has not applied for authorization to manage RCRA programs in lieu of EPA. In accordance with
RCRA Section 3006 (b), Idaho's authorized program must be equivalent to the federal RCRA program;
however, an Idaho statutory provision prohibits regulations more stringent than EPA's (Idaho Code
Section 39-4404, Hazardous Waste Management Act of 1983.)
4.4 RISK-BASED SCREENING
When no promulgated standards or criteria are available for cleanup levels, facility-specific RBCs should
be calculated for the remaining COPCs. Before RBCs are calculated, hazardous constituents can be
screened from further consideration if they are present at concentrations below significant health concerns.
Generic RBCs calculated using residential scenario assumptions can be used to screen constituents. This
screening process should be performed as follows:
2. List risk-based concentrations of each constituent, using PRGs
calculated by EPA Region 9 (described in Section 4.5 and included
in Attachment E) (EPA 1996c)
List maximum concentration of each constituent in each medium
S HtPA'.Rl"n*8%TASK««EVISBrjnNA|.\|i.tASTIit WpOM5l-ft«»»8llo7tSRlini«?W SUmuae
4-15
-------
3.
Eliminate constituent from screening if the maximum concentration
is:
• Less than 1E-6 cancer risk screening value (Region 9
carcinogenic PRGs are based on 1 E-6 screening value so do
not need to be altered) or
• Less than 0.1 hazard index (HI) screening value (Since the
Region 9 noncarcinogenic PRGs are based on an HI of 1.0,
the PRG should be divided by 10 to meet 0.1 HI screening
value)
4.
Include remaining constituents for further consideration in
calculating cleanup levels
RBCs are provided only for soil and tap (drinking) water and are based on ingestion, inhalation, and
dermal contact (soil only) exposure pathways. Constituents present in other media or exposure pathways
should not be screened using this method. In addition, if a constituent is retained for consideration in the
risk analysis in one media, it generally should be retained in all media of concern to address possible
constituent migration and multiple exposure routes for the constituent.
As indicated in Step 3, the default screening level at which carcinogenic constituents can be eliminated is
based on a 1E-6 cancer risk. This screening level should be adequately protective of the cumulative risks
that may result from multiple facility-related carcinogens. In accordance with the Region 10 Supplemental
Risk Assessment Guidance for Superfund policy (EPA 1996f), Step 3 also shows that the screening
concentration for noncarcinogens should be based on an HI of 0.1, rather than 1.0. The screening level of
0.1 is conservatively protective of cumulative effects that could occur when multiple noncarcinogenic
hazardous constituents with similar toxic endpoints are present.
For purposes of assessing site risks, it may be assumed that if no single sample maximum value exceeds a
screening concentration as described above, total exposure to the constituent is not of concern for human
health.
I ftEPAUtli»NUTAS&^VIS£2\FfNAUMASTEllWFDlttMtM«l«iniMilJ/lVTOV Sfcirmue
-------
Aluminum, calcium, tron) magnesium, potassium, and sodium are not associated with toxicity to humans
under normal circumstances. No quantitative toxicity information is available for these elements from
EPA sources. Unless these elements have promulgated standards or criteria or unless site-specific factors
dictate, they generally can be eliminated from consideration during development of cleanup levels.
4.5 EXPOSURE ASSUMPTIONS
The exposure assumptions presented in this
section are used to determine the magnitude of a
potential chemical dose, which is the amount of a
given chemical entering the human body during a
specified time.
The default residential and industrial exposure assumptions recommended in this document to calculate
RBCs are presented in Table 4-2. These exposure assumptions, which were primarily taken from EPA
guidance documents, address human health concerns and are consistent with current federal and Region 10
CERCLA guidance. With the exception of the dermal exposure factors, the Table 4-2 factors are
consistent with those recommended in the most recent version of the Region 9 PRGs (see Attachment E,
[EPA 1996c]). Region 9 PRGs are updated annually and are available on the World Wide Web at
http://www.epa.gov/region09/waste/sfund/prg/index/html and on the California Regional Water;Board's
Bulletin Board System at (510) 286-0404 (PRG2ND96.ZIP).
Table 4-3 compares the exposure pathways considered in soil, water, and air risk-based concentration
calculations for EPA Regions 3 and 9. Region 3 has calculated risk-based concentrations for soil, water,
and air that are similar to the EPA Region 9 PRGs. The Region 3 values were previously recommended by
Region 10 for screening purposes; however, the use of Region 9 PRGs for risk-based screening is currently
being recommended, primarily because they are more comprehensive in that they take into account more
exposure pathways.
The exposure assumptions used to calculate
cleanup levels include body weight,
inhalation and ingestion rates, skin surface
area, absorption fractions, exposure
frequency and duration, and volatilization
and particulate emission factors (PEF).
5 WMMSI-ftlinlanKmltfll'm Wimm 4- 1 I
-------
(THIS PAGE INTENTIONALLY LEFT BLANK)
S iEM*li»«rT*SKI'«VlSEW1HAUM*S1I* WPPtlSl-»PMlItlHM»«»ll
-------
TABLE 4-2
STANDARD DEFAULT EXPOSURE FACTORS
REGION 10 RCRA RISK-BASED CLEANUP LEVEL GUIDELINES
Category
\ Symbol
Definition (units)
| Default
[1
Toxicity Factors
CSFo
CSFi
RfDo
Cancer slope factor oral
(mg/kg-dH
Cancer slope factor inhaled
(mg/kg-d)-!
Reference dose oral (mg/kg-d)
w
EPA 19%g.
EPA 1997a.
Section 4.6.2
RfDi
Reference dose inhaled (mg/kg-d)
-
Target Risks and Hazards
TR
Target cancer risk
10*
„
THQ
Target hazard quotient
I
„
Body Weight
BWa
Body weight, adult {kg)
70
EPA 1989c
BWc
Body weight, child (kg)
15
EPA 1989c
¦aging Time
ATc
Averaging time - carcinogens (days)
25550
EPA 1989c
ATn
Averaging time - noncarcinogens
(days)
ED*365
Dermal
SAa
Surface area exposed, adult
See Tabic 4-6
-
SAc
Surface area expose, child
See Table 4-6
-
AF
Adherence factor
See Tabic 4-6
--
ABS
Skin absorption
See Tabic 4-5
-
Inhalation
IRAa
Inhalation rate - adult (m'/day)
20
EPA 1991b
JRAc
Inhalation rate • child (m'/day)
10
EPA 1989c
Water ingestion
IRWa
Drinking water ingestion - adult
(L/day)
2
EPA 1989c
IRWc
Drinking water ingestion - chili
(L/day)
1
Gal/EPA 1994
Soil Ingestion
IRSa
Soil ingestion - adult (mg/day)
100
EPA 1991b
IRSc
Soil ingestion - child (mg/day)
200
EPA 1991b
IRSo
Soil ingestion - occupational
(mg/day)
50
EPA 1991b
Exposure frequency and
Duration
EFr
Exposure frequency - residential
(days)
350
EPA 1991b
EFo
Exposure frequency - occupational
(days)
250
EPA 1991b
EDr
Exposure duration - residential
(years)
30*
EPA 1991b
EDc
Exposure duration - child (years)
6
EPA 1991b 1
S 7A&K* RE VT$C2'.F)MAL\MA$T£ll WPtMM-f: X'lv 'KM, r \n
-------
TABLE 4-2 (Continued)
STANDARD DEFAULT EXPOSURE FACTORS
REGION IIRCRA RISK-BASED CLEANUP LEVEL GUIDELINES
Category
Symbol
Definition (units)
Default
Refei
Exposure Frequency and
Duration
EDO
Exposure duration - occupational
(years)
5
EPA 1991b
Age-adjusted Intake Rates'1
Age-adjusted factors far
carcinogens:
1
IFSadj
Ingestion factor, soils
(Img«yr]/Ikg~cJJ)
114
EPA 1991c
\
<
SFSadj
Skin contact factor, soils
([mg*yr}/[kg*d|)
503
EPA 1991c
IrthFadj
Inhalation factor ((m'«yr}/[kg-d})
11
EPA 1991c
IFWadj
Ingestion factor, water
([l«yr)/fkg-dj)
II
EPA 1991c
Fate mid Transport Models
VPw
Volatilization factor for water
(L/m')
0.5
EPA 1991c
PEF
Particulate emission factor (m'/kg)
Chemical-specific
(Table 4-8 )c
EPA 1996b
VPs
Volatilization factor for sol (rn'/kg)
Chemical-specific
(Table 4-7)c
EPA 1996b
sat
Soil saturation concentration
(mg/kg)
Chemical-specific
(Table 4-l0)c
EPA 1996b
S®«rce: Modified from U.S. Environmental Agency (EPA) 1996c
Notes:
a Exposure duration for lifetime residents is assumed to be 30 years total. For carcinogens, exposures arc combined for
children (6 years) and adults (24 years),
b Intake rates determined by analogy to age-adjusted soil ingestion factor published by EPA (1991c).
c Section 4.5.2 and Tables 4-7.4-1, and 4-10 are presented in EPA's Interim Guidelines for Developing Risk-based Cleanup
Levels at RCRA sites in Region 10 (this report).
mg/kg
cm3
m1
L
Cal/EPA
Milligram per kilogram
Square centimeter
Cubic meter
Liter
California Environmental Protection Agency
K v
lV
i
-------
TABLE 4-2
STANDARD DEFAULT EXPOSURE FACTORS
REGION 10 RCRA RISK-BASED CLEANUP LEVEL GUIDELINES
Category
Symbol
Definition (anits)
fault
Reference
Toxicity Factors
CSFo
Cancer slope factor oral
(mg/kg-d H
-
EPA 1996g.
EPA 1997a.
CSFi
Cancer slope factor inhaled
(mg/kg-d)-1
_
Section 4,6.2
RfDo
Reference dose oral (mg/kg-d)
_
RfDi
Reference dose inhaled (mg/kg-d)
_
Target Risks and Hazards
TR
Target cancer risk
,0-6
*•
THQ
Target hazard quotient
1
Body Weight
BWa
Body weight, adult (kg)
70
EPA 1989c
BWc
Body weight child (kg)
15
EPA 1989c
Averaging Time
ATc
Averaging time - carcinogens (days)
25550
EPA 1989c
ATn
Averaging time - noncarcinogens
(days)
ED'365
SAa
Surface area exposed, adult
See Table Jbft.—-
-Surface area expose, chilsU--"~
See Table 4-6
AF
jbdhoKrtlxfoctor ¦—
—>S«Tablc 4-6
—
—"""
ABS
Skin absorption
Sec Tabled?""—
<-
Inhalation
IRAa
Inhalation rate - adult (mVday)
20
EPA 1991b"
IRAc
Inhalation rate - child (mVday)
10
EPA 1989c
Water Ingestion
IRWa
Drinking water ingestion • adult
(L/day)
2
EPA 1989c 1
IRWc
Drinking water ingestion • child
(L/day)
I
Cal/EPA 1994
Soil Ingestion
IRSa
Soil ingestion - adult (mg/day)
V 100
EPA 1991b •
IRSc
Soil ingestion - child (mg/day)
EPA 1991b
IRSo
Soil ingestion • occupational
(mg/day)
/\
EPA 1991b
Exposure Frequency and
Duration
EFr
Exposure frequency - residential
(days)
350
EPA 1991b
EFo
Exposure frequency - occupational
(days)
250
EPA 1991b
EDr
Exposure duration - residential
(years)
30*
EPA 1991b
EDc
Exposure duration - child (sears)
6
EPA 1991b
" Sfi s^TTP i*PD •«> f i V'l *44*1** 4-18
-------
TABLE 4-2 (Continued)
STANDARD DEFAULT EXPOSURE FACTORS
REGION 10 RCRA RISK-BASED CLEANUP LEVEL GUIDELINES
! Category
Symbol
Definition (anits)
Default
Reference
Exposure Frequency and
Duration
EDo
Exposure duration - occupational
(years)
25
EPA 1991b
Age-adjusted Intake Rates'1
Age-adjusted factors for
carcinogens;
(FSadj
Ingestion factor, soils
([mg-yrj/fkg-dj)
114
EPA 1991c
SFSad]
Skin contact factor, soils
([mg-yr]/[kg-d])
503
EPA 1991c
InhFadj
Inhalation factor (fm'-yrMkg-tff)
11
EPA 1991c
IFWa#
Ingestion factor, water
<[|.yrl/[kg-d])
1-1
EPA 1991c
Fate and Transport Models
VFw
Volatilization factor for water
(Urn1)
0.5
EPA 1991c
PEF
Particulate emission factor (mJ/kg)
Chemical-specific
(Table 4-8)'
EPA 1996b
VFs
Volatilization factor for soil (m3/kg)
Chemical-specific
(Table 4-7)c
EPA 1996b
sat
Soil saturation concentration
(mgflcg)
Chemical-specific
(Table 4-10)c
EPA 1996b 1
1!
Soorce: Modified from U.S. Environmental Agency (EPA) 1996c
Notes:
a Exposure duration for lifetime residents is assumed to be 30 years total. For carcinogens, exposures are combined for
children (6 years) and adults (24 years),
b Intake rates determined by analogy to age-adjusted soil ingestion factor published by EPA (1991c).
c Section 4.5.2 and Tables *4-7. 4-8. and 4-10 are presented in EPA's Interim Guidelines for Developing Risk-based Cleanup
Levels al RCRA sites in Region 10 (this report).
mg kg Milligram per kilogram
cm: Square centimeter
m ' Cubic rotter
L Liter
C. 1 PA California Environmental Protection Agency
4,m
-------
TABLE 4-3
U.S. ENVIRONMENTAL PROTECTION AGENCY REGION 3 AND REGION 9
SOIL, WATER, AND AM EXPOSURE PATHWAYS
REGION 1# RCRA RISK-BASED CLEANUP LEVEL GUIDELINES
Medium
Pathway
Region 3*
Re]
Water
Ingestion
Yes
Yes
Inhalation of volatiles
No
Yes
Soli
Ingestion
Yes
Yes
Inhalation of particulates
No
Yes
Inhalation of volatiles
No
Yes
Dermal contact
No
Yes
Air
Inhalation
Yes
Yes
Notes:
a EPA 1996h
b EPA 1996c
S ftEPAftlittMKTASKrJtfVlSErZMAI^MASTER llmxic
-------
TABLE 4-4
WASHINGTON STATE MODEL TOXICS CONTROL ACT
CLEANUP LEVEL EXPOSURE FACTORS
REGION 1® ECEA RISK-BASED CLEANUP LEVEL GUIDELINES
Method B
Method C
ctbod C (Iadvstrial)
Expoinre Ftelor
No»-
CtrtiMgeu
Carcinogens
Noo-
CtrcintgetB
CirtiMgm
Non-
Cardnogeits
C#rcii*«l$
W«er Ingestion
Intake Rale (L/day)
I
2
2
2
NA
MA
Exposure Frequency*
—
_
_
—
HA
NA
Exposure Duration (year)
—
30
_
30
NA
NA
Body Weight (kg)
16
70
70
70
NA
NA
Averaging Time (year)
—
75
_
75
NA
NA
Unit Conversion Factor
(mg/kg)
1,000
1,000
1.000
WOO
NA
NA
Inhalation Coneclion factor
2 (VOCs)
2 (VOCs)
2 (VOCs)
2 (VOCs)
NA
NA
Summiry Factor
Nor-VOCs
VOCs
16.000
8,00#
8.75E-2
43IE-2
35,000
17,50®
8.75E-1
4.38E-1
NA
NA
NA
NA
Soil & Dost Ingestion
Intake Rate (mg/day)
200
200
100
100
S#
50
Exposure Frequency*
1.0
1.0
as
05
04
0,4
Exposure Duralion (year)
—
6
—
6
—
20
Body Weight (kg)
16
16
16
16
7©
7®
Averaging Time (year)
_
15
—
75
—
75
Unit Conversion Factor
(mg/kg)
to*
10"*
10*
lO*
10*
W*
Gastrointestinal Absorption
1.0
10
1.0
1.0
1.0
Summary Factor
10,000
1 0
320,000
40
3.500.000
131.3 J
Inhalation
Intake Rale (m'/day)
10
20
20
2®
NA
NA
Exposure Frequency1'
—¦
—¦
—
-
NA
NA
Exposure Duration (year)
—
30
_
3®
HA
NA
Body Weight (kg)
!6
70
70
70
NA
NA
Averaging Time (year)
—
75
—
75
NA
NA
Unit Conversion Factor
(mg/kg)
1,00®
1.000
1,000
1.00®
NA
NA
Absorption Percentage
1.0
1.0
1.0
1.0
NA
NA
Summary Factor
1,600
7 5E-3
3-50®
7 5E-2
NA
NA I
Notes:
a Target hazard quotient is I 0 for all cleanup levels, Target Risk levels are 10"* for Method B awl 10"' for Method C
b Exposure frequency is presented as a fraction of a year For example, 0-4 refers to an exposure frequency of 146 d*ys per year.
Ecology Washington Deparutwm of Ecology
— Value is present in numerator and denominator of equation and therefore, does not affect calculation
L Liter VOC Volatile organic compound
MA Not applicable kg Kilogram
mg Milligram
S urn *i<«n« r*si-r«;viitnnKAi.i«*nui wromMunowotwu/iim*
-------
Exposure parameters dictated by MTCA {Methods B and C) are presented in Table 4-4. For soil
exposures, MTCA only considers the ingestion pathway. For water exposures, MTCA considers both
ingestion and volatile organic compound inhalation pathways. MTCA rules do not have to be followed for
EPA-lead corrective actions, but they should be considered In terms of preventing the need of future,
additional corrective action under state authorities. No other EPA Region 10 states have developed
specific exposure assumptions for RCRA facility RBC calculations; exposure parameters recommended by
the EPA guidance should be used.
Additional information on dermal absorption factors is presented in Section 4,5,1, Information on how
fate and transport models are used to incorporate hazardous constituent migration into cleanup levels is
presented in Section 4.5.2.
As stated in Section 4.1, other scenarios including recreational, agricultural, and trespasser may be more
appropriate at a given facility. An exposure parameter source for these and other scenarios is EPA's
Exposure Factors Handbook 1989d, which EPA is updating [the EPA (1996e) update was currently
available on the Internet at the time of this printing]. The Exposure Factors Handbook summarizes the
current literature regarding human exposures to contaminated media via a variety of specific exposure
conditions (for example, inhalation rates based on light, medium, and heavy activities). Values from the
handbook can be used in lieu of the default exposure factors when reliable facility-specific exposure
information is available. The Draft Exposure Assessment Guidance, Attachment C of Guidance for
Performing Screening Level Risk Analyses at Combustion Facilities Burning Hazardous Waste
(EPA 1994i), further described in Section 4.5.2.4, also includes exposure parameters for human food chain
exposure pathways, such as garden produce and fish ingestion rates.
New Policy on Evaluating Health Risks to Children (EPA 1995e) requires that exposures to infants and
children be considered separately from adults. Children may be more or less sensitive to specific
constituents. They may also experience different types and rates of exposures; therefore, separate risk and
hazard estimates should be made for infants and children, or it should be clearly stated why this is not
done (for example, demonstrate that infants and children are not expected to be exposed to the constituent
IMinSt\TASK»vR£VlS£j\nNALU*4 ASTER
-------
of concern), EPA's Exposure factors Handbook(\9S9d) and the draft I996e update include information
on infant and children exposure rates,.
4.5,1 Dermal Absorption Factors
Dermal absorption of chemicals from soil and water across the skin and into the blood stream may occur.
The rate of absorption may be estimated by dermal absorption factors (for chemicals in soil) and dermal
permeability constants (for chemicals in water). Both dermal absorption factors (ABS) and dermal
permeability constants are used to calculate absorbed doses of chemicals via the dermal exposure route.
Few chemical-specific ABS values are available from EPA. Table 4-5 presents recommended ABS
values. References for these values include literature sources and Dermal Exposure Assessment:
Principles and Applications (EPA 1992d). The ABS value units are percentages (that is, percent absorbed
through skin).
Differences in soil characteristics may affect chemical desorption from soil For example, EPA (1992d)
compiled an ABS range of 0 J§1 to 0.03 for dioxins based on experimental data and recommended that the
lower end of the range could be used for soils with high organic content (dioxin less available to desorb)
and the higher end of the range for soils with low organic content (dioxin more available to desorb).
Limited experimental data are available lo assign constituent-specific ABS values based on soil
characteristics. To the extent that they are available and scientifically justifiable, constituent-specific ABS
values identified in the literature should be used.
Default ABS values for volatile organic compounds (VOC) .recommended by Region III Technical
Guidance Manual, Risk Assessment: Assessing Dermal Exposure from Soil (EPA 19950 tnay be used for
calculating cleanup levels if constituent-specific values are not available: 0.0005 for volatiles with a vapor
pressure equal to or greater than benzene (approximately 95.2 mm mercury) (Skowronski et al. 1988;
Franz 1984) and 0.03 for volatiles with a vapor pressure lower than benzene.
An EPA workgroup has drafted but not yet published a supplementary Superfund risk assessment guidance
specific to the dermal pathway; when available it should be referenced.
v wroui-HlmMlriJwij/lirtVl IIWMK 4-23
-------
Exposure parameters dictated by MTCA (Methods B and C) are presented in Table 4-4. For soil
exposures, MTCA only considers the ingestion pathway. For water exposures, MTCA considers both
ingestion and volatile organic compound inhalation pathways. MTCA rules do not have to be followed for
EPA-lead corrective actions, but they should be considered in terms of preventing the need of future,
additional corrective action under state authorities. No other EPA Region 10 states have developed
specific exposure assumPtions for RCRA facility RBC calculations; exposure parameters recommended by
the EPA guidance should be used.
Additional information on dermal absorption factors is presented in Section 4.5.1. Information on how
fate and transport models are used to incorporate hazardous constituent migration into cleanup levels is
presented in Section 4.5.2.
As stated 10 Section 41»other scenarios including recreational, agricultural, and trespasser may be more
appropriate at a given facility. An exposure parameter source for these and other scenarios is EPA's
Exposure Factors Handbook 1989d, which EPA is updating [the EPA (1996e) update was currently
ava,lable 00 Ae Intemet at time of Aii printing]. The Exposure Factors Handbook summarizes the
.- -X current literature regarding human exposures to contaminated media via a variety of specific exposure
¦ conditions (for example, inhalation rates based on light, medium, and heavy activities). Values from the
handbook can be used in lieu of the default exposure factors when reliable facility-specific exposure
information is available. The Draft Exposure Assessment Guidance, Attachment C of Guidance for
Performing Screening Level Risk Analyses at Combustion Facilities Burning Hazardous Waste
(EPA I994i), further described in Section 4.5.2.4, also includes exposure parameters for human food chain
exposure pathways, such as garden produce and fish ingestion rates.
New Policy on Eval*»ing Health Risks to Children (EPA 1995e) requires that exposures to infants and
chiIdren be considered separately from adults. Children may be more or less sensitive to specific
constituents. They may also experience different types and rates of exposures; therefore, separate risk and
hazard estimates should be made for infants and children, or it should be clearly stated why this is not
done (for example, demonstrate that infants and children are not expected to be exposed to the constituent
4-22
-------
of concern). EPA's Exposure Factors Handbook (1989d) and the draft 1996e update include information
on infant and children exposure rates.
4.5.1 Dermal Absorption Factors
Dermal absorption of chemicals from soil and water across the skin and into thej>lood stream may occur.
The rate of absorption may be estimated by dermal absorption factors (for chemicals in soil) and dermal
permeability constants (for chemicals in water). Both dermal absorption factors (ABS) and dermal
permeability constants are used to calculate absorbed doses of chemicals via the dermal exposure route.
Few chemical-specific ABS values are available from EPA. Table 4-5 presents recommended ABS
values. References for these values include literature sources and Dermal Exposure Assessment:
Principles and Applications (EPA 1992d). The ABS value units are percentages (that is, percent absorbed
through skin).
Differences in soil characteristics may affect chemical desorption from soil. For example, EPA (1992d)
compiled an ABS range of 0.001 to 0.03 for dioxins based on experimental data and recommended that the
lower end of the range could be used for soils with high organic content (dioxin less available to desorb)
and the higher end of the range for soils with low organic content (dioxin more available to desorb).
Limited experimental data are available to assign constituent-specific ABS values based on soil
characteristics. To the extent that they are available and scientifically justifiable, constituent-specific ABS
values identified in the literature should be used.
Default ABS values for volatile organic compounds (VOC) recommended by Region III Technical
Guidance Manual. Risk Assessment Assessing Dermal Exposure from Soil (EPA 19950 may be used for
calculating cleanup,levels if constituent-specific values are not available: 0.0005 for volatiles with a vapor
pressure equal to,Or greater than benzene (approximately 95.2 mm mercury) (Skowronski et al. 1988;
Franz 1984) and 0.03 for volatiles with a vapor pressure lower than benzene.
An EPA workuroup has drafted but not yet published a supplementary Superfund risk assessment guidance
specific to the dermal pathway; when available it should be referenced.
4-23
A'-MJS? P -"Mm we
$
-------
i
I
TABLE 4-5
RECOMMENDED DERMAL ABSORPTION FACTORS FOE SOIL
REGION 10 RCRA RISK-BASED CLEANUP LEVEL GUIDELINES
1 Compound
Dermal Absorption
Factor
Reference
Arsenic
0.03
Wester et al. (1993a)
Cadmium
fif
0.01
Wester etal. (1992a)
EPA (I992d)
Chlordane
0.04
Wester etal. (1992b)
2,4-D
0.05
Wester et al. (1996)
DDT
0.03
Wester etal. (1990)
TCDD
Low Organic Soil (<10%)
High Organic Soil (>10%)
0.03
0.001
EPA (1992d)
EPA(1992d)
Other Dioxins and Dibenzofurans
0.03
EPA (1992d)
PAHs
0.13
Wester etal. (1990) |
RGBs
0.14
Wester etal. (1993b)
EPA (1992d)
Pentachlorophenol
0.25
Wester etal. (1993c)
Generic Defaults
Volatile organic compounds with vapor
pressure i benzene (95.2 millimeters mercury)
0.0005
EPA 1995f, Skowronski et al.
1988
Volatile organic compounds with vapor
pressure < benzene (95.2- millimeters mercury)
0.03
EPA 1995f
Semivolatile organic compounds
0.1
Ryan et al. (1987)
Inorganic Compounds
0.01
Ryan et al. (1987)
Sources: EPA 1997b, EPA 1992d
Noln:
EPA U S Environmental Protection Agency
2.4-D 2..4-dichloropbct»xy acetic acid
DDT Dichlofodiphcnyltrichlorocthanc
TCDD Tclrachlorodibcnzo-p-dioxin
PAH Pot) nuclear aromatic hydrocarbon
PCD Pol chlorinated biphcnyl
Table revised 10/16/98
> rmi TASKrKCvtSOfHMtWMSlBi whmsmw w> r u i."1 , mmm 4-24
-------
To evaluate dermal contact with constituents in water, dermal absorption across the skin barrier is
determined using constituent-specific dermal permeability constants, expressed in units of centimeters per
hour. Equations for calculating dermal permeability constants are presented in Dermal Exposure
Assessment: Principles and Applications (EPA 1992d); EPA recommends calculating permeability
constants for organic compounds using the Potts and Guy equation presented on pages 5-36 through 5-38
(EPA 1992d). EPA recommends the use of the measured permeability constants for inorganic compounds
presented in Table 5-3 of the dermal exposure assessment report (EPA 1992d). The dermal exposure to
constituents in the water pathway was not incorporated into EPA Region 9 PEG equations. Equations for
assessing this pathway are included in Dermal Exposure Assessment: Principles and Applications
(EPA 1992d). Reduced equations for calculating risks or hazards resulting from dermal contact with
constituents in water lave been incorporated in the Section 4.7.2 RBC calculation equations. Adult and
child residential exposures (during showering or bathing) are considered.
When evaluating the dermal contact exposure pathway (for both soil and water) the total surface area of
body exposed must be estimated. For showering and bathing, whole-body surface area is assumed. For
soil exposures, portions of the body (for example, hands, arms, lower legs, face, and neck) are assumed to
contact soil. The duration of exposure must also be estimated (for example, assume a 15-minute-per-day
showering time). EPA-recommended defaults for dermal contact exposure factors are presented in
Table 4-6.
4.5.2 Fate and Transport Models
Table 4-1, discussed in Section 4.1, lists potential exposure pathways for human receptors. Many of the
exposure pathways result from contamination migrating from one medium to another. For example, soil
contamination may migrate into groundwater, subsequently causing exposure to persons using the
contaminated aquifer as a drinking water source and/or discharge to surface waters, which may have both
human health and ecological impacts, depending on use. Cleanup levels for the primary medium may
require an adjustment to be protective of hazardous constituent migration into the secondary medium.
Certain fate and transport modeling equations have been standardized for this purpose.
5 flCFA.RI « i*!T ASM*EV?SEi\f JVALXMASTER ICjunUM
-------
See
TABLE 4-5
RECOMMENDED DERMAL ABSORPTION FACTORS FOR SOIL
REGION 10 RCRA RISK-BASED CLEANUP LEVEL GUIDELINES
Dermal Absorption
Factor
Compound
Reference
Arsenic
Wester et al. (1993a)
Cadmium
Wester et al. (! 992a)
EPA (I992d)
Chlordane
Wester et al. (1992b)
Wester et al. (1996)
Wester et al. (1990)
TCDD
Low Organic Soil (10%)
EPA (1992d)
EPA (1992d)
Other Dioxins and Dibenzofurans
EPA (I992d)
Wester et al. (1990)
Wester et al. (1993b)
EPA (!992d)
Pentachlorophenol
Wester et al. (1993c)
Generic Defaults
Volatile organic compounds with vapor
j Pressure a benzene (95,2 millimeters mercury)
0.0005
^PA 1995f, Skowronski et al.
I9g8
Volatile organic compounds with vapor
pressure < benzene (95.2 millimeters mercury)
0.03
Semivolatile organic compounds
0.1
Ryan et al. (1987)
Inorganic Compounds
Sources: EPA 1997b. EPA I992d
Noim:
EP-\ I S environmental Proteciion Agency
2.4-D 2 4-dichlornphenoxy acetic acid
DDT Dichlnrodiphem imcbloroethanc
JX'DD Ti'trachlorodibenzo-p-dioHin
P \It S'nhnuclear aromatic h>dnreartoo
Pt B PnhcMorinated biphenyl
0.01
- 1 . ' ^ V ' ' M ¦ .
4-24
-------
To evaluate dermal contact with constituents in water, dermal absorption across the skin barrier is
determined using constituent-specific dermal permeability constants, expressed in units of centimeters per
hour. Equations for calculating dermal permeability constants are presented in Dermal Exposure
Assessment: Principles and Applications (EPA 1992d); EPA recommends calculating permeability
constants for organic compounds using the Potts and Guy equation presented on pages 5-36 through 5-38
(EPA 1992d). EPA recommends the use of the measured permeability constants for inorganic compounds
presented in Table 5-3 of the dermal exposure assessment report (EPA 1992d). The dermal exposure to
constituents in the water pathway was not incorporated into EPA Region 9 PRG equations. Equations for
assessing this pathway are included in Dermal Exposure Assessment: Principles and Applications
(EPA 1992d). Reduced equations for calculating risks or hazards resulting from dermal contact with
constituents in water have been incorporated in the Section 4.7.2 RBC calculation equations. Adult and
child residential exposures (during showering or bathing) are considered.
When evaluating the dermal contact exposure pathway (for both soil and water) the total surface area of
body exposed must be estimated. For showering and bathing, whole-body surface area is assumed. For
soil exposures, portions of the body (for example, hands, arms, lower legs, face, and neck) are assumed to
contact soil. The duration of exposure must also be estimated (for example, assume a 15-minute-per-day
showering time). EPA-recommended defaults for dermal contact exposure factors are presented in
Table 4-6.
4.5.2 Fate and Transport Models
Table 4-1, discussed in Section 4.1, lists potential exposure pathways for human receptors. Many of the
exposure pathways result from contamination migrating from one medium to another. For example, soil
contamination may migrate into groundwater, subsequently causing exposure to persons using the
contaminated aquifer as a drinking water source and/or discharge to surface waters, which may have both
human health and ecological impacts, depending on use. Cleanup levels for the primary medium may
require an adjustment to be protective of hazardous constituent migration into the secondary medium.
Certain fate and transport modeling equations have been standardized for this purpose.
. v . i *.i \ : HSU M*nT!K » »' *¦*>* - Mi-*-*
4-25
-------
I
FO
T\
TABLE 4-6 (amended li/lfifflg)
RECOMMENDED DEFAULTS FOR DERMAL EXPOSURE FACTORS*
REGION 10 RCRA RISK-BASi-P » mm»»d , cSSS™TORS
water Contact
Soil Contact
Swimming
Central
Cental
Central
site-specific
Event tfme and
frequency
10 minutes evert
1 event/day
350 tfays/year
15 minutes/went
1 event/day
350 days/year
Site-specific
60 minute
1 evenVmorth
Site-specific
350 events/year
Exposure duration
9} " :
6 pars-child
30 pare
9 years
30 years
Skin surface area
18,000 cm*-adufl*
6,500 cm1 -chiltf
18,000 cm2 - adt#
6,500 en#- chief
-adulf-
2,31 cm2-chit#
2,500 enr®- occupation#
Soil-lo-skin
adherence rale'
0.1 mgfcirf - event - adult
0.2 mg?rf-evert-child
01 nrfcm8 - evant»occ.'
child and adult - none*1
0.2 mnfar? - went - occ.'
«»w draft of l he
Assumes total body surface area far adult and child
ESSSSftSSS^
Skin surface area has no upper value since online body weight*pr ca^^ts 2l5 a,ms' w ^ 'e0!-
Value as established for a gardener.
vlT^eZTS^f'"""181b"S"dl""*9l"»"dl'amH:
No! applicable.
-------
I
I
Models thai address volatile and particulate emissions from soil into air are described In Section 4.5.2.1.
Fate and transport assumptions that estimate the transfer ofVOC from water into indoor air during
household water use are identified in Section 4.5.2,2, A model and partition equation that address
migration of constituents from soil to groundwater are described In Section 4.5.2.3. Partition equations
that address migration of constituents from soil into food chain organisms are discussed in
Section 4.5.2,4.,
Several of the models discussed in this section are recommend by EPA's Soil Screening Guidance:
Users * Guide (EPA 1996b), These include models for estimating volatile and particulate emissions from
soil, and a model for estimating soil-to-groundwater constituent migration, EPA's soil screening levels
and associated models were developed for screening purposes, For use during early investigative
processes such as the RFI. The use of conservative, facility-specific soil and aquifer parameter* will
result m the calculation of health-protective soil screening levels. The facility must adequately justify all
facility-specific parameters used to calculate soil screening levels.
Alth0ugh tlle EPA soil screening guidance (1996b) was developed for the CERCLA program, it can be
considered for RCRA corrective action facilities. EPA does not intend for soil screening levels to serve
as national cleanup standards. The screening levels arc very conservative and can be used to determine
that the soil-to-air or soil-to-groundwater pathways are either not significant or that further assessment of
these pathways is warranted. The soil screening levels could be used as cleanup levels if conditions at
the facility are representative of those assumed during the development of the screening levels. Higher
cleanup levels that are stilt fiea th protective may be identified using facility-specific fate and transport
models.
4,5.2.1 SoIl-to-Air
Volatilization and particulate emission factors (VF and PEF), which are described in the following
section, are used in the soil RBC calculations to address long-term inhalation exposures. Equations for
deriving these factors are presented in Tables 4-7 and 4-8. Section 4.7 and Exhibit 4-1 present how the
factors are incorporated into RBC calculations. VFs should be estimated for VOCs, while PEfs should
be est,mated for compounds that may exert significant toxicity via dust inhalation. VOCs are defined in
* ! **>.< nttr/ii.u.aiAiTW wmi»»m»UMur
4-2?
-------
TABLE 4.6
RECOMMENDED DEFAULTS FOR DERMAL EXPOSURE FACTORS*
REGION 10 RCRA RISK-BASED CLEANUP LEVEL GWDEUNES
Water
Contact
Soil Contact
Bathing
Swimming
Cenlral
Central
Iral
site-specific
^cut lime jnd
IrcquciKv
tO minutes event
event/day
J50 days/year
15 minuiesfcyeni
I evenl/day
350 days/year
Site-specific
60 minute
vent/month
Site-specific
350 events/year
I MMisiirc dilution
V years - adult
<> years - child
30 years
9 years
Skin surface area
18.00(1 tin - adult"
6,000 cm' - child*
22,000 cm1 - adult'
7,500 cm1 - child*
18.000 cut4-adult*
6Jflocm1 • child*
5,700 cm - adult1
2,900 cm1 - child*
6.600 cm' - adult'
3,;E PA <' 9"b). With background data and rationales for the defaults derived from EPA informal
' lsor> Jraft of the Exposure Factors Handbook (1996e draft update of EPA 1989d)
Assumes total body surface area for adult and child
Square centimeter
milligrams
Not applicable
S IMMMVISllflNU'MASnit Durw
-------
Models that address volatile and particulate emissions from soil into air are described in Section 4.5.2J»
Fate and transport assumptions that estimate the transfer of VOC from water into indoor air during
household water use are identified in Section 4.5.2.2. A model and partition equation thai address
migration of constituents from soil to groundwater are described in Section 4.5.2.3. Partition equations
that address migration of constituents from soil into food chain organisms are discussed in
Section 4.5.2.4.
Several of the models discussed in this section are recommend by EPA's Soil Screening Guidance:
Users' Guide (EPA 1996b). These include models for estimating volatile and particulate emissions from
soil, and a model for estimating soil-to-groundwater constituent migration. EPA's soil screening levels
and associated models were developed for screening purposes, for use during early investigative
processes such as the RFI. The use of conservative, facility-specific soil and aquifer parameters will
result in the calculation of health-protective soil screening levels. The facility must adequately justify all
facility-specific parameters used to calculate soil screening levels.
Although the EPA soil screening guidance (1996b) was developed for the CERCLA program, it can be
considered for RCRA corrective action facilities. EPA does not intend for soil screening levels to serve
as national cleanup standards. The screening levels are very conservative and can be used to determine
that the soll-io-air or soil-tcv-groundwater pathways are either not significant or that further assessment of
these pathways is warranted. The soil screening levels could be used as cleanup levels if conditions at
the facility are representative of those assumed during the development of the screening levels. Higher
cleanup levels that are still health protective may be identified using facility-specific fate and transport
models.
4.5.2.1 Soil-to-Air
Volatilization and particulate emission factors (VF and PEF), which are described in the following
section, are used in the soil RBC calculations to address long-term inhalation exposures. Equations for
deriving these factors are presented in Tables 4-7 and 4-8. Section 4.7 and Exhibit 4-1 present how the
factors are incorporated into RBC calculations. VFs should be estimated for VOCs, while PEFs should
be estimated for compounds that may exert significant toxicity via dust inhalation. VOCs are defined in
. (Ipip . >. 1
-------
TABLE 4-7
DERIVATION OF THE VOLATILIZATION FACTOR
REGION I© RCRA RISK-BASED CLEANUP LEVEL GUIDELINES
Q/C x {3,14 x DA x 7) x 10'4 (m2lcm2)
VF (m3lkg) 1 -± ! -
(2 * pb x Da)
where
„ KB^D/Y' ~ 6™DJIn*\
A PA - B» * W
Parameter
Definition (units)
Default
VF
volatilization factor (m'/kg)
- |
Da
apparent diflusivity (craVs)
— I
Q/C
inverse of lie mean concentration at the center of a
0.5-acre-square source (g/m'-s per kg/m5)
68.81 (Los Angeles) or facility-specific 1
(Table 4-9) , . |
T
exposure interval (s)
9.5 x 10* (30 years) J
ft
dry soil bulk density (g/cm5)
1.5 )
0.
air-filled soil porosity (LjL^j)
n - 0W = 0.28
II11
total soil porosity (L^^/L^)
1-(ft/p.) = 043
k
water-filled soil porosity (1^,/L^J
0.15
P.
soil particle density (g/cms)
2.65
D,
diflusivity in air (cmVs)
chemical-specific*
H'
diiuensionless Henry's law constant
chemical-specific*1 b
Dw
diflusivity in water (cmVs)
chemical-specific'
K,
soil-water partition coefficient (cnvVg) ¦ KJk
(organics)
chemical-specific*
K«
soil organic carbon partition coefficient (cmJ/g)
chemical-specific*
L
fraction organic carbon in soil (g/g)
0.006 (0.6%) or facility-specific, if
available
Source? US, Environmental Protection Agency 1996b
Notes:
The Henry's Law constant used in the VF equation is tfimenstonJesi, end can be converted from a Henry's Law constant expressed in writs
of atmosphere-cubic meter per mole by multiplying by 41.
a
See AMacteneftl H ctnJ
Square centimeter
b
Dimensionless Hcnry'i Law constant mJ
Cubic meter
Lb,
Volume of air in litm s
Second
Volume of soil in Bias m1
Square meter
Pore volume in liters
Volume of water in liters
e
Gram
cm*
Cubic centimeter
kg
Kilogram
¥ \lTSjmW8AlL£YWlW4) _. »K2\131 * 1 *H lOTfrWH/l 0 IJ wfatt
4-28
-------
TABLE 4-8
DERIVATION OF THE PARTICULATE EMISSION FACTOR
REGION 1® RCRA RISK-BASED CLEANUP LEVEL GUIDELINES
PEF (m3/kg) - QIC x 3,600 s,h
0,036 x (l-F) x (UJUtf x F{x)
Parameter
DeflofcRxi (until)
Default j
PEF
particulate emission factor (mVkg)
1.32x10'
Q/C
inverse of mean concentration at the center of a
0.5-acre-square source (g/mJ-s per kg/niJ)
90,80 - Minneapolis or
facility-specific (Table 4-9) ||
¥
fraction of vegetative cover (uiutlcss)
0.S (50%) J
u„
mean annual wind speed (m/s)
4.69 ||
Ut
equivalent threshold value of wind speed at 7 in (m/s)
11.32 ||
F(x)
function dependent on U JUt derived using data published by
Cowherd el &1. (1985) (unit!ess)
0.194
Source: Modified from U.S. Environmental Protection Agency 19961}/
Notes:
Tic defaults presented in this figure are intended to calculate a PEF that is adequately protective at most facilities. Cowherd el ai
(1985) present method* for site-specific measurement of the parameters necessary to calculate a site-specific PEF.
g Oram
kg Kilogram
m Meter
til1 Square meter
m3 Cubic meter
s Second
4-29
-------
EPA guidance as having a Henry's Law Constant greater than 10 s atmosphere-cubic meter per mole and a
molecular weight less than 200 grams per mole (EPA 1991c). The models discussed in this section concern
modeling hazardous constituents from soil to outdoor air. Soil to indoor air may also need to be considered;
however, models regarding this migration pathway are not described in this document. A model developed by
P.C. Johnson and R.A. Ettinger (1991) can be used to predict the intrusion rate of constituent vapors into
buildings through breaches such as foundation cracks. EPA Region 10 risk assessors should be consulted if
soil to indoor air is a potential exposure pathway.
VolalHiMtiaiiJEMQr
The soil-to-air VF (referred to as VF,) is used to define the relationship between the concentration of the
constituent in soil and the flux of the volatilized constituent to air. The VF, equation presented in EPA's Soil
Screening Guidance: Users' Guide (1996c) should be used when calculating soil screening levels for VOCs
and can also be used to calculate risk-based concentrations. This equation, which is presented in Table 4-7, is
used to incorporate VOC inhalation exposures into Region 9 soil PRGs (1996d). The VF, equation
calculates the maximum flux of a constituent from contaminated soil and considers soil moisture conditions.
Chemical-specific parameters that may be used to calculate VF% including difiusivity in air, dimension less
Henry's Law constant, difiusivity in water, and soil organic-carbon partition coefficient, are presented in
EPA's Soil Screening Guidance: Users' Guide (1996c) and in Attachment F (Chemical Properties Table C
from EPA [1996c]). These chemical-specific parameters arc used to calculate Region 9 soil PRGs (EPA
1996d) and are presented in the electronic version of the PRGs (accessible through the World Wide Web
address cited in Section 4.5 and presented in Attachment C).
The dispersion factor (Q/C) used in the VF, equation was derived from a modeling exercise using a full year
of meteorological data for 29 U.S. locations selected to be representative of the range of meteorological
conditions across the nation. The results of these modeling runs have been compiled for nine climatic zones
and sources of 0.5 to 30 acres (Table 4-9). A dispersion factor of 68.81 grams per square-meter second per
kilogram per cubic meter (g/nr-s per kg/m3) (Los Angeles) is used by Region 9 to determine a default VF,
and may be used for screening purposes (EPA 1996c). To develop a facility-specific VF„ place the facility
into a climatic zone and choose a dispersion factor that best represents the site's size and meteorological
conditions.
MWMUUrwiW) mriJl llarat.
4-3®
-------
TABLE 4-9
DISPERSION FACTOR VALUES BY SOURCE AREA, CITY, AND CLIMATIC ZONE
REGION 10 RCRA RISK-BASED CLEANUP LEVEL GUIDELINES
Zone am# City
Zone 1
Q/C (g/m1-# per kg/m*) |
1 Acres
2 Acres
5 Acrei
1® Acres
30 Acres I
Seattle, WA
82.72
72.62
64.38
55,66
50.09
42.86 1
Salem, OR
73,44
64,42
57,09
4933
44.37
37.94
Zone II
Fresno, CA
62.00
54J7
48.16
41.57
37.36
31,90
Los Angeles, CA
68.8!
60.24
53.30
45.93
41.24
35.15
• San Francisco, CA
89.51
78,51
69.55
60,03
53.95
46.03
ZoneIH
Las Vegas, NV
95,55
83,87
74.38
64,32
57.90
• 49.56
Phoenix, AZ
64.04
56.07
49.59
42.72
38.35
32.68
Albuquerque, MM
84.18
73.82
65.40
56.47
50.77
43,37
Zone IV
Boise, ID
69.41
60.88
53.94
46.57
41.87
35.75 1
Wumemucca, NV
69,23
60,67
53,72
46.35
41,65
35.55 J
Sail Lake City, UT
78.09
68.47
60.86
52.37
47,08
40.20
Casper, WY
100,13
87.87
77,91
67.34
60.59
51.80
Denver, CO
75.59
66.27
58,68
50.64
45.52
38.87
Zone V
Bismarck, ND
83.39
73.07
64.71
55.82
50,16
42,79
Minneapolis, MN
90.80
79.68
70.64
61.03
54.90
46.92
Lincoln, NE
81.64
71.47
63.22
54.47
48.89
41.65
Zone VI
Little Rock, AR
73.63
64.51
57.10
49.23
44.19
37.64
Houston, TX
79.25
69.47
61.53
53.11
47.74
40.76
1 Atlanta, GA
77.08
67.56
59.83
51.62
46.37
39.54
PAUa£RMBAJCLEy\WLVF4},B»VUMt|O0il07S«U/4^i/|0i)w\M>
-------
TABLE 4-9 (Continued)
QUALITY CONTROL VALUES BY SOURCE AREA, CITY, AND CLIMATIC ZONE
REGION 10 RCRA RISK-BASED CLEANUP LEVEL GUIDELINES
Zone and City
Q/C (g/raJ-s per kg/mJ)
©.5 Acre
1 Acres
2 Acres
5 Acres
10 Acres
30 Acre
Atlanta, GA
77.08
67.56
59.83
51.62
46.37
39.54
Charleston, SC
74.89
65.65
58.13
50.17
45.08
38.48
Raleigh-Durham, NC
77,26
67.75
60.01
51.78
46.51
39.64
Zone ¥11
Chicago, IL
97.78
SSMt
76.08
65.75
59.16
50.60
Cleveland, OH
83.22
73.06
64.78
55.99
50.38
43.08
Huntington, IN
53.89
47.24
41.83
36.10
32.43
27.67
Harrisburg, PA
81.90
71.87
63.72
55.07
49.56
42.40
Zone ¥111
Portland, ME
74.23
65.01
57.52
49.57
44.49
37.88
Hartford, CT
71.35
62.55
55.40
47.83
43.00
36.73
Philadelphia, PA
90.24
79.14
70.14
60.59
54.50
46.59
Zone IX
Miami, FL
85.61
74.97
66.33
57.17
51.33
43.74
Source: Modified from U.S. Environmental Protection Agency 1996b
Notes:
g Gram
kg Kilogram
rtr Square meter
m1 Cubic meter
s Second
0 C Inversion of mean concentration at the center of a source
S REPAJU!»>J* ?ASK*-*EV1SEjnW*UMAS7CR WPD-nMUmivroiiniZflt/W/* JSwftysc
4-32
-------
Because of its reliance on Henry's Law (which provides a measure of the extent of chemical partitioning
between air and water) the VF model is applicable only when the constituent concentration in soil water
is at or below saturation (that is, there is no free-phase constituent present). This corresponds to the
constituent concentration in soil at which the adsorptive limits of the soil particles and the solubility
limits of the available soil moisture have been reached. Above this point, pure liquid-phase constituent
can be expected to exist in the soil. Table 4-10 presents the soil saturation concentration equation
(originally presented by Soil Screening Guidance: User's Guide [EPA 1996b]).
In addition, EPA Region 9 (1996c) las calculated soil saturation concentrations for VOCs, and reported
these concentrations as PRGs when they exceed the saturation limit (this is designated with a "SAT'
qualifier in the tables),
Inhaiation of fugitive dusts is a consideration for nonvolatile constituents in surface soils. The PEF •
relates the concentration of constituent in soil to the concentration of dust particles in air and represents
an annual average emission rate based on wind erosion. The PEF equation presented in EPA's Soil
Screening Guidance: Users' Guide (1996b) should be used when calculating soil screening levels for
compounds know to exert significant toxicity via the fugitive dust inhalation pathway (Table 4-8) and
can also be used to calculate risk-based concentrations. This equation is also used by EPA Region 9
(1996c) when calculating soil PRGs. The Q/C used in the PEF equation was derived from a modeling
exercise using a full year of meteorological data for 29 U.S. locations selected to be representative of the
range of meteorological conditions across the nation. The results of these modeling runs have been
compiled for nine climatic zones and sources of 0.5 to 30 acres (Table 4-9). A dispersion factor value of
90.80 g/nr-s per kg/mJ (Minneapolis) is used by Region 9 to determine a default PEF and may be used
for screening purposes (EPA 1996c). To develop a facility-specific PEF, place the facility into a climatic
zone and choose a dispersion factor and wind speeds that best represent the site's size and meteorological
conditions.
S \REPA«1»»tSi\TASWr.REVlSE2\nNAL\MASTER wmiSi-RiMMlKTOsstl/Jim* iSttrttx
-------
TABLE 4-1#
DERIVATION OF THE SOIL SATURATION LIMIT
REGION 10 RCRA RISK-BASED CLEANUP LEVEL GUIDELINES
C„, = £ (If A ~0.~ H' e.)
p.
Parameter
Definition (units)
Default
soil saturation concentration (mg/kg)
calculated
S
solubility in water (mg/L-water)
chemical-specific*
Pt»
dry soil bulk density (kg/L)
1.5
Kd
soil-water partition coefTicieni (L/kg)
x fw (chemical-specific1)
fraction organic carbon in soil (g/g)
0.006 (0.6%) or facility-specific, if available
®w
walcr-filled soil porosity (L^L**)
0.15
H'
dimension less Henry's law constant
chemical-specific*
0,
air-filled soil porosity (L^/L^,,)
n - 0W = 0.28
ii
total soil porosity
I - (Pt/P.) ~ 0.43
P.
soil particle density (kg/L)
2.65
Source: Modified from U.S. Environmental Protection Agency 1996b
Notes;
a
See Attachment H
mg/kg
Milligram per liter
L
Liter
S
Gram
^•water
Volume of water
^"SOll
Volume of soil in liters
Laif
Volume of air in liters
^-pote
Pore volume in liters
REPVJtimftrTASKrNgVtSErtfNALWASTEft WTOMM'ftimitoTWnlWim* "taw
4-34
-------
Both EPA (1996b) and EPA Region f (1996c) acknowledge that when soil ingestion and fugitive dust
inhalation are evaluated together, the risks and hazards associated with ingestion are significantly greater
than those associated with inhalation. Exceptions are the metals chromium (hexavalent form) and
cadmium; therefore, the inclusion of the fugitive dust inhalation pathway can be limited to these two
metals or other compounds known to exert significant toxicity via dust inhalation. Default PEF
modeling assumptions can normally be assumed; however, if site conditions are such that higher fugitive
dust emissions than the defaults are likely (for example, dry, dusty soils, high average annual wind
speeds, vegetative cover less than 50 percent) and cadmium or hexavalent chromium is present in surface
soils, site-specific parameters should be used in the PEF equation (EPA 1996b).
4.5.2.2 Household Water-to-indoor Air
A groundwater-to-indoor air VF (VFW) of 0.005 x 1,000 L/mJ is used to define the relationship between
the concentration of the constituent in household water and the average concentration of the volatilized
constituent in air (EPA 1991c). In the derivation, all uses of household water were considered (for
example, showering, laundering, dish washing). It was assumed that the volume of water used in a
residence for a family of four is 720 L/day, the volume of the home is 150,000 L, and the air exchange
rate is 0.25 mThour. Furthermore, it is assumed that the average transfer efficiency weighted by water
use is 50 percent (that is, half of the concentration of each chemical in water will be transferred into air
by all water uses. Note: the range of transfer efficiencies extends from 30 percent for toilets to
90 percent for dishwashers). The VF„ is used in the groundwater RBC calculation equation (presented in
Section 4.7.2) to assess volatilization of constituents from tap water into indoor air. Use of the EPA
(1991 c) water-to-air Vfw results in a conservative estimation of volatilized constituent concentrations.
Updated estimates of water volume use, house volume, and air exchange rate are presented in the 1995
* draft revised Exposure Factors Handbook (EPA 1996e) and may be used to recalculate a Vfw if
warranted by facility-specific conditions. In addition, the intrusion rate of vapors through building
foundations into enclosed spaces may be predicted using a model developed by Johnson, et al. (1991).
4-1S
S REPA TASKiTJREVlSEJ\f1NAL\MAST1EX ^ ** *
-------
4.5.2 J Soil-to-Groundwater Estimations
EPA's Soil Screening Guidance: Users' Guide (EPA 1996b) recommends a dilution factor model and
soil/water partition equation for estimating soil screening levels that are protective of groundwater. The
approach requires that groundwater constituent concentrations at the downgradient edge of contaminated
soil not exceed MCL or risk-based groundwater cleanup levels (that is, cleanup levels for residential
use). The method is applied in step-by-step fashion. A groundwater cleanup level and dilution factor are
identified. The dilution factor is based on facility-specific aquifer characteristics, including hydraulic
conductivity, hydraulic gradient, mixing zone depth, and source length. The dilution factor is multiplied
by the groundwater cleanup level to determine a target soil ieachate concentration. The target soil
leachate concentration represents a constituent concentration that, upon dilution in groundwater, will not
result in an exceeded groundwater cleanup level directly beneath the site.
The target soil leachate concentration can be compared directly to leach test results for the site. EPA
(1996b) provides some guidance on tie availability of leach tests and suggests using the EPA Synthetic
Precipitation Leaching Procedure [Method 1312 from the current edition of SW-S46 (EPA 1986b)].
EPA (1996b) also provides a soil/water partitioning equation for converting the target soil leachate
concentration into a total soil concentration. The equation requires that site-specific soil parameters,
including fraction organic carbon, soil porosity, and soil density, be determined. Default soil parameters
are also proposed. The total soil concentration can be used as a soil screening level.
The previous procedures assume an infinite source of contamination. Because this assumption can
violate mass balance considerations, such as for small sources, EPA (1996b) also presents a model for
calculating mass-limit soil screening levels. The mass-limit soi! screening level represents a soil
constituent concentration that is still protective of groundwater cleanup levels when the entire volume of
contamination leaches to groundwater over an assumed 30-year exposure duration. The mass-limit soil
screening model can be used when the area and depth of the source are known or can be estimated
reliably. Both standard and mass-limit soil screening levels should be calculated, and the higher of the
two values should be selected (EPA 1996b).
As previously stated, the soil screening concentrations are to be used in preliminary facility
investigations and assume residential exposure circumstances, EPA did not intend that they would serve
5 4-36
-------
as national cleanup standards. More detailed fate and transport models can be used to back-calculate soil
cleanup levels that are protective of groundwater,. Additional facility-specific data are required for these
models. Multimedia models are available that simulate the hazardous constituent transport through both
the vadose zone and the aquifer (for example, MEPAS, GWSCREEN, ROAM, RES RAD, and
MULT1MED). Other models may only simulate vadose zone transport (for example SOLUTE,
BIOPLUME, and AT123D). In this situation, two models should be used to simulate multimedia
transport, and a mass balance conservation approach should be used to connect the models. Qualified
hydrogeologists should be consulted when selecting a fate and transport model, and the use of the model
should be subject to the approval of regulatory personnel overseeing the corrective action or closure
activities. Further modeling and/or monitoring to assess groundwater discharges to surface water may be
required on a facility-specific basis.
Washington State MTCA regulation (WAC 173-340-740) requires that soil cleanup standards be
protective of groundwater. Historically, MTCA has required that the soil concentration be equal to or
less than 100 times the groundwater cleanup standard unless it could be demonstrated that a higher soil
concentration is protective of groundwater at the facility. At this printing, Ecology was preparing to
propose a number of changes to MTCA regulations, including ways to calculate protection of
groundwater from soil contamination. Ecology's web site should be consulted for proposed and final
rules (http://www.wa.gov/ecology/tcp/cleanup.html). For further information, contact Charles San Juan
of Ecology at (360) 407-7191.
4.5.2.4 Food Chain Exposure Pathways
Migration of contamination into human food chain pathways may require consideration in facility-
specific situations. For example, if soil contamination is present in areas that are likely to be used for
home gardening, this pathway should be considered when setting soil cleanup levels. EPA RCRA
guidance for assessing indirect exposures at hazardous waste combustion facilities provides useful
intermedia hazardous constituent partitioning equations for estimating constituent migration into the
food chain. The draft Guidance for Performing Screening Level Risk Analyses at Combustion Facilities
Burning Hazardous Wastes (EPA 1994i) compiles and streamlines intermedia partition equations
proposed in earlier RCRA combustion guidance documents (EPA 1990, 1993d). The EPA {19941}
document contains RCRA program guidance issued by the Office of Emergency and Remedial Response,
-------
while the EPA (1990, 1993d, and I995g) reports are technical support documents. In addition, the draft
Estimating Exposure to Dioxin-Like Compounds (EPA 1994J) presents similar intermedia partition
equations for estimating dioxin uptake into plants, animals, and fish. EPA regional risk assessors should
be contacted for further information on the availability of updated documents.
While much of the RCRA combustion guidance series addresses the fate of constituent air emissions,
partitioning equations presented for secondary migration pathways can be used to derive soil and water
RJBCs. The EPA (1994ij) guidance documents address the primary food chain pathways, including
constituent migration into garden produce, fish, beef, and milk. Intermedia constituent partitioning
equations for migration pathways, such as soil-to-root-vegetable and water-to-fish, are recommended.
For example, the equations estimate the concentration of a constituent in a secondary media, such as a
garden plant, that results from constituent uptake into the plant from soil. Similarly, estimates can be
made of constituent concentrations in fish resulting from constituent uptake from water or sediment, or
constituent concentrations in beef cattle and dairy cattle milk resulting from constituent uptake from soil
and plants. These equations rely on constituent-specific biotransfer or bioconcentration factors that
represent the ratio of constituent concentrations in the secondary media (for example, garden produce) to
constituent concentrations in the primary media (for example, soil).
These equations can be used in a HHRA to first calculate constituent concentrations in food chain
pathways resulting from air, soil, or water contamination. Human ingestion rates for these food chain
pathways (for example, garden produce or fish ingestion rates) are then estimated. The dose and
resulting risk or hazard are then calculated. Food chain pathways are not typically considered when
calculating risks or RBCs, and their relevance should be determined on a case-by-case basis. Methods
for assessing food chain exposure pathways are typically conservative and may result in RBCs that are
lower than RBCs based on direct contact. They should not be evaluated at every facility, but should be
considered on a case-by-case basis where the food chain pathway is known to be a complete exposure
pathway. At sites where the plant ingestion pathway may reasonably exist, screening level estimates
may be developed using the EPA Region 10 ASARCO plant uptake data for arsenic, cadmium, and lead.
For other contaminants, applicable portions of EPA (1994i) should be used. Special situations where
food chain organisms such as fish or shellfish are consumed at a subsistence level, such as in Alaska or
for Native American populations, should be incorporated into RBCs. EPA Region 10 risk assessors
should be consulted regarding rates of food consumption for such specific situations (for example, for
ASM REVISE2\flNALVMASTE& Dam*)*
-------
Puget Sound or Columbia River fisheries). EPA Region 10 risk assessors should also be contacted to
confirm the selection of appropriate partitioning equations for use in determining either facility risks and
hazards or facility-specific soil and water RBCs. EPA Region 10 has decided to assume that 10 percent
of the inorganic arsenic in seafood is organic (see Attachment 1).
4.6 TOXICITY ASSESSMENT
The toxicity assessment summarizes the toxicologic basis for all chemical-specific toxicity data using
available dose-response information. Toxicity assessments should be conducted as described in
Chapter 7 of Risk Assessment Guidance for Superfimd, Volume /» Human Health Evaluation Manual,
Part A, (EPA 1989c). The following sections present an overview of the types of dose-response
information used to characterize carcinogenic and noncarcinogenic dose responses (Section 4.6.1 %
sources of EPA toxicity values (Section 4.6.2), and methods for assessing chemicals without EPA
toxicity values (Section 4.6.3).
4.6.1 Dose-Response Information
In developing HHRA methods, EPA recognizes fundamental differences between carcinogenic and
noncarcinogenic dose-response variables used to estimate risks. Because of these differences, human
health risk is characterized separately for the carcinogenic and noncarcinogenic effects related to
hazardous constituents. Some analytes have both carcinogenic and noncarcinogenic effects, although in
many cases, EPA has published toxicity criteria for only one of them. It should not be assumed that in
all cases this is the more sensitive type of toxic effect. Information sufficient to characterize and/or
quantify the dose-response relationship may be lacking (for example, in the case of endocrine disrupters,
or when no appropriate animal models exist).
Human epidemiologic data provide the strongest evidence of a positive association between hazardous
constituents and human health effects; however, human health effects data adequate to develop
quantitative dose-response relationships are available for only a few chemicals. As a result, toxicity
information obtained from nonhuman mammalian experiments is often used to predict human
dose-response relationships and to develop chemical-specific toxicity criteria. Animal toxicity data are
typically derived from studies in which animals are exposed to relatively high doses of a chemical. In
S R€PA\RSmurTASKi'Jl£V1SE^nNAHMASTEH Stefrtwc
-------
contrast, the chronic exposures evaluated in the HHRA are for much lower doses. In addition, the
animals are exposed for relatively short periods of lime compared with chronic exposure risk
assessments, which typically assume humans will be exposed for a lifetime. Both of these contribute to
the uncertainties ie HHRA (see Section 4.1).
4,6,1.1 Toxicity Information for Carcinogenic Effects
Currently, the key dose-response variable used to evaluate carcinogenic effects is the cancer potency
factor (CPF), which is derived from carcinogenicity studies (typically conducted at high doses). To
evaluate the probability of developing cancer at the lower doses more typically encountered by the
public, the EPA-recommended linearized, multistage model is applied to the data. This mathematical
model expresses individual excess cancer risk as a function of exposure and is based on the conservative
assumption that even a single, low-dose exposure to a carcinogen may result in cancer. The CPF,
expressed as risk per milligrams per kilogram per day [(mg/kg/day)'1], quantitatively defines the
relationship between dose and response.
In HHRA, chemical-specific CPFs are multiplied by the lifetime average daily dose (LADD) of a
hazardous constituent from a given exposure route to assess the upper-bound cancer risk associated with
that dose. Carcinogenic risk is expressed as the probability of an individual in a population developing
cancer (for example, one in a million or 1E-6). Chemical-specific CPFs can be incorporated into RBC
equations along with dose information to back-calculate RBCs that correspond to selected target risks
(Section 4.7).
EPA assigns weight-of-evidence classifications to potential carcinogens. Under this system, chemicals
are classified as belonging to one of six groups (EPA 1997a):
S RI'»»!K\TaSK*'ACV"JSE3 JtNALWASTER
-------
Group A - chemicals for which sufficient data exist to support,
a causal association between exposure to the agent and the
induction of cancer in humans
Group B » Probable carcinogens:
Group B1 - chemicals for which there is limited
evidence of carcinogenicity from human exposure
studies but sufficient evidence of carcinogenicity from
animal studies
Group B2 - chemicals for which there is inadequate
evidence of carcinogenicity from human exposure
studies but sufficient evidence of carcinogenicity from
animal studies
Group C - chemicals for which there is limited evidence of
carcinogenicity from animal studies; possible carcinogens
Group D - chemicals for which the carcinogenicity database is
inadequate
Group E - chemicals exhibiting no evidence of a carcinogenic
response in humans or animals
For HHRAs, carcinogenic risks are evaluated only for chemicals with weiglt-of-evidence classifications
of A, B t, B2, and C.
EPA issued Proposed Guidelines for Carcinogen Risk Assessment to incorporate advances in scientific
knowledge into the risk assessment process (EPA 1996i). When finalized, this will essentially redefine
EPA's approach to human cancer risk assessments, In particular, classification of a chemical as a
carcinogen will involve all available evidence, including structure-activity relationships and comparative
metabolism and toxicokinetics. The primary effects of this change in procedure will be to decrease the
uncertainty in the toxicity assessment and to allow for risk estimates that more adequately reflect the
scientific understanding of a specific chemical's role in the process leading to cancer. Other proposed
changes include replacing the current A through E classification scheme for the weight-of-evidence by
three classifications, with some subdivisions. These new classes will be exposure route-specific. After
the proposed guidelines are finalized, the EPA intends to reevaluate carcinogens on an individual basis,
probably a few per year; therefore, classification changes will be phased in over time. Project managers
should ensure that facilities utilize and reference the most current information.
f[ 4 «i »i*T4SKrMVt<&]\FI-AL\MUTERWMX(*l-ft|iMlin»»ltninV< Mmmm
-------
4.6.1.2 Toxicity Information for Noocarcinogenic Effects
The key dose-response variable used In quantitative HHRA of noncarcinogenic effects is the chronic
reference dose (RID). The chronic RfD, expressed in units of milligrams per kilogram-day for a specific
chemical, is an estimate of the daily exposure to the human population (including sensitive subgroups)
that is likely to be without an appreciable risk of deleterious effects during a lifetime (EPA 1989c).
Chronic RfDs consider exposures that occur over about 10 percent or more of a lifetime. RfDs are
usually based on the relationship between the dose of a noncarcinogen and the frequency of systemic
toxic effects in experimental animals or humans. It is a specific assumption of this method that there is a
threshold intake rate below which toxic effects do not occur. The threshold of observed effects is
divided by an uncertainty factor or factors to derive an RfD that protects the most sensitive members of
the population. The uncertainty factors are usually multiples of 10, and each factor represents a specific
area of uncertainty inherent in extrapolation from the available data.
Uncertainty factors are applied to data in the following cases (EPA 1989c):
•
To account for variation in the general
population (to protect sensitive
subpopulations)
a
To extrapolate the data from animals to
humans
0
To adjust for using a subchronic study rather
than a chronic study
•
To adjust for using a lowest-observable-
ad verse-e fleet-level instead of a no-
observable-ad verse-e fleet-level in developing
an RfD
•
To account for database deficiencies
A modifying factor ranging from 1 to 10 is also applied to the RfD to address uncertainties in the
scientific studies used to develop RfDs.
S *EPA\* |tsiil\T4SKS«%1SeW1MALWASTER WKftlS|-fttnM|tfmNM2/*ff)7« »)mvuc
4-42
-------
Once an MP for a compound has been verified by EPA, it is used to characterize the likelihood of
noncarcinogenic hazards resulting from long-term chemical exposures at a facility. In HHRA, the RfD is
compared to the average daily dose (ADD) calculated in the exposure assessment to determine whether
chronic effects might occur, The ratio of the ADD and the RID is called the hazard quotient (HQ). If tie
predicted ADD exceeds the RfD, the HQ is greater than 1.0, and there may be concern for potential
noncancer effects (EPA 1989c). HQs for individual constituents can be added to calculate an exposure
pathway or site hazard, referred to as the hazard index (HI). According to Risk Assessment Guidance for
Superfurtd, Volume 1, Human Health Evaluation Manual, Part A, the addition of HQs from all hazardous
constituents is appropriate as a screening-level approach (EPA 1989c). If the resulting HI is greater than
1,0, however, it would be appropriate to calculate new His for hazardous constituents with similar
critical effects and mechanisms of action. Further guidance on segregating COPCs by critical effects and
mechanisms of action is presented in Section 4.7.4. Chemical-specific RfDs can be incorporated into
RBC equations to back-calculate RBCs that corresponded to target hazards (Section 4.7).
EPA has also derived inhalation reference concentrations (RfC), which are estimates of daily exposures
(via inhalation) to the human population including sensitive subgroups that are likely to be without
appreciable risk of deleterious effects. RfCs arc generally reported as a concentration in air (milligrams
per cubic meter). For purposes of using standard RBC equations, however, RflCs can be converted to a
corresponding inhaled dose (milligram per kilogram-day) by dividing by 70 kilograms (an assumed
human body weight), multiplying by 20 cubic meters per day (an assumed human inhalation rate), and,
preferably, adjusting by an appropriate, chemical-specific absorption factor. This conversion, however,
may often be technically incorrect, and the appropriateness of doing this must be evaluated on a case-by-
case basis (EPA 1997a). RfCs can be used in screening risk assessments to determine whether a
constituent may contribute significantly to the HI; however, the appropriateness of RfC conversions and
their use in baseline HHRA's should be verified by a EPA Region 10 risk assessor (EPA 1997a) prior to
their use.
4.6.2 U.S. Environmental Protection Agency Toxicity Factors
Toxicity factors (RfDs, RfCs, and CPFs) are not always readily available, so multiple sources may need
to be consulted. The EPA Region 10 hierarchy of sources for RfDs, RfCs, and CPFs is as follows:
% XE?AJRI«HisMASKrjt£V1SEIifmM.\MASTER
-------
1. Integrated Risk Information System (IRIS) on-line database (EPA I996g). IRIS is the
preferred EPA source for toxicity information. It provides RfDs, RfCs and CPFs thai
have been reviewed and verified by agency-wide work groups. Supporting discussion
and references also appear in each chemical file. IRIS User Support at (513)569-7254
can provide information about how to access IRIS. IRIS is available to EPA Region 10
personnel on its automaxx menu and to the public on EPA's web site at
http://www.epa.gov/iris.
2. Health Effects Assessment Summary Tables (HEAST) (EPA 1997a). HEAST
summarizes all currently available toxicity factors developed by National Center for
Environmental Assessment (NCEA) and a bibliography of Health Effects Assessments
and related documents. These documents contain supporting information for toxicity
values developed by EPA NCEA. The HEAST tables are revised quarterly. Toxicity
factors that appear in IRIS do not appear in HEAST.
3. Provisional values developed by NCEA. Region 10 risk assessment staff should be
contacted to obtain information from NCEA.
4. Agency for Toxic Substances and Disease Registry (ATSDR) minimal risk levels
(MRL). These MRLs are developed using an approach that is generally consistent with
RfD methodology. These may be available for acute, intermediate, or chronic exposure
durations and are potentially useful for situations of short-term exposure, for which
verified RfDs are seldom available. They can be found in ATSDR Toxicity Profile
documents (in the Health Effects Summary section, in the text and/or on the
"thermometer" chart). MRLs can also be found on the Internet at
http://atsdrl .atsdr.cdc.gov:8080/mrls.html. Concurrence with use of MRLs for a specific
situation should be sought from Region 10 risk assessment staff.
5. Region 10 risk assessors may have access to additional toxicity numbers from other
sources that may be appropriate for a given circumstance.
In addition to the RfDs, RfCs, and CPFs, the EPA weight-of-evidence classifications and the types of
cancers observed in animal testing should be presented for all carcinogenic hazardous constituents, while
tiie confidence levels, critical effect and target organs, and uncertainty and modifying factors associated
with the available RfDs should also be presented. The identification of critical effects and target organs
becomes important during risk characterization when segregating COPCs for HI calculations.
Since carcinogenic chemicals may also cause noncarcinogenic health effects, RfDs (when available)
should be compiled for carcinogenic chemicals and used to evaluate potential noncarcinogenic effects for
these chemicals. It should not be assumed that noncarcinogenic effects are negligible or even less
important than cancer risks just because RfD or RfCs are not available.
s Rf.rA R |t»u« TASM«EV1SE}*1NAL;MASTER WPOM5I R tmHMTumiJ/lStMnttC
4-44
-------
4,43 Chemicals or Exposure Pathways With No U.S. Environmental Protection Agency
Toxicity Values
This section identifies key constituents that do not have toxicity values in the EPA IRIS and HEAST
databases. Recommendations for what toxicity values may be used to evaluate these constituents are
presented. Methods for evaluating exposure pathways with limited or no EPA toxicity values are also
presented. If it is determined that provisional or other alternative values may be used, their uncertainties
must be discussed in the risk assessment document (specifically in the risk characterization section) and
should be considered In risk management decisions.
4.6 J.I Polyauclear Aromatic Hydrocarbons
Poiynuclear aromatic hydrocarbons (PAH) are a very extensive group of organic compounds that contain
at least two benzene rings. Sources of PAHs include petroleum and coal tar products. The major source
of PAH releases to the environment is combustion of these products (EPA 1982). PAHs may be found at
hazardous waste sites that used or generated these products, such as coal gasification plants, coal tar
generators, power plants, wood treaters that used creosote, and coke manufacturers. Some PAHs are
carcinogenic and some are not, Benzo(a)pyrene is currently the only carcinogenic PAH for which a CPF
has been verified by EPA. It is recommended that toxicity equivalency factors (TEF) based on the
relative potency of each PAH compound to that of benzo(a)pyrene be used to evaluate the toxicity of the
remaining carcinogenic PAHs on the target compound list (see following list). The TEFs presented here
were recommended by EPA's NCEA and are from the Provisional Guidance for Quantitative Risk
Assessment of Polycyclic Aromatic Hydrocarbons (EPA 1993e). Concentrations of specific carcinogenic
PAHs should be multiplied by their respective TEFs to calculate their concentrations relative to
benzo(a)pyrene potency. This benz(a)pyrene equivalent concentration should then be used for risk
characterization. These TEFs are not in IRIS, and therefore are not necessarily accepted by all states and
may be revisited by NCEA in the future.
-------
POLYNUCLEAR AROMATIC HYDROCARBON
TOXICITY EQUIVALENCY FACTORS
Compound
TEF
Benzo(a)pyret»e
1.0
Benzo(a)anthracene
0.1
Benzo(b)fluoranthene
0.1
Benzo(k)fluoranthene
0.01
Chrysene
0.001
Dibenzo(ath)anthracene
1.0
lndeno(l,2,3-cd)pyrene
0.1
4.6.3.2 Chlorinated Dioxins and Furans
Dioxins and furans are created during combustion processes, such as during incineration of wastes or
during the burning of fossil fuels. Dioxins and furans may also be created during the manufacture of
chlorine and chlorinated products (for example, chlorinated phenols, chlorinated benzenes, PCBs,
phenoxy herbicides, and other compounds), and during paper manufacturing involving chlorine
bleaching (EPA 1994j). EPA has verified a CPF for 2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD); it is
the only chlorinated dioxin or furan with a verified CPF. TEFs based on the relative potency of each
dioxin and furan congener to that of TCDD have been developed by EPA and are presented in Interim
Procedures for Estimating Risks Associated with Exposures to Mixtures of Chlorinated
Dibenzo-p-Dioxins and Dibenzofurans (CDDs and CDFs) and 1989 Update (EPA 1989f). The TEFs are
listed on the following page. Concentrations of specific dioxin and furan congeners should be multiplied
by their respective TEFs to calculate their concentration relative to 2,3,7,8-TCDD potency (example
follows list of TEFs). This 2,3,7,8-TCDD equivalent concentration should then be used for risk
characterization. At the Dioxin *97 conference held in Indianapolis in August 1997, the World Health
Organization presented an abstract in which a new TEF scheme for dioxins, furans, and dioxin-like PCBs
was delineated for humans/mammals and separately for fish and for birds. This abstract is included as
Attachment J. EPA may at some point decide to adopt some or all of these TEFs once the World Health
* R Jt'TASKK.HEVTSEJ'ilNM.vMASTER
-------
Organization publishes its complete report. A regional risk assessor should be consulted to determine
EPA's TEF policies for PCBs and dioxins/furans when they are COPCs at a given site,
EPA published a draft comprehensive reassessment of dioxin toxicity in 1994 and is currently finalizing
that reassessment. Several chapters of the draft reassessment are available on the Internet at
http://www.epa.gov/docs/exposure/. Revised chapters are placed on that site as they become available.
It is expected that the reassessment will be finalized in early 1998. While dioxin and furan
noncarcinogenic effects are not insignificant, EPA has not quantified such effects.
DIBENZO-P-DIOXIN AND DIBENZOFURAN
TOXICITY EQUIVALENCY FACTORS
Compound
TEF
Dioxins
2,3,7,8-T etrachlorodibenzo-p-d iox in (TCDD)
1.0
Pentachlorinated dibenzo-p-dioxin (2,3,7,8
0.5
chlorines)
Hexachkmnated dibenzo-p-dioxin (2,3,7,8
0.1
chlorines)
Heptachlorinated dibenzo-p-dioxin (2,3.7,1
0.01
chlorines)
Octachlorinated dibenzo-p-dioxin
0.001
Other chlorinated dibenzo-p-dioxins
0
Furans
2,3,7,8-Tetrachlorodibenzofuran
0.1
1,2,3,7.8-Pentachiorodibenzofuran
0.5
2,3,4,7,8-Pentachlorodibenzofuran
0.05
Hexachlorinated dibenzofuran (2,3,7.8
0,1
chlorines)
Heptachlorinated dibenzofuran (2,3,7,8
0.01
chlorines)
Octachlorinated dibenzofuran
0.001
Other chlorinated dibenzofurans
0
S«PA«BWW'.TASK««im«:HnN*l,*M*SIW
-------
EXAMPLE OF CALCULATING 2A?,8-TCDD EQUIVALENT
CONCENTRATION USING TOXICITY EQUIVALENCY FACTORS
Analyte
Soil
Concentration
(mgflkg)
TIF
2A7.8-TCD1
Equivalent
Conceatratio
(mg/kg)
Octach formated dibenzofuran
2,3,7,8-Tetrachlorodibenzofiiran
1.5E+2
2.0E+1
0.001
0.1
1.5E-1
2.0E+0
Total 2,3,7,8-TCDD equivalent
concentration
2.1E+0
4,6.3 J Polycblorinated Bipbeayls
The following discussion of PCBs is based on PCB carcinogenicity Information presented in PCBs:
Cancer Dose-Response Assessment and Application to Environmental Mixtures (EPA 1996j). PCBs are
mixtures of synthetic organic chemicals. The primary PCB molecule consists of two six-carbon rings
with one chemical bond joining a carbon from each ring. Commercial mixtures of PCBs have from one
to ten chlorines attached to the other carbons on the two rings. There are 209 possible arrangements of
chlorines on the two rings; these molecular arrangements are referred to as congeners. PCB congeners
with the same number of chlorines are called isomers. Commercial PCB mixtures manufactured in the
United States cany the trademark "Aroclor" (for example, Aroclor 1016, 1242, 1248, 1254, and 1260).
Each of these Arociors is made up of mixtures of PCB congeners, ranging from congeners will four
chlorines or less (for example, Aroclor 1016) to congeners with five to nine chlorines (for example,
Aroclor 1260).
PCBs are classified by EPA as probable human carcinogens, but PCB mixtures differentially contribute
to excess cancer risk. Certain PCB mixtures (Aroclor 1254) and congeners (see following list) have
demonstrated tumor-promoting activity. Congener information is useful when evaluating the potential
for a PCB mixture to cause cancer, and as discussed below, is used to select a CPF appropriate to the
mixture and to the exposure pathway(s).
1 A£PAAHn(R'TASKMl£V|5E2VRNAL\M4STEIt WHftm«ltKH«HMMyiJ'YTff SUaum
4-48
-------
When the two six-carbon rings that make up a PCB molecule are aligned on the same plain, the molecule
is referred to as being "coplanar." Certain coplanar PCB congeners have toxicity mechanisms that are
similar to that of dioxin (see list of these congeners on page 4-52). As discussed below, concentrations
of these PCB congeners can be converted to equivalent concentrations ofTCDD, and TCDD toxic
equivalent risks can be calculated.
' - — ¦
POLYCHLORINATED BiPHENYL MIXTURES AND
CONGENERS THAT TESTED POSITIVE FOE
TUMOR-PROMOTING ACTIVITY
Mixture
Congener
Arock* 1254
2,4,2',4'-Tetrachk>robiphenyl (TECB)
Kanechior 400
2,4,2', 5-TECB
Kanechlor 500
2,5f2\5'-TECB
Clophen A50
3,4.3',4'-TECB
2»3,4»3",4"-Pcntacllorobiphenyl (PECB)
2,4,5,3',4-PECB
3,4tS,3',4*-PECB
2,4,5r2',4,t5,-Hexachk)robiphenyI
Source: EPA 1996]
PCB mixtures in soil, sediment, air, water, aid biota media may differ from the parent commercial
mixture initially released to the environment, based on partitioning, bioaccumulation, and transformation
processes. Anaerobic and aerobic biodegradation may result in the removal of chlorines and the
breaking of carbon rings; however, PCB congeners are persistent, particularly those with a high chlorine
content. These more chlorinated congeners can absorb onto soil and sediment particles and become
concentrated in fish and animal fat. The make-up of a bioaccumulated PCB mixture can therefore vary
from that of its parent commercial mixture and may contain a higher percentage of more persistent,
highly chlorinated congeners,
Bioaccumulated mixtures appear to be more toxic than commercial mixtures (Aulerich et al. 1986). This
toxicity is not necessarily based on chlorine content only; both tie number and position of chlorines is
important. Partitioning of more toxic PCB congeners in environmental media may result in increased
toxicity to exposed humans compared to the toxicity of the parent commercial mixture. For example, the
i 1EM «W^T»iMW>«»W«»yiWWIW>mK1»flM«> HUB WtlwmmffW WHimi 4-49
-------
cancer potencies of PCB mixtures in the food chain, soil, or sediment are predicted to be greater than the
potency of more water-soluble PCBs.
Historically, the CPF for PCB was based on commercial mixture toxicity, EPA has developed new
procedures to evaluate environmental mixtures of PCBs: PCBs: Cancer Dose-Response Assessment and
Application to Environmental Mixtures (EPA 1996]). This document presents updated toxicity
information that can be used to evaluate the carcinogenic risks from environmental PCB exposure.
These procedures and new PCB slope factors were incorporated into EPA's IRIS in October of 1996.
This update approach presents and describes the following:
I
A range of upper-bound and central cancer potency factors for
PCB mixtures, depending on:
The effect of environmental processes on the
mixture
The route and timing of human exposures to
the mixture
Available information on the site-specific
types of PCB congeners
The range of potency observed for commercial mixtures must be used to represent the potency of
environmental mixtures since no toxicity data on environmental mixtures are available. The range
reflects experimental uncertainty and variability of commercial mixtures, but not human heterogeneity or
differences between commercial and environmental mixtures. As noted, environmental processes alter
mixtures through partitioning, transformation, and bioaccumulation, which may decrease or increase
toxicity. The overall effect can be considerable, and the potency range observed from commercial
mixtures may underestimate the true range for environmental mixtures. Limiting the potency of
environmental mixtures to the range observed for commercial mixtures reflected a choice to base
potency estimates on experimental results, rather than apply safety factors to compensate for lack of
information. EPA addressed this issue by developing CPFs that consider the make-up and fate of
environmental mixtures. EPA (1996j) presents a range of PCB central and upper-bound CPFs with three
reference points. Each reference point or CPF has criteria that should be met for that CPF to be used.
S R£PA'Mi»!*.TA$KrftEWS£2i;lNAL
-------
Criteria include the human exposure pathway evaluated and specific information on the congener
composition of the mixture that must be obtained through environmental sampling.
A tiered approach is used that allows different types of information in estimating the potency of
environmental mixtures of PCBs. Total PCBs or congener or isomer analyses are recommended. The
first (default) tier is invoked when congener information is limited. Only the high-risk CPF should
normally be used in the first tier since it is not possible to demonstrate the absence of persistent, dioxin-
like, or tumor-promoting congeners without such analysis. The lowest-risk CPFs cannot be used without
specific information on the congener composition of the mixture.
The second tier is invoked when congener or isomer information is available from sampling and
analysis; it can be used to farther refine a potency estimate that was chosen to reflect an exposure
pathway. The lowest upper-bound CPF (0.07 [mg/kg/day]'1) may be used if congener or isomer analyses
verify that congeners with more than four chlorines comprise less than 0.5 percent of total PCBs. The
higher CPF (2.0 [mg/kg/day]*1) should be used if dioxin-Iike, tumor-promoting, and persistent congeners
are present. When dioxin-Iike coplanar congener concentrations are available, the use of CPFs for PCBs
may be supplemented by the calculation of TCDD toxic equivalent risks. Under this method, PCB
congeners that are not dioxin-Iike are evaluated using the appropriate PCB CPF, while PCB congeners
that are dioxin-Iike are evaluated using the TCDD CPF. TEFs for the dioxin-Iike PCB congeners (see the
following list) would be used to estimate TCDD toxic equivalent concentrations by multiplying the
concentrations of individual dioxin-Iike PCB congeners by the TEFs. TCDD toxic equivalent
concentrations from all dioxin-Iike PCB congeners are then added together. This sum is used to
calculate a lifetime average daily dose of dioxin, which Is then multiplied by the dioxin CPF to estimate
dioxin-Iike PCB risk. Section 4.6,3.2 describes possible changes in the TEFs for dioxin-Iike PCBs based
on World Health Organization studies.
-------
DIOXIN-L1KE POLYCHLORINATED BIPHENYLS
AND TOXICITY EQUIVALENCY FACTORS
No*-Ortho
Congener*
TEF
Mooo-Ortbo
Congeners
TEF
D»-Ortfco
Coagcpcrs
TEF
3,4.3 \4'-Tctrschkwobipbcnyl {TECS)
0.0005
23.43\4'-PECB
0.0001
23.4,5,2'3',4-HPCB
0.000i
3,4,53',4'-Pcnt»cMorob«pbenyi (PECB)
0.1
2J,4J.4'-PECB
00005
23,4.5,2'.4 \5'-HPCB
000001
3.4.S,3'.4,.y-Hauchk>robipheayl
0.01
2,4,53V~'-PECB
0.0001
(HXCB)
3,4^X4*-rcCB
0.000i
2,3,4,5 J',4-HXCB
0 0005
23,43\4\5'-HXCB
0.0005
2,4,5X4',5'-IKCB
0.00001
23,4.5,3*. 4*,S-
0.0001
Hcpttcfciorobiphcayi
CHPCB)
Source: EM 1995j
Table 4-11 summarizes the range of CPFs for PCBs, indicating how exposure pathway and
congener/isomer information is used to select CPFs. EPA (!996j) presents three examples that show
how additional information regarding the types ofPCB congeners present affects the CPFs used to
evaluate risk and the subsequent risk estimates. Example 3 shows how PCB congener information,
specifically regarding dioxin-like congeners, can be incorporated into the risk estimates. Essentially,
lifetime average daily doses and risk estimates would be calculated for the dioxin-like and nondioxin-
like portions of the mixture,
PCBs: Cancer Dose-Response Assessment and Application to Environmental Mixtures (EPA 1996])
summarizes uncertainties associated with the proposed PCB CPFs. These include uncertainties inherent
in the experimental information used to derive CPFs and uncertainties associated with applying the
S AEPA M t'»tl«-T*SKrREVISE2\FTNALMASTEll WTO\iSi4U«J*imwn8im«?« 55a«srtuc
-------
TABLE 4-11
MANSES OF HUMAN POTENCY AND CANCER POTENCY FACTORS
FOE ENVIRONMENTAL MIXTURES OF PCBS
REGION 10 RCRA RISK-BASED CLEANUP LEVEL GUIDELINES
Risk and
Persistence
Central
Cancer potency
factor
(mg/kg/day)*1
Upper-bound
Cancer potency
factor
(ng/kg/dif)"1
Site-Specific
Criteria for Use
High*
I
2
Food chain exposure, ingestion of
contaminated sediment or soil; inhalation of
dusts or aerosols; presence of dioxin-like,
tumor-promoting, or persistent congeners;
early life exposure
Low
0.3
0.4
Ingestion of water soluble congeners, vapor
inhalation, dermal contact (if no absorption
factor is applied)
Lowest
0.04
0.07
Congener or isomer analyses verify that
congeners with more than four chlorines
comprise less than 0.5 percent of total PCBs.
Soiree: U.S. Environmental Protection Agency I996j
Note;
a In the absence of congener-specific analytical information, the slope factor of two should normally be used
since it is not possible to demonstrate the absence of dioxin-like, tumor-promoting, or persistent congeners
without such analysis.
I REPAJltirrA$MJlEVlSE2if1NAI.\MASTER i5%rnsm
-------
experimental information to human environmental exposures. Examples of uncertainties associated with
human exposures include the following:
• It is crucial to recognize that commercial PCBs tested
in laboratory animals were not subject lo prior
selective retention of persistent congeners through the
food chain. For exposure through the food chain, risks
can be higher than those estimated by EPA (I996J).
• PCBs persist in the body, providing a continuing
source of internal exposure after external exposure
stops. There may be greater-than-proportional effects
from less-than-iifetime exposure, especially for
persistent mixtures and for early-life exposures.
When planning for the collection of PCB samples, lie ability of the laboratory to analyze for PCB
congeners or isomers should be verified, and such analyses requested if relevant and appropriate based
on facility conditions. For example, congener specification will be critical when it is suspected lower
persistence and lower risk PCB congeners are present but verification is required. Likewise,
confirmation that tumor-promoting ordioxin-like PCB congeners are present will require specific PCB
analyses. Draft EPA method 8082 or similar analyses can be used to detect specific PCB congeners.
4.6.3.4 Manganese
A chronic, oral RfD is available on IRIS. It was revised in November 1995. The RfD reflects the total
dietary intake of manganese. Manganese is a naturally occurring element and is present in the normal
diet to a certain extent. When assessing the exposure to manganese from sources other than food, the
narrative accompanying the IRIS value advises the use of a modifying factor of 3. This is especially
important for the protection of infants who may be adversely affected, such as if they are fed formula
made up with water which contains elevated levels of manganese. EPA Region I, in a technical bulletin
called Risk Updates, published a description of how the modifying factor should be applied in risk
assessments (See Attachment K). Region I Risk Updates are available on the Internet at
http://www.epa.gov/regionO 1/remed/riskupdates.html The use of the modifying factor should always be
used when determining acceptable levels of manganese in groundwater which is being used or may be
-- KTP* R«'¦>!* T*SM REVtSi:>5|N«MvMASTEft WPE*1 >».«
-------
used in the future as drinking water. A site-specific determination should be made regarding the use of
the modifying factor for soil If it is a reasonable assumption that infants will not be exposed to the soil,
the use of the modifying factor would not be necessary.
4.63.5 Total Petroledm Hydrocarbons
EPA verified toxicity values are not available for total petroleum hydrocarbons (TPH) or for most of the
hundreds of individual chemicals that comprise petroleum products. Cleanup frameworks for TPH that
have been developed and adopted by individual Region 10 states should generally be followed for
petroleum releases in those states. The assessment of risks posed by TPH releases should always include
at a minimum the measurement of benzene and the carcinogenic PAHs (see box in Section 4.6.3.1 for
list). The leaching potential to ground and surface waters and vapor releases to ambient and enclosed
breathing areas such as basements through structural breaches should be considered where applicable.
The federal Water Pollution Prevention Act (also known as the Clean Water Act) prevents the discharge
of oil to navigable waters of the United States such that it causes a film or sheen or causes a sludge or
emulsion beneath the surface of the water (Clean Water Act 310(a)(1); 311(b)(3); and 40 CPK 110.3).
Each state in Region 10 is addressing TPH cleanups differently. For RCRA facilities where TPH
releases are an issue, An EPA Region 10 underground storage tank technical expert in the groundwater
protection unit should be consulted [(206) 553-1587] for the status of each state's TPH cleanup program.
EPA's Office of Underground Storage Tanks is also developing guidance for TPH cleanups on Native
American lands; however, the state in which the land is located may have a more sophisticated or more
pertinent TPH cleanup framework. The Native American stakeholders should be consulted in making
decisions about which cleanup framework to follow for petroleum releases on their lands.
4.63.6 Dermal Toxicity Factors
No RfDs or CPFs are available for the dermal route of exposure. In some cases, however, risks and
hazards associated with dermal exposure can be evaluated using an oral toxicity factor. This
route-to-route extrapolation assumes that the toxicity of a hazardous constituent is the same whether it is
4-55
-------
experimental information to human environmental exposures. Examples of uncertainties associated with
human exposures include the following:
* « crucial to recognize that commercial PCBs tested
in laboratory animals were not subject to prior
selective retention of persistent congeners through the
food chain. For exposure through the food chain, risks
can be higher than those estimated by EPA (1996j).
* PCBs persist in the body, providing a continuing
source of internal exposure after external exposure
stops. There may be greater-than-proportionai effects
from less-than-lifetime exposure, especially for
persistent mixtures and for early-life exposures.
When Plannin8 for the collection of PCB samples, the ability of the laboratory to analyze for PCB
congeners or isomers should be verified, and such analyses requested if relevant and appropriate based
on faci,lty conditi°ns- For example, congener specification will be critical when it is suspected lower
persistence and lower risk PCB congeners are present but verification is required. Likewise,
confirmation that tumor-promoting or dioxin-like PCB congeners are present will require specific PCB
analyses, Draft EPA method 8082 or similar analyses can be used to detect specific PCB congeners.
4.6.3.4 Manganese
A chronic, oral RfD is available on IRIS, it was revised in November 1995. The RfD reflects the total
dietary intake of manganese. Manganese is a naturally occurring element and is present in the normal
diet to a certain extent. When assessing the exposure to manganese from sources other than food, the
narrative accompanying the IRIS value advises the use of a modifying factor of 3. This is especially
important for the protection of infants who may be adversely affected, such as if they are fed formula
made up uith water which contains elevated levels of manganese. EPA Region I, in a technical bulletin
called Risk Updates, published a description of how the modifying factor should be applied in risk
assessments (See Attachment K). Region 1 Risk Updates are available on the Internet at
finp - uww.epa.gov regionO! remedriskupdates.html. The use of the modifying factor should always be
llscd uhen determining acceptable levels of manganese in groundwater which is being used or may be
~ * » k. J ? nv.. \u HP 'APj> :« V . 4-54
-------
used in (he future as drinking water. A site-specific determination should be made regarding the use of
the modifying factor for soil. If it is a reasonable assumption that infants will not be exposed to the soil,,
the use of the modifying factor would not be necessary.
. Use MT-'ft
4.6.3,5 Total Petroleum Hydrocarbons ^
EpX^v^rified toxicity values are not available for total petroleum hydrocarbonsj^TPH) or for most of the
hundredsoNqdividual chemicals that comprise petroleum products. Clem#frameworks for TPH that
have been developttjand adopted by individual Region 10 stales should generally be followed for
petroleum releases intkoK states. The assessment of risks pose^tfy TPH releases should always include
at a minimum the measurem^h^of benzene and the carcinog?fiic PAHs (see box in Section 4.6.3.1 for
list). The leaching potential to groitndand surface wat^rtand vapor releases to ambient and enclosed
breathing areas such as basements through^tnicturaH)reaches should be considered where applicable.
The federal Water Pollution Prevention Act (ate known as the Ctean Water Act) prevents the discharge
of oil to navigable waters of the United States suclW it causes a film or sheen or causes a sludge or
emulsion beneath the surface of the water (Clean WaterXrti310(a)(l)j 311(b)(3); and 40 CFR 110.3).
Each state in Region 10 is addrfcssing TPH cleanups differently. Fb^RCRA facilities where TPH
releases are an issue. An EPA Region 10 underground storage tank technical expert in the groundwater
protection unit should^ consulted [(206) 553-1587] for the status of eachW's TPH cleanup program.
EPA's Office ofjtfnderground Storage Tanks is also developing guidance for TRH cleanups on Native
American latfds; however, the state in which the land is located may have a more sophisticated or more
pertinmt/flPH cleanup framework. The Native American stakeholders should be consisted in making
deepens about which cleanup framework to follow for petroleum releases on their lands.
fx, Z.r
4.6.3.6 Dermal Toxicity Factors
X-c V
Mo RfDs or CPFs are available for the dermal route of exposure. In some cases, however, risks and
hazards associated with dermal exposure can be evaluated using an oral toxicity factor. This
route-to-route extrapolation assumes that the toxicity of a hazardous constituent is the same whether it is
X
4-55
-------
oral or dermal route. In certain instances, it may not be appropriate to use oral toxicity factors to evaluate the
dermal pathway (for example, when a toxicant is known to exert a specific point-of-contact [skin] effect). A
risk assessor can assist with the evaluation of the appropriateness of a route-to-route extrapolation.
Exposures via the dermal route are calculated and expressed as absorbed doses, while most oral toxicity
factors are expressed as administered doses. An administered dose is the dose that is presented to a persons
"exchange surfaces" or points of contact with the external world, including the mouth, skin, and nose. An
absorbed dose in the portion of the administered dose that actually enters the general circulation of the body.
For example, because the skin is an effective barrier to many chemicals, only a portion of the dose
administered on the skin's surface will be absorbed through the skin into the blood stream. When evaluating
dermal exposure to contaminants in water or soil, it may be necessary to adjust an oral toxicity value based on
an administered dose to one based on an absorbed dose using a chemical's oral absorption efficiency. This
section discusses the method for making this adjustment. If the oral toxicity factor is used unadjusted, the
resulting risk or hazard estimates will be less conservative because adjusted values are more protective than
unadjusted oral values (see examples below).
Information concerning absorption efficiencies may be found in various chemical-specific documents
including ATSDR toxicological profiles and Health Effects Assessments. Another source of absorption
efficiencies is a list compiled by the Health Sciences Research Division of the Oak Ridge National
Laboratory. This list is included as Attachment L. A Region 10 risk assessor should be consulted before
the adjustment of oral toxicity factors is considered.
1 The oral absorption efficiencies listed in the box to the right are recommended in Supplemental Guidance to
RAGs: Region 4 Bulletins, Human Health Assessment
and may be used as interim default values in the absence
of chemical-specific values (EPA 1995h). An exception
to the default value for inorganics should be made for
cadmium. IRIS (EPA 1996g) states that the RfD for
cadmium (based on drinking water) assumes an oral
absorption efficiency of 5 percent. A risk assessor should be consulted for a value for arsenic.
• VOCs - 80 percent
• SVOC - 50 percent
• Inorganic Chemicals - 20 percent
-------
As shown in the box to right, an oral CPF,
expressed as an administered dose, is convert:©!
to an absorbed dose by dividing the oral CPF by
the oral absorption efficiency.
Likewise, an oral RfD, expressed as an
administered dose, is converted to an absorbed
dose by multiplying the oral RfD by the oral
absorption efficiency (either determined from
literature or assumed) in the species on which
the oral RfD is based.
4.6.3.7 Inhalation Toxicity Factors
Inhalation toxicity factors (RFC) are available for only a select number of hazardous constituents.
Provisional inhalation toxicity values may be available from various sources such as EPA's NCEA or
chemical-specific ATSDR toxicity profiles. EPA Region 10 risk assessors may be consulted as to the
availability of inhalation toxicity factors.
For cases where EPA-derived toxicity factors are not available for the inhalation route of exposure but arc
available for the oral route, a Region 10 toxicologist should be contacted for guidance on route-to-route
extrapolation. If an oral toxicity factor is used to evaluate inhalation risks or hazards, the uncertainty
associated should be discussed. In certain instances, it may not be appropriate to use oral toxicity factors to
evaluate the inhalation pathway (for example, when a toxicant is known to exert a specific point-of-contact
[lung] effect). If it is recommended that route-to-route extrapolation not be considered and if no provisional
toxicity values is available, the hazardous constituent in question should be discussed qualitatively, and its
absence should be discussed in the uncertainty section.
4.7 CALCULATION OF RISK-BASED CONCENTRATIONS
EPA intends that contaminated RCRA sites be remediated in a manner consistent with available, protective,
risk-based media cleanup standards (FR 19449. May 1,1996). When appropriate promulgated standards and
Example: An oral CPF, unadjusted for absorption,
equals 1.6(mgflcg/day)"1. Information (or
an assumption) indicates a 20 percent oral
absorption efficiency; therefore, the
adjusted CPF would be 1.6 (mg/kg/day)1 /
0.20 « 8 (iDglcg/day)'1.
iMMple:
An oral RfD, unadjusted for absorption,
equals 10 mg/kg/day. Information (or an
assumption) indicates a 20 percent oral
absorption efficiency, therefore, the
adjusted RfD would be 10 mg/kg/day x
0.20 •* 2 mg/kg/day.
4-57
-------
promulgated standards and criteria do not exist, protective media cleanup standards can be developed
using a facility-specific HHRA approach. The primary goal of HHRA is to calculate the chemical dose
that a person may receive when exposed to contaminated media and to determine the type and magnitude
of toxic effects that are known lo occur at that dose level. The exposure and toxicity assessment methods
described in Sections 4,5 and 4.6 are the major risk assessment tools used to accomplish this goal.
Standardized risk characterization equations that incorporate exposure aid toxicity information can be
adjusted to calculate RBCs that correspond to selected target cancer risks or HQs.
The following subsections describe how target risks and HQs are selected (Section 4.7.1), what HHRA
equations are used to calculate RBCs {Section 4,7.2 and Attachment E), how RBCs are calculated
(including an example) (Section 4,7.3), how to calculate RBCs for multiple hazardous constituents
(Section 4.7.4), how fate and transport models are used to determine RBCs for hazardous constituents in
one media that may migrate into another media (Section 4.7.5), and how RBCs can be estimated for lead
(Section 4.7.6).
4.7.1 Selection of Target Risks and Hazard Quotients
EPA's RCRA program risk reduction goal is to reduce the threat from facility-related carcinogenic
hazardous constituents such that, for any medium, the excess risk of cancer to an individual exposed over
a lifetime generally falls within a range from 1E-6 to JE-4 (that is, 1 in one million to 1 in one hundred
thousand) (FR 19449, May 1, 1996). Available risk-based media cleanup standards are thus considered
protective if they achieve a level of risk that falls within the 1E-6 to 1E-4 cancer risk range. Program
implementors and facility owners/operators should generally use 1E-6 as a point of departure when
initially developing target site-specific media cleanup levels. For noucarcinogens, the HI should
generally not exceed 1.0 (FR 19449, May 1, 1996).
Washington State MTCA regulations prescribe target risks of IE-6 for Method B cleanup levels and
1E-5 for Method C or industrial cleanup levels for carcinogens, and HQ of 1.0 for all scenarios.
Cleanups and closures at EPA-lead RCRA sites in Washington should typically be at least as
conservative as MTCA requires to avoid the need for further action later. The Oregon State approved
RCRA corrective action authorization package states that it will rely upon EPA risk assessment guidance
documents to determine appropriate cleanup levels at RCRA facilities, including the use of a IE-6 target
ASKI'JIE VJSEWIN*L\M ASTtK WKfcm-ftlMiiaiffiaMl/ll/W/V SSi
4-58
-------
cancer risk and a 1J target HI. Idaho Stale relies on EPA's RFI guidance (EPA 1989a) for setting
cleanup levels at RCRA facilities and bases residential land use cleanup levels on target cancer risks and
HQs of 1E-6 and 1.0, respectively, Idaho las also indicated that industrial land use cleanup levels may
be based on a upper-end target cancer risk of 1E-4 [Tetra Tech EM Inc. (Telra Tech) 1996bj.
As previously described in the introduction to Chapter 4, data collection and evaluation, exposure
assessment, toxicity assessment, and risk characterization steps are used during HHRA to determine
facility risks and hazards.
During an exposure assessment, a daily dose is calculated by estimating the
amounts of a hazardous constituent that a person may intake from soil,
water, or air during typical residential, industrial, or other land use activities.
In a toxicity assessment, data on toxic effects that occur at known dose
levels are compiled, and toxicity values are derived from this information.
CPFs are used to estimate an upper-bound probability of a person
developing cancer if exposed to a specific dose of a constituent over a
lifetime. Toxicity values for noncancer health effects (RfDs) predict the
dose below which no toxic effects are expected to occur. Outputs from the
exposure and toxicity assessments are combined to complete the risk
characterization step.
For a carcinogen, a risk estimate represents an estimate of the incremental probability that an individual
could develop cancer over a lifetime as a result of site-related exposure to that carcinogen (EPA 1989c).
It does not include risks associated with exposure to the same or other carcinogens that are not facility-
related (that is, background concentrations or occupational exposures).
These excessive lifetime cancer risks are calculated using equation 4-1:
Lifetime cancer risk = LADD x CPF _ (4-1)
4.7.2
Risk-Based Concentration Calculation Equations
where
LADD -
CPF =
Lifetime average daily dose (mg/kg/day)
Cancer Potency Factor (mg/kg/day)'1
-------
The LADD expresses the hazardous constituent dose for the facility based on exposure information.
This dose is compared to the constituent's toxicity per unit dose to calculate the risk from exposure to the
constituent attributed to releases from the facility.
For noncarcinogens, the potential for individuals to develop noncarcinogenic effects is evaluated by
comparing an assumed intake developed over a specific exposure period to an RfD developed over a
similar exposure period. This comparison takes the form of a ratio called an HQ and is expressed in
equation 4-2:
where
HQ
HQ
ADD
RfD
ADD/RfD
Hazard quotient (unitless)
Average daily dose (mg/kg/day)
Reference dose (mg/kg/day)
(4-2)
The ADD and RfD are calculated over the same exposure period. The ADD expresses the constituent
dose for the facility based on exposure information. The facility dose is compared to the RfD to estimate
the likelihood of health effects.
Risk can be expressed as a lifetime excess cancer risk or as a noncarcinogenic hazard.
RBCs are calculated by the same methods used to calculate risks. The risk equations are simply reversed
to solve for a daily constituent dose that is equivalent to a selected target risk or hazard level. The
concentration of a hazardous constituent released from a facility that would cause such a dose is
determined based on the site-specific exposure conditions assumed (for example, see Table 4-2) or
measured. This concentration may serve as the basis of a cleanup level unless multiple hazardous
constituents are present in concentrations above screening levels (see Section 4.7.4 for adjustments based
on multiple constituents).
Examples of equations that may be used to calculate RBCs are presented in Exhibit 4-1. The equations
are reduced versions of the PRG equations recommended by EPA Region 9 (1996c). The Region 9
equations are derived from standard EPA equations used to calculate risks or hazards. The standard
». aim
-------
equations have been rearranged to solve for the soil or water constituent concentration that corresponds
to target risk or HQ levels. The fall EPA Region 9 equations and exposure assumptions are presented in
Attachment E.
The ultimate selection of target risks and hazards are risk management decisions. EPA guidance for risk
assessment (1989c, 1995a) anticipates that the risk assessment will proceed independent of and before
risk management decisions. In Washington State, where MTCA is followed, many risk management
decisions are already made by virtue of the exposure assumptions and levels of protectiveness required
by Methods B and C.
Summary, reduced RBC equations are presented in Exhibit 4-1 for soil and water media. The reduced
equations consider the primary direct exposure pathways for these media, including ingestion, inhalation,
and dermal contact for soil and water. The exposure assumptions for each exposure pathway (presented
in Table 4-2) were incorporated into the original Region 9 risk equation to arrive at the reduced
equations. When available and determined to be appropriate to the facility assessment, chemical-specific
parameters, including ABS values (Section 4.5.1), VF„ PEFs (Section 4.5.2.1), and toxicity values
(Section 4.6), must be entered into the reduced equations to calculate RBCs. The exposure pathway for
dermal contact with constituents in water is not included in the Region 9 (1996c) PRG equations. This
exposure pathway was incorporated in the Exhibit 4-1 groundwater equations. Depending on site-
specific conditions, it may be determined that one or more exposure pathways are incomplete (that is,
there are no actual or potential receptors to the constituents in questions, via a specific exposure
pathway). In this event, the bracketed portion of the Exhibit 4-1 equations that correspond to that
exposure pathway can be dropped from the equation.
Exhibit 4-2 presents the same reduced equations as Exhibit 4-1; however, in Exhibit 4-2, the equations
are adjusted to solve for risk or HQ levels. Soil and water constituent concentrations and chemical-
specific parameters (the same as those noted for Exhibit 4-1) can be entered into the Exhibit 4-2
equations to calculate risks and hazards for specific constituents. The soil and water constituent
concentrations should represent the average concentration levels to which a human receptor could be
exposed (this is defined as the exposure point concentration). Typically, the 95th percent upper
confidence limit on the arithmetic mean (95th UCL) constituent concentration for the area of exposure is
SWEf'Ami*M«\tASKlMt£V1S£nFlNAL\MASTOt WPDMSI«llMi«ia»ftMI3AinrM'S$jinvuc
-------
EXHIBIT 4-2
REDUCED RISK AND HAZARD INDEX EQUATIONS
REGION 10 RCRA RISK-BASED CLEANUP LEVEL GUIDELINES
Soil Equations
For soils, equations are based on three exposure routes (ingestion, skin contact, and inhalation)'.
Equation 1: Combined Exposures to a Single Carcinogenic Constituent in Residential Soil (adults and
children)
R = C (mg/kg) x
(1.1E-4 x CPFJ * (5E-4 x ABS x CPF+
II * CPFi
———
73
Equation 2: Combined Exposures to a Single Noncarcinogenic Constituent in Residential Soil {children)
ABS
HQ = C (mg/kg) x
' 2E-4
r
+
* 4£-4 +
/ \
1®
**L.
*P| *
15.6
Equation 3: Combined Exposures to a Single Carcinogenic Constituent in Industrial Soil (adults)
R = C {mg/kg)
f (CPFm x 5E-5) + (CFF### i ABS x J£-3) ~ {CPFJVF, x 20) ]
286
Equation 4: Combined Exposures to A Single Noncarcinogenic Constituent in Industrial Soil (adults)
HQ * C (mg/kg) x
£rgiMwater.,,Ema||flM
5£ -5
*>P.
+ \JHS-xlE-3
RfD
20
W, * VF,
102.2
for groundwater, equations are based on three exposure routes (ingestion, skin contact, and inhalation of
volatiles).'
Equation 5: Combined Exposures to A Single Carcinogenic Constituent in Water (adults and children)
f (1.1 x CPF, ) + (5.5 x CPF. ) + (2.6 x Kx CPF, ) 1
R = C (ng/L) x L—— 'JL L. ll 1 t±l
73.000
4=64
-------
EXHIBIT 4-2 (Continued)
REDUCED RISK AND HAZARD INDEX
REGION 10 RCRA RISK-BASED CLEANUP LEVEL GUIDELINES
Equation i: Ingestion and Inhalation Exposures to A Single Noncarcinogenic Constituent in Water
(adult) r.
Air Equations
Equation 7: Inhalation Exposures to a Single Carcinogenic Constituent in Air (adults and children)
C (jigfm3) x CPF.
R = -
6,600
Equation 8: Inhalation Exposures to a Single Noncarciaogenic Constituent in Air (adults)
73,000
gg . C <&*.*> .
MJDI x 3,650
Source: Modified from U.S. Environmental Protection Agency I996g
Notes.
HO
CPF,
• ?F
k
\BS
C
R
VF,
! 'D,
RIB,
..g/kg
mg/L
..g/ra'
Volatile chemicals are defined as having a Henry's Law Constant (atm-mVmol) greater than 10'' and a molecular wei;
less than 200 grams/mol, Use Vf, for volatile chemicals and PEF for nonvolatile chemicals
Derma! absorption factor
Concentration
Cancer risk
Hazard quotient
Cancer potency factor, oral
Cancer potency factor, oral, adjusted
Cancer potency factor, inhalation
Chemical -specific permeability coefficient (centimeters per hour)
Volatilization factor, soil
Reference dose, oral
Reference dose, oral, adjusted to account for percent gastrointestinal absorption of chemical
Reference dose, inhalation
Microgram per lifer
Milligram per liter
Microgram per cubic meter
4=65
-------
used as the exposure point concentration. EPA's (1992e) Supplemental Guidance to RAGs: Calculating
the Concentration Term details how to calculate 95th UCL exposure point concentrations for soil.
The Supplemental Guidance to RAGs: Estimating Risk From Groundwater Contamination (EPA 1993a)
details how to calculate groundwater exposure point concentrations. Both documents are included as
Attachment M. Additional information on determining exposure point concentrations and points of
compliance is provided in Section 73,
An example of a constituent-specific calculation using the Exhibit 4-1 equations is presented in the
following section. The same example can be applied to the Exhibit 4-2 equations, except that a
constituent concentration is used to calculate a risk term instead of using a target risk term to calculate a
RBC.
4,7.3 Examples of Risk-Based Concentration Calculations
As an example, the following assumptions have been made about a hypothetical RCRA facility:
An RPI with adequate site characterization has been conducted
Methylene chloride, an industrial solvent, has been determined
to be a COPC in soil The concentration of methylene chloride
used in a facility HHRA was 2,100 mg/kg, representing the
95th percent upper confidence limit of the arithmetic mean.
Calculated cancer risks and noncarcinogenic hazards exceeded
target risk and target hazard levels
Soil ingestion, dermal contact with soil, and inhalation of
volatiles from soil have been determined to be the direct
exposure pathways of concern
A residential land use scenario has been determined to be
appropriate
A target excess cancer risk of 1E-6 and a target hazard of 1.0
have been identified by the program for this facility
a l«»im TASK«'JIEV1SE2.FINAJ.VMASTER WKM 91 *lt I Ml IW»M 2/11 Attn SSsmtoe
-------
No promulgated federal standards for methylene chloride in soil are available. An RBC can be
. calculated using equations I and 2 in Exhibit 4-1.
The following paragraphs discuss the specific information for methylene chloride that is needed to
calculate carcinogenic and noncarcinogenic RBCs. Sections 4.7,3J and 4,7,3,2 present example RBC
calculations based on carcinogenic and noncarcinogenic effects, respectively.
Tie IRIS (EPA 1996g) and HE A ST (EPA 1997a) databases indicate that methylene chloride has toxicity
values for both carcinogenic and noncarcinogenic effects. The oral CPF is 0.0075 (mg/kg/day)"1, the
inhalation CPF is 0.0016 (mg/kg/day)-1, the oral RID Is 0.06 mg/kg/day, and the inhalation RfC is
3.0 milligrams per cubic liter,, converted to an inhalation RID by multiplying by 20 cubic meters and
dividing by 70 kilograms (no absorption factor available). As discussed in Section 4.6.3, the oral CPF
and RfD are adjusted to account for the oral absorption efficiency of methylene chloride when assessing
the dermal exposure pathway. Following the recommendations In Section 4.6,3, a default oral absorption
efficiency of 80 percent may be assumed for VOCs In tie absence of chemical-specific information. The
oral CPF is adjusted for 80 percent oral efficiency to 0.0094 (mg/kg/day)"' to arrive at a dermal CPF.
The oral RfD Is adjusted for 80 percent oral efficiency to 0,048 mg/kg/day to arrive at a dermal RfD.
An ABS is needed to estimate the amount of methylene chloride that is absorbed across the skin. In the
absence of a methylene chloride-specific ABS value, a default ABS value for methylene chloride of
0.05 percent (0.0005) is used. This value is based on recommendations from Region III Technical
Guidance Manual, Risk Assessment: Assessing Dermal Exposure from Soil (EPA 19950- EPA Region 3
recommends two ABS values for VOCs: 0.05 percent for volatiles having a vapor pressure equal to or
greater than benzene (approximately 95.2 millimeters mercury) and 3 percent for volatiles having a
vapor pressure lower than benzene (EPA 19950- Methylene chloride has a vapor pressure of 349
millimeter mercury; therefore, the ABS value of 0.05 percent is used.
Since methylene chloride is a VOC, a VF is derived to assess inhalation exposures, A VF of 2,815 m3/kg
is derived following the method described in Section 4.5.2.1.
-------
Using the Exhibit 4-1 equations (derived from EPA Region 9 [1996c] equations) and the
chemical-specific input factors, the following RJBCs can be calculated for carcinogenic and
noocarcioogenic effects of methylene chloride in groundwater.
4 J3.1 Carcinogenic Effects
The equation for calculating RBCs in soil is as follows:
73 x TR
(4-3)
C (mg/kg)
f 11 x CPF.)
(l.I£-4 * CPF J * (5E-4 i ABS x CPF+
,
(See Table 4-2 and Attachment E for farther explanation of the individual terms in the equation).
Continuing with the example of methylene chloride, and substituting the chemical-specific val >r
cancer CPFs, ABS, VF, and target risk, results in the following equation;
c («#%) - —
(4-4)
(l.l£-4 x 7.5E-3) + (5£-4 x 5E-M x 9.4£-3) + 1 " * "6E~3
f 11 x L6E'
| 2.8£+3
The RBC is as follows:
= 10 3 mZ/k* (l"5)
4.73.1 Noncarcinogeiiic Effects
The equation for calculating health-based RBCs in soil for noncarcinogenic effects is as follows:
„ , „ 4 15.6 x THQ
C (mglkg) = - 7—————
2E-4 ABS x 4E-4
W, I i w.
+*i
10
*•/*>. x FFJ
1 •*EFA *fnl »TA«r.M V1SWWAA1MASTT* ManAac
4-68
-------
Substituting the appropriate value for the RfD, ABS, and VF results In the following equation:
. 15.6 x I * ' *
mg S ( 2E-4 ) + ( 5C-04 i 4£-4| + ( IP \
{ 6.0E-2) * [ 4ME-2 J | 8.6£-l x 2.8E*3J
The RBC is as follows:
c<-*w = 1-877 »«¦% (4'8)
The RBC calculated based on carcinogenic effects is iowcr than that based on noncarcinogenic effects. It
is also lower thai the soil saturation limit concentration of 2,500 mg/kg calculated using tie
equation presented in Table 4-10, In such cases, the lower RBC should be selected as the
health-protective level,
It should be noted that these RBC calculations do not account for potential migration of methylene
chloride to groundwater. If discharge of methylene chloride to groundwater or surface water cannot be
ruled out, the Section 4,5,23 methodology for evaluating these pathways should be followed. Either a
soil screening level or a facility-specific soil-to-ground water RBC should be developed and compared to
the previously noted direct contact RBCs before determining a final RBC.
4.7.4 Adjustment of Risk-Based Concentrations for Multiple Hazardous Constituents
When developing RBCs for facilities at which multiple carcinogens are of concern, the target risk level
or range must still be met. Risks from constituents having carcinogenic effects are assumed to be
additive according to Risk Assessment Forum Review of **Guidelines on Health Risk Assessment of
Chemical Mixtures " (EPA 1997c); therefore, target risk levels for individual constituents may need to be
adjusted downward to ensure that the tola! residual facility risk is within the target risk range. This
adjustment should be done on a facility-specific basis, depending on tie number and nature of hazardous
S »f>« Rt»nU«OTflnraic
4-69
-------
COPCs present. Exhibit 4-3 demonstrates how RBCs for carcinogens can be adjusted lo achieve an
acceptable target risk range.
In general, cleanup levels for the constituent(s) contributing most significantly to total risk can be
adjusted to keep total risk in the target risk range. For example, five carcinogens may be present; four
representing total risks in the 10"4 range and one representing risk in the 10'3 range. Adjusting the
cleanup level for the single constituent contributing most of the risk may be adequate to achieve a total
facility risk in the target range. This would be more practicable if a corrective measure could be selected
that is especially effective for the constituent constituting the highest risk (for example, if the constituent
is highly vulnerable to bioremediation); however, this method may not be practicable in all situations. In
that case, the target risk from carcinogens can be established overall and the remediation would continue
until the combined risk from the mixture of the residual carcinogens had been reduced to the target risk
or lower.
HQs from constituents having similar systemic toxic effects and similar mechanisms of action are also
assumed to be additive. The IRIS (EPA 1996g) and HEAST (EPA J997a) databases provide information
on the types of toxic effects that are the basis for each RfD. Hazardous constituents with similar
noncarcinogenic toxic affects should be grouped together to determine cleanup levels for multiple
constituents. For example, many organic solvents typically affect the central nervous system or the liver.
The target HQ level for individual constituents in each toxic effects group should be adjusted downward
to ensure that the total residual facility HI is less than 1.0. This adjustment can be done on a
facility-specific basis, depending on the number of hazardous constituents present. As with carcinogens,
instead of adjusting the target HQ for individual constituents, the remediation results could be monitored
to ensure that the target HI is achieved or surpassed.
Exhibit 4-4 demonstrates how cleanup levels for noncarcinogens can be adjusted to achieve an
acceptable target HI. As was explained for carcinogens, adjusting the cleanup levels for the chemicals
presenting most of the hazard may achieve an acceptable total hazard.
A regional risk assessment specialist should be consulted to confirm how chemicals are grouped
according to types of toxic effects.
-f BfcPA ft I > M T A K £ V| SE F) N AL\M A$T?R WPDv^s-R m>|sa
4-70
-------
EXHIBIT 4-3
ADJUSTING CLEANUP LEVELS FOR CARCINOGENIC COMPOUNDS
REGION 10 RCRA RISK-BASED CLEANUP LEVEL GUIDELINES
Assume that five carcinogenic hazardous constituents are present in groundwater at the point
of compliance and that each presents the following cancer risks at current concentrations and
at MCL concentrations.
Constituent
Current
Contaminant
Concentration
Excess
Cancer
Risk
Maximum
Contaminant
Level (MCL)
fcg/L) .
Excess
Cancer
Bisk at
MCL
Benzene
50
i x ir
5
1 x 10J
l.2-Dichk>rocthitne
(1,2-DCA)
75
6x icr4
5
4 x 10*s
1,1-Dichtoroethene
(U-DCE)
14
2 x JG"4
7
i x itr*
Tetrachloroethenc
SO
4 * icr»
5
4 x 10*
Trichloroetbene
50
• 3 x Ws
5
3 x 10*
Total Cancer Risk
licit1
2*ir
If a total target risk in the low range is required, the attainment ofcleanup levels below
MCLs will be necessary. One approach may be to set an overall target risk that must be met
by corrective action and monitor constituent concentrations during remediation until the
target risk is achieved. A second approach may be to set target cleanup levels for each
constituent. For example, by setting cleanup levels for each constituent at one-tenth of their
respective MCLs, an overall target risk of2x J(fs would be achieved if the cleanup goals are
met. The ability of remedial technologies to meet clean levels must be confirmed and may
dictate the targeted reductions of constituent concentrations Likewise, analytical methods
capable of detecting constituents at the targeted cleanup levels must be available. For the
purposes of this example, EPA method 8260 is capable ofdetecting the constituents at
0. i pg/L and lower concentrations.
4.71
S W:PA JMwtiiiTASKlt^VISEOTNAUMASTEA WTO\ 151-R S WH IWJiKh E J/1mm Simvm ^ ' 1
-------
COPCs present. Exhibit 4-3 demonstrates how RJBCs for carcinogens can be adjusted to achieve an
acceptable target risk range.
In general, cleanup levels for the constituents) contributing most significantly to total risk can be
adjusted to keep total risk in the target risk range. For example, five carcinogens may be present: four
representing total risks in the 10"6 range and one representing risk in the 10"3 range. Adjusting the
cleanup level for the single constituent contributing most of the risk may be adequate to achieve a total
facility risk in the target range. This would be more practicable if a corrective measure could be selected
that is especially effective for the constituent constituting the highest risk (for example, if the constituent
is highly vulnerable to bioremediation); however, this method may not be practicable in all situations. In
that case, the target risk from carcinogens can be established overall and the remediation would continue
until the combined risk from the mixture of the residual carcinogens had been reduced to the target risk
or lower.
HQs from constituents having similar systemic toxic effects and similar mechanisms of action are also
assumed to be additive. The IRIS (EPA 1996g) and HEAST (EPA 1997a) databases provide information
on the types of toxic effects that are the basis for each RfD. Hazardous constituents with similar
noncarcinogenic toxic affects should be grouped together to determine cleanup levels for multiple
constituents. For example, many organic solvents typically affect the central nervous system or the liver.
The target HQ level for individual constituents in each toxic effects group should be adjusted downward
to ensure that the total residual facility HI is less than 1.0. This adjustment can be done on a
facility-specific basis, depending on the number of hazardous constituents present. As with carcinogens,
instead of adjusting the target HQ for individual constituents, the remediation results could be monitored
to ensure that the target HI is achieved or surpassed.
Exhibit 4-4 demonstrates how cleanup levels for noncarcinogens can be adjusted to achieve an
acceptable target HI. As was explained for carcinogens, adjusting the cleanup levels for the chemicals
presenting most of the hazard may achieve an acceptable total hazard.
A regional risk assessment specialist should be consulted to confirm how chemicals are grouped
according to types of toxic effects.
- uma 7Tl» *Pf> -*> *<-• 4-70
-------
"Tex"' "3re
EXHIBIT 4-3
ADJUSTING CLEANUP LEVELS FOR CARCINOGENIC COMPOUNDS
REGION 10 RCRA RISK-BASED CLEANUP LEVEL GUIDELINES
Assume that five carcinogenic hazardous constituents are present in groundwater at the point
of compliance and that each presents the following cancer risks at current concentrations and
at MCL concentrations.
Constituent
Current
Contaminant
Concentration
(pg/L)
Excess
Cancer
Risk
Maximum
Contaminant
Level (MCL)
(Pg/L) •
Excess
Cancer
Risk at
MCL
Benzene
SO
1 x I©*4
5
I x 10°
1.2-Dichlorocthane
(1.2-DCA)
75
6 x 10-*
5
4 x 105
!,l-Dkhloroethenc
(1,1-DCE)
14
2x I04
7
I x 10"*
Tetrachloroethcnc
50
4x I04
5
4 x 10"*
Trichloroethene
50
• 3 x 10"J
5
3 x 10*
Total Cancer Risk
I * 10°
2 x 10"4
If a total target risk in the low 104 range is required, the attainment of cleanup levels below
MCLs will be necessary. One approach may be to set an overall target risk that must be met
by corrective action and monitor constituent concentrations during remediation until the
target risk is achieved. A second approach may be to set target cleanup levels for each
constituent. For example, by setting cleanup levels for each constituent at one-tenth of their
respective MCLs, an overall target risk of2x Ht* would be achieved if the cleanup goals are
met. The ability of remedial technologies to meet clean levels must be confirmed and may
dictate the targeted reductions of constituent concentrations. Likewise, analytical methods
capable of delecting constituents at the targeted cleanup levels must be available. For the
purposes of this example. EPA method 8260 is capable of detecting the constituents at
0.1 fjg/L and lower concentrations.
4-71
-------
Toa; c i ^ H2 (,ip -b c4»te
EXHIBIT 4-4
ADJUSTING CLEANUP LEVELS FOE
NONCARCINOGENIC HEALTH EFFECTS
REGION 10 RCRA RISK-BASED CLEANUP
LEVEL GUIDELINES
Assume that four hazardous constituents are present in groundwater at
the point ofcompliance, and cause similar noncarcinogenic health
effects. Initial cleanup levels (MCLs) and equivalent hazards quotients
are as follows:
Constituent
Cleanup
Level (/ig/L)
Equivalent
Hazard Quotient
1,2-Dichlorobenzene
600
1.6
1,1-Dichloroethane
810
1.0
1,2-DIettIoraettaene
70
1.3
Ethylbenzeae
700
0.5
Hazard Index
4.4
If a hazard index of 1.0 is required, cleanup level for the first three
constituents can be reduced by one order of magnitude, resulting in ¦
hazard quotients of 0.16, 0.10, and 0.13, respectfidly, and a hazard index
of 0.89. A second approach could be to monitor ongoing remediation
until a target hazard index of 1.0 is achieved or surpassed.
4.7.5
Risk-Based Concentrations Based on Hazardous Constituent Migration
Hazardous constituent migration across media can be incorporated into RBC calculations. As described
in Section 4.5.2.1, VFs and PEFs may be used to incorporate VOC and particulate air emissions into soil
RBCs h is recommended that these be considered. Section 4.5.2.3 describes a partition equation.
models- and criteria that be used to calculate soil RBCs that are protective of groundwater. Soil
RBCs protective of groundwater should be compared to soil RBCs calculated based on direct soil
exposures. The lower of the two RBCs should be selected.
Section 4.5.2.4 describes sources of partition equations that can be used to estimate hazardous
constituent migration from primary media such as soil and groundwater into the food chain. An example
-------
of how the partition equations may be used follows. At a site where home gardening is an important
exposure pathway, constituent migration from soil to garden root crops may be estimated using the
partition equations in Guidance for Performing Screening Level Risk Analyses at Combustion Facilities
Burning Hazardous Wastes. Attachment C, Draft Exposure Assessment Guidance (EPA I994t). Using
the HHRA equations presented by EPA (!994i), the amount of the root crop ingested by a home gardener
would then be estimated, and the risk associated with plant ingestion would be calculated. Soil RBCs
determined after constituent migration into plants has been considered can be back-calculated using the
same HHRA, equations. Chemical-specific parameters required for the calculation include oral toxicity
values and soil-to-plant partition coefficients. Partition coefficients for common hazardous waste
combustion facility constituents are included in the EPA (1994i) combustion guidance (that is, for
dioxins, PAHs, PCBs, nitroaromatics, phthalates, and certain metals). Some of the primary sources of
plant partition coefficients cited by the combustion guidance document series include A Review and
Analysis of Parameters for Assessing Transport of Environmentally Released Radionuclides Through
Agriculture (Baes et al. 1984) for inorganic compounds and Bioconcentration of Organics in Beef Mitt
and Vegetation (Travis and Arms 1988) for organic compounds. Evaluation of Dredged Material
Proposedfor Ocean Disposal, Testing Manual (EPA I991d) also provides octanol-water partition
coefficients for many chemicals that can be converted to plant partition coefficients using equations
recommended by Travis and Arms (1988).
As previously noted, human food chain exposures are typically not significant pathways of concern at
RCRA facilities, so indirect exposures to food chain organisms require consideration only on a
case-by-case basis. Methods for incorporating food chain exposures into RBCs may also be conservative
(for example, garden produce partitioning equations) and may overestimate actual site risks. If they are
significant exposure pathways and are not included, however, risks would be underestimated.
4.7.6 Cleanup Levels for Lead
Cleanup levels for lead are not calculated using standard EPA risk assessment equations. Rather, EPA
has developed an integrated exposure uptake biokinetic model (IEUBK) that predicts lead blood levels in
«n*sn » r
4-73
-------
EXHIBIT 4-4
ADJUSTING CLEANUP LEVELS FOR
NOMCARCMOGEN1C HEALTH EFFECTS
REGION I® RCRA RISK-BASED CLEANUP
LEVEL GUIDELINES
Assume that four hazardous constituents are present in groundwater at
the point of compliance, and came similar nomarcimgmic health
effects. Initial cleanup levels (MCLs) and equivalent hazards quotients
are as follows:
Constituent
Cleanup
Level Cig/L)
Equivalent
Hazard Quotient
1,2-Dichiorobenzene
600
1.6
1,1-Dichloroethane
81®
1.0
1,2-Dichloroethcne
70
1.3
Ethylbenzene
700
0.5
Hazard Index
4,4
If a hazard index of 1.0 is required cleanup level for the first three
constituents can be reduced by one order of magnitude, resulting in
hazard quotients of 0.16, ft 1ft and ft 13, respectfully, and a hazard index
of 0.89. A second approach could be to monitor ongoing remediation
until a target hazard index of 10 is achieved or surpassed
4.7.5
Risk-Based Concentrations Based on Hazardous Constituent Migration
Hazardous constituent migration across media can be incorporated into RBC calculations, As described
in Section 4.5,2.1, VFs and PEFs may be used to incorporate VOC and particulate air emissions into soil
RBCs. It is recommended that these be considered. Section 4.5.2.3 describes a partition equation,
models, and criteria that can be used to calculate soil RBCs that are protective of groundwater. Soil
RBCs protective of groundwater should be compared to soil RBCs calculated based on direct soil
exposures. The lower of the two RBCs should be selected.
Section 4.5.2.4 describes sources of partition equations that can be used to estimate hazardous
constituent migration from primary media such as soil and groundwater into the food chain. An example
4 REM Rlmir.TASKraEVlSEIiFINAUMASIEII W»tSi.Rlt«i»irai
-------
of how the partition equations may be used follows. At a site where home gardening is an important
exposure pathway, constituent migration from soil to garden root crops may be estimated using the
partition equations in Guidance for Performing Screening Level Risk Analyses at Combustion Facilities
Burning Hazardous Wastes, Attachment C, Draft Exposure Assessment Guidance (EPA 1994i), Using
the HHRA equations presented by EPA (1994i), the amount of the root crop ingested by a home gardener
would then be estimated, and the risk associated with plait ingestion would be calculated. Soil RBCs
determined after constituent migration into plants has been considered can be back-calculated using the
same HHRA equations. Chemical-specific parameters required for the calculation include oral toxicity
values and soil-to-plant partition coefficients. Partition coefficients for common hazardous waste
combustion facility constituents are included in the EPA (19941) combustion guidance (that is, for
dioxins, PAHs, PCBs, nitroaromatics, phthalates, and certain metals). Some of the primary sources of
plant partition coefficients cited by the combustion guidance document series include A Review and
Analysis of Parameters for Assessing Transport of Environmentally Released Radionuclides Through
Agriculture (Baes et al, 1984) for inorganic compounds and Bioconcentration of Organics in Beef Milk,
and Vegetation (Travis and Arms 1988) for organic compounds. Evaluation of Dredged Material
Proposed for Ocean Disposal, Testing Manual (EPA 199 Id) also provides octanol-'water partition
coefficients for many chemicals that can be converted to plant partition coefficients using equations
recommended by Travis and Arms (1988).
As previously noted, human food chain exposures are typically not significant pathways of concern at
RCRA facilities, so indirect exposures to food chain organisms require consideration only on a
case-by-case basis. Methods for incorporating food chain exposures into RBCs may also be conservative
(for example, garden produce partitioning equations) aid may overestimate actual site risks. If they are
significant exposure pathways arid are not included, however, risks would be underestimated,
4,7.6 Cleanup Levels for Lead
Cleanup levels for lead are not calculated using standard EPA risk assessment equations. Rather, EPA
has developed an integrated exposure uptake biokinetic model (IEUBK) that predicts lead blood levels in
4-71
5 \REPMR!»MK TASK».REVlSE}'JnN/^\MA5TER ^ 1
-------
children. The model was developed for children because they are more sensitive to lead effects. The
IEUBK is considered superior to the use of a RID in a standard HHRA equation because the model
(I) recognizes the multimedia nature of lead exposures, (2) incorporates important absorption and
pharmacokinetic information, and (3) considers the potential distributions of exposures and risk likely lo
occur al a facility. The model allows for lie incorporation of water, soil/dust, air, dietary, paint and
maternal-blood lead concentration levels into a lead dose estimate.
The Memorandum Regarding Revised Interim Soil Lead Guidance for CERCLA Sites and RCRA
Corrective Action Facilities (EPA 1994k) recommends a residential soil lead screening level of
400 mg/kg. This screening level is back-calculated using the IEUBK model, assuming a blood lead
concentration of 10 micrograms per deciliter (jUg/dL) in children and standard defaults for water, air,
diet, paint, and maternal-blood lead levels. This target blood level is based on analyses conducted by the
Centers for Disease Control and EPA that associate blood lead levels of 10^g/dL or higher with health
effects in children. No comparable soil lead screening level has been set for a nonresidential adult; the
IEUBK. model is designed specifically for children.
EPA is currently considering what industrial soil lead levels may be appropriate. An interim approach is
recommended by EPA (1996k) in Recommendations of The Technical Review Workgroup for Leadfor
An Interim Approach to Assessing Risks Associated with Adult Exposures to Lead in Soil. This document
recommends the use of a biokinetic slope factor, which relates the increase in typical adult blood lead
concentrations for a women of child-bearing age to average lead uptake from contaminated soils. The
document then recommends the use of a proportionality constant, which relates a fetal blood lead
concentration at birth to a material blood lead concentration. Using a fetal blood lead goal of 10 pg/dL
or less, a soil RBC can be calculated that should not result in an exceeded of the fetal blood lead goal.
Using default parameters summarized in the EPA (1996k) review, RBCs ranging from 743 mg/kg to
1,738 mg/kg can be calculated, depending on the baseline level of adult blood lead and the blood lead
standard deviation associated with adults exposed to similar on-site lead concentrations. EPA (1996k)
should be consulted to confirm a facility-specific industrial soil lead RBC. The Region 10
representatives on the EPA Superfund technical review group should be consulted for information on
current lead policy.
S ftEPA IU *»tn TASK* REVSSEi'-F&NA.l'.MASTER WPOmi'RIMiUWftliAll/UnW
4-74
-------
The EPA memorandum recommends using the IEUBK model on a site-specific basis to develop media
cleanup standards at RCRA facilities where site data support modification of model default parameters
(1994k).
4 J UNCERTAINTY ANALYSIS
Once the HHRA process and the step-by-step process for developing cleanup levels are understood,
facility and regulatory officials should be prepared to answer the questions listed in Exhibit 4-5 for a
contaminated facility undergoing corrective action or clean closure under RCRA.
EXHIBIT 4-5
RISK ASSESSMENT AND CLEANUP LEVEL DETERMINATION PROCESS
REGION 10 RCRA RISK-BASED CLEANUP LEVEL GUIDELINES
" " ' ' -"-rm'-T-f "ttttttt-i nt nri ii n
• Have receptors and exposure pathways that are likely affected by facility
contamination been identified?
• Are data of sufficient quantity and quality available to determine whether and to
what extent on-site and off-site remediation is necessary to protect the health of the
maximally exposed individual?
• Have contaminants of concern been identified?
• Are promulgated standards or criteria available? If so, are they sufficiently
protective?
• Are there state methodologies, regulations, or policies that should be considered?
• Is the corrective action or clean closure being considered subject to a RCRA-
authorized state program?
• Have current and future land use scenarios been identified?
• Have risk-based concentrations been calculated using published RfDs and CPFs for
any of the COPCs?
• Have risk management decisions regarding target risk and hazard levels been
identified by the lead regulatory agency?
• Have cleanup levels been adjusted (as appropriate) to account for multiple
hazardous constituents?
K*lft£V!SE3*FmAl\MAyrcft 33aittuc
-------
RBCs are associated with varied levels of uncertainty. Uncertainty can be defined as a lack of precise
knowledge of the qualitative or quantitative truth. As part of an Weal risk analysis, a complete
uncertainty analysis would provide a risk manager with the ability to estimate risk for each individual in
a given population in both actual and projected scenarios of exposures; it would also estimate the
uncertainty in each prediction in quantitative, probabilistic terms (see Chapter 6). But even a less
exhaustive treatment of uncertainty will serve a very important purpose: it can reveal whether the
deterministic risk estimate overestimates or underestimates risk and if so, to what extent.
Uncertainties associated with all sections of the human health-based analysis should be discussed in all
HHRAs. This section discusses uncertainty in qualitative terms; for a discussion of quantitative
uncertainty, see Chapter 6 (Probabilistic Risk Assessment Methods). Additional uncertainty discussions
can be found in the National Research Council's Science and Judgement in Risk Assessment (Chapter 9)
(1994) and in EPA Administrator Carol Browner's memorandum of risk assessment (EPA 1995a)
(Attachment A). The effect that each element of uncertainty las on the final cleanup levels should be
included. For example, an assumption that future land use will be industrial will lead to less
conservative (that is, higher) cleanup levels.
4.8.1 Data Uncertainty
Issues contributing to uncertainty in data evaluation include the identification of COPCs, data quality,
data useability, adequacy of quantitation limits (relative to target risk levels), and data coverage.
All data thai are used to evaluate compliance with a cleanup level should be reviewed with respect to
data quality standards presented in the site-specific quality assurance project plan. Data that do not meet
the data quality standards should be qualified as necessary, and the uncertainty associated with these data
should be discussed. Uncertainty associated with data quality may result in an underestimation or
overestimation of hazardous constituent concentrations, depending on the specific data quality issue.
By the time the uncertainty section is prepared, data should have been reviewed to ensure that the
detection or quantitation limits achieved for the various analyses were sufficient to evaluate site-specific
hazardous constituents at concentrations equal to or less than the hazardous constituent-specific cleanup
levels. If a chemical's detection or quantitation limit is greater than its cleanup level, one cannot
S?.;- Pa .v? t «• iK TaSKX*EVISEAl-.MASTER HtrnVuc
4-76
-------
determine whether the cleanup level has been met. Detection and quantitation limits should be compared
to cleanup levels before chemical analyses. Chapter 3 of this guidance provides additional information
on data evaluation procedures, while Chapter 7 describes how below detection limit data should be
handled in HHRA.
Data coverage refers to the ability of the sampling plan and associated sample results to completely
characterize the contamination. (Were sufficient data collected to evaluate all potential exposure
pathways? Do the sampling locations adequately represent actual or potential exposure condition's?) for
example, uncertainty would arise if groundwater sample points were not located downgradient of a
source.
Data quality indicators were discussed in Section 3.2: completeness, representativeness, comparability,
precision, and accuracy. Any problems associated with any of the data quality indicators should be
discussed in the uncertainty section, Other issues that could adversely affect data quality are blank
contamination, matrix interferences, sample holding times, and sample preservation.
4.8.2 Exposure Assessment Uncertainty '
Uncertainty is inherent in the evaluation of exposure pathways and in the assumptions used to estimate
exposure doses. Human activity patterns and individual characteristics (for example, body weight) can
vary significantly within a given population. The degree of uncertainty depends to a large extent on the
amount and adequacy of facility-specific data available. Typically, the most significant areas of
uncertainty for the exposure assessment include exposure pathway identification, exposure assumptions,
assumptions of steady-state conditions, environmental chemical characterization, and modeling
procedures. These areas of uncertainty are described as follows:
~ Exposure pathway identification: To the degree that actual or future human activity
patterns are misrepresented, uncertainty is introduced into the cleanup levels. In most
cases, there is uncertainty regarding future land use at a site. This uncertainty must be
considered when evaluating exposure estimates developed under the future land use
scenarios.
4.77
-------
Exposure assumptions: Standard default assumptions for population characteristics,
such as body weight and surface area, life expectancy, period of exposure, and exposure
characteristics such as frequency, duration, amount of intake or contact, and degree of
absorption or soil adherence may not accurately represent exposure conditions. The
exposure assumptions used may overestimate or underestimate actual exposure.
Assumption of steady-state conditions: Estimated future exposure doses are based on art
assumption of steady-state conditions. The inherent assumption is that future chemical
concentrations are the same as those measured during sampling. This assumption
ignores the effects of various fate-and-transport mechanisms, which will alter the
composition and distribution of most chemicals present in the various media. The
assumption of steady-state conditions usually results in overestimations of future
chemical concentrations and exposure doses.
Hazardous constituent characterization: It is impossible to completely characterize the
nature and extent of hazardous constituents in the environment. Instead, the various
environmental media are sampled to estimate hazardous constituent concentrations and
to identify hazardous constituents actually present as a result of releases at the facility.
Because no sampling can completely and accurately characterize environmental
conditions, the exposure dose calculation will be somewhat uncertain. Uncertainties are
introduced into exposure dose calculations during collection, analysis, and evaluation of
environmental chemical data. Two areas of uncertainty that should be addressed in a
risk assessment are the assumption of uniform concentrations in an exposure area and
the treatment of nondetection results.
Modeling procedures: Modeling assumptions are used to determine hazardous
constituent concentrations in outdoor air resulting from VOC and particulate emissions
and VOC concentrations in indoor air generated by household water use. The numerous
assumptions included in these models introduce uncertainty to the degree that they do
not reflect actual conditions. Use of models may overestimate or underestimate actual
environmental concentrations.
Analytical and numerical models are also used to estimate soil-to-ground water
contaminant migration. There are many soil-to-groundwater models available that have
been developed by government agencies, universities, and the private sector. Models
can simulate different fate and transport processes and migration mechanisms such as
advection, dispersion, diffusion, retardation, chemical reactions, and microbial reactions.
In addition, different models are designed to account for various phases of contamination
{for example, vapor, water, or nonaqueous phase liquids) and various dimensions (one-,
two-, or three-dimensional). Some of the models simulate complex geologic and
hydrogeologic systems while other models are more appropriate for simple geology and
groundwater flow conditions. The models use assumptions to simplify the real-world
conditions and describe these conditions with mathematical equations (usually partial
-------
.
4.8,3 Toxicity Assessment Uncertainty
RfDs and CPFs must be viewed in light of uncertainties and gaps in lexicological data, for instance,
direct information concerning toxic effects in humans is often limited t® historical cases of accidental or
industrial exposures. Animal studies conducted with specially bred homogeneous species are typically
extrapolated to a heterogenous human population. The reliance on animal studies introduces
uncertainties regarding effects on humans including sensitive subpopulations and differences in
physiological characteristics between the animal species studied and humans, such as target organs,
metabolism, organ sensitivity, and detoxification capabilities.
In addition, high-dose, short-term (acute) animal studies may not be applicable to low-level, long-term
(chronic) exposures that humans are more likely to experience. Likewise, the quality of the animal study
may introduce additional uncertainty if, for example, accepted scientific protocols were not employed.
The uncertainties discussed previously are addressed by dividing tie no-observable-adverse-effect level
(NOAEL) for a hazardous constituent from animal studies by uncertainty factors of 10 to 10,000 to
obtain RfDs. The NOAEL is the highest level of a hazardous constituent evaluated in a study that does
not cause statistically significant differences between the experimental and control animals. The lowest-
observable-adverse-effect level (LOAEL), on the other hand, is the lowest level of a hazardous
constituent evaluated in a study that causes statistically significant differences between experimental and
control animals. Uncertainty factors are applied to data in the following cases (EPA 1989c);
differential equations). When a model is applied at a site, these assumptions should be
carefully evaluated against the site-specific conditions. A sensitivity analysis should be
performed to identify the most sensitive parameters. The uncertainty of the models can
be determined by evaluating the agreement between model assumptions and site
conditions as well as the accuracy of input parameters.
4.70
* AEFAUIiHlft'TASKtVUVlSUOTNAiAMASnEJl WPC*l5l'filmitniUm*]S)im» iiimw ^ ' 7
-------
•
To account for variation in the
general population (to protect
sensitive subpopulations)
•
To extrapolate data from animals to
humans
m
To adjust for using a NOAEL from a
subchronic study rather than a
chronic study
m
To adjust for using a LOAEL
instead of a NOAEL in developing
an RfD
A modifying factor ranging from 1 to 1© is also applied to the RID to address uncertainties in the
scientific studies used to develop RfDs. Published RfDs already contain the necessary uncertainty and
modifying factors.
Uncertainty associated with determining chemical carcinogenicity is reflected in the weight-of-evidence
classification groups assigned to carcinogens. In addition, CPFs are derived from the low-dose end of
the dose-response curve. The studies are usually conducted at the high-dose end of the curves. The
selected 95 th upper confidence limit of the slope of the dose-response curve is considered an
upper-bound toxicity value (that is, there is only a 5 percent chance that the probability of a response
could be greater than the estimated value on the basis of the experimental data and model used).
The use of oral toxicity factors to evaluate dermal exposures is associated with uncertainty. The use of
oral toxicity factors as surrogates for this pathway is necessary because no dermal toxicity factors have
been approved by EPA. Most of the uncertainty associated with the use of surrogate toxicity factors
exists because the constituents in question are not known to exhibit similar toxicologies! effects (that is,
degree of toxicity, target organ) during dermal contact and the oral pathway. In addition, dermal
absorption assumptions add to uncertainty. Using surrogate toxicity factors is more conservative than
ignoring the dermal pathway and allows for a quantitative cleanup level rather than a qualitative
discussion.
s •R|fVJ«l'»M*'TASKl\]RE\1Sei\nN/aWA!rreM\W558-»5i»nw>,»*«5J/%l/»)7w
4-80
-------
4.8.4 Cleanup Level Uncertainty
Since the cleanup level calculations incorporate Information from all the previous processes, the
uncertainties associated with the data evaluation, exposure assessment, and toxicity assessment sections
will all directly affect the cleanup level uncertainly.
S KrM Jtl'IHIT.TASKrXEVISEl'JINAUMASTt* WFtMH-HIWIlMinM J/1117") )>»**»« 4-31
-------
Simulation Rwalts
Using probabilistic simulations, an outcome is calculated repeatedly for a predetermined number of
iterations, producing an associated PDF. The following forecast chart (Exhibit 6-5) displays the PDF for
the B(a)P RBC.
EXHIBIT 6-5
CASE STUDY PROBABILITY DENSITY FUNCTION FOR
BENZO(A)PYRENE RISK-BASED SOIL CONCENTRATIONS
REGION 10 RCRA RISK-BASED CLEANUP LEVEL GUIDELINES
Cell BIB
.025"
Forecast RBC
Frequency Chart
9,888 Trials Shown
244
183
HllUJtlimdn
12
0.50
1.75
3.00
5.50
Figure 6-5 shows that the PDF is lognormal in shape, and the numerical output of the PRA indicates that
the PDF has a minimum value of 7.4E-01 mg/kg and a maximum value of 7.72E+00 mg/kg. Output PDFs
are often highly non-Gaussian (nonnormal) in shape for two reasons. First, some or all of the exposure
factor inputs may not have normal or even symmetric distributions. Second, since the exposure factors
enter the formula by multiplication and division, even if all the factors have normal distributions, the
results will not (Thompson et al. 1992).
As noted in Section 6.2.2, a sensitivity analysis should be performed to identify critical input variables.
The sensitivity analysis determines the degree to which a specific exposure factors will affect the final
outcome. A sensitivity analysis could have been performed before the PRA simulation so that critical
exposure factors could be identified. For this case study, however, PDFs were available for all of the
exposure intake variables, so they were all included in the simulation, and the sensitivity analysis was
5 Hiitt-.TASKr«VBEl*tNAl-«ASm »>w 6-25
-------
performed during the simulation. With Crystal Ball software, sensitivity is calculated by computing
Spearman rank correlation coefficients (a common statistical measure of dependency) between every
exposure factor and outcome calculation while the simulation is running (Decisioneering 1993). A
positive correlation coefficient indicates that an increase in the exposure factor is associated with an
increase in tie outcome. A negative correlation coefficient indicates that an increase in the exposure factor
is associated with a decrease in the outcome (Decisioneering 1993). The larger the absolute value of the
correlation coefficient, the stronger the relationship; however, caution should be applied when interpreting
the sensitivity analysis in simulations where exposure factors are correlated. For example, if a highly
sensitive exposure factor were correlated with an insensitive one, the insensitive exposure factor would
likely have a high sensitivity will regard to the outcome (Decisioneering 1993).
Exposure factors that are identified as being highly sensitive, contributing a high degree of uncertainty to
the outcome, may be further refined so as to decrease the effect on the final outcome. Likewise, it may not
be necessary to spend PRA resources on those factors that have little effect on the risk outcome.
In the Exhibit 6-6 sensitivity chart, exposure factors are listed on the left side, beginning with the exposure
factor with the highest sensitivity.
In this simulation, adult exposure duration has the highest sensitivity ranking and can be considered the
most important assumption in the model. Likewise, the child exposure frequency and the child soil
ingestion rate have the lowest sensitivity rankings and can be considered the least important assumptions in
the model. Considering this, collection of facility-specific exposure durations could be prioritized over the
collection of child soil ingestion rates if further data collection was deemed necessary. Confidence must
also be established in the child exposure frequency and child soil ingestion rate information used in the
sensitivity analysis. For example, if it is believed that facility-specific child soil ingestion rates are
significantly underestimated, further data collection and a second sensitivity analysis may be warranted.
Descriptive statistics shown in Table 6-1 were calculated for the B(a)P RBC PDF. Population percentiles
arid corresponding RBCs are also determined for the PDF, as presented in Table 6-2.
is MFA • TA»*t¥BO»NAtlrm ssmmm
-------
EXHIBIT 6-6
CASE STUDY SENSITIVITY CHART FOR EXPOSURE PARAMETERS
REGION I® RCRA RISK-BASED CLEANUP LEVEL GUIDELINES
Sensitivity Chart
Target Forecast: Risk Based Concentration (mg/kg)
Adult Exposue Duration (years)
-..73
Adult Body Weight (kg)
43
Chid Exposure Duration (years)
-.42
¦¦Hit
Adufl Ingestion Rale (mg sol/day)
-.14
¦
Child Bo# Weight (kg!
,13
¦
Adult Exposure Frequency (day/year)
-.06
1
Chid Exposure Frequency (day/year)
•03
(
Chid Ingestion Rale (mg soi/daf)
-.03
1
•0.5 0 0.5
Measured by Rank Correlation
S ftff»A'J?j(sur.TASi;rREviS£J\FlNAiTM aster wpmnmmwtfrnmmiffl?* Htmvm
6-27
-------
TABLE 6-1
DESCRIPTIVE STATISTICS FOE BENZO(A)PYlENE
RISK-BASED CRITERIA PROBABILITY DENSITY FUNCTION
REGION IB RCRA RISK-BASED CLEANUP LEVEL GUIDELINES
Statistics
V
Trials
10,000
Mean
9E+00
Median
2.41E+00
Standard Deviation
tE+00
Variance
>£+00
Skewncss
9.6E-0I
Kurtosis
3.99E+00
CoefF. of Variability
3.9E-01
Range Minimum
7.4E-01
Range Maximum
7.72E+00
Range Width
6.98E+00
Mean Std. Error
1.00E-02
Definitions:
Trials
Mean
Median
Mode
Standard deviation
Variance
Skegness
ktiriisNis
Coefficient of variability
Mean standard error
Number of iterations
Arithmetic average of the risk-based concentration (RBC)
Value midway between the smallest RBC value and largest RBC value
Value (if exists) that occurs most often in the data set
Measuremeni of variability of the data set: square root of the variance
Average of the squares of (he standard deviations of a number of observations from the mean
value; square of ihe standard deviation
The measure of the degree of deviation of a curve from Ihe norm of an asymmetric distribution.
The greater the degree of skewness. the more points of the curve lie to either side of Ihe curve. A
normal distribution curve, having no skewness. is symmetrical (Decisioneering 1993)
The measure of the degree of peakedness of a curve. The higher Ihe kurtosis. the closer Ihe
points of Ihc curve lie to the mode of curve, A normal distribution curve has a kurtosis of 3
(Decisioneering 1993)
A measure of relative variation that relases the standard deviation to ihe mean
(Decisioneering 1993)
The standard deviation of ihc distribution of possible sample means. This statistic describes the
accuracy of the simulation (Decisioneering 1993)
•» 1 ewsta JWMJMASTE* Mwnue
6-28
-------
TABLE 6-2
POPULATION PERCENTILES FOR BENZO(A)PYRENE
RISK-BASED CRITERIA PROBABILITY DENSITY FUNCTIONS
USING I.0E-06 AS THE TARGET RISK
REGION 10 RCRA RISK-BASED CLEANUP LEVEL GUIDELINES
Percentile
mg/kg
0.0%
7.4E-01
2.5%
1.2E+I
5.0%
1.3E+I
50.0%
2.4E+00
95.0%
4.5E+00
97.5%
5.0E+00
100.0%
7.7E+00
For example, 5 percent of the population exposed to 1.3E+00 mg/kg of B(a)P would have a carcinogenic
risk of IE-06.
6.4.2 Deterministic Risk Assessment
A B(a)P RBC was also calculated using the deterministic risk assessment approach for residential exposure
via soil ingestion. The algorithm used to calculate the soil RBC is the same as that previously presented
for the Monte Carlo simulation. RME exposure factors used for the calculation are presented in Table 6-3.
6.4.3 Summary of Results
Using the deterministic risk assessment approach, the calculated B(a)P RBC is 2.2E-0I mg/kg for
residential RME exposure via soil ingestion. Because of compounding conservatism in the deterministic
risk assessment approach, this B(a)P RBC falls well below the 1 percentile of the RBC distribution
determined using the Monte Carlo simulation. Thus, information about the uncertainty surrounding the
conservative assumptions in the deterministic risk assessment provided in the PRA will be useful to risk
managers during the decision-making phase of the process,
8 REP4W8«#US'TASKI«£V*$£i'nNAL\MASTZR 55«m«
-------
TABLE 6-3
/
EXPOSURE FACTOMS FOE RESIDENTIAL
EXPOSURE INGESTION OF SOIL
REGION 1# RCRA RISK-BASED CLEANUP LEVEL GUIDELINES
ictor
RME Value
TR
= Target Risk
1.DE-OS*
CSFo
= Cancer Slope Factor for B(a)P (mR/kg-day)"1
7.3
IR
= Ingestion rate (mg/day) (EPA 1993h)
Adult
Child
100
200
EF
= Exposure frequency (days/year) (EPA 19931)
350
ED
= Exposure duration (years) (EPA 1993h)
Adult
Child
24
6
CF
= Conversion factor (kg/mg)
I.0E-06
BW
= Body Weight (kg) (EPA 1991b)
Adult
Child
70
15
AT
= Averaging time (days)
Carcinogenic
25550
Note:
a Point of departure target cancer risk
*• ftEPtJl i* i*-U ICfftEVISO'JINAtWASTER
6-30
-------
6,5
WORK PLANS, REPORTS, AND PRESENTATIONS
There are several minimal requirements for PRA work plans and reports to ensure sufficient data quality.
In fact, it is important that the facility submit a work plan for EPA review before doing the PRA
simulation. The work plan should include exposure variables for human receptors. Guidance for
application to environmental receptors is not available at this time. The work plan should describe the
software to be used, the exposure routes and models, and input probability distributions and references.
The following good principles of practice should be used to select input data and distributions for the PRA
work plan (EPA 1997d, Items 11 through 16* page 17),
• A complete and thorough description of tie exposure model and its equations should be
provided.
• The presentation of tie deterministic point estimate should always accompany a PRA
analysis.
• Where possible, areas of uncertainty should be identified accompanied by an explanation
of how it will be dealt with in the report.
• Sensitivity analysis should be used to identify model structures, exposure pathways, and
model input assumptions and factors that make important contributions to the assessment
endpoint and its overall uncertainty and variability.
• Probabilistic assessment should be restricted to significant pathways and variables.
• Sufficient data should be used to support the choice of input distributions for model input
factors.
• Surrogate data can be used to develop distributions when they can be appropriately
justified.
• Data should be collected to develop input distributions for the exposure model following
the basic tenets of environmental sampling. Furthermore, particular attention should be
given to the quality of information at the tails of the distribution.
• Expert judgment may be used to select appropriate input distributions, but the reasons and
justification for subjective analysis should be included in detail.
S RtP* »{'»H* TASKa-^E\lSEl"/!KAl.MASTER WPI*151.ft|t«1 J/Vt.">?A» S
6-31
-------
Presentation of PRA simulation results should be tailored to the targeted audience. Entirely different types
of reports are needed for scientific and nonscientific audiences. For example, descriptive and less detailed
summary presentations may be appropriate for the nonscientific public. Graphs and tables showing and
describing each input distribution, distribution of risk for each exposure route, and distributions of total
risk should be included.
s TASK* REVISE: fINALMASTCR 93«rtMC
6-32
-------
n
o
¦e
C
o
m
-------
CHAPTER 7
DETERMINATION OF COMPLIANCE WITH TARGET CLEANUP LEVELS
Following the selection of target cleanup levels and corrective actions to treat or remove a contaminated
media, additional sampling Is performed to determine whether the cleanup levels have been attained. This
sampling Is performed using the same data quality objectives (DQO) and data quality assessment (DQA)
procedures described in Chapter 3 of this guidance. The DQO and DQA steps are summarized in
Sections 7.1 through 7.6. Section 7.7 provides additional guidance on how to handle below-detection-limit
(BDL) sampling results when calculating constituent concentrations.
7.1 DATA QUALITY OBJECTIVES STEPS 1 THROUGH 3
Steps 1 through 3 of the DQO process, as described in Chapter 3, apply directly to compliance
determinations, with the exception that cleanup levels are considered that may differ from the screening
levels or preliminary background levels considered during the Resource Conservation and Recovery Act
(RCRA) facility investigation (RFI) stage. The decision rule is likely to be very similar to the Chapter 3
example (that is, have constituent levels been reduced to a concentration below the selected cleanup level
concentration? If so, no further action would be called for; if not, further remedial action may be
required). The remaining Chapter 7 sections provide additional information on DQO Steps 4 through 7
and the DQA as they apply to determination of compliance with target cleanup levels.
7.2 STEP 4: STUDY BOUNDARIES
In Step 4 of the DQO process, the spatial and temporal boundaries of the problem are defined. To
determine compliance with cleanup levels, facility-specific points of compliance that were earlier
established in the formal RCRA corrective action process should be used. A point of compliance is the
location or locations at which media cleanup levels are to be achieved. No final U.S. Environmental
Protection Agency (EPA) RCRA corrective action regulations defining points of compliance have been
promulgated. In an update to the proposed Subpart S rule, EPA (Federal Register [FR] 19450, May 1,
1996) recommends that points of compliance be determined on a site-specific basis and notes that program
implementors and facility owners have routinely established points of compliance in the following manner:
5 REM'RHMI»',T*SKl-JII«SEIW«.W*rflS WPDil JKIWumwMI/tOTW Sbattac
7-1
-------
For air releases, the location of the most exposed receptor or
other specified point(s) of exposure closer to tie source of
release (for example, the unit boundary)
For surface water, at the point at which releases could enter
the surface water body; if sediments are affected by releases
to surface water, a sediment point of compliance Is also
established
Soil points of compliance are generally selected to protect
human and ecological receptors against direct contact or food!
chain exposures and to protect other media from cross-media
transfer
For groundwater, throughout the area of contaminated
groundwater or at and beyond the boundary of the waste
management area encompassing the original sources of
groundwater contamination when waste is left in place
Washington Stale Mode! Toxics Act (MICA) regulation (Washington Administrative
Code [WACJ 173-340) requires points of compliance for sites conducting cleanup under RCRA authorities-'1
as shown below.
MTCA Points of Compliance
Compliance with air cleanup levels should be attained in the
ambient air throughout the site. A conditional air point of
compliance may be set at the property boundary of an industrial
facility (WAC 173-340-750).
Compliance with surface water cleanup levels should be the point
or points where hazardous substances are released to surface
waters of the state, unless a dilution zone is authorized
(WAC 173-340-730).
For soil cleanup levels based on groundwater protection, the point
°r points of compliance is soils throughout the site. For soil
cleanup levels based on direct contact human exposures, the point
of compliance is soils throughout the site from the ground surface
to 15 feet below the ground surface (WAC 173-340-740).
S ^PAJ^MiiUTASKlAEVISerflNALXMASTER WJWISl-RHitlllnTniniyii/vTrtSSj^yj^ 7»2
-------
MTCA Points of Compliance
Continued
Compliance with groundwater cleanup levels should be
determined for each groundwater monitoring well or other
monitoring points such as a spring (WAC 173-340-720).
In summary, the study boundaries for compliance decisions are primarily set by points of compliance. The
points of compliance should be determined following EPA guidance or, in Washington, the specified
MTCA regulation.
7.3 STEPS: DEVELOP DECISION RULE
As described in Clapter 3, Step 5 requires that a decision rule be developed to define the conditions that
would cause the decision maker to choose among alternative actions. The decision rule is an "if... then"
statement that incorporates the information determined during PQO Steps 1 through 4. An example
decision rule for a compliance decisions follows: if the parameter of interest (average concentration of
constituent of concern) within the study area (the point of compliance) is less than the cleanup level (a
standard, criterion, risk-based, or background concentration) following remediation, then alternative action
A (no further remedial action) should be taken; otherwise, alternative action B (remove additional
contamination) should be taken.
Note that for the previous example, the average concentration of the constituent was selected as the
statistical parameter to compare with the cleanup level. The statistical parameter selected (for example, the
mean, median, or an upper percentile constituent concentration) will be a measurement of the
contamination present within tie study boundaries. Because a receptor is assumed to move randomly
across an exposure area overtime, spending equivalent amounts of time in each location, EPA's
Supplemental Guidance to RAGs: Calculating the Concentration Term {EPA 1992e) {Attachment M)
recommends the use of the true mean lo characterize long-term exposures in a specific study area. To be
reasonably sure that the comparison value is at least as large as the true site mean, EPA (1992e)
recommends use of the 95th upper confidence limit on the arithmetic mean (95th UCL) for the exposure
point concentration and details how to calculate it. The 95 UCL is defined as a value that, when calculated
repeatedly for randomly drawn subsets of site data, equals or exceeds the true mean 95 percent of the time.
7-1
s TASkimvurmnAj-'MArrt* s ' ->
-------
The 95 UCL of the mean Is used because it is not possible to know the true mean, particularly with limited
sampling data. As sampling data become less limited, uncertainly decreases, and the UCL moves closer to
the true mean (EPA 1992e). Other statistical parameters thai characterize the population may be relevant
For example, an upper percentile of the distribution of constituent measurements in the study area (for
example, the 95th percentile) may be compared to a cleanup level to determine whether a subpopulation
(for example, a potential hot spot) is present. Statistical outlier tests may also be used to identify hot spots.
Statistical Analysis of Ground-Water Monitoring Data at RCRA Facilities»Interim Final Guidance
(EPA 1989k), Guidance for the Data Quality Objectives Process (EPA 1994b), and Determination of
Background Concentrations of Inorganics in Soils and Sediments at Hazardous Waste Sites (EPA 1995c)
provide more detailed guidance on the selection of an appropriate statistical parameter to compare with a
cleanup standard.
Washington State MTCA regulation (WAC 173-340-740) requires that upper percentile constituent
concentrations be used to evaluate compliance with soil cleanup levels based on short-term or acute toxic
effects. For cleanup levels based on chronic or carcinogenic effects, MTCA requires that the mean soil
concentration generally be used to evaluate compliance unless large variations in constituent
concentrations occur. To address hot spots, the MTCA rule also specifies that no single sample
concentration exceed two times the cleanup level and that less than 10 percent of the sample
concentrations exceed the cleanup level.
7.4 STEP 6: SPECIFY TOLERABLE LIMITS ON DECISION ERRORS
Step 6 requires that the decisions maker's tolerable limits on decision errors be specified. As noted in
Chapter 3, the true value of the population parameter being measured (for example, the average constituent
concentration) can never be exactly defined because of sampling design and measurement design errors. A
decision error occurs when the data mislead the decision maker into concluding that the parameter of
interest is on one side of a cleanup level when the true value of the parameter is on the other side of the
cleanup level. As described in Chapter 3, tie EPA DQO guidance (1994b) explains how the probability of
decision'errors can be controlled by adopting a scientific approach that incorporates hypothesis testing.
for example, following remediation, the decision maker may want to know whether a hazardous
constituent is present at a solid waste management unit (SWMU) at an average concentration that is now
S TASKX\ft£VTSE:vnNAL\MA5TEK Jlpraluc
-------
below the cleanup level. Because the extent of contamination is well delineated and is believed to have
been completely removed, the decision maker may view the consequence of deciding that the average
concentration is greater than the cleanup level when it is actually less than the cleanup level as a more
severe decision error than concluding the concentration is less than the cleanup level when it is actually
greater (for example, more cleanup would be required at significant cost when ail parties believe the
cleanup was successful). The null hypothesis would then be that the average concentration is less than the
cleanup level, A conclusion that the concentration is greater than the cleanup level when it is actually less
would be a false positive error, while a conclusion that the concentration is less than the screening level
when it is actually greater would be a false negative error. The decision maker then sets allowable decision
error probabilities at points below (false positive errors) and above (false negative errors) the cleanup level,
starting at the boundaries of the grey area near the cleanup level where the consequences of errors are
minor. As described in Section 3,1.6, the grey area is bound by the action level and the concentration
where the decision maker wants to begin to control false negative error. For this example, the grey area
extends from the action level to a concentration above the action level where a false negative error rate is
assigned. Error will not be controlled at the concentration range within the grey area, based oil the
minimal consequences of making an error or the expense of collecting enough samples to control error.
Null and alternative hypotheses may be predetermined by regulations. For example, the MTCA cleanup
regulations (WAC 173-340) recommend a confidence interval approach for evaluating compliance that
requires a one-tailed test of the null hypothesis that the true media concentration exceeds the cleanup level.
A tolerable false positive error probability of 5 percent is specified.
Detailed examples of DQO evaluations developed by EPA (EPA 1994b and 1996a) are included in
Attachment C.
7.5 STEP 7: OPTIMIZE THE DESIGN FOR OBTAINING DATA
As noted in Chapter 3, Step 7 includes identifying a resource-effective data collection (sampling) strategy
for generating data that are expected to satisfy the DQOs. The sampling strategy typically will focus on the
sampling design, the sample size, and the analytical methods that will be required to meet the DQOs. A
primary requirement of Step 7 will be to define a statistical method for testing the Step 6 hypothesis and a
sample size formula that corresponds to the statistical method and the sample design. EPA has published
HEP* *lu>lrT*$KrAEVlS£OTKWMA5TOI WPtMSl-MiiJiiowMIrtlBW Mmwbi
7-5
-------
several guidance documents on the selection of appropriate statistical models for determining sample sizes
and sample designs. These documents are listed and briefly described in the following paragraph: similar
statistical models can be used in both corrective action investigations and compliance determinations.
EPA has published severs! guidance documents that specify mathematical models for testing statistical
hypotheses. The previously noted Methods for Evaluating the Attainment of Cleanup Standards,
Volume J: Soils and Soil Media (EPA 1989b) and a follow-up document Statistical Methods for
Evaluating the Attainment of Cleanup Standards, Volume 3: Reference-based Standards for Soils and Soil
Media (EPA 1992i) describe statistical models for testing soil cleanup level attainment hypotheses. EPA
(1989b) describes how to determine whether a mean or upper percentile site concentration is statistically
less than a cleanup standard. The document also describes how the statistical models can be used to
determine the sample size required to meet allowable decision error probabilities. The statistical models
may assume that the data conform to a certain distribution type (for example, normal or lognormal) or that
the data sets being compared (that is, site and background) have equal or unequal variances. Generally, the
variability of constituent concentrations, the tolerable probability of error, and the size of the grey zone will
have the greatest effect on the number of required samples. Guidance on designing a sampling plan is also
presented by EPA (1989b). Parametric and nonparametric tests for comparing facility concentrations to
cleanup levels are presented. The parametric tests are used when the distribution of contamination is
known or assumed to be normal or lognonnal. Otherwise, nonparametric tests (no distribution assumed)
should be used (see Chapter 6 for information on distribution types). EPA (1992i) provides additional
guidance on determining soil cleanup level achievement. This document focuses on two nonparametric
statistical tests and a hot spot measurement comparison in addition to addressing other statistical data
analysis issues, such as treatment of below quantitation limit data. In the more recent Geostatistical
Sampling and Evaluation Guidance for Soils and Solid Media (EPA 1996a) review draft document, EPA
proposes detailed guidance on using average, upper percentile, or hot spot facility data to determine
compliance with cleanup levels. The document outlines sampling plans as well as scenarios and provides
guidance on evaluating decision errors and uncertainty versus sampling costs. Guidance for sampling
design and sample sizes for verifying the cleanup of PCB spills is provided in the EPA Office of Toxic
Substances Verification of PCB Spill Cleanup By Sampling and Analysis (EPA 1985). This document
describes sampling on a hexagonal grid centered on the cleanup area to determine residual PCB
concentrations. The methodology described in this document can be applied to the cleanup of other
constituents released to soils as well.
I REM*milrTASKattEVUEWMALMUTEK 15mm* "7-6
-------
Statistical Analysis of Groundwater Monitoring Data at RCRA Facilities, Interim Final Guidance
(EPA 1989k) and a follow-up addendum (EPA 1992k) describe statistical models for testing groundwater
cleanup level attainment hypothesis. EPA (1989k) describes how to determine whether contamination in a
given well exceeds background or target cleanup level concentrations. Sampling sizes are recommended,
and parametric and nonparametric tests are described. Tie EPA addendum (1992k) provides methods for
determining whether constituent concentrations are normally distributed or whether unequal variances in
constituent concentrations occur between wells. The document also focuses on nonparametric tests (that
is, no distribution is assumed) for comparing compliance well data to background or target cleanup
concentrations and provides recommendations for handling nondetect data. The EPA Statistical Methods
for Evaluating the Attainment of Cleanup Standards, Volume 2: Groundwater (1992h) document also
provides guidance on selecting statistical tests for determining sample sizes.
Data Quality Objectives Decision Error Feasibility Trials (DQO/DEFT) Users Guide, Version 4,0
{EPA 1994c) software package can be used to iterate through one or more DQO steps to identify a sample
design that will meet the budget and generate adequate data. The DEFT software allows the user to
change DQO constraints such as limits on decision errors or the grey region and evaluate how these
changes affect sample sizes (and resulting costs) for several basic sample decisions.
7.6 DATA USEABILITY AND DATA QUALITY ASSESSMENT
The output of the DQO process will be a sampling strategy that defines sampling design, sample numbers,
and analytical methods. Steps should also be taken to assure that the data collected are useable. Section 3.2
of Chapter 3 describes methods that should be followed to assure that the sample data collected are of
acceptable quality to use in compliance decisions, following the data useability determination, the DQA
process briefly introduced in Section 3.3 of Chapter 3 should be performed. The purpose of the DQA is to
verify DQO assumptions, complete statistical comparisons of target cleanup level concentrations with the
levels measured through field sampling, and determine whether compliance with cleanup goals has been
achieved. Guidance for Data Quality Assessment, Practical Methods for Data Analysis (EPA 1996d)
provides specific details on how to perform the DQA.
S REP& ft h«hh TA&M-*EVISE2'FtNAl .MASTER Mmmuc
7-7
-------
7.?
DETECTION LIMITS
Facility constituent concentrations must be determined during either the risk assessment or a compliance
determination. As noted in Section 7.3, the 95th UCL Is typically calculated for risk assessment and
compliance purposes. Other statistical parameters, such as a 95th percentile, may require calculation when
making hot spot or background determinations. Supplemental Guidance to RAGS: Calculating ike
Concentration Term (EPA 1992e) provides a method for calculating the 95th UCL. The State of
Washington's Statistical Guidance for Ecology Site Managers (Washington Department of Ecology 1992)
describes methods for calculating a variety of statistical parameters, including means, median, and
percentiles.
When calculating statistical parameters, environmental data sets often contain samples for which no
constituents have been detected. In this situation, tie only information available on the constituent
concentration is the detection limit. Data sets that contain below detection limit (BDL) data are known as
censored data sets. Such data sets are problematic when used to calculate statistical parameters such as 95th
UCLs because of uncertainty in the actual concentration of the constituent in the BDL samples. Methods for
incorporating BDL data into calculations of average facility constituent concentrations have been
summarized in a quality assurance course module prepared by the National Center for Environmental
Research and Quality Assurance {EPA 19970- The EPA (19970 course module is presented in
Attachment P, and the options are summarized briefly as follows:
•
—
Throw away or otherwise ignore BDL data
m
Set all BDL data at zero
•
Set ail BDL data at the detection limit
m
Set all BDL date at some value (for example,
one-half the detection limit)
•
Use a statistical approach to evaluate BDL
data
% 'M^AtthUMbTASKt-ftEVISEroiNAUMAroR Siwiswe
7-8
-------
As noted in Attachment P, application of the first three methods may result in overestimations or
underestimations of the true mean and variance. If BDL data are not used when calculating statistical
parameters, the true mean may be overestimated and variability may be underestimated. If BDL data are
all set at zero, the true mean may be underestimated and variability overestimated. If BDL data are all set
at the detection limit, the true mean may be overestimated and variability underestimated.
The fourth method, using one-half the detection limit for BDL data, is frequently assumed in risk
assessments. The one-half detection limit method simply estimates a concentration half-way between that
assumed by Method 2 (zero) or Method 3 (the detection limit). When using any of the first 4 methods
(referred to as substitution methods), the resulting bias in estimating mean and variance is small when tie
BDL data make up less than 15 percent of the data set. When the BDL data make up between 15 percent
and 50 percent of the data, however, the biases increase, and statistical approaches such as Cohen's
adjustment, a trimmed mean, or the Winsorized mean and standard deviation cart be applied {EPA 1997f).
Use of the statistical approaches will reduce the bias associated with BDL data and the use of the
substitution methods. Attachment P presents a further description of the substitution and statistical
methods for handling BDL data.
Guidance for Data Quality Assessment, Practical Methods for Data Analysis (EPA 1996d) provides
additional specific guidance on the statistical methods. Risk Assessment Guidance for Superfund.
Volume I. Human Health Evaluation Manual, Part A (EPA 1989c) recommends that sample quantitation
limits be used as the detection limits of first choice when applying BDL methods. If sample quantitation
limits are not available, contract-required or method detection limits should be used.
S PfPAjH «*l*TA5»;K\*E\1SEmNAi .MASTER
7-9
-------
z
n
pi
w
-------
REFERENCES
American Society for Testing and Material (ASTM). 1995a, Standard Guide for Developing
Conceptual Site Models for Contaminated Sites. Standard E-I689. Philadelphia, Pennsylvania.
November.
ASTM. 1995b. Standard Guide for Risk-Based Corrective Action Applied at Petroleum Release Sites,
Standard E-1739. Philadelphia, Pennsylvania, November.
ASTM. 1996. Draft Standard Guide for Probabilistic Risk Assessment and Handling Uncertainties in
Risk Assessments for Contaminated Sites. ASTM, Philadelphia. February.
Aulerich, RJ., R.K. Ringer, J. Safronoff. 1986. Assessment of Primary Versus Secondary Toxicity of
Aroclor 1254 to Mink Arch. Environ. Contain. Toxicol. 15:393-399.
Baes, C.F., R..D. Sharp, A.L. Sjoreen, R.W. Shor. 1914. A Review and Analysis of Parameters for
Assessing Transport of Environmentally Released Radionuclides Through Agriculture. Oak
Ridge National Laboratory. ORNL-5786. September.
California Environmental Protection Agency (Cal/EPA). 1994. Preliminary Endangerment Assessment
Guidance Manual. Department of Toxic Substances Control. January
Cal/EPA. . 1996a. Guidance for Ecological Risk Assessment at Hazardous Waste Sties and Permitted
Facilities - Part A: Overview, Department of Toxic Substances Control. Sacramento, California.
July.
Cal/EPA. ! 996b. Guidance for Ecological Risk Assessment at Hazardous Waste Sites and Permitted
Facilities - Part B: Scoping Assessment. Department of Toxic Substances Control. Sacramento,
California. July.
Cowherd, C.G., G Muleski, P. Engelhart, and D. Gillette. 1985. Rapid Assessment of Exposure of
Particulate Emissions from Surface Contamination. Prepared for U.S. Environmental Protection
Agency, Office of Health and Environmental Assessment, Washington, D.C. EPA/600/8-85/002.
National Technical Information Service (NTIS) PB85-192219 7AS.
Daniel, Wayne W. 1983. Biostatistics: A Foundation for Analysis in the Health Sciences, Third Edition.
John Wiley and Sons, Inc.
Decisioneering, Inc. 1993. Crystal Ball Version 3.0 User Manual. Denver.
Franz, T.J. 1984. Percutaneous Absorption of Benzene. Advances in Modern Environmental Toxicology.
Volume 6, Applied Toxicology of Petroleum Hydrocarbons. Princeton Scientific Publish, Inc.
Pages 61-70.
Gilbert, Richard. 1987. Statistical Methods for Environmental Pollution Monitoring. Van Nostrand
Reinhold Company, Inc.
1 HEP* JUtastrTftSKritEVISI&FMALWASTEft Slnw
R-l
-------
Gilbert, Richard. 1993. Letter Regarding Background Comparison Methods for Rocky Flats Plant, To
Beverly Ramsey, Systematic Management Services, Inc. July 30,
Johnson, P.C., and R.A. Ettinger. 1991. "Heuristic Modelfor Predicting the Intrusion Rate of
Contaminant Vapors into Buildings. " Environmental Science and Technology: Volume 25,
Number 8, pages 1445 -1452.
Kissel, J.C., Richter, K.Y. and Fcnske, R.A. Unpublished. Dermal Soil Exposure: Two Investigation of
Soil Contact and Skin Coverage with the Aid of a Fluorescent Tracer. (Unpublished).
Kissel, J.C., Richter, K.Y. and Fenske, R.A. 1996. Field Measurement of Dermal Soil Loading Attributed
to Various Activities: Implications for Exposure Assessment, Risk Analysis 16: 115-125. b
Lloyd, K.J., K.M. Thompson, and D.E. Burnt aster. 1992. Probabilistic Techniques for Backcalculating
Soil Cleanup Targets In: Superfund Risk Assessment in Soil Contamination Studies, ASTM
STP1158 (Eds. K.B. Hoddinott, American Society for Testing and Materials, Philadelphia).
Long, E.R. and L.G. Morgan. 1990. The Potentialfor Biological Effects of Sediment-Sorbed
Contaminants Tested in the National Status and Trends Program. Technical Memorandum NOS
OMA 52. March.
Massachusetts Department of Environmental Protection (MDEP). 1995. Guidance for Disposal Site Risk
Characterization. Interim Final Policy. WSC/ORS-95-141. Boston, Massachusetts. July.
Mckone, T.E., and K.T. Bogen. 1992. "Uncertainties in Health-Risk Assessment: An Integrated Case
Study Based on Tetrachloroethylene in California Groundwater." Regulatory Toxicology and
Pharmacology, Volume 15, pages 86-103.
Morgan, M. Granger and Max Henrion. 1990. Uncertainty, A Guide To Dealing With Uncertainty in
Quantitative Risk and Policy Analysis. Cambridge University Press.
National Research Council. 1994. Science and Judgement in Risk Assessment, Prepublication Copy.
National Academy Press.
ODEQ. 1996. "Guidance for Ecological Risk Assessment." Draft. November.
Remington, Richard D., and M. Anthony Schork. 1970. Statistics with Applications to the Biological and
Health Sciences. Prentice-Hall, Inc.
Ryan, E.A., et al, 1987. "Assessing Risk From Dermal Exposure at Hazardous Waste Sites." In
Superfund '87: Proceedings of the Eight National Conference: November 16-18; Wash. D.C.,
G. Bennett and J. Bennett (eds). The Hazardous Materials Control Research Institute,
pp. 166-8.
Sigal, L.L. and G.W. Suter II. 1989. Potential Effects of Chemical Agents on Terrestrial
Resources. Environmental Professionals. Volume 11. Pages 376-384.
% R£PA TASK* REVISE: Jr5WAX..MASTtR
R-2
-------
Skowronski, G.A., R.M. Turkall, and M.S. Abdel-Rahman. 1988, Soil Adsorption Alters Bioavailability
of Benzene in Dermally Exposed Male Rats, American Industrial Hygiene Association Journal 49
(10). Pages 506-511.
Smith, Roy L. 1994. Use of Monte Carlo Simulation for Human Exposure Assessment at a Superfund
Site, Risk Analysis, Volume 14, Number 4. Pages 433-439.
Suter 11, G.W. 1993. Ecological Risk Assessment, ISBN 0-87371-875-5, Lewis Publishers. Chelsea,
Michigan.
Tetra Tech. 1996a. Record of Telephone Conversation between Neil Morion of Tetra Tech and Randy
Steger of Idaho Division of Environmental Quality. August 1.
Tetra Tech. 1996b. Record of Telephone Conversation Between Paul Racette of Tetra Tech and Randy
Buckindorf of Alaska Department of Environmental Conservation. October 24.
Thompson, K.M., D.E. Burmaster, and E.A.C. Crouch. 1992. Monte Carlo Techniques for Quantitative
Uncertainty Analysts in Public Health Risk Assessments. Risk Analysis, Volume 12, No. I,
pages 53-63.
Travis, C.C., and A.D. Arms. 1988. Bioconceniration of Organics in Beef Milk, cmd Vegetation. In
Environmental Science and Technology. Volume 22. Pages 271-274.
U.S. Environmental Protection Agency (EPA). 1979. Guidelines and Specifications for Preparing
Quality Assurance Program Plans. QAMS-004/80.
EPA. 1980. Interim Guidelines and Specifications for Preparing Quality Assurance Project Plans.
Office of Monitoring Systems and Quality Assurance. Office of Research and Development.
QAMS-005/80. December.
EPA. 1982. An Exposure and Risk Assessment for Benzo(a)pyirne arid Other Polycyclic Aromatic
Hydrocarbons, Volume IV, Final Draft Report, Office of Water. October.
EPA. 1985. Verification of PCB Spill Cleanup By Sampling and Analysis, Office of Toxic Substances.
EPA-560/5-85-026. August.
EPA. 1986a. Quality Criteria for Water 1986, U.S. Environmental Protection Agency, Office of Water,
EPA440/5-86-001. May 1.
EPA. 1986b. Test Methods for Evaluating Solid Waste. SW-846, Office of Solid Waste and Emergency
Response. November plus updates.
EPA. 1988. Guidance for Conducting Remedial Investigations and Feasibility Studies Under CERCLA,
Interim Pinal. Office of Emergency and Remedial Response. EPA/540/G-89/004. October.
EPA. 1989a. RCRA Facility Investigation (RFI) Guidance, Interim Final. EPA 530/SW-89-0231.
Office of Solid Waste and Emergency Response.
S REP* * i *< * TASK*REVISE:-FINAl'AUSTER WtW*t-R SSaiMU*
R-3
-------
EPA, 1989b. Methods for Evaluating the Attainment of Cleanup Standards. Volume J: Soils and Solid
Media. Office of Policy, Planning, and Evaluation. EPA/230/02-89/042. February.
EPA, 1989c. Risk Assessment Guidance for Superfund, Volume I, Human Health Evaluation Manual,
Part A, Baseline Risk Assessment, Office of Solid Waste and Emergency Response. OSWER
Directive No. 9285.7-01 A. EPA 540/1-19/002.
EPA, 19894 Exposure Factors Handbook. Office of Health and Environmental Assessment.
EPA/600/8-89/043. http://www.epa.gov/ORDZWebPubs/exposure/
EPA. 1989e. Clean Closure Guidance Manualfor Hazardous Waste Management Units,
EPA/530-8W-87-022. August.
EPA. 1989f. Interim Procedures for Estimating Risks Associated with Exposures to Mixtures of
Chlorinated Dibertzo-p-Dioxins andDibenzofurans (CDDs and CDFs) and J989 Update.
EPA/625/3-89/0 J 6. March.
EPA. 1989g. Risk Assessment Guidance for Superfund. Volume 2: Environmental Evaluation Manual.
Interim Final. EPA/540/1-89/001 A. Office of Solid Waste and Emergency Response.
Washington, D.C. March.
EPA. 1989h. Ecological Assessment of Hazardous Waste Sites: A Field and Laboratory Reference.
EPA/600/3-89/013. Office of Research and Development. Corvallis, Oregon.
EPA. 19891- The Nature and Extent of Ecological Risks at Superfund Sites and RCRA Facilities.
EPA-230-03-89-043. Office of Policy Analysis. Washington, D.C. June.
EPA. I989j. Ecological Risk Assessment of Superfund Sites: An Overview. ECO Update. Intermittent
Bulletin. Volume 1, Number 2. Office of Emergency and Remedial Response. Washington, D.C.
EPA. 1989k. Statistical Analysis of Ground-Water Monitoring Data at RCRA Facilities, Interim Final
Guidance. Office of Solid Waste. April.
EPA. 1990. Methodology for Assessing Health Risks Associated with Indirect Exposure to Combustor
Emissions, Interim Final. Office of Health and Environmental Assessment. EPA/600/6-90/003.
January.
EPA. 1991a. EPA Region 10 Supplemental Risk Assessment Guidance for Superfund. August.
EPA. 1991 b. Risk Assessment Guidance for Superfund. Volume I: Human Health Evaluation Manual.
Supplemental Guidance: Standard Default Exposure Parameters. OSWER directive 9285.7-05.
EPA. 1991c. Risk Assessment Guidance for Superfund, Volume /, Human Health Evaluation Manual,
Part B, Development of Risk-based Remediation Goals. Office of Solid Waste and Emergency
Response. OSWER Directive No. 9285.7-01 B. EPA/540/R-92/003,
EPA. 199 Id. Evaluation of Dredged Material Proposed for Ocean Disposal, Testing Manual. Office of
Water. EPA/503/8-91/001. February.
s mfh ftimiiffTASK»«EVlSEOTNAL\MA5TEII WFOUJI-fttwiiiifliifMirtl/WV $}*jn\uc
-------
EPA, 1992a. Interim Final Guidelines for Developing Risk-Based Cleanup Levels at RCRA Sites in
Region 10, EPA Region 10. March,
EPA. 1992b, Final Guidance for Data Useability in Risk Assessment (Part A). Office of Solid Waste ami
Emergency Response. Publication 9285.7-09A/FS. May,
EPA. 1992c. Exposure Assessment Guidelines. Federal Register, May 29, 1992. 57 FR 22888-22931.
EPA. I992d. Dermal Exposure Assessment: Principles and Applications. Office of Health and
Environmental Assessment. EPA/600/8-91/01 IB. January.
EPA. !992e. Supplemental Guidance to RAGS: Calculating the Concentration Term, Office of Solid
Waste and Emergency Response. Publication 9285.7-081.
EPA. !992f. Framework for Ecological Risk Assessment. EPA/630/R-92/001. Risk Assessment Forum.
Washington, D.C. February.
EPA. 1992g. Guidance for Data Useability in Risk Assessment (Part B). Publication number
9285.7-09B. Office of Emergency and Remedial Response. Washington, D.C. May.
EPA. 1992b. Statistical Methods for Evaluating the Attainment of Cleanup Standards. Volume 2:
Groundwater. Office of Policy, Planning, and Evaluation.
EPA. 1992i. Statistical Methods for Evaluating the Attainment of Cleanup Standards. Volume 3:
Reference-Based Standards for Soils and Solid Media. Office of Policy, Planning, and
Evaluation. EPA/230/R-94/004. December.
EPA. 1992j. Guidance on Risk Characterization for Risk Managers and Risk Assessors. Memorandum
from F. Henry Habicht 11, Deputy Administrator. Office of the Administrator. Washington, D.C.
February.
EPA. 1992k. Statistical Analysis of Ground-Water Monitoring Data at RCRA Facilities, Addendum to
Interim Final Guidance. Office of Solid Waste. July.
EPA. 1993a. Supplemental Guidance to RAGS: Estimating Risk from Groundwater Contamination.
Office of Solid Waste and Emergency Response. December.
EPA. 1993b. Quality Assurance Project Plans for RCRA Ground-Water Monitoring and Corrective
Action Activities. July.
EPA. 1993c. RCRA Public Involvement Manual. EPA 530-R-93-006.
EPA. 1993d. Addendum to Methodology for Assessing Health Risks Associated with Indirect Exposure to
Combustor Emissions, External Review Draft. Office of Health and Environmental Assessment.
EPA/600/AP-93/003. November.
EPA. 1993 e. Provisional Guidance for Quantitative Risk Assessment of Polycyclic Aromatic
Hydrocarbons. EPA/600/R93/089. July.
1- REPA R i<» i s TA k*AFVT,ErJtNAt\MASTER
-------
EPA, J993f. Wildlife Exposure Factors Handbook - Volume 1 of 2, EPA/600/R-93/187a. Office of
Research and Development. Washington, D C. December.
EPA, 1993g. Wildlife Exposure Factors Handbook - Appendix: Literature Review Database - Volume 2
of 2. EPA/600/R-93/187b. Office of Research and Development. Washington, D.C. December.
EPA. 1993h. Superfund's Standard Default Exposure Factors for the Central Tendency and Reasonable
Maximum Exposure.
EPA. 1994a. RCRA/CERCLA Interface-Interim Final Guidance. EPA Region 10 Memorandum.
August 3.
EPA. 1994b. Guidance for the Data Quality Objectives Process. EPA QA/G-4. Quality Assurance
Management Staff. Washington, D.C. September.
http://www.epa.gOv/r 1 Oearth/offices/oea/aaindex.htm
EPA. 1994c. Data Quality Objectives Decision Error Feasibility Trials (DQO/DEFT) Users Guide,
Version 4.0. EPA QA/G-4D. September.
http://www.epa.gov/rlOearth/offjces/oea/aaindex.htm
EPA. 19944 EPA Requirements for Quality Assurance Project Plans for Environmental Data
Operations, Draft Interim Final. EPA QA/R-5. August.
http://www.epa.gOv/r 1 Oearth/offices/oea/aaindex.htm
EPA. !994e. Requirements for Quality Management Plans (QA/R-2). August.
http://www.epa.gOv/r 1 Oearth/offices/oea/aaindex.htm
EPA. !994f. USEPA Contract Laboratory Program National Functional Guidelines for Organic Data
Review Publication 9240.1-05. February.
http://www.epa.gOv/r I Oearth/offices/oea/aaindex.htm
EPA. 1994g. USEPA Contract Laboratory Program National Functional Guidelines for Inorganic Data
Review. Publication 9240.1-05-01. February.
http://www.epa.gov/rlOearth/offices/oea/aaindex.htm
L I'A 1994h. Region 8 Superfund Technical Guidance RA-03: Evaluating and Identifying Contaminants
of Concern for Human Health. Region 8, Denver. September.
EPA. 19945, Guidance for Performing Screening Level Risk Analyses at Combustion Facilities Burning
Hazardous Wastes. Attachment C. Draft Exposure Assessment Guidance. Office of Emergency
and Remedial Response. December 14.
EPA. 1994j. Estimating Exposure to Dioxin-Like Compounds. Office of Research and Development.
EPA/600/6-88/005B. http://www.epa.gov/docs/exposure/
EPA, 1994k. Memorandum Regarding Revised Interim Soil Lead Guidance for CERCLA Sites and
RCRA Corrective Action Facilities. From Elliott P. Laws, Assistant Administrator. To Regional
Administrators I-X. July.
i 'SEPWHiwImTMKrWVtSgWIWiUMASTl* Wm>lSl«H«»»»tnWlJBIfWf SSamw R"6
-------
EPA. 19941 Ecological Risk Assessment Guidance for Superfund: Process for Designing and
Conducting Ecological Risk Assessments, Review Draft, Environmental Response Team. Edison,
New Jersey. September.
EPA. 1994m. Region 8 Superfund Technical Guidance: Use of Monte Carlo Simulation in Performing
Risk Assessments. Region 8, Denver. December.
EPA. 1995a. EPA Risk Characterization Program, Memorandum from Carol Browner, EPA
Administrator. March 21.
EPA. 1995b. Region 5 Ecological Data Quality Levels. Draft Final, Prepared for the Office of RCRA by
PRC Environmental Management, Inc. Chicago, Illinois. September.
EPA. 1995c. Determination of Background Concentrations of Inorganics in Soils and Sediments at
Hazardous Waste Sites. Office of Solid Waste and Emergency Response. EPA/540/5-96/500.
December.
EPA. 1995d. Land use in the CERCLA Remedy Selection Process, Memorandum from Elliott P. Laws.
Office of Solid Waste and Emergency Response Directive No. 9355.7-04. May 25.
EPA. 1995e. New Policy on Evaluating Health Risks to Children. Office of the Administrator.
October 20.
EPA. I995f. Region III Technical Guidance Manual, Risk Assessment: Assessing Dermal Exposure from
Soil. Office of Superfund Programs. December.
http://www.epa.gov/reg3hwmd/riskmenu.htm
EPA. 1995g. Memorandum Regarding Further Issues for Modeling the Indirect Exposure Impacts from
Combustor Emissions. From Matthew Lorber, Exposure Assessment Group and Glenn Rice,
Indirect Exposure Team, Office of Research and Development. To Addressees. January 20.
EPA. I995h. Supplemental Guidance to RAGS: Region -i Bulletins, Human Health Risk Assessment.
Waste Management Division. November.
EPA. 1996a. Geostatistical Sampling and Evaluation Guidance for Soils and Solid Media, Review Draft.
Office of Solid Waste. February.
EPA, 1996b. Soil Screening Guidance: User's Guide. Office of Solid Waste and Emergency Response.
Publication 9355.4-23. April, http://www.epa.gov/superfund/oerr/soil/index.html
EPA. 1996c. Region IX Preliminary Remediation Goals (PRGs) 1996. August 1.
http://www.epa.gov/region09/waste/sfund/prg/index.html
EPA 1996d. Guidance for Data Quality Assessment. Practical Methods for Data Analysis, EPA QA/G-9,
QA96 Version, EPA/600/R-96/084. July.
http://www.epa.gov/rl Oearth/offkes/oea/aaindex.html
D 7
S REfAUBMIETASKMEVlSK-.flNAUMASTS* WMSI-ittmltwlWrtMil Ijmmat ,v" '
-------
EPA. 1996e. Exposure Factor Handbook, Internet Download, EPA/600/P-95/002A. August,
EPA. 1996f. Supplemental Risk Assessment Guidance for Superfund. Draft, EPA Region 1 § Risk
Evaluation Unit Seattle, Washington. February.
EPA, 1996g. Integrated Risk Information System (IRIS). On-line database.
http://www.epa.gov/ORD/dbKes/iris/ (monthly updates)
http://www.epa.gov/ngigpm3/iris/index.htnil (web prototype)
EPA. 1996h. Risk-Based Concentration Table, January-June 1936. U.S. Environmental Protection
Agency, Region 111. January,
http ://www.epa. gov/reg3 h wmd/riskmenu.htm
EPA. 1996i. Proposed Guidelines for Carcinogen Risk Assessment. Office of Research and
Development. EPA/600/P-92/003C. April.
http://www.epa.gov/ORD/WebPubs/carcinogen/
EPA. 1996j. PCBs; Cancer Dose-Response Assessment and Application to Environmental Mixture,
National Center for Environmental Assessment. Office of Research and Development.
EPA/600/P-96/001A. September.
http://www.epa.gov/ORD/WebPubs/pcb/
EPA. 1996k. Recommendations of the Technical Review Work Group for Lead for an Interim Approach
to Assessing Risks Associated with Adult Exposures to Lead in Soil. December.
EPA. 19961. Proposed Guidelines for Ecological Risk Assessment. Published in CFR Volume 61,
Number 175. Risk Assessment Forum. Washington, D.C. EPA/630/R-95/002B. September.
http://www.epa.gov/ORD/WebPubs/ecorisk/
EPA. 1997a. Health Effects Assessment Summary Tables (HEAST), FY 1997 Update. Office of Solid
Waste and Emergency Response, EPA-540-R-97-036. NTIS No. PB97-921199. July.
EPA. 1997b. Risk Assessment Guidance for Superfund Volume I: Human Health Evaluation Manual,
Supplemental Guidance, Dermal Risk Assessment Interim Guidance. Office of Emergency and
Remedial Response. June 19.
EPA. 1997c. Risk Assessment Forum Review of "Guidelines on Health Risk Assessment of Chemical
Mixtures NCEA-C-0148. June 20.
EPA. 1997d. Guiding Principles for Monte Carlo Analysis. EPA/63Q/R-97/001. March.
http://www.epa.gov/ncea/mcpolicy.htm
EPA. I997e. Policy For Use of Probabilistic Analysis in Risk Assessments at the U.S. Environmental
Protection Agency. Office of Research and Development. May 15.
http://www.epa.gov/ncea/mcpolicy.htm
EPA. 1997f. Quality Assurance Course Module 492, Dealing with Data Below Detection Limits,
National Center For Environmental Research and Quality Assurance. Downloaded March.
5 REMMBM-WSKrjtetfBOWNMJMMTE* WWM5MlmittoMlltJBwm IMm
-------
EPA. 1997g. Rules of Thumb for Superfund Remedy Selection. OSWER 9355.0-69. August.
EPA. 1997h. The Role of CSGWPPs in EPA Remediation Programs, OSWER Directive 9283.1-09.
April 4.
Washington Department of Ecology (Ecology). 1992. Statistical Guidance for Ecology Site Managers.
Washington State Department of Ecology, Toxics Cleanup Program. August.
Ecology. 1994. Guidance for Clean Closure of Dangerous Waste Facilities, Publication No. 94-! 11.
August.
Ecology, 1995a. Washington State Comprehensive State Groundwater Protection Program, Core
Program Assessment Document, July.
Ecology. 1995b. Summary of Guidelines for Contaminated Sediments. WDOE publication 95-308.
March.
Ecology. 1996. Model Toxics Control Act Cleanup Levels and Risk Calculations (CLARC11} Update.
Publication No. 94-145. February.
Wester, R.C., H.I. Maibach, D.A.W. Bucks, L. Sedik, J. Melendres, C.L. Laio, S. DeZio. 1990.
Percutaneous absorption of [14CJDDTand [I4C]benzo(a)pyrene from soil Fund. Appl.Toxicol.
15:510-516.
Wester, R.C., H.I. Maibach, L. Sedik, J. Melendres, S. DeZio, M. Wade. 1992a. In vitro percutaneous
absorption of cadmium from water and soil into human skin. Fundam. Appl. Toxicol. 19:1-5.
Wester, R.C., H.I. Maibach, L. Sedik, J. Melendres, C.L. Laio, S. DeZio. 1992b. Percutaneous
absorption of [14C]chlordane from soil. 1. Toxicol. Environ. Health 35:269-77.
Wester, R.C., Hi. Maibach, L. Sedik, J. Melendres, M. Wade. 1993a. In vivo and in vitro percutaneous
absorption and skin decontamination of arsenic from wafer and soil. Fundam. Appl. Toxicol.
20:336-40.
Wester. R.C., H.l. Maibach, L. Sedik, J. Melendres, M. Wade. 1993b. Percutaneous absorption of PCBs
from soil: in vivo Rhesus monkey, in vitro human skin, mid binding m powered human stratum
corneum. J. Toxicol. Environ. Health 39:375-82.
Wester, R.C.. H.l. Maibach, L. Sedik, J. Melendres, M. Wade, S. DeZio. 1993c. Percutaneous absorption
of pentachlorophenolform soil. Fundam. Appl. Toxicology 20:68-71.
Wester. R.C.. J. Melendres, F. Logan, X. Hui, H.I. Maibach. 1996. Percutaneous absorption of
2.4-dichlorophenoxyacetic acid from soil with respect to the soil load and skin contact time: in
vivo absorption in Rhesus monkey and in vitro absorption in human skin. J. Toxicol, Environ.
Health 47:335-44.
^ K! ?'**<««»»» ?A*KR 8£VlSE2.f$NAl MASTER WPJKJ Sl.ftioriieottxr
R-9
-------
Wentsel, R.S., T.W. LaPoInt, M. Simini, D, Ludwig, and L. Brewer. 1994. Procedural Guidelines for
Ecological Risk Assessments at U.S. Army Sites - Volume 1. Report Number ERDEC-TR-221.
Aberdeen Proving Ground, Maryland. December.
5 REPAIR t»!U"TA5K*\REV15EJ"JrIMAt.\MASTER WPOU5l
-------
INDEX
Name Page
ABS (see dermal absorption factors)
Alaska Department of Environmental Quality (ADEC) 2-2
Air ................... 4-9,4-12,4-13,4-14,4-17,4-20,4-27,4-28,4-30 thru 4-35,4-72.4-78, 7-2
Ambient water qualify criteria 4-13, 4-14, 5-15, 5-17
Analytical methods 3-8,3-11,4-70, 6-6, 7-5, 7-7
AWQC . {see ambient water quality criteria)
Cancer potency factor 4-40 thru 4-53,4-68,4-69
Combustion [[[ 4-22, 4-37, 4-44,4-45,4-72
Conceptual site model [[[ 3-3, 5-2, 5-9, 5-18
Constituents of potential concern (COPCs).................... Chapter 1, 3-1, 3-14, 4-1, 4-15, 4-43
COPCs (see constituents of potential concern)
CPF (see cancer potency factor)
Data quality assessment (DQA).............. 3-13, 7-1, 7-7, 7-9,
Data quality objectives (DQO) 3-1,3-7, 3-8, 5-22, 7-1, 7-4, 7-7
Data useability ... 1-5, 3-9, 4-75, 7-7
Decision errors .................................................. 3-6,3-7, 3-8, 7-4, 7-6, 7-7
Decision role [[[ 3-2, 3-5,3-6, 7-1, 7-3
Dermal absorption factors .................................... 4-23,4-24,4-25,4-62 thru 4-67
Dermal permeability 4-23, 4-25
Detection Limits [[[ 3-11, 7-8, 7-9
Dioxin 4-23, 4-24, 4-38,4-46,4-47, 4-48,4-51, 4-53
Ecological adversity[[[ 5-26
Ecological risk assessment .......................... 1-6,1-8, Chapter 5, 6-1,6-8, Attachment N
Ecology (see Washington State Department of Ecology)
Exposure assessment .... ...................................... 4-1,4-2, 4-22,4-25, 4-73
Exposure assumptions .............................. 1-7,4-14,4-17,4-22, 4-60,4-76,4-77, 5-16
Exposure pathways 4-1,4-2, 4-3,4-5,4-16, 4-17,4-20,4-22,4-25, 4-36,
4.37,4-44, 4-60, 4-65,4-72,4-74,4-76, 5-1 thru 5-18, 5-29, 6-11, 6-12, 6-13, 6-31, Attachment N
Fate and transport 3-4,4-19,4-25,4-27, 4-36,4-57,4-77, 5-5, 5-9, 5-16,
5-20, 5-29, 6-9, Attachment N
Groundwater 3-5,4-2, 4-3,4-4,4-6 thru 4-11,4-13, 4-14,4-25,4-27,
4-34, 4-35, 4-36,4-53,4-60,4-65,4-67,4-68, 4-70,4-71,4-76,4-77, 5-9, 6-12, 7-2, 7-3, 7-7
Hazard index ........................... 4-15,4-16,4-42,4-43, 4-57,4-63, 4-69, 4-71, 5-26, 6-11
Hazard quotient .................................. 3-11,4-18,4-21,4-42, 4-59,4-71, 5-23, 5-25
Henry's Law [[[ 4-28, 4-30, 4-33
HHRA ................................................. (see human health risk assessment)
HI , (see hazard index)
-------
INDEX
Mme * Page
Inhalation reference concentrations , „... > 4-42,4-43,4-44
Land use 1-6,2-1,3-3,4-5 thru 4-9,4-13,4-14,4-58,4-65,4-74,4-75, 4-76
Lead .................. 2-1,4-22,4-38,4-57,4-73, 4-74,4-75, 5-19,6-13, 6-17, 6-18, Attachment N
Lowest adverse effects level (LOAEL) » 4-78, 4-79, 5-22
Maximum contaminant levels (MCLs) .4-10,4-13, 4-14.4-70,4-71
Measurement endpoints 5-2, 5-13, 5-14, 5-25, Attachment N
Model Toxics Control Act 2-3,4-6,4-7,4-10 thru 4-14, 4-21,4-22,4-36,4-57,4-60, 7-2 thru 7-5
Monte Carlo simulation , 5-29, Chapter 6
MICA (see Model Toxics Control Act)
No observed effect level (NOAEL) 4-78,4-79. 5-4, 5-15, 5-21, 5-22, 5-24
Oregon Department of Environmental Quality (ODEQ).,,,............ 2-2,4-8,4-10,4-15, 5-1, 5-6
PAHs ., (see polycyclic aromatic hydrocarbons)
Particulate emission factors ................,................ 4-17,4-19,4-27,4-29,4-30, 4-34
PCBs (see polychlorinated biphenyls)
PEF . (see particulate emission factors)
Polychlorinated biphenyls 4-24,4-46,4-47, 4-48 thru 4-53,4-72
Polycyclic aromatic hydrocarbons 4-24,4-45,4-54,4-73, Attachment N
Preliminary remediation goals (PRG) 3-11,4-15, 4-17,4-25,4-59, 4-60
Problem formulation 1-6, 5-2, 5-4, 5-6, 5-9, 5-10, 5-15, 5-16, 5-22, 5-27
Quality assurance project plan (QAPP) 3-10,3-15
Quantitation limits 3-11, 3-13,3-16,4-75,4-76, 7-9
RJEJCs (see risk-based concentrations)
KCRA facility investigation (RFI)............. 3-1, 3-3, 3-4,4-1,4-27.4-57,4-65, 5-4, 5-5, 5-22, 7-1
Reference Dose ................ 4-41 thru 4-44,4-53,4-56,4-59,4-66, 4-68,4-69,4-73,4-79, 6-11
Rrcs (see inhalation reference concentration)
KfD . .., ,... (see reference dose)
RFI (see RCRA facility investigation)
Risk characterization .................... 1-9, 3-15,4-1,4-2,4-43, 4-44,4-45,4-57,4-58, 5-1, 5-2,
5-22 thru 5-27,6-1, Attachment N
Risk management..... 1-4, 1-10, 1-11,4-7,4-44,4-60, 4-74, 5-2, 5-15, 5-24, 5-25, 5-27, Attachment N
Risk-based screening Level (RBSL) 1-8, 4-15, 4-17
Risk-based concentrations ...... Chapter I, 3-4, 3-14, 4-1,4-2,4-14,4-15,4-17,4-30,4-37, 4-38, 4-39,
4-42,4-57,4-59,4-60,4-66 thru 4-69,4-71 thru 4-75, 0-21, 6-26, 6-27
Safe Drinking Water Act 1-3,4-10,4-13
Soil cleanup levels .................................................. 4-5, thru 4-8, 4-14,7-2
Solid waste management units 3-3,3-5, 3-6,3-7, 5-3, 7-4
Surface water 4-2,4-3, 4-5 thru 4-9,4-11 thru 4-14,4-36,4-68, 5-9, 7-2
SWMU (see solid waste management units)
2,3,7,8-Tetrachlorodibenzo-p-dioxin (see dioxin)
1-2
-------
INDEX
Mimf Eagfi
Total petroleum hydrocarbons 4-55
Toxicity reference values (TRV) 5-4, 5-15, 5-21, 5-23, 5-24, 5-25
TPH (see total petroleum hydrocarbons)
TRV (see toxicity reference values)
Uncertainly .......... 1-3, 1-9, 1-10,4-40 thru 4-43.4-49,4-56, 4-74 thru 4-80, 5-6, 5-25, 5-26, 5-27,
5-29, 5-30, 6-1, 6-2, 6-5 thru 6-9,6-13, 6-19, 6-26, 6-31, 7-4, 7-6, 7-8, Attachment N
VF [[[ (see volatilization factor)
Volatile organic compounds (VOC) 4-23, 4-27,4-66, 4-71, 4-77
Volatilization factor 4-27,4-28,4-30, 4-34,4-66,4-67,4-68
Washington Department of Ecology (Ecology) ....... 2-3,4-8 ,4-10,4-13,4-14,4-21,4-36, 5-17, 7-8
-------
>
p
o
PI
-------
ATTACHMENT A
U.S. ENVIRONMENTAL PROTECTION AGENCY
CAROL BROWNER MEMORANDUM ON RISK ASSESSMENT
% REP A .ft M*»i R'TASKR^REVlSErnNAL^MASTf R WPtftlSI'RtmtlltilwftHJ/WJftl fcjpmtoc
-------
* mm ±
I ^22^ ° UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON. DC, 20460
TKf AOMtNISTRATO*
HAR 2 113S5
RECEIVED
MAR 2 8 1995
SUBJECT: EPA Kiik Characterization Program - ' OFFICE OF
rn- REGIONAL ADMINISTRATOR
Asaociate Athwart ratocs
Regfcaal Administrators
General Counsel
Inptctor
MaafaopepeopieaRnMof
eoviraimi^il awes tod^ than m die pit and their lew! of and irterest in
iiiieiiaiiivAeieiiiinaimmtokiitt W®iowff^ffitt»i»pia©§f»IBcfci$k3t
oafyinteR8ledmlaiowiiigiidialEPAtlmdEidxKit&paiticiibriiiD^I)iitalsolwwwB60)iBet~
our coodusons.
\ ,
Muni and mrwi fay ^ripehnllfim in «»fW n# mn»«h «rfhwtM*irw> tn
aflow them to pfepcoJerftf ami and aiakBjiiilpBenliaboat the ij^ScM»of erawoomeirtil
lftivaRtoaooeeed.aBd build our
credibility and stature a» i leader m mmmmtMd. protection fbr the next centuiy»EPA must be
responsive and resolve to «» openly and fc% co^ninicite to the pAfie die cempfaoMes and-¦
dbfcuges, of comronnieatal itomoiaiailoig ia Hm ftcc ©Ctae^ic vuertaintr.
AtknsweflitttaiDiimco^l^pcoideloiinMfeuiiiiMeofGPAiiiisl
hettjerimriamiMrithaliaAffarflmdaciaioaSf as wdl as our confidence in the data, the scicace
polcyjuclfpi^toifefcwemiie,,Mi theincatiiMf imilieMifmrfenljis©. liorto'toacKewe
~hi< Artier undentandsogi we must in^ww® the way in wydj we claiiirtrfze aid communicate
awranmertal risk Wc must embrace c«tak fundamental values
-------
-2-
I
* so that we nqr bepn tic process of changing the way in wKch wc interact with each other, the
- public, and key stakeholders oo environmental ride issues. I need your Wp to ensure that these
values are embraced and Am -we change the w»y we do business.
Warn," we must adopt as values transparency in oar liwnoiilig process and cfaurity ®
comnuniGalion with cadi other and the pile RgBdmg amDDaiil * and the laicxitainfies
This menu that we must iliBy, ©paly,
1^1 etejHy rihaara/ierhae mint In Aiiflg so, we wiB disdoie the idiarfiic, anafysea, imwlaMies,
aasnmptiotti and adence'poiiciet wMA uotel© a* dedAns aathcy am made tinxgliaiit the
ritlr MttmMrt «nJ risk ntanagemaftf pfociMSi. 1 want to be surethat key science pofiqf issues
are identified jnavdi during the ri&tasesanent process that poficgrnrioMiafBfidljrnvan and
eqg«ged in theteiedkn of adencepoficy option* and that thdr choices and thniatknde for-
those cMces w daufcr acticidated and vMMe in oar ce^mnictftoni about eoiiMpertal risk
- - -I undentand that aone may be concerned dm# additional chaBaqges and dbpotefc. I ,
However, I atroqBfarbeBam that
imMm thk chmnem to m. more open dbarinmnakiag process wiB lead to -«QH8- niaaninttfiil cubic
paiticipatiiML letter nJbmatioa fir iediiQOBialnift mpmmi i/mmm, andjnom puISc support
aadrapectftrEPApoaitiaiisaDdttedriani Themvdnnadiariiviiiilioilieathn
conptaiiiea and dalteqge* we then in oalriqgdedawiit in the ftcnof uncertainty. IwetDsidig
this change at eaaaiitid to the long tans ancceaa of tHsAgency. • • .
• . OiAy in coBMmi^lQa tls© neias that
im«i«w«im««irf«dfiwri^p«y«MeaBCiHiawhnnw»ta8»lidcnianaaBBBantactio«L Wemst
' tifothdlminm miA find legitimate waweto heto the pubBc better coapwhaad thn idatfae
¦ : -. * , -
gMQiMf ^ barrage (^liptnaqr In deriakwmakktg and clarity in cogMmicMiott wM Ikciy
leed to more outtide questioning ofourawnnptiona and adeneepolide* we awst be morn •
v»on»n> about enswg. Ait our core assumptions and science policies aze consistent sod
comparable WW programs, weS grounded ia adence, and that'they M wM»® a "zone of
reasonnbieBeas." - ¦
-------
WMle I believe thai the American public eqpects us to or on the side of protection in the fee of
scientific uncertainty, I do not want our assessnents to be unrealsticallj cooservrfve. We cannot
lead the figte for envfemnental protection ta© the next century unless we use common in
all we do.
These mm values of transparency, clarity, consistency, and reasonableness need to guide
each of us in our day-to-day woric; from the toadcologist rsviewing the individual cancer study, to
the exposure and risk assessors* to iheri& anger,, and tfarou^ to the uhunatededsionmakBr. I
Yon need to Wicw m tie
. Importance of this change and convey your beleS to your managere and staff through your words
aiiiiciommorfalbfthecliigBtoocciif. Yonalseikeeito play aataagrriiofe in developing
the implementing po&des and procedures for your programs.
I am issuing the attadari EPA lU&CfaaiacteriatfkMPoBcjr and Guidance today. Ivlew
' these documents as tanMIng blocks for the development of your prograii-speoic policies ami
procedures. Tte Science Policy Counc3(SPC) plans to adopt the same basic approach to -
llat is, the Counl unl lorn * Advisoiy Group
wffl woA wftfc • IroM T«m ^ rf i«i «wy
CMfice and Rcgioalfiach Program Office and each Region wiB be asked by the Adviscxy Group
to develop propwt and repoo-speclie policies and proceiiHW ft* risk cfcaiiclefiaiien
consisted wfthtte,vafaes€>£ttm«p«a^»cl««%f€an»iit«n^ft»ii«»M»«liiieB«if and ¦
coosisteot with Ae tttacfaecl polcy and guidance.
I recognize that as you develop yOTPiognm-specifopcrficies and procedures you-are
Bkdy to need aidittoaal tools to i^pieme* this policy. I want ym to identify these needed
tools mi I warty*
drat popam and regtoMpeofic poSdqt, procedures, and mgilaKotaiion plaw to be developed
and toibaiittedlo the Adris^Gi^-ftr review by no later than May 30; 1995. You will he
cortacteishor%lf fc Sit Steering Committee to ottata the names of jour nominees to the
Implemeotatiom Team.
Attachments
-------
March 1995
POLICY FOR RISK CHAEACTIIIZATION
at the U.S. Environmental Protection Agency
INTRODUCTION .
Many EPA policy decisions are based in part on the results of risk assessment, an
analysis of scientific information on existing and projected risks to human health
arid the environment As practiced at EPA, risk assessment makes use of many
different kinds of scientific concepts and data (e.g., exposure, toxicity, epidemiology,
ecology)/ all of which are used to "dMacterize" the expected riak'assodated with a
parfioilar agent or action in a particular environmental context Wormed use of
relaMe "scientific information from many different sources is a central feature of the
risk assessment process.
RelaMe information may or may not be available for many aspects of a risk ¦
assessment.. Sdentffic uncertainty is a fact of life for the risk assessment process, and
agency ¦tramagera almost always must mate decisions using assessments, that are not
as definitive In all important areas as would be destraMe. They therefore need to .
understand the strengths and the limitations of each assessment, and to .:
communicate ttus information to all participants and the public
i . ¦
IMs policy reaffirms the principles and guidance found in the Agency's 1992 policy
(Guidance on Msk Characterization for Mslt Managas and Risk Assessora* February
26,1992). That guidance was based on EPA's risk 'assessment guidelines, which are
products of peer review and public comment. The 1994 National Research Council
(MRQ report, "Sdence and Judgment in Risk Assessment,* addressed the Agency's -
approach to risk assessment, including the 1992 risk characterization policy. The
' NRC statement acampanjntg the report skied,"... EPA's overall approach to ,
assessing risks is fundammfaJty sound despite ofte-heart, .criticisms, but the
Agency must more clearly establish the scientific and policy basis for risk estimates
and better describe the uncertainties in its estimates of risk."
This policy statement and associated guidance for risk characterization is designed to
ensure that critical information from etch stage of a risk assessment is used in
forming condtusions about risk'and that this inibimatioin is. communicated from
risk assessors to risk managers (policy makers), torn middle to upper management,
and from the Agency to the public. Additionally, the policy will provide a basis for
greater clarity, transparetey, reasonableness, aid consistency in risk assessments
across Agency programs. While most of the discussion and examples in this policy
are drawn from health risk assessment, these values also 'apply to ecological risk
assessment. A parallel effort by the Risk Assessment Forum to develop EPA
ecological risk assessment guidelines will include guidance specific to ecological risk
characterization.
-------
fnligy Statement
Each risk assessment prepared in support of decision-making at EPA should
¦ include a risk characterization that follows the principles and reflects the values
outlined In this policy. A risk characterization should be prepared in a manner that
is clear, transparent, reasonable arvd consistent with other risk characterizations of
'• similar scope prepared across programs in the Agency. Further, discussion of risk in
all EPA reports, presentations, decision packages, and other documents should be
. substantively consistent with the xisk characterization. The nature of the risk
" characterization will depend upon the information available, the regulatory,
application of the risk information, and the resources (including time) available. In
all cases, however, the assessment should identify and discuss all the major issues
' associated with determining the nature and extent of the risk and provide ¦ ••
commentary on any consteaMte Uniting fuller exposition* - - -
Kj^LAracf^itBlriLChaadBfetfian / . ' ¦ / " •
ir
- Bridging risk assessment and risk management.,/As risk
assessment and risk management, risk characterizations should be clearly presented,
and separate from any risk managwifflt conad«atioiis. Msk management options
should be developed using the risk characterization and should be based on
consideration of all relevant factors, scientific aid nonscientific.
« s
- . • Discussing confidence and uncertainties. Key scientific conicepts, data and
methods (e.g., use of animal or human data for extrapolating from, high to low
doses/use of pharmacokinetics data, exposme pathwayv'sampliag methods, ' .
availability of chemical-specific information, quality of data) shoul&be discussed.
To ensure transparency, risk characterizations should include a statement of
- ' confidence urthe assessment 'Out identifies all major isieoSafatjes along with
comment on "their influence o» the assessment, consistent with the' Guidance on
Risk Characterization (attached). • -
Presenting several types of risk information. Information should be
presetted on the tinge of exposures derived from ©fjosmi scenarios and on the use
of multiple risk descriptors (e.g., central tendency, higfrend of individual risk, .
population risk, important subgroups, if known) consistent with terminolojgy m the
Guidance on Risk Characterization, Agency risk, assessment guideline, and
program-specific guidance. In decision-making, risk managers should use nsk
- information appropriate to their program legislation. , • • ,
EPA conducts many types of risk assessments, including screening-level
assessments of new chemicals, in-depth assessments of pollutants such, as dioxin
2
-------
and environmental tobacco smoke, and site-specific assessments for hazardous
waste site. An iterative approach, to risk assessment, beginning with screening
techniques, may be used to determine if a more comprehensive assessment is
necessary. The degree to which confidence and uncertainty are addressed in a risk
characterization depends largely on the scope of the assessment. In general, the
scope of the risk characterization should reflect the information presented in the
risk assessment and program-specific guidance. When special drcumstarices (e.g.,
lack of data, extremely complex situations, resource limitations, statutory deadlines)
preclude a full assessment, such circumstances should be explained and their impact
on the risk assessment discussed.
Risk Characterisation in Context . .
Risk assessment is based on a series of questions that the assessor asks about
scientific information that is relevant to human and/or environmental risk. Each.
question rails for analysis and interpretation of the available studies, selection of the
concepts and data that are most scientifically reliable and most relevant to the'
problem at hand, and scientific conclusions regarding the question presented. For
example, health risk assessments involve the following questions:
Hazard Identification - What is known about the capacity of an environmental
agent for causing cancer or other adverse health effects in humans, laboratory
animals, or -wildlife species? What are the related uncertainties and science *
policy choices? . " , .
Dose-Response Assessment — What is known about the biological mechanisms
- and dose-rwponse relationships underlying any effects observed in the laboratory
or epidemiology studies providing data for the assessment? What are the
• related uncertainties and science policy choices?
s *
Kimmure A«a»ssmml - What is known about the principal paths, patterns, and
magnitudes of human or wfldlife exposure and numbers of-persons or wildlife-
species likely to be tecposed? What are the related uncertainties and science
policy choices? '
*
Corresponding principles and questions for ecoioj^Ql risk assessment are being
discussed as part of the effort to develop ecological risk gutddms.
Risk characterization is the summarizing step of risk assessment The risk ^
characterization integrates information from the preceding components of the risk
assessment and symtteizes an ©veraO! conclusion about risk that is complete,
informative and useful for decisionmakers.
3
-------
Risk characterizations should clearly highlight both the confidence and the
uncertainty associated with the risk assessment.. For example, numerical risk I _
estimates should always be accompanied by descriptive information carefully
selected to ensure an objective and balanced characterization of risk in risk- -
assessment reports and regulatory documents. In essence, a risk characterization
conveys the assessor's judgment as to the nature and existence of (or lack of) human
health or ecological risks.. Even though a risk characterization describes limitations
in an assessment, a balanced discussion of reasonable condusioiis and related
uncertainties'enhances, rattier than detects, from the overall credibility of each '
assessment. '
"Risk characterization" is not synonymous with "risk communication." This
risk characterization policy addresses the interface betofcte'risk assessment and risk
management" Risk communication, in contrast, emphasizes the -process of .
exchanging information and opinion with the public — including individuals^ *
groups, and other institutions. The development of a risk assessment may involve
risk communication. For exainple, in the case of site-specific assessments for - •'
hazardous waste sites, discussions with the public may influence the exposure
pathways included in the risk assessment. While the final risk assessment . • _
document (including the risk xdhararterization) is available to the public, the risk
communication process may be better served by separate risk uifonxiation ( -
documents designed for particular audiences. : - ¦¦¦
t . - 4 , ^ "
Promoting QiritgXMipMhil|EjaA£gBMgig * ¦ ¦
There are several reasons that the' Agatcy should stave for greater c!anty,
consistency and comparability in risk assessments. One reason is to minimize
confusion. For example, many people have not understood that a risk estimate of • .
one in a million for an "average" individual is not comparable to another one in a '
million risk estimate for the "most exposed individuaL" Use of such apparently .
similar estimates without further explanation leads to misunderstandings about the
relative significance of risks and the protectiveness of risk reduction actions. « -
EPA's Exposure Assessment Guiddlnes provide standard descriptors of _
exposure and risk.. Use of these tenas to afl Agency risk assessments wfli premote
consistency and comparability. Use of several descriptor*, rather than a single
descriptor, wifl enable EPA to present a fuJler picture of risk that cacresponds to the
range of different exposure conditions encountered^ by various individuals and
populations exposed to most environmental chemicals. -
4
-------
tfigalEligsl
This policy statement and associated guidance on risk characterization do not
establish or affect legal rights or obligations. Rather, they confirm the importance of
risk characterization as a component of risk assessment, outline relevant principles,
and identify factors Agency staff should consider in implementing the policy.
The policy and associated guidance do not'stand alone; nor do they establish a
binding norm that is finally determinative of the Issues addressed. Except where
otherwise provided by law/the Agency's decision on conducting a risk assessment in
any particular case is within the Agency's discretion. Variations in the application
of the policy and associated guidance, therefore, are not a legitimate basis for
delaying or complicating action on Agency decisions.
Aiadkakiliiy • ,
Except where otherwise provided by law and subject to the limitations on the
policy's legal effect discussed above, this policy applies to risk assessments prepared
by EPA and to risk assessments prepared by others that are used in support of EPA
decisions.
EPA will consider the prmcijples in this policy in evaluating assessments
submitted to EPA to complement or chaHeitge Agency assessments. Adherence to
this Agency-wide policy will improve understanding of Agency risk assessnteitte,
lead to more informed decisions, and heighten the credibility of both assessments
and decisions. , '
Assistant Adnunisteaters and Regional Administrators are responsible for
implementation of this policy within their organizational units. The Science Policy _
Council (SPC) is organizing Agency-wide implementation activities. Its
responsibilities include promoting consistent interpretation, assessing Agency-wide
progress, working with ©eternal groups on risk characterization issues -and methods,
and developing recommendations for revisions .of the policy and guidance, as
necessary.
Each Program and Regional office will develop office-specific policies and
procedures for risk characterization that are consistent with this policy and the'
associated guidance. Each Program and Regional office will designate a risk .
manager or risk assessor as the office representative to the Agency-wide Implementa-
5
-------
tion Team, which will coordinate-development of office-specific policies and
procedures and other implementation activities. The SFC will also designate a
small cross-Agency Advisory Group that will serve as the liaison between the SPC
and the Implementation Team. •
In ensuring coordination and consistency among EPA offices, the
Implementation Team will take into account statutory and court deadlines, resource
implications, and existing Agency and program-specific guidance on risk
assessment. The group will work closely with staff throughout Headquarters and
Regional offices to promote development of risk tittaracterizattons that present a full
* and complete picture of risk that meets the needs of the risk managers.
• • ' tttiUISI
; APPROWD: (L£M4^ DATE: '
1 Carol M. Brownfe, A3ministrator
6
-------
ELEMENTS TO CONSIpERWHEN DRAFTING EPA RISK '
•• " CHAEACTmiZATIONS
March 1995.
. Background - Risk Characterization Principles
There are a number of principles which form the basis lor a risk characterka
• Risk assessments should be transparent, in that flue condusfons drawn fr
stiestce are identified separately nom policy judgements, and the use of d
fates or methods and the use of assumptions In the risk assessment are
'articulated. ' * ""
, • Risk characterizations should indude a summary of the key issues and
• ^ 'condusions of each. of the oftfif ctaqxatoils of Ac risk assessment, as- wi
describe the likelihood of harm. The summary should include a descript
the overall strengths and the limitations {including uncertainties) of the
assessment and conclusions.
• Risk characterizations should be consistent fa. general format, but fecogn
unique Aaracteristio of each specific situation.
• Risk characterizations should include, at least in a qualitative sense, a di
. of how a specific risk and its context compares with other similar risks. 1
be accomplished by comparisons with other chemicals or situations in w
Agency has decided to act, or with other situations wM ch the public may
familiar with. The discussion should highlight the limitations of such
. comparisons. • ¦ *
• Risk diaractaizatimi is a key component "of risk communicatkm, wMA is
interactive process involving exchange of Information 'and export opinion
among individuals, groups and institutions. - , "
The following outline is a guide and formatting aid for developing risk
characterizations for chemical risk assessments. Similar outlines will be de
for other types of risk characterizations, induding site-specific assessments i
ecological risk assessments. A common format wifl assist risk managers in
evaluating and using risk characterization.
The outline has two parte. The first part tacks "the risk assessment to bring
its major conclusions. The second part draws all of the information togethc
characterize risk. The outline represents the expected findings for a typical
chemical assessment for a single chemical. However, exceptions for the
-------
circumstances of individual assessments exist and should be explained as part of the
risk characterization. For example, particular statutory requirements, court-ordered
deadlines, resource limitations, and other specific factors may be described to explain
why certain elements are incomplete.
•s
This outline does not establish or affect legal rights or obligations. Rather, it
confirms the importance of risk characterization, outlines relevant principles, and
identifies factors Agency staff should consider in implementing the policy. On a
¦ continuing basis, Agency mmmfgemmt is expected to evaluate the policy as well as
the results of its application throughout the Agency and undertake revisions as
necessary. Therefore, the policy does not stand alone; nor does it establish a binding
norm that is finally determinative of the issues addressed. Minor variations in its
application from one instance to another me appropriate and expected; they thus are
not a .legitimate basis for delaying or complicating action on otherwise satisfactory
scientific, technical, and regulatory products. ' - '
PART, ONE
SUMMARIZING MAJOR CONCLUSIONS IN RISK CHARACTERIZATION
L*' Characterization of Hazard Identification
A.. What is the key toxicological study (or studies) that provides the basis for
health concerns? ¦ ¦ • :
- Howgoodisthekeystucfy? > . ;
- Are the data from laboratory or field studies? In single species or
multiple spedes?. "
- If the hazard Is carcinogenic, comment on issues such as: observation of
. single or multiple tumor sites; occurrence of benign or malignant -
tumors; certain tamor types not linked to carcinogenicity; .use of the
maximum tolerated- dose (MTD): » -
- If the hazard is oAer than carcinogenic, what endpointe were observed,
and what is the basis, for the critical effect?''
- Describe other studies that support this finding.
- Discuss any valid studies which conflict with this finding.
B. • Besides the health effect observed in' the key study, aie there other health
endpoints of concern?
- What are the significant data gaps?
C Discuss available epidemiological or clinical data. For epidemiological
studies:
- What types of studies were used, i.e., ecologic, case-control, cohort?
2
-------
n.
- Describe th© I6®16® to W^ic^ exPosures were adequately described,
accounted foe.65"6 t0 W confounding Actors were adequately
Describe the degree to which other causal factors were excluded.
D" Wh" ***« ««
I n^?^-reie
-------
— What assumptions or uncertainty' factors were used?
— What is the confidence in the estimates?
- For carcinogenic hazards:
— What dose-response model was used? IMS or other linear-at-low-
. dose modeI, a biologically-based model based on metabolism data,
or data about possible mechanisms of action?
— What is the basis for the selection of the particular • dose-response
model used?. Are there other models that could have been used
. ~ "wth equal plausibility and scientific validity? What is the basis for
selection of the model used In this instance?
C Discuss the route and level of exposure observed, as compared to expected '
human exposures. - _
" ~ Are the available date from the same route of exposure as the expected
human exposures? If not, are pharmacokinetic data available to
extrapolate across route of exposure?
- How far does one need to extrapolate from the observed data to
environmental exposures (one to two orders of magnitude? multiple
orders of magnitude)? What is the impact of such an extrapolation?
' D. If adverse health affects have been observed in wildlife species, characterize
dose-response information using the process outlined in A-C.
m. Characterization of Exposure
A. What are the most significant sources of environmental exposure?
- Are there data on sources of exposure from different media? What is the
relative contribution of different sources of exposure?
- What are the most significant environmental pathways for exposure?
B. Describe the populations that were assessed, including as the genera!
population, highly exposed groups, and highly susceptible groups.
C Describe the basis for the exposure assessment, including any monitoring,
modeling, or other analyses of exposure distributions such as Monte-Carlo
or krieging.
D. What are the key descriptors of exposure?
- Describe the (range of) exposures to: "average" individuals, "high end"
individuals, general population, high exposure group{s), children^
susceptible populations.'
- How was the cental tendency estimate developed? What factors and/or
metftods were used In- developing1 this estimate?
- How was the high-end estimate developed?
4
-------
" ^t?re™Sle^Tf^y° «* they?
assessment eXp0SUre? HoW »» *V """atel for £ the
E. Is there reason to be mngmad. about: cnm,.i,n.~ ¦— ,
' heme of ethnic, radii, or socioeconomic P ^oslres
E w;MrfSe hedtft ®ffectS ha;ve been observed in wildlife species dhararte * '
wrldMe exposure by discussing the Levant issues ^T5£$T?35£
G' -m^^J%°^etC0"dUSiTS and Ae following:
«bu4S; appIOaCheS' U' mod^ monitor*! probabffity
°f efch' md fte ^ge of most reasonable values- and
confidence m the results obtained, and the limitations to the' resute.
part two
RISK CONCLUSIONS AND COMPARISONS
IV. Risk Conclusions t
A' ** ha2aid ««*-** lden,';e P°Uq' °ptiOTS to each of &e toee ™jor analyses?
What axe the alternative approaches evaluated?
- What are the reasons for the choices made?
• Risk Context
A. What are the qualitative characteristics of the hazard fe e voluntary vs
from iechn.ol«3gicaj vs. natural, etc)? Comment on If any
from studies of ask perception that relate to this hazard or similJtazttfe
B. What are the alternatives to tfais hazard? How do the risks compare?
-------
C. How does this risk compare to other risks?
L- How does this risk compare to other risks In this regulatory program,, or
other similar risks that the EPA lias made decisions about? *
2. Where appropriate, can this risk be compared with past Agency
decisions, decMons by other federal or state agencies, or commoa risks
with which people may be familiar?
3. Describe the limitations of making these comparisons. •
D. Comment on significant community concerns which influence public
perception of risk?
VI. Existing Risk Information
Comment on other risk assessramte* that have been done on tins chemical by
EPA, other federal agencies, or other organizations. Are there significantly
different conclusions that merit discussion? .
VII. Other Information -
Is there other information that would be useful to the risk manager or the
public in this situation that has not been described above?
6
-------
»5»
I
V
m
8
^prO^
y
March 1995
IMPLEMENTATION PROGRAM
FOR
THE EPA POLICY ON RISK CHAIUCTERIZATION
Introduction
The EPA Science Policy Council (SPC) is implementing the Administrator's policy on
risk characterization - through a year-long program of activities that will involve
risk assessors and risk managers in the practice of folly characterizing risk. This
interactive. approach calls for inter- and intra-office activities to gain experience
with the fundamentals of the policy and to resolve issues that were identified daring .
Agency-wide review of early drafts! Implementation will include program-specific
guidance development; case, study development; and risk characterization workshops
and rountables for risk assessors and managers. "
A SPC-sponsored "advisory group" will plan and execute these implementation
activities. This advisory group will organize an "implementation team" composed of
representatives from the program offices, _ regions and ORD laboratories and centers.
This team will work closely with the advisory group to coordinate implementation '
activities within their offices.
Program Guidance '' Development
Risk characterizations often differ
according to the type of assessment
involved. The aim is to woit closely with
the Program Offices and Regions to
identify and address their specific risk "
characterization needs and, where
appropriate, to develop assessment-
specific guidance. * .
This program updates and implements the
risk characterization guidance issued in
early 1992. The policy features a paper
entitled "Elements to Consider When
Drafting EPA Risk Characterizations."
This paper outlines generic elements for
characterizing risk, • and provides a
prototype for assessment-specific
guidance. Program and Regional
offices will use this paper to identify
and address risk characterization issues
associated with specific assessment
types that' differ from the general
prototype (e.g., site-specific. and .
ecological risk assessments). Lessons •
learned from the case studies,
roundtables and workshops (discussed
below) will also contribute to program-
specific guidance development.
Case Studies
Today, when asked to provide an
example of a "good" risk
characterization, few people can
identify good examples, let alone
examples that others would agree are
RJSIC CmEACtERKATION
1995
IMPLEMENTATION SCHEDULE
IP re £ p.* i® ® isp*© din c Ouldance !>*~« lop n> cot
...
! \ t V i | i r T~ > ! » ! t ! t t
-r—r—J—
K. i. i _ t_ ! i i 1
R ound tables and WorKSDops
: J ¦!!!!! ! : j : > ; •
rM
-------
-m
i
i
of high quality. As a first step,,.a selected
number of risk characterization case
smiles will be developed for use as.,
teaching tools. ' A "case study" will, be am
exercise to improve aa existing risk
characterization, using- the Information
available in an existing risk assessment.
While based on actual risk assessments,
identifying information (e.g., site
identification information) will be
removed to avoid any implied judgement
Ronndtables * and .. Workshops
as to the adequacy of the original risk
assessment and risk characterization.
Examples of case studies may include a
chemical assessment, a site-specific
assessment^ and a screening-level
assessment. The case studies will be
developed by the risk characterization
. advisory group, working in consultation
with the implementation team, and will
be used for discussion at the first risk
characterization ' workshop.
EPA decision-makers will be invited' to participate In roundtabie discussions on risk
characterization. In addition, a minimum of two workshops are planned for EPA, risk
assessors md risk managers.
° Risk Decision-maker Ronntahles . on • Risk ¦ Characterization - The goal will ,
be to determine the types of risk characterization information needed by managers
for effective risk-based decision-making.
* Risk Characterization Workshop / - Will focus on identifying the qualities of
"good" risk characterizations, . program-specific plans and guidance development,
and case' studies. . - . .
Characterization Workshop Ii - Risk assessors and risk managers will
meet to wrap-up program-specific plans and guidance, and discuss any necessary
updates to " the agency-wide risk characterization ' guidance..
-------
GUIDANCE
FOR
RISK CHARACTERIZATION
i
U.S. Environmental Protection Agency
Science Policy Council
February, If 95
-------
CONTENTS
t
I. Hie Risk Assessment-Risk Management Interface
II. Risk Assessment and Risk Characterization
III. ; Exposure and Risk Descriptors .
-------
PREFACE
TWs guidance contains principles for developing 'and describing EPA risk
assessments, with a particular emphasis on risk characterization. Hie current
document is an update of the guidance issued with the Agency's 1992 policy
(Guidance on Risk Characterization for Risk Managers and Risk Assessors, February
26,1992). The guidance has not been substantially revised, but includes some
clarifications and changes to give more prominence to certain issues, such as the
need to explain fte use of default assumptions.
As in the 1992 policy, some aspects of this guidance focus on risk'
assessment, but the guidance applies generally to human health effects (eg.,
neurotoxicity, developmental toxicity) and, with appropriate modifications, should
be used in all health risk assessments. TMs document has not been revised to
specifically address ecological risk assessment, however, initial guidance for
ecological risk characterization is included in EPA's Framework for Ecological Risk
Assessm«tte (EPA/630/R-92/001). Netfhff does this guidance address in detail the
' use of risk assessment information (e.g., infozmatkm front the Integrated Risk
Information System (IRIS)) to generate site- or media-specific risk assessments.
Additional program-specific guidance will be developed to enable implementation,
of EPA's Risk Characterization Policy. Development of such guidance will be
overseai by the Science Policy Council and will involve risk assessors and risk
managers front aooss the Agency.
-------
I THE RISK ASSESSMENT-RISK MANAGEMENT INTERFACE
Recognizing that for. many people the term risk- assessment has wide meaning,, the
National Research Council's 19© report on risk assessment In the federal
, government distinguished between risk assessment and risk management.
"Broader uses of the tent [risk assessment! than oms also embrace analysis of
perceived risks, comparisons of risks associated with different regulatory
strategies, and occasionally analysis of the economic and social implications of
regulator dedsions - ^^^Aat^MMgnAaskmaM^imt
(emphasis added). Q) . . . . . .
¦;«. ¦ .
In 1984, EPA atdosed .these distinctions between risk assessment and risk
management for Agency use (2), and later relied on them in developing risk..
¦ assessment guidelines (3). In 1994, the NRC reviewed the Agency's approach to and
use of risk assessment apd issued an extensive zeport on their findings (4). This
distinction suggrate that EPA participants m the process on be grouped into two
main ofegoiies, each with somewhat different responsibilities, based on their roles,
-with respect to risk assessment and risk management-
and Risk Managqni -
Within the Risk Assessment category there is a g^oup that devdops'diemical-
spedfic risk assessments by collecting, analyzing, and synthesizing scientific data to
produce the hazard identification, dose-response, and exposure assessment portion -
of the risk assessment and to characterize risk.. This group relies in part on Agency
risk assessment guidelines to address sdetoe policy lames and sdeftific '
mcertarafies. Generally, this group includes scientists and stotistidans in the Office
of Research and Development; the Office-of Ftevoitioft, Pesticides and Toxics and
other program offices; the Carcinogen Risk Assessment Verification Endeavor
(CRAVE); and the Retrace Dose (RfD) and'Reference Concentration (RfC)
Workgroups*? * . '
Another group gaierates site- or medM-spedfic risk assessmoits for use In
regulation development or site-specific decisionmaking. ¦ These assessors rely on
existing databases (e.g., IMS, ORD Health. Assessment Documents/ CRAVE and
RfD/RfC Workgroup documents, and' program-specific toxicity information) and
media- or site-spedfic exposure information in developing risk assessments. TMs
group also relies in part on Agency risk assessntmt guidelines and program-specific
guidance to address science policy issues and scientific uncertainties. Generally, this
group includes scientists and analysts in program offices, regional offices, and the
Office of Research and Development.
1
-------
Risk managers, as a separate category, integrate the risk characterization with other
considerations specified. in applicable statutes to make and Justify regulatory
decisions. * Generally, tills group includes Agency managers and dedislon-ntakers.
Risk managers also play a role in determining the scope of -risk assessments. The
risk assessment process involves regular interaction between risk assessora and risk
managers, wtth overlapping responsibilities at various stages in the overall process.
Shared responsibilities include initial decisions regarding the planning'and conduct
of an ass«snt«it, discussions as the assessment develops, decisions regarding new
data needed to-'Complete an assessment and to address significant unoertaifttie. At
critical fratctraes in the assessment such consultations shape the nature of, and
schedule far, the assessment External experts and members of the public may also
play a role in determining the scope of the assessment; for example, die public is
often, concerned about certain chemicals or exposure pathways in the development
of site-specific risk assessmmts. - ' • *
The following guidance outlines principles for those who generate, review, use, and
integrate risk assessments for decision-making. " v
1* Rbk assesson and ride nmtageis should be sensitive to distinctions between
xiskimessmentandiiAmaMigeBMaiL * ' .
The major participants In the risk assessnwt process have many shared
responsibilities. Where responsibilities differ, it is important that participants
confine themselves.to tasks in their areas of responsibility and not inadvertently
obscure differences between- risk assessment and risk management -; *
s * e ' -
the generating of fee ameMtngnfr. distinguishing between risk assessment and
• risk management means that' scientific information is selected, evaluated, aid
presetted .without considering Issues such as cost feasibility, or how the scientific
analysis might influence the regulatory or site-specific decision. Assessors are
charged with (1) generating a credible, objective, realistic, and scientifically balanced
analysis; (2) presenting information on hazard, dose-response, exposure and risk;
and (3) explaining confidence n each assessment by clearly delineating strengths,
imcertaiitties and assumptions, along with the impacts of these factors (e.g.,
confidence limits, use of conservative/non-conservative assumptions) on the
overall assessment They do not make decisions on the acceptability of any risk
level for protecting pubic health or selecting procedures for reducing risks. - ,
For ^m-fl£JM_as«smgilmd-farAcgion=makm who Integrate these'
assessments into regulatory or site-specific decisions, the distinction between risk
assessment and risk management means refraining from influencing the risk
2
-------
description through consideration of other factors — e.g., the regulatory outcome -
and from attempting to shape the risk assessment to avoid statutory constraints,
meet regulatory objectives,, or serve political purposes* Such management
considerations are often legitimate considerations for the overall regulatory decision
(see next principle), but they have no role in estimating or describing risk.
However, dedsioft-ntaleets and risk ass®sors participate in an Agency process that
establishes policy directions that determine the overall nature and tone of Agency
risk assessments and, as appropriate, provide policy guidance on difficult and
controversial risk assessment issues. Matters such as risk assessment priorities,
degree of conservatism, and acceptability of particular risk levels are reserved for
decision-makers who are charged with making decisions regarding protection of
public health. - .
2. The risk assessment product that is, the risk characterization, is only one of
- several kinds of information used for regulatory decision-making.
* - *
Risk characterization, the last step in risk assessment, is the starting point for risk
" management considerations and the foundation for regulatory decision-making, but
it is only one of several important components in such decisions. ¦ As the last.step in
- risk assessment, the risk characterization identifies and highlights the noteworthy
risk conclusions and related, uncertainties. Each of flhe environmental laws
administered by EPA calls for consideration ol other factors at various stages in the
regulatory process. As authorized by different statutes, decision-makers evaluate
technical feasibility (e.g., treatability, detection limits), economic, social, political, and
legal factors as pail of the analysis of whether or not to regulate and, if so, to what
extent. Thus, regulatory decisions are usually based on a contbinatiofi of the
technical analysis used to develop-the risk assessment and information from other
fields. ' ' ¦
For this reason, risk assessor and managers should understand that tjtie regulatory
" decision is usually not determined solely by the outcome of the risk assessment For
example, a regulatory decision on the use of a particular pesticide considers not only"
, the risk level to affected populations, but also the agricultural benefits of its use that
may be important for the nation's food supply. Similarly, assessment ififofts may
produce -an RfD for a particular chemical, but other considerations may result in, a -
regulatory level that is more or less protective than the RfD itself.
For decision-makers, this means that societal considerations (e.g., costs and benefits)
that, along with the risk assessment, shape the regulatory decision should be
described as fully as the sdmfiffc information set forth in the risk characterization.
Information on data sources and analyses, their strengths aid limitations,
confidence in the assessment, uncertainties, and alternative analyses are as
important here as they are for the scientific components of the regulatory decision.
Decision-makers should be able to expect, for example, the same level of rigor from
the economic analysis as they receive from the risk analysis. Risk management
decisions involve numerous assumptions and uncertainties regarding technology,
economics and social factors, which need to be explicitly identified for the
decision-makers and the public.
3
-------
n. W8K CHARACTERIZATION
A. Defining Risk Characterization In the Context of Risk Assessment
EPA risk assessmoit principles and practices draw on many sources Obvious
sources include the environmental laws administered by EPA, the National
Research Council's 1983 report pn risk assessment'^), the Agency's Risk Assessment
Guidelines (3), and various program specific guidance (e.g., the Risk Assessment
Guidance for Superfund). Twenty yeara of EPA experience in developing, .
defending, and enforcing risk assessment-based regulation is another. Together
these various sources stass the toportai» of a dear explanation of Agency
processes for evaluating hazard^ dose-response, and other data that
provide the scientific foundation for characterizing risk.
This section focuses on two requirements for full characterization of risk. First, the
characterization should address qualitative and quantitative features of the
assessment - Second, it should. identify the important strengths and uncertainties in
the assessment as part of a discussion of the confidence in the assessment This
.. emphasis on a full description of all elements, of the assessment draws attention to
the importance of the qualitative, as well as the quantitative, dimensions of the
assessment, The 1983 NRC report carefully distinguished qualitative risk
assessment from quantitative assessments, preferring risk statements that are not
•" strictly numerical.
«
Hie term risk assessment is often given narrower and broader meanings
than we have adopted here. For some observers, the term is synonymous
with quantitative risk assessment and emphasizes reliance on numerical
results. Our broader definition includes quantification, but also includes
• qualitative expressions of risk. Quantitative estimates of risk are not always .
feasible, and they may be eschewed by agencies for policy reasons. (I)
EPA's Exposure Assessment Guidelines define risk characterization as the final step
in the risk assessment pieces that
• Integrates die individual characterizations from the hazard identification, dose-
response, and exposure assessments;
. • Provides an evaluation of the overall quality of the assessment and the degree.
of confidence the authors have In the estimates of risk and conclusions drawn;
• Describes risks to individuals and populations in terms of extent and severity of
probable harm; and
• Communicates results of the risk assessment to the risk manager. (5)
Particularly critical to full characterization of risk is a frank and open discussion of
the uncertainty in the overall assessment and in each of its components. The
uncertainty discussion is important for several reasons.
4
-------
1. Information from different sources carries different kinds of mcerfalitty and
knowledge of these differences is important what uncertaliities are combined
for characterizing risk
2. The risk assessment process, witk management input, involves decisions
regarding the collection of additional data (versus living with uncertainty); in
the risk characterization, a discussion of the uncertainties will help to identify-
where additional information could contribute significantly to reducing
- uncertainties In risk assessment- '
3. A clear and explicit statement of the strengths and limitations of a risk
• assessment refukes a clear aid explicit statement of related uncertainties.
A discussion ofjuncertainty requires onmmgjtt on sittii fowii** as the quality and
- quantity of avapabfe data, gaps buhe data base fa specific chemicals, quality of the
• measured data, use of default masnmpttais,;. incomplete 'imdtastandmg of general
biological phenomena, and scientific judgments or science policy positions that* were
employed to bridge Infaimatipii gaps*
In short, broad agreement exists on the importance of a full picture of risk,
parfkrtarly including a statement of confidence la the assessntoii and the ¦ ¦
associated uncertainties. Ito section dlscii»«'fafoimaticm content-and uncertainty
assets of risk Aaractaratioiv whfle Section HI discusses various descriptors used
in risk characterization.
B. Guiding Principles .
1. - He risk characterization Jategiates the infematiott faint the Juand- '"
• . . identification, dote-Mspons* and aqposofe as»e«®meiil% using a cmnbinaftioii of
¦ qualitative information, quantitative information, and information regarding
uncertainties.. . ' ** - , '¦ *i * • ¦- '¦ • ' •
Risk assessment Is based on a series of questions that the assessor asks about the data
and the implications of the data for ham* risk. Each. question calls for analysis and
interpretation of the available studies, selection of the data that are most
scientifically reliable and most relevant to Ac problem at hand, and scientific
conclusions legending die .question presented. As suggested below, beau* the
questions and analyses are-oomptal a complete characterization includes several
' different kinds of information, carefully selected for reliability and relevance.
a. Hazard Identification — What is known about the capacity of an environmental
agent for causing cancer (or other adverse effects) in humans and laboratory'
animals? * *
f
5
-------
Hazard identification is a qualitative description based on factors such as the kind
and quality of date on humans or laboratory animals, the availability of ancillary
information (e.g., structure-activity analysis, genetic toxicity, pharmacokinetics)
from other studies, and the weighfrofche-evidence front all of these data source.
For example, to develop this description, the issues addressed include:
1) the nature, reliability, and consistency of the particular studies in humans and
in laboratory animals;
. ¦»
2) the available information on the mechanistic basis for activity; and
3) experintoital animal responses and tiheir relevance to human outcomes.
Utaie issues mate clear that the fask of fitzarf identification is Aara€fariized by
describing the full range of available information and the implications of that
information for human health. ' •' . • -•
r
• b- Dose-Response Assessment - What is known about the biological mechanisms
• ¦ and* dose-response relationships underlying any effecto observed In the
laboratory or epidemiology studies providing data for the assessment? ,
The dose-response assessment examines quantitative relationships between •
eqpstrts (or dose) and effects In the studies used to identify and define effects of
concern. Bus information is late used .along with "ml world", ©cposuie
infoimaMai (see Wow) to develop estimates- of ft* HkeSltood of adverse efifecte in -
populations potentially at risk. It should be noted that, in practice, hazard
identification for dewelopiitental toxicity- and'Other non-cancer health ellecte is
usually done in conjunction with an evaluation of dose-response relationships,
since the determination of whether there is a Jbazand is often-dependent on wbetfter
a dose response relationship is present (6) Abo, Ac framework developed by EPA
for ecological risk assessment does not distinguish between hazard identification
and doSe-response assessment but father alb for a -"characterization of ecological
effects," (7) •
Methods for establishing dose-response relationships often depotd on various ¦
assumptions used in Hen of a complete data base, and the method chosen can''
strongly influence the owesall'assessment The Agency's risk assessment guMeirtes
often identify so-called "default assumptions" for use in the absence of other
information. The risk assessment should pay careful attention to the choice of a
high-to-low dose exteapoktfoit procedure. As a result an assessor who is
characterizing a dose-response relationship considers several key issues:
1) the relationship between extrapolation, models selected and available
information on biological medKanisms;
6
-------
2) how appropriate data sets were selected from those that show the range of
possible potencies both in laboratory animals and humans;
3) the basis lor selecting interspecies dose scaling factors to account for scaling
doses from experimental amntals to humans;
» •
4) the .correspondence between the expected route(s) of exposure and the exposure
route(s) utilized in the studies forming the basis of the dose-response
assessment,, as well as the interrelationships of potential effects from different
exposure routes;
5) the correspondence between the expected duration of exposure and' the
exposure durations in the studies wed In forming the basis of the dose-response
assessment, c.gv dhfonlc studies would be used to ass«s long-term, cwmaJative
.exposure-concentrations, wMle acute studies would be .used in assessing peak
levels of exposure^ and • " ' . ,
6) the potential for differing susceptibilities among population subgroups.
The Agency's Integrated Risk Information. System (KB), is a repository for such
mfamtafiofi for EPA. EPA program offices also maintain program-specific
databases, such as the OSWER Health Effects Assessment Summary Tables (HEAST). *
IRIS includes data summaries representing Agency coiistttws cm specific ctemcals, ¦ - -
based on a careful review of the scientific issues listed above? For specific risk
assessntoite based on data from My source, risk assesns should carefully review
" -the information presorted, emphasizing confidence in the data and uncertainties '
(see subsection 2 Wow). Speefficatty#,wtieii KB data are used, -the KB statement of
* confidence should be included as an explicit part of the risk characterization for
hazard -and dosfe-response information. . _ _ . ¦. .' . .
c- Exposure Assessment — What is known about the principal paths, patterns, and -
magnitudes of human exposure and numbers of pesoits who may be exposed?
Hie expose assessment examines a wide range of exposure parameters pertaining
to the environmental scenarios of people who may be exposed to Ac agent under
study. The information considered for the exposure assessment indudes -
monitoring studies of chemical concentrations in environmental media, food, and
other materials; modeling of environmental fete and transport of contaminants;
and information on different activity patterns of different population subgroups.
Aa assessor who Aaracteriz« exposure should address several-issues:
1) The basis for the values and input parameters used for each exposure8 scenario.
If the values are based on data, there should be a discussion of the quality,
purpose, and representativeness of the database For monitoring data, there
should be a discussion of the data quality objectives as they are relevant to risk
7
-------
assessment, including the appropriateness of the analytical detection limits. II
models are applied, the appropriateness of the models and information on their
validation should be presorted. When assumptions are made, the source and
general logic used to develop the assumptions (e.g., program guidance, analogy,
professional judgment) should be described.
2) The confidence in the assumptions made about human behavior and the
relative UkeMJiood of the different exposure scenarios.-
3) The major factor or factors (e.g., concentration, body uptake, duration/frequency
of exposure) thought to account for the greatest uncertainty in the exposure
estimate, due either to sensitivity or lade of data.
4) IbM between- Ac aqposme information and fl* risk desmptora cBscusseci
-* in Section niofthis'Appendix. S|pedfically,. the risk assessor needs to discuss , -
the ctMitecfioit betwem the conservatism or non*»nservatisin of the
• data/assumptions used in'the scenarios and' the choice of descriptors.
5) Ofcer information that may be important for the particular risk assessment
For example, for many assessments, other sources and bs&kgrotmd lev A in the '
•' environment may contribute significantly io population exposures and should
be discussed. -
2) The ride characterization indudes a discuMim of uncertainty and variability.
In the risk characterization, cowdiwiofis about bawd and dose response am
integrated'with those from the exposure assessment In addition, confidence about
' these cwiduslmia, induding information about-the uncertainties associated with
each aspect of the assessment in/the final risk summary, is highlighted.. In ft*
preview assessnwt steps and,in Ac risk cttamcteriaafion# the risk assessor must
• dMrogimfa between gariafaflitg and uncertainty. ¦ • • • *-i ¦' *'• - ' .
Variability arises foont true heterogeneity in characteristics such as dose-response
, differences within a population, or differences in contaminant levels in the'
x environment The'vunes of some ¦fatjables used-in an assessment change with
time and space, or across' the population whose exposuse is bong es&mted. _
.Assessments should address the*resultiftg variability to doses received by members
of the target population. Individual exposure, dose, and risk can vary widely in a
large population. Ilie central tendency and high end individual risk descriptors
(discussed Sot Section M Wow) are mtertded to capture the variability in exposure,
lifestyles, and other factors that lead to a distribution of risk across a population.
Uncertainty, on the other hand, represents lack of knowledge about factors such as
adverse effects or contaminant levels which may be reduced with additional study.
Generally, risk assessments carry several categories of uncertainty, and each merits
8
-------
consideration. Measurement uncertainty refers to the usual error that accompanies
scientific measmemente—standard statistical techniques can often be used to express
measurement uncertainty. A substantial amount of uncertainty is often inherent In f
environmental sampling, and assessments should address these uncertainties.
There are likewise uncerfalnties associated with-the use of sdmtific models, e.g.,
¦" dose-response models, models of environmental fate and transport. Evaluation of
model uncertainty would consider the scientific basis for the model and available
empirical validation. • • .
A different Mud of uncertainty stems front data gapsthat is, mtintates or
assumptions used' In the assessment Often# the data gap is broad, such as, the. -
absence of information on the effects of exposure to a chemical on humans or on
the biological mechanism of action of an agent The risk assessor should include a
statement of confidence that reflects the degree to which the risk assessor believes
that the estimates or assumptions adequately fill the data gap. Per some common ¦ ¦ ,
and. important; data gqw, Aga»^orp»^aBi-^ed&riAa^MmaitgttIdiiia . ,
^ provides default assumptions or values.. lisk-assewots should carefully consider aM
* available data^before deciding to rely on default assumptions. If defaults are used,
' the risk assessment should reference flic Agatcf guidance that explains the default
assumptions "or- values. . ¦ * : ¦ • - , - *
Often risk assessors and manages simplify discussion of risk issues by speaMftg.oiiy
of the numerical components of an assessment That is, they refer to the alphas
numeric weightaMheevktence dassification; wit risk, the Mk-spetiiiic dose or the
- qi* for cancer risk, and the RfD/RfC for health effects other than cancer, to the
exclusion of other information bearing on the risk case -However/since every
assessment carries• uncertainties, a amplified numerical presentation of risk is ,
¦ always incomplete and often misleading., For this reason, 'the MRC (1) and EPA risk
assessment guidelines (2) call for "characterizing? risk to include qualitative , .
information, a related numerical risk estimate and a discussion of uncertainties, •
limitations, and assumptions^efault aiui otherwise.-.
Qualitative information ©ft methodology, alternative interpretations, and working
assumptions (including defaults) is an important component of risk
characterization. For example, specifying that animal studies rather ton human
studies were used In an assessment rib others flat the risk estimate is based on _
assumptions about human response to a particular chemical rattier than human
data. Information that human exposure estimates are based on tie subject*
presence in the vidnity of a chemical accident rather than issue measurements
defines blown and unknown' aspects of the exposure component of the study. -
Qualitative descriptions of this kind provide crucial information that augmoits
imderetandirtg of numerical risk estimate. Uncertainties such as these are expected
in scientific studies and in any risk assessment based on these' studies. Such
9
-------
uncertainties do not reduce the validity of the assessment Rather, they should be
MghMghted along with -other important risk assessment conclusions to inform
others fully oil the results of the assessment.
Ik many cases, assessors must choose among available date, models# or assumptions
in estimating risks. Examining the impact of selected, plausible alternatives on the
conclusions of the assessment is an important part of, the uncertainty discussion.
The key words are "selected" and "plausible;" listing all alternatives to a particular
assumption, regardless of their merits would be superfluous. Generators of the
assessment using best professional judgment, should outline the strengths aid
. weaknesses of the plausible alternative approaches.^
An adequate description of the process of alternatives selection involves several
aspecte. .. -
a. A rationale for the choke.
•b. Ofsctissiott of the effects of alternatives selected cm -the assessment
c Comparison with other plausible alternatives, where appropriate. *
The degree to which variability and uncertainty are addressed depends largely on
tibe scope of the assessment and the resources available. for example, Ate Agency
does not expect an assessment to evaluate and assess may conceivable exposure
scenario for every possible pollutant to examine afl suscep^tible populations
potentially at risk, or to characterize every possible environmental scenario to
• estimate the cause and "effect relationships between exposure to pollutants and
adverse health effects. Rather, the discussion of uncertainty and variability should
reflect the type and complexity of the risk assessment, with the level of effort for
¦ analysis and discussion of uncertainty corresponding to the level of effort for the .
assessment ¦ " • . '• - : ' ,
*
3.- -Well-balanced risk dhaadoizatioai present risk amdonons and information
regarding the strengths and limitations of the assewntoit fin* other risk
• assessors, EPA dedskm-makexs, and the public.
The risk assessment process calls for identifying and highlighting significant risk
conclusions and Mated uncertainties partly to assure full communication among
risk assessors and partly to assure that decision-makers are fully informed. Issues
are identified by atcknowledging noteworthy qualitative and quantitative factors that
make a difference in the overall assessment of hazard and risk, and- tend in the
ultimate regulatory decision. The key word is "noteworthy." Information that
lln cases where risk assessments within an Agatcy program routinely address similar sets of
alternatives, program guidance may be developed to streamline and simplify the discussion of these
alternatives.
10
-------
significantly influences the analysis-Is explicitly noted — in all future presentations
of the risk assessment and in the related decision. Uncertainties and assumptions
that strongly influence confidence in the risk estimate also require special attention.
Numerical estimates should not be separated from the descriptive information that
is integral to risk characterization. Documents and presentations supporting
regulatory or site-specific decisions should include both the numerical estimate and
descriptive information; in short reports, this information can be abbreviated. Fully
visible information assures, that important features'of the assessment are
immediately available at each level of review for evaluating whether risks are •
acceptable or unreasonable. V ; - ¦
-------
11 EXPOSURE ASSESSMENT ANB RISK DESCRIPTORS •
A. Presentation of Risk Descriptors
The results of a risk assessment are usually communicated to the risk manager in
the risk characterization portion of the assessment This communication is often
accomplished through risk descriptors which convey information and answer
questions about risk, each descriptor providing different information and insights.
Exposure assessment plays a key role in developing these risk,d«oiptois since each
descriptor is based in part on the exposure distribution within the population of
interest. .
Hie following guidance outlines the different desmptois in a convenient order that
' should not be consteued as a Meraictty of importance. These descriptors should be
used to describe risk m a variety of ways for a given assessmMt consistettt with the '
asses»t«it's purpose, the data, available, and the infanitafioa the risk manager -
needs. Use of a range of descriptors instead of a single descriptor enables Agency
programs' to present a picture of risk that corresponds to the range of different
exposure conditions encounteied for most environmental .chemicals. TMs analysis,
in tern, allows risk ntanagewfo identify populations at greater arid lessor risk arui to
shape regulatory solutions accordingly.
Agency risk assessments will be expected to address or provide descriptions of (1)
individual risk that include the coital tendency and high aid portions of the risk
distribution, (2) population risk, and (3) important subgroups of the population,
such as highly exposed or highly susceptible groups. Assessors may also use *
additional descriptors of risk as needed when these add to the clarity of the
presentation. With title exceptioa of assessments where particular descriptors clearly ¦'
do not apply, some faint of tee three types of descriptors should be routinely
developed and presented for Agency risk^assessmeittsi. In other cases, where a
descriptor would be relevant, but the program lacks the data or methods to develop
it, the program office should design and implement a plan, in coordination with
other EPA offices, to meet these assessment needs. WMle gaps continue to exist,
risk assessors should mate their best efforts to address each risk descriptor, and at a
minimum, should briefly discuss the lack of data or methods. Finally, presenters of
risk assessment-information should be prepared to routinely answer questions by
risk managers concerning these descriptors.
It is essential that presenters not only communicate the results of die assessment by
addressing each of the descriptors where appropriate, but that they also'
^Program-specific guidance will need to address these situations. For example, for site-specific
assessments, the utility and appropriateness of population risk estimates witt be determined based on
the available data and program guidance.
12
-------
communicate their confidence that these results portray a reasonable picture of the
¦ actual or projected exposures. Bus task will usually be accomplished by frankly
commenting on the key assumptions and parameters that have the greatest impact
on the results, the basis or rationale for choosing these assumptions /parameters,
and the consequences of choosing other assumptions.
• B. Relationship Between Exposure Descriptoni and Risk Descriptors
. In the risk assessment process, risk is estimated as a function of exposure, with the
xisk of adverse aJ&cts Increasing as exposure increases. Information on the levels of
exposure experienced by different members of the population is key to
• undeistemfcig the range of risks 'that may occur. .¦ Risk assessors and risk managers
should keep ixr mind, however, that exposure is not synonymous with risk. -
Differences'among individuals-in absorption rates,'susceptibility, or other factors*
mean that individuals'with the same level of exposure may be at different levels of
• risk. In moist cases, the state of the science is not: yet adequate to define distributions
of factora as popoMfion, susceplfflMltf: llie guidance principles-below'discuss a •
¦ variety of risk descriptors that primarily reflect differences in estimated exposure. If
-. a full description of the range of susceptibility in the population cannot be
' ? presented, an effort should be:made to identify subgroups that, for various reasons,
may be particularly susceptible , ' ¦. ¦
C. Guiding Principles , :
1. Information about the distribution of individual exposures is important to
communicating the results of a risk assessment.
^ " ' * »
"Die risk manager is generally interested in awww to questions such, as tfbe
following: ¦ " ¦ ' " ¦
• Who are the people at the highest risk? ¦
• What rislclevels aie they subjected to?
• What are they doing, where do they live, etc, that might be putting them at this
higher risk?
• What is the average risk for individuals in the population of interest?
Individual exposure and risk descriptors are intended to provide answers to tftese
questions so as to illuminate the risk management decisions that need to be made.
In order to describe the range of risks, both high end and central tendency
13
-------
descriptors are used to convey the variability in risk levels experienced by different
individuals in the population.
a. High end descriptor
For the Agency's purposes, high end risk descriptors are plausible estimates of the
individual risk for those persons at the upper end of the risk distribution. Given
limitations in current mdostaitdlitg of variability in individuals' sensitivity to
toxins, high end descriptors wfll usually address high end exposure or dose (herein
referred to as exposure for brevity). The intent of these descriptors is to convey
estimates of exposure In the upper range of the distribution, but to avoid estimates
which are beyond the to* distribution. Conceptually, high end ©cposuje
. exposure above about the 90th percentile of the population distribution, but not
higher than the*'individual In the population who has the highest ecpostne. When
large populations are assessed, a large number of individuals may be, included
within the ''high end"- (eg., above 90th or 95th percentile) and information cm the
range'of exposures received by these individuals should be presented: •
High end descriptors are intended to estimate the exposmra that are expected to
occur-in smalt but definable, "high end" segment* of the subject population^ The-
• individuals with these-exposures may be members Of a speoal population segment
or mdi¥idiMto in the general popuIaMofi who .are highly exposed because of the
inherent stochasticnature of the factors which give rise to expostne, Where
- dfflferatces In sensitivity fiui be Icfattified within- the population# high end estimates
addressing sensitive individuals or subgroups can be developed.
In those few cases in which the complete date on the population distributions of
exposures and doses are available, high end exposure or dose estimates on be
¦ represented by repotting exposures or doses ataaetof selected percentiles of the
distributions, such as the 90th, 95th, and 98th percentile. High end exposures or -
doses, as appropriate,, can then be used to calculate high end risk estimates.
* . - '
Irt the majority of cases where the complete distributions are not available, several
methods help estimate a high end exposure or dose. If sufficient information about
the variability in chemical' comcatteatiotis, activity patterns, or. other factors are
available, the distribution may be estimated through the use of appropriate
modeling (e.g., Monte Carlo simulation or parametric statistical methods). The
3High aid estimates focus an estimates of exposure in the exposed popukticw. Bounding
estimate, on the other ttand# are constructed to be equal to or peater than the highest actual risk in
the population (or the highest risk that couM be expected in a future scenario). A "wocst case scenario"
refers to a combination of events and conditions such that^ taken together, produces the highest
conceivable risk. Although it is possible that such an exposure, dose, or sensitivity combination might
occur in a given population of Interest tt* probability of an mdivMual imving this combination of
events and conditions is usually anal, and often so small that such a combination will not occur in a
particular, actual population.
14
-------
determination of whether available information is sufficient'to support the use of
probabilistic estimation methods requires careful review and documentation by the
risk assessor. If the input distributions are based on limited data, the resulting
distribution should be evaluated carefully to determine whether it is an
improvement over more traditional estimation techniques. If a distribution is
developed, it should be described with a Series of percentiles or population . '
frequency estimate, particularly in the high end range. Hie assessor and risk
manager should be aware, however, that unless a great deal is known about
exposures and doses at the high end of the'distribution, these estimates will involve
considerable uncertainty which the exposure assessor will need to describe. Note
that in this context, the probabilistic analysis addresses variability of exposure-in the
population. Probabilistic techniques may also be applied to evaluate uncertainty in
estimate (see section 5, below). However, it is generally Impropriate to combine "
distributions reflecting both uncertainty and. variability to get a single overall
distribution. Such a result is not readily interpretable for the concerns of *
environmental decision-making. * ¦' ! *
If only limi^'informaticm on the distribution of the exposure or dose factors is
available, the assessor should approach estimating the high end by identifying the
most sensitive variables and'using Mgh end values for a subset of these variables,
¦ leaving others at their central valoes.*' M doing this, the assessor needs to avoid
combinations of. parameter values thafaie mconsistBit (e.g., low body weight used
^ m combination with high dietary nttake rate), and mist keep In mind the ultimate'
objective of being within the distribution of actual expected exposures and doses,
and not beyond it '
If very little data are available on the ranges for the various variables, it will be -,
difficult to estimate exposures or doses and associated risks in the high end with ,
. much confidence. One method that has'been used in'such cases is to .start with a
bounding estimate and "bade ©IF; the limits used until the" auilriiiation of
^ parameter values is, in the judgment dl the assessor, within the distribution of .
expected exposure, and stil ies wittdn the upper 10% of pmoiis exposed
Obviously, ttus method results in a large uncertainly and requires explanation.
• ' * «
b. Central tendency descriptor
Central tendenqr descriptors generally reflect central estimates of exposure or dose.
The descriptor addressing central tendency may be based on either the arithmetic
mean exposure (average estimate) or the median exposure (median estimate), either
^Maximizing all variables will in virtually all ass result In an estimate that is above the
actual values-seen in the population. When. the principal paramete» of the dose equation, eg.,
concentration (appropriately Integrated over time), intake rate, arid duration, are broken out into sub-
components, it may be necessary to use maximum tallies far more Hum two of these sub-ccanp««t
parameters, depending on a sensitivity analysis.
15
-------
oI which should be clearly labeled. Hie average estimate, used to approximate the
arithmetic mean, can often be derived by using average values for all the exposure
factors.5 It does not necessarily represent a particular individual on the distribution.
Because of the skewness of typical exposure profiles, the arithmetic mean may differ
substantially from the median estimate (I.e., 50th percentile estimate, which is equal
to the geometric mean for a log normal distribution). The selection of which
descriptors) to present in the risk characterization will depend on the available data
and the goals of the assessment When data are limited, it may not be possible to
consixud true' median or mean estimate, but it is still possible to construct
estimates of cental tendency. Hie discussion of the use of probabilistic techniques
in Section 1(a) above also applies to estimates of central tendency. " ...
2. _ Information about population exposure leads to another important way to •
dtesmberislL ' . ¦' • *
. • »
Population risk refers to an assosmait of the extent of harm for the population as a
whole. In theory, it can be calculated by summing, the individual risks for all
individuals within the subject population. This task, of course, requires a great deal
more information than is normally, if ever, available.
The kinds.of questions addressed by descriptors of population risk include the
following: _ ¦ •
• How many caw of a particular health effect might be probabilistically estimated
in this population for a specific time period?
• For non-carcinogens, what portion of the population, is ' within a specified range
of some reference level; e.g.# ©cceedance of ti* RfD (a, dose), the RJC (a
concentration), or other health concern level? .
• For carcinogens, what portion of the population is above a certain risk level,
such as 10-6? . ,
These questions can lead to two different descriptors of population risk.
a. Probabilistic number of .cases . ,
The first descriptor is the probabilistic number of health effect caw estimated in the
population of interest over a specified time period. This descriptor can be obtained
either by (a) summing the individual risks over all the 'individuals In the
population, e.g. using an estimated distribution of risk in the population, when
SThis holds true when variables are added (e.g.» exposures by different routes) or when
independent variables are multiplied (e.g., concentration x intake). However, it would be incorrect for
products of correlated variables, variables used as divisors, or for formulas involving exponents.
16
-------
such information Is available, or (b) through the use of a risk model that asmimpg a
linear non-threshold response to exposure, such as many carciitogwiic Models. In
these calculations, data will typically be available to address variability in individual
exposures. If risk varies linearly with exposure,-multiplying the mean risk by the
population size produces an estimate of the number of cases.6 At the present time,
most cancer potency values represent plausible upper bounds on risk. When such a
value is used to estimate numbers of cancer cases, it is important to understand that
the result is also an upper bound. As with other risk descriptors, this approach may
not adequately address sensitive subgroups for which different dose-response curve
or exposure estimates migfrt be needed. • - . ¦ ' ¦ r
- *
Obviously, the more information one has, the more certain the estimate of this risk
descriptor, but inherent tmcertatofies-lii risk assessment methodology place
limitations on the accuracy of the estimate. The discussion of uncertainty involved
in estimating .the nmnberof:;caseS'ShouId Indicate that this descriptor is not to be
of cases In Itepopidirtfari (which is a " ' ¦
statistical prediction based cm a great deal of empirical data).
In general, it should be recojpized that when small populations ase exposed,
population risk estimates may be very small.. For example, if 100 people are exposed
to an individual Mfetfrae'cancef risk of 10-*, the ecpecieii number of cases Is 0.01~ In
such situations, individual risk estimates will usually be a more meaningful
parameter for decision-makers.. - ' '
b. Estimated percentage of population with risk greater than some level
For non-cancer effects, we generally have not developed the risk assessment ¦
techniques to die point of knowing how to add risk probabilities, so a second'
descriptor is usually more appropriate: An estimate of the percentage of the
population, or the number of persons, above a specified level of risk or within a
specified range of some reference level, eg., exceedance of the RfD or the RfC> ' ¦
LOAHL, or other spedfie level of interest'- This descriptor must I* obtained through
measuring or 'simulating the population distribution. '
* * » t,
3. Information about the distribution of exposure and risk for different subgroups
of the population are important components of a risk assessment * ' -
¦ A risk manager might also ask potions about the distribution of the risk burden
among various segments of die subject population such as the following: How do
exposure and risk impact various subgroups?; and, what is the population risk of a
^However, certain Important cautions apply (see EFA'a Exposure Assessment Guidelines). AJso#
this is not appropriate for non-caxdnogenic elfecte or for oilier types of ameer models. For nan-linear
cancer models, an estimate of population risk must be calculated using the distribution of individual
risks.
17
-------
particular subgroup? Questions about the distribution of exposure and risk among
such population segments require additional risk descriptors.
a. Highly exposed
HigWy exposed subgroups can be identified, and where possible, characterized and
. the magnitude of risk quantified. This descriptor is useful when there is (or is
expected to be) a subgroup experiencing significantly different exposures or doses
from tot of the larger population. These sub-populations may be identified by age,
sex' lifestyle, economic factors, or other demographic variables. For example, -
toddlers who play In contaminated soil and high fish ansumos represmt sub-
populations that may have "greater exposures to certain agents.
b. Highly susceptible
i
HigMy susceptible subgroups can also be identified, and if possible, characterized and
the magnitude of risk quantified, This descriptor is useful when the sensitivity or
susceptibility to the'effect for specific subgroups is (or is expected to be) significantly
different from- that of the larger population* In order to calculate risk for these • '
subgroups, it will sometimes-be necessary to use a different dose-response
relationship; e.g., upon exposure to a chemical, pregnant women, elderly people,
children, and people with certain illnesses may each be more sensitive than the
population as a whole. For example, children are thought to be both highly exposed
and highly susceptible to the effects of environmental lead. A model has been
developed that uses data on lead concentrations in different gnvirnrrniAnfeil media
to predict the resulting blood lead levels in children. Federal agencies are worldztg
together to develop specific guidance on blood lead levels that present risks to
children. - •
.
It is important to note, however, that the Agency's current methodologies for
developing reference.doses and reference Concentrations. (MPs and RfCs) are* •
designed to protect sensitive populations,; If date on sensitive human populations
are available (and there is confidence in the quality of the data), then the RfD is set at
the dose level at which no adverse effects are observed in the sensitive population
(e.g., RfDs for fluoride and nitrate). If no such data are available (for example, if the
RfD is developed using data from humans of average or unknown sensitivity) then
an additional 10-fold factor is used to account for variability between the average
human response and the response of more sensitive individuals.
Generally, selection of the population segments is a matte of either a priori interest
in the subgroup (e.g., environmental justice considerations), in which case the risk
assessor and risk manager can jointly agree on which subgroups to highlight, or a
matter of discovery of a sensitive or highly exposed subgroup during the assessment
18
-------
process. In either case, once identified, the subgroup can be treated as a population
in itself, and characterized In the same way as the larger population using the
descriptors for population and individual risk.
-1
4. Situation-specific information adds perspective on possible future events or
regulatory options.
What if...? questions can be used to examine candidate risk management options.
For example, consider the following:
• What if a pesticide applicator applies this pesticide without using protective
equipment?
• What if this site becomes residential in the future?
• What risk level will occur if we set the standard at 100 ppb?
Answering these "What if...?" questions involves a calculation of risk based on
spcofec combinations of factors postulated within die assessment?. The answers to
these What if...?" questions do not, by themselves, give information about how
~ - likely the combination of values might be in the actual population or about how
many (if any) persons might be subjected to the potential future risk. However,
information on the likelihood of the postulated scenario would also be desirable to
include in the assessment. ¦
When addressing projected changes for a population (either expected future
developments or consideration of different regulatory options), it is usually
appropriate to calculate and consider all the risk descriptors discussed above. When
cental tendency or high end estimates are developed for a future scenario, these
descriptors should reflect reasonable expectations about future activities. For
example, in site-specific risk assessments, future scenarios should be evaluated
when they are supported by realistic forecasts of future land use, and the risk'
descriptors should be developed within that context
5. An evaluation of the uncertainty in the risk descriptors is an important
component of the uncertainty discussion in the assessment
Risk descriptors are intended to address variability of risk within the population and
the overall adverse impact on the population. In particular, differences between
high end and central tendency estimates reflect variability in the population, but not
the scientific uncertainty inherent in the risk estimates. As discussed above, there
^Some programs routinely develop future scenarios as part of developing a risk assessment
Program-specific guidance may address future scenarios in more detail than they are described here.
19
-------
will be uncertainty in all estimates of risk. These uncertainties can include
measurement uncertainties, modeling uncertainties, and assumptions to fill data
gaps. Risk assessors should address file impact of each of these factors on the
confidence in the estimated risk values.
Both qualitative and quantitative evaluations of uncertainty provide -useful
information to users of the assessment The techniques of quantitative uncertainty
analysis are evolving rapidly and both the SAB (8) and the NRC (4) have urged the
Agency to incorporate these techniques into its risk analyses. However, it should be
noted that a probabilistic assessment that uses only the assessor's best estimates for
distributions of population variables addresses variability, but not uncertainty. .
Uncertainties in the estimated risk distribution need to be separately evaluated.
20
-------
REFERENCES
1, National Research Council. BjsjLA^esamfinLmJfafi-Efideial Government:
Management the Process. 1983.
2. US. EPA. Risk Assessment and Management Framework for Decision Making
3. US. EPA. "Risk Assessment Guidelines." 51 Federal Register, 33992-34054,
September 24,1986. • . '
4. National Research Council. Science and Judgement in Risk Assessment. 1994.
5.' U.S. EPA "Guidelines for Exposure Assessment*' 57 Federal Register, 22888-
22938, May 29,1992.
6. US. EPA- "Guidelines for Developmental Toxicity Risk Assessment" 56 Federal
Register, 67398-63826, December 5,1991. •
7. US. EPA. Framework for Ecological Risk Assessment. 1992. ¦ ¦,
8. Loehr, RA., and MatoosM, G.M., Lett® to Carol M. Browner, EPA
Administrator, Re: Quantitative Uncertainty Analysis for Radiological
• Assessments. EPA Science Advisory Board, July 23,1993 (EPA-SAB-RAC-COM-
93H006).
21
-------
ATTACHMENT B
RCRA/CERCLA INTERFACE-INTERIM FINAL GUIDANCE,
EPA REGION 10 MEMORANDUM, AUGUST 3
S 8EPA T -*.SK*'R£V1S£2'RNAtLMASTER WPD'}*>S-f* |ww*it|.|n/24rrt.!l :»*rw4ae
-------
Attachment
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
REGION 10
1200 Sixth Avenue
Seattle, Washington 98101
August 3, 1994
Reply To
Attn Of:
HW-124
memorandum
SUBJECT: RCRA/CERCLA Interface - Interim^Final
FROM:
Michael Gearheard, Chief
Waste Management Branch
fir
Carol Rushin, Chief
Superfund Remedial Branc,
0^—
TO;
^ James Everts, Chief \/p MJOfV (/
Uy Superfund Response/Investigation^ Branch
George C. Hofer, Chief Jl /jlmjP#
Federal Facilities Superfun^/Br
Hazardous Waste Division
INTRODUCTION
Over the years there have been some areas of confusion
between the RCRA and CERCLA programs, as one might expect when
you have two programs dealing with hazardous waste but using.two
separate statutes and sets of regulations and guidance. The
RCRA/CERCLA Interaction Workgroup was formed to identify areas
where the two programs have routinely overlapped, and there is
reasonable expectation that a consolidated approach would result
in efficiencies. A summary of their findings is attached. (If
you want the full report please call Sharon Smith at 3-6637). We
want to thank the Workgroup members for all their hard work and
great product we now have to work with.
Judi Schwarz
Christy Ahlstrom Brown
Dave Croxton
Marcia Bailey
Bill Adams
Thor Cutler
Nancy Harney
Ed Jones
It was particularly significant to us that the Workgroup
members all felt that they had learned as a result of their
experience and that it clearly demonstrates the benefit of cross-
program teaming.
Wrtm an ftacycJM Fnpiir
-------
2
This nemo is intended to give greater clarity to and set out
our expectations as how these potential overlaps should be
handled. If you have general situations that are not covered
here, please call Judi Schwarz as she is quite knowledgeable in
both program areas. We branch chiefs would also welcome a call
as a major part of our job is to work cross-program issues.
This guidance will not answer every question that.may cone
up in developing a site-specific comprehensive and non-
duplicative solution. We encourage creative solutions and
discourage rigid interpretation of the regulations and guidance,
but we also recognize that some solutions nay require input fro*
the section, and branch chiefs. We encourage you to involve us
early in the process rather than becoming frustrated by a
situation.
We are issuing this guidance as an interim final product.
We want to begin to apply this approach, but realize that we may
not have thought of all of the implications and problems that may
arise. If necessary, this guidance can be revised in the future
to reflect what we will learn about integrating our two EPA
programs. Other revisions may also be necessary if and when this
approach is applied to EPA RCRA/State clean-up program or
CERCLA/State RCRA program overlaps.
PARITY POLICY - THE BIG PICTURE
We are committed to doing everything we can to avoid
duplication between the Region 10 RCRA and CERCLA programs. To
this end, we believe RCRA and CERCLA program managers and staff
need to: (1) be knowledgeable about both programs, (2) maintain
close communication where sites may involve, both programs and
develop a coordinated strategy such as a site management plan for
such sites, and (3) recognize that either program will produce
substantially equivalent cleanup outcomes. Actions are set forth
below to help achieve these goals.
Because we believe that the environmental outcome reached at
a site managed under CERCLA or RCRA will be similar, we are
declaring parity between RCRA corrective action and CERCLA
remedial action decisions. Parity means that a site-specific
decision under one program will be considered equivalent to a
decision under the other program. The CERCLA and RCRA corrective
action programs rely on similar risk-based approaches, and they
address remediation of past activities/practices. Under parity,
one program will normally not recheck or re-open unit-specific
decisions made by the other program.
-------
3
Declaring parity between the RCRA and CERCLA programs at
RCRA regulated units1 themselves is not as simple. Overlaps
between these two programs are most likely to occur at facilities
that have interim status for either operating, or closing units or
have illegal RCRA regulated units (e.g., they managed hazardous
wastes in a unit without having achieved interim status). For
those RCRA regulated units that have not yet been permitted,
there may be RCRA interim status requirements,such as groundwater
monitoring, financial assurance, and closure and post-closure
care that will still have to be met by the facility even if
CERCLA is involved at the site. There nay also be issues of
regulatory compliance or violations. ' However, through the steps
outlined in this memo, the requirements of both programs can and
should be net with a single coordinated approach.
In summary, for RCRA corrective action and Superfund
remedial actions, the actual environmental results achieved
through cleanup'are expected to be environmentally equivalent.
In addition, application of RCRA closure and post-closure
1 "RCRA 101" - Definitions:
A regulated unit is a unit, such as a landfill, surface impoundment,
storage area, etc., within a RCRA Treatment, Storage, or Disposal Facility (or
TSDF), that managed RCRA hazardous waste at any time after the appropriate
regulation went into effect. (Generally this means any time after November
1980.) A facility cannot be a TSDF unless it has at least one RCRA regulated
unit. A Solid Haste Management Unit (or SWMU) is any discernible unit at a
TSDF at which solid wastes have been placed at any time, irrespective of
whether the unit was intended for the management of solid or hazardous waste.
Regulated units are a subset of SWMUs. Areas of Concern (AOC) are areas
within a facility that are known or suspected to be contaminated'by hazardous
constituents but which were not a location of solid waste management. A one
time product spill is an example.
. All TSDFs are required to eventually have RCRA permits or go through
closure of the regulated units. Permits must include long-term post-closure
care for units that cannot clean close. Until EPA or the delegated state has
issued chat permit, or until the facility is clean closed, the TSDF is subject
co the interim status requirements found in 40 CFR 265.
Corrective action is required for all releases from SWMUs. For
facilities that are being permitted, corrective action requirements must be
part of the permit. Corrective action can also be required at interim status
facilities through administrative orders.
ft facility that only generates RCRA hazardous waste and stores this
hazardous waste for less than 90 days in tanks or containers under certain
conditions is a hazardous waste generator subject to the RCRA generator
standards found in 40 CFR 262 and is conditionally exempt from the storage
permit requirements. Such an exempt storage facility is generally not subject
to RCRA corrective action under a RCRA 3008(h) order. However environmental
problems could be addressed through RCRA's imminent hazard order authority
(Section 7003) or RCRA's investigation authority (Section 3013.}
Please note that these definitions are somewhat simplified and should -
not be* relied upon to give the correct answer in all situations.
-------
4
processes and Superfund remedial actions are also expected to
achieve environmentally, equivalent results. The only exception
to this Is that some proposed/potential Superfund No Further
Action decisions at regulated units may require additional steps
to ensure that the RCRA clean closure standards are also met.
PROCEDURES
In
Clean-up decisions will continue to be routinely presented
in documents such as fact sheets. Statement of Basis,
Decision documents, and Records of Decision (RODs). However,
these decision documents and the related public notices and
Proposed Plans should explain that the selected action will
satisfy the requirements for remediation under both statu -
addition, as long as the clean-up action fits into the pari y
categories described above, the two hazardous waste programs will
not cross-check decisions. The workgroup did recommend that
informal peer review be used to inform and educate (see below).
Either program may perform or postpone some or all of the
cleanup areas as long as there is no duplication and the decision
is reflected in the site-specific coordinated strategy. The
decision should be based on, among other things, tilting,
resources and environmental priority, and should involve
consultation with the other program.
Where RCRA corrective action is being incorporated into
Superfund activities, the need for action at all identified Soli
Waste Management Units (SWMUs) and Areas of Concern.(AOCs) will
be considered by the Superfund program. Whenever possible, the
RCRA Facility Assessment (RFA) should be started and completed
early enough so that the results of the RFA can be factored into
the Superfund process in a timely and coordinated manner. If tor
some reason, one or more SWMUs are not identified or considered
in the superfund evaluation process, the RCRA program may choose
to evaluate the need for further investigation at such units, •
particularly where there is or will be a RCRA permit.^for
regulated activities. Superfund should document their
evaluations of SWMUs even if they are not addressed in the ROD.
BUT
That is not the whole story. The existence of one or more
RCRA regulated units at an PPL site raises several additional
concerns regarding that regulated unit that must be addressed.
These are:
1. if the regulated unit is not closing, it must obtain a
RCRA permit. The permit must, by law, include site-
* wide corrective action. At NPL sites, this corrective
action requirement can be met in the permit by
referencing a legally enforceable CERCLA agreement
-------
5
(e.g., see the permits for Elmendorf AFB, and Fort
Wainwright).
2. if the regulated unit is closing by clean closure,
certain requirements apply and must be considered even
though the environmental outcome reached under RCRA and
CERCLA is substantially equivalent.- The Regulated Unit
Checklist that will be developed and the discussion
below on public participation requirements for closing
regulated units will give guidance on how we can
satisfy both programs1 requirements.
3. If the regulated unit will be closed as a landfill -
i.e., with waste in placje - then different procedural,
tilling and substantive requirements may apply. Three
of these requirements are outlined below. The
Regulated Unit Checklist will provide a complete list.
In these situations, the respective site managers need
to develop a more detailed site-specific coordinated
strategy.
USING CERCLA TO CLOSE INTERIM STATUS REGULATED UNITS
The purpose and scope of the RCRA regulatory provisions for
closure and post-closure of regulated hazardous waste management
units/activities are not identical in scope or purpose to RCRA
Corrective Action or Superfund cleanup provisions. There may be
regulatory obligations and schedules applicable at the facility,
as well as procedural requirements, which make it wore difficult
to declare universal parity when we want CERCLA activities to
achieve RCRA closure of regulated units. Nonetheless, we must
strive for reduction or elimination of duplication of effort.
The following addresses the major potential differences in
approach or scope and identify considerations necessary to
determine that Superfund activities satisfy regulated unit
requirements.2
2-CERCLA 101- - Applicable or Relevant and Appropriate Requirements
(ARARs>.
Any remedial action selected under CERCLA must meet two "threshold"
criteria: protectiveness, and ARARs. To comply with the "applicable" part of
ARARs requirement, a remedy must meet all substantive promulgated
environmental requirements that would apply if the site was not a Superfund
site. Under the "relevant and appropriate" part of the ARARs requirement, a
remedy must meet all substantive promulgated environmental requirements that
fit the circumstances at the site, even though those requirements would not
apply if the site was not a Superfund site. If Superfund decides that
remedial action is necessary at a RCRA regulated unit, the substantive parts
of the RCRA regulations, such as landfill closure requirements, would be
applicabie and would have to be followed. Superfund could also require
additional actions beyond the RCRA regulations if necessary to meet the
"protectiveness" threshold criteria.
-------
6
1. Groundwater Monitoring Requirements (40 CFR 265 subpart, F)
The RCRA groundwater monitoring requirements for interim
status regulated land-based units are designed to detect unit
specific releases from units that have yet to leak or to assess
the nature, rate and extent of releases which_have been detected.
For regulated units that have leaked, there nay be little reason
to continue to strictly apply the interim status groundwater
requirements in those cases where the contamination has .been
successfully assessed and where the priority is to remediate such
contamination or to monitor the effectiveness of the remedial
actions. Both Superfund and RCRA Corrective Action groundwater
monitoring are oriented towards effective remediation.
Therefore, in the future, we expect that for each facility where
program overlaps occur, groundwater detection monitoring
requirements will be designed to meet the requirements of both
programs, so that the RCRA groundwater requirements at leaking
regulated units can be sufficiently addressed by CERCLA.
2. Closure (40 CFR 265 Subpart G)3
The respective site managers should develop a coordinated
strategy to determine which authority/program will address which
closing units and to insure that either the closure plan approval
process or CERCLA proposed plan and ROD are designed to satisfy
the. respective administrative and procedural requirements. A
potential Superfund No Further Action decision at a regulated _
unit is a special case that requires early cross-program planning
as part of this coordinated strategy.
Both programs have public participation requirements.^ The
public notices and proposed plans should be written to satisfy
the requirements of both programs. It may be also appropriate to
include a discussion of this joint approach in documents seeking
. public comment.
n
with the consent of the programs and the facility, Superfund
may manage the closure and post-closure care of a RCRA regulated
unit.* In such a case, an enforceable Superfund process may allow
the RCRA program to delay formally processing a post-closure
permit,
.a 1
3. Financial Assurance Instruments (40 CFR 265 Subpart H)
3 "RCRA 101", continued - Types of permits:
Regulated units that are operating have to get an operating RCRA permit.
Regulated units that are closing {and this is the majority of regulated units)
either go through "clean" closure or "landfill closure." Clean closed units
require no further controls or actions. "Landfill" closures are all other
closures. Currently, "landfill" closure units are required to have a post-
closure permit to ensure long-tern care.
-------
7
This RCRA regulatory requirement is not applicable at
federal facilities. For non-federal facilities, facility
owners/operators should be made aware that even if CERCLA is
taking responsibility for investigation or remediation of a
facility with RCRA regulated units, the self-implementing '
regulations of RCRA, including financial assurance, still apply.
Depending on the financial viability of the owner/operator, it is
sometimes not possible for them to meet this requirement.
Violations may be addressed through various RCRA enforcement
mechanisms.
IMPLEMENTATION
Identification of facilities/sites where program overlap nw
occur. To aid in identifying where areas of overlap and
therefore duplication of effort may occur, attached is a
list of facilities which appear on both the CERCLA NPL and
the RCRA treatment, storage, disposal facility (TSDF) list.
This is not a static list so as new RCRA TSDFs are
discovered, or as new sites are listed on the NPL, the RCRA
staff should call David Bennett to see if this site is also
on the NPL and the CERCLA staff should call Patricia Hanley
to see if the site is a TSDF.
Develop site-specific cooi^jjiafcfifl_gfeOfcgg3L^ha£-ggvg££-JU3fl
integrates the points above. Site managers working on a
facility on this list are expected to seek out their
counterpart in the other program and develop a written
coordinated strategy for the site. Such strategies should
be reviewed by both programs' site managers at least
annually and updated as needed. Our state counterparts
should also be involved in the development of the site-
specific strategies, where appropriate.
The site-specific coordinated strategy may be fairly simple
at those sites where CERCLA is handling only SWMUs and AOCs
that may require corrective action. A more detailed plan
may be appropriate where CERCLA is involved with any
regulated units.
When developing a coordinated strategy, it may be helpful to
keep in mind that the differences between approaches to site
remediation between the programs may be primarily a factor
of the individual project managers rather than program
specific differences. The workgroup observed that
differences were more likely to be based on individual
practices or philosophies, and less likely due to statutory
or regulatory requirements. The workgroup also felt that
similar differences exist among sections of the same branch
and among individuals within the same section. In other
-------
8
words, there is frequently room for flexibility when it is
needed.
Management: follow-up. Managers should support and follow-up
on the development of the site-specific strategies.
Creative and effective approaches should-be shared-^ In
addition, managers should regularly ask questions lite: Are
there any RCRA regulated units or CERCLA operable .units at
this site? Where are these in relationship to our program's
concerns? How are we coordinating our approach? What else
night we be doing at this site (or in this document) to
avoid re-work?
Remedy Selection Tii formation Exchanges^ it is important
that we do more to share information between our programs on
our remedy selection decisions and our reasons for these
decision. Three steps are planned.
Regular presentations. There will be three remedy
selection presentation meetings a year, at which each
clean-up branch would present one site or facility
decision, to be followed by a discussion period. A
technical staff group is working on setting up these
presentations. 1
Decision summaries... All future remedy/corrective
action decisions will be summarized and distributed on
the LAN to HMD staff. These information exchanges will
focus for now on clean-up level and action (or the "go
do something") levels. The same technical staff group
is working on setting up a format for these summaries.
informal Peer Review. We encourage informal peer
review at the appropriate decision points in the RCRA
and CERCLA remedial processes. This review is intended
to inform and educate both staffs; it is not intended
to force either program to gain technical/regulatory
approval from the other. Such a macro-scale peer
review would not include review of detailed documents.
One way to increase such informal peer review is to
invite staff from other HWD programs to attend internal
briefings or discussions. Jtidi and other workgroup
members are available to help identify interested
staff.
¦state Delegated RCRA Programs and Other SfcafcB_Clfian=
up/Remedial Programs. The workgroup was created and this
memo was written to address EPA Region 10 HWD concerns.
However, this memo and our general approach will be shared
with our state counterparts and discussed during our annual
' meetings.
-------
9
The states may wish to participate in a similar approach
either with EPA (e.g., at sites where the state has the RCRA
lead and EPA has the CERCLA lead), or between two state-only
programs. For now, a site-by-site decision is the
recommended approach. Such mutual state-EPA decisions
should be memorialized in writing. .
OTHER ACTIONS TO IMPROVE RCRA/CERCLA INTERACTION
Training. RCRA program managers will provide RCRA
orientation training for all interested Superfund staff and
managers. In addition, training which addresses the
implementation and ramifications of this policy will be
provided for CERCLA and RCRA Program Managers and staff.
Regulated Unit Checklist. RCRA program managers and staff
will develop and distribute to the CERCLA program a
checklist of what, according to the regulations, has to
happen for closure, and where necessary, post-closure care
at regulated unit. This list will form the basis of
negotiations between the programs. This list will also be
the minimum list of items that must be addressed in the
coordinated site strategy.
A schedule which outlines when these steps will be taken and
by whom can be found in the second attachment.
We all recognize that improving RCRA/CERCLA^ interaction will
be a continuing process. It is our hope that this memorandum and
approach will make it easier for Regional RCRA and CERCLA staff
to reduce duplication of effort in the context of remedial
decision-making. Suggestions for further improvements aie always
welcomed and should be sent to any of us or to Judi Schwarz.
Attachments:
Summary of workgroup results
List of TSDF/NPL sites (June 1994 draft)
Implementation and next steps schedule
NOTICE: The policies set out in this document are intended solely as guidance
to EPA Region 10 personnel; they are not final EPA actions and do not
constitute rulemaking. These policies are not intended, nor can they be
relied upon, to create any rights enforceable by any party in litigation wi.th
the United States. EPA officials may decide to follow the guidance provided
in this document, to act at variance with the guidance, based on an analysis,
of specific site circumstances. EPA also reserves the right to change the
guidance at any time without public notice.
-------
RCRA/CERCLA Interaction Workgroup Results - 2/94
ISSUES (in general priority order)
Issue Mo. l - Closure of Regulated Units: Where superfund
is requiring remedial action at a regulated unit, is it
enough to satisfy RCRA closure requirements?
<• s
Issue No. 2 - Closure of Regulated Units: Where superfund
has made a "No Action" decision at a regulated unit, what
next?
Issue No. 3 - Removals at RCRA facilities are short-
circuiting the RCRA process.
Tisr XX
Issue No. 4 - Different Corrective Action and Remedial
Action approaches create uncertainty.
Issue No. 5 - Groundwater Monitoring at Regulated Units:
Different Groundwater monitoring requirements at regulated
units result in inefficiencies.
Issue No. 6 - No action while site transfers from one
authority to another.
Issue No. 7 - superfund-type actions at sites with operating
Regulated Units. ¦
Issue No. 8 - Superfund investigations at facilities with
regulated units are not always coordinated with RCRA - and
RCRA inspections and other actions are not always
coordinated with Superfund.
Issue No. 9 - The definitions of "Site" vs. "Facility"
result in different universes under focus.
GENERAL OBSERVATIONS AND RECOMMENDATIONS
%
- Region 10 lacks a comprehensive list of sites/facilities
which have the potential for dual program regulation. Such a
list should be developed.
- There is a need for a formal statement of "parity" between
the clean-up programs.
-------
'£t
v r.• v
,60 sr«,
-------
June 1994 - draft
LIST OP CERCLA AMD TSD FACILITIES
Washington:
U.S. Bremerton Naval Complex (was Navy Puget Sound Naval
Shipyard)
USAF Fairchild
U.S. Amy "Fort Lewis
U.S. Navy NUWES Keyport
U.S. Navy Fuel Department NSC Puget Sound
Kaiser Aluminum and Chemical Corp. Mead Plant
Port Hadlock
U.S. DOE Hanford
McChord AFB
BPA Ross Complex
Oregon;
Teledyne Wah Chang
Umatilla
Mate®.
Mountain Home AFB
Union Pacific Railroad Company
INEL
Eastern Michaud Flats (i.e., includes FMC)
Alaska;
NAS ADAK
Eielson AFB
Elmendorf AFB
Fort Richardson
Fort Wainwright
Other RCRA/NPL overlap Sites
Oregon;
Martin Marietta
-------
«
3
C
C
(8
-o
jC
o
If
u
4J)
•M
m
u
-O n
•
*> *>
at *
to 13
Q<|
«M 3
XI
01
«
.3
C
C
¦a
>»
>.
¦Q
m
¦W o
® e
rH *t
ft nt
o H
O 3
u
o
0)
14
©
4J
>•4
V4
?
4J
•*4
6
u
0)
^4
<
S
u
a
ec
m
TJ
o
c
c
ill
(0
(ft P*
x e
a* o
cc u
^ i
HI
<1
u
ss
w
4J
c
01
*0
0
V
E
m
1
m ^
cc «
5h
jd
m
4J I
m m
*4 JJ
0*4
!*
U 3
e
o
Pt«M
s
o. •
ec
a
*0 u
c c
« 4
m a
x
o» .
PC o
«
c
•
Ai
Jtf
—I
u
3
m
•
*» -•
-~
c
O u
m
m A
m 0
® *i
m o
O
c
o c
«4 —»
¦u
0 O*
4J C
c
® *»
B U
# <8
u u
o>»
# -
• u
u «
X •
H >.
C A
« JO
n m
3 Q
w
» U
J3 £
- O
o.
3 -
C
JJ o
• u
« v«
»
0»JS
c o
-4 JJ
® 3
03 X
c
«
o*
•0
C fa
£
o •
<-5 j
0 <• 3
X «0
c
o
—<
iJ
m
E
u
0
»
c
o
u
c
®
Bl
at
u
a
3
0»
®
0£
>,
O
*»
*
43 »
M
e -•
Vt -s.
O O
fc r-l
TJ «
e >4
4J ®
m c *>
> o
o e v»
a » 3
a
Ot*>
>. 6 -*
jo -* tz
Q, m » a «
a
u c < ¦
• a «
» jC o ®
u K (J
o» c
C -• -O «
C ***^
q ~l 14 <-4
as o.
w e
u -x o
4 OiCV O
e 3 «
y o *o
O >4 >.C
6» 0»J5 <
D»
C
0
o>
1
c
o
o
"O
«
4J
m
™ u
? ai
y
u u
"4 (0
U |
3
i n
a
3
O ^4
M H4
14 «
o
< w
a »
o •
« Vi
« ®
o *»
c
S"
u *o
« c
at a
c
o «
>»J=
14
8) C
> 0
a o
u
®
e
a
e
w
o
B
U
9
®
«
c
m
x
«
jj
w
14
3
o
ft
3
3
0
><
c
a
*r
m
m
p.
*1 T>
a •
3 TJ
o»»
3 •
•
u
a
m
m
"O
C
«
•a
u
«s
»
£
o
»
o
Q,
®
Vl
<0
£
tn
<
J
o
o:
u
o
Vl
o
O •
g
o
az >«
u
o»—I
C -t
t
u
m
5 H4
6 o
c c
0 o
«#°t »«4
» 4J
vl 3
> XI
*4
O Vl
JJ
1 m
•w
TJ
o»
o* o»
C
*» 3
• o
3 t-i
O
3 O
« T
•a
c
m
«
•o
c
Si
o
X
o
c
4
n
CD
-I c
M "
« r
a*
o
u
«
»
«
•a
o
•a o»
m 9
« *»
z «
W U
4J
0
ff» II
rt
c
o
u
X
o*
u
o
>
rt
a
•o
c
M
Tt
U
m
9
£,
Vl
It
m
o
0
M
IN
u
a>
s-l
s«H
0 c
X
On O
o
B
s U
J£
O
-------
ATTACHMENT C
EXAMPLES OF DATA QUALITY ASSESSMENT APPLICATIONS
^ hm'juii'» .* task* mvw.2 fin*lmastw v.ppim-5m
-------
-------
Appendix 1 from EPA (1994a)
Guidance for tie Data Quality Objectives Process, EPA QA/G-4
* m-pA i srvssi: hnaumaster 3*nt.u«
-------
APPENDIX B
DQO CASE STUDY: CADMIUM-CONTAMINATED
FLY ASH WASTE.
Introduction
Thjs appendix presents a functional, but realistic example of the DQO outputs for a
decision that could fce made within the Resource Conservation and Recovery Act (RCRA)
hazardous waste management program. The example is intended to illustrate the types of
outputs that are common to the DQO Process. It is not intended, however, to represent the
policy of the RCRA program for actual situations that may be similar to the example. Please
consult with a knowledgeable representative within the RCRA program office about the
current policy for making waste classification decisions for fly ash or other types of
hazardous waste.
The case study has been chosen because it is simple and straightforward, and because
the outputs are uncomplicated. Although some of the outputs from this example may seem
intuitive, this is not often the case in practice. For many studies, the DQO Process is
complicated and thought-provoking. Even so, some steps will require more effort than others.
Keep in mind that all of the steps in the DQO Process are necessary to develop a data
collection design. Once the first six steps have been completed and thoroughly thought-out,
then development of the most resource-effective data collection design can proceed.
Background
A waste incineration facility located in the Midwest routinely removes fly ash from its
flue gas scrubber system and disposes of it in a local sanitary landfill. Previously it was
determined that the ash was not hazardous according to RCRA program regulations. The
incinerator, however, recently began treating a new waste stream. The representatives of the
incineration company are concerned that the waste fly ash could now contain hazardous levels
of cadmium from the new waste sources. They have decided to test the ash to determine
whether it should be sent to a hazardous waste landfill or continue to be sent to the municipal
landfill. They have decided to employ the DQO Process to help guide their decision making.
Cadmium is primarily used as corrosion protection on metal parts of cars and electrical
appliances. It is also used in some batteries. Cadmium and cadmium salts have toxic effects
for humans through both ingestion and inhalation exposures. Ingestion exposure usually
causes mild to severe irritation of the gastrointestinal tract, which can be caused by
concentrations as low as 0.1 mg/kg/day. Chronic (long-term) inhalation exposure can cause
increased incidence of emphysema and chronic bronchitis, as well as kidney damage.
EPA QA/G-4
47
September
-------
Under the current Code of Federal Regulations, 40 CFR. Part 261, a sold waste can
be considered "hazardous" if it meets specific criteria of ignitability, corrosivity. reacuvity,
and toxicity One method that is used to determine if a solid substance, such as iy ash,
2cm the criteria for toxicity under the RCRA program iqulrtms hmmi a
sample" of the waste and, perform a Toxicity Characteristic Jxaciuiig ftoccdure CTCLP)
described in 40 CFR. Pt 261, App, D. During this process, the sold fly ash will be
"extracted" using an acid solution. The extraction liquid (the ICLP ieacbaie) will then be
subjected to tests for specific metals and compounds. For ibis ample, theoolywnamB
with the concentration of cadmium to the leachate. He primary benefit of the DQO Process
wifl be to establish the data collection design needed to determine if the waste is hazardous
under RCRA regulations within tolerable decision error rales.
As a precursor to the DQO Process, the incineration company has conducted a pilot
study of the fly ash to determine the variability in the concentration of cadmium between
loads of ash leaving the facility. They have determined thai each load is fairly homogeneous.
There is a high variability between loads, however, due to the nature of the waste stream.
Most of the fly ash produced is not hazardous and may be disposed of in a sanitary landfill.
Thus, the company has decided that testing each individual waste load before it leaves the
facility would be the most economical, Then they could send loads of ash that exceeded the
regulated standards to the higher cost RCRA landfills and continue to send the others to the
samtaiy landfill.
POO Development
The following is a representative example of the output from each step of the DQO
Process for the fly ash toxicity problem.
State the Problem — a description of the problcm(s) and specifications of available
resources and relevant deadlines for the study.
(1) Identify the members of the planning team — The members of the planning team will
include the incineration plant manager, a plant engineer, a statistician, a quality
assurance officer, an EPA representative who works wilMn the RCRA program, and a
chemist with sampling experience.
(2) Identify the primary decision maker — There will not be a primary decision maker,
decisions will be made by consensus.
(3) Develop a concise description of the problem — The problem is to determine which
loads should be sent to a RCRA landfill versus a sanitary landfill.
(4) Specify available resources and relevant deadlines for the study — While the project
will not by constrained by cost, the waste generator (the incineration company) wishes
to hold sampling costs below $2,500. They have also requested that the waste testing
.be completed within 1 week for each container load.
e* QA/CV4
48
September JW4
-------
Identify the Decision — a statement of the decision thai will use environmental data and the
actions that could result from this decision.
(1) Identify the principal study question — Is the if ash waste considered hazardous
under RCRA regulations?
(2) Define alternative actions thai could result from resolution of the principal study
question —»
(a) The waste fly ash could be disposed of in a RCRA landfill.
(b) The waste fly ash could be disposed of in a sanitary landfill.
(3) Combine the principal study question and the alternative actions into a decision
statement — Decide whether or not the fly ash waste is hazardous under RCRA and
requires special disposal procedures.
(4) Organize multiple decisions — Only one decision is being evaluated.
Identify the Inputs to the Decision — a list of the environmental variables or characteristics
that will be measured and other information needed to resolve the decision statement.
(1) Identify the information that will be required to resolve the decision statement To
resolve the decision statement, the planning team needs to obtain measurements of the
raHminm concentration in the leachate resulting from TCLP extraction.
(2) Determine the sources for each item of information identified — The fly ash should be
tested to determine if it meets RCRA regulated standards for toxicity using the test
methods listed in 40 CFR, Pt. 261, App. n. Existing pilot study data provide
information about variability, but do not provide enough information to resolve the
* decision statement.
(3) Identify the information that is needed to establish the action level — The action level
will be based on the RCRA regulations for cadmium in TCLP leachate.
(4) Confirm that appropriate measurement methods exist to provide the necessary data
Cadmium can be measured in the leachate according to the method specified in 40
CFR. PL 261, App. EL The detection limit is below the standard.
Define the Boundaries of the Study — a detailed description of the spatial and temporal
boundaries of the problem, characteristics that define the population of interest, and any
practical considerations for the study.
EPA QA/C-4
49
September W*
-------
,.. characteristics that define the papulation of interest — Fly ash waste from
the hazardous waste incinerator will be analyzed The fly ash should not t*m«cd
the hazard rxceDl watcr thai is used for dust control. Each load of
""h" hould fUl at least 70% of the waste trailer. In cases where the trailer is filled less
to 70%. the trailer must wait on-site until more ash is produced and Gils the trailer
10 the appropriate capacity.
(2) Define the spatial boundary of the decision statement —
U) Define the geographic area to which the decision statement applies. Decisions
wm apply <° eLh container load of fly ash waste.
fhl When appropriate, divide the population into strata that have relatively
feLireneoL characteristics Stratification is not necessary since the waste ash l.
relatively homogeneous within each container.
(3) Define the temporal boundary of the decision statement —
(al Determine the timeframe to which the decision statement applies. It will be
that the sampling data represent both the current and future concentration
of cadmium within the ash.
(b) Determine whm m collect dam k tie tracks, the waste does not pose
a to humans or the envinHunenL Additionally. since lie fly ash ,s nor
subject to change, disintegration, or alteration, the, decision about the waste _
characteristics does not warrant any temporal constraints. To expedite decision
making, however, the planning team has placed deadlines on sampling aod
leoortiiiit The fly ash waste wiM be tested within 48 hours of being loaded onto
wEte hauling trailers. The analytical results from each sampling round should be
completed and reported vtifila 5 working days of sampling. Untd analysis is
complete, the trailer cannot be used
(4) Define the scale of decision making — The scale of decision making will be each
container of waste ash,
(5) Identify practical constraints on dam collection — The most important practical
( } cSScm. mm could interfere with the study is tie .Mfty to take samples from the
fly ash that is stored in waste banting trailers.. Although the: twtas have open access,
special procedures and methods will have to be implemented for the samples to be
representative of the entire depth of the ash. It has ten suggested IhM core samples
ma™' pneta! solution'to this problem. To get addtional
truck and to minimize the cost, compositing of core samples has been sugges .
1PAQWG4
50
September 1994
-------
Develop a Decision Rule — to define the parameter of interest; specify the action level and
integrate previous DQO outputs into a single statement that describes a logical basis for
choosing among alternative actions.
(1) Specify the statistical parameter that characterizes the population of interest — The
planning team is interested in the true mean concentration of cadmium in the TCLP
leachate for each container.
(2) Specify theaction level for the study — The action level for the decision will be the
RCRA regulatory standard for cadmium of 1.0 mg/L in the TCLP leachate.
(3) Develop a decision rule (an "if-then..." statement) — If the mean concentration of
cadmium from the fly ash leachate in each container load is greater than 1.0 mg/L
(using the TCLP method as defined in 40 CFR 261), then the waste will be considered
hazardous and will be disposed of at a RCRA landfill. If the mean concentration of
cadmium from the fly ash waste leachate is less than 1.0 mg/L (using the TCLP
method as defined in 40 CFR 261). then the waste will be considered non-hazardous
and will be disposed of in a sanitary landfill.
Specify Tolerable Limits on Decision Errors — the decision maker's; tolerable decision
crror rates based on a consideration of the consequences of making a decision error.
(1) Determine the possible range of the parameter of interest — From analysis of records
of similar studies of cadmium in environmental matrices, the range of the cadmium
concentrations is expected to be from 0-2 mg/L. Therefore the mean concentration is
expected to be between 0-2 mg/L for this investigation.
(2) Identify the decision errors and choose the null hypothesis —
(a) Define both types of decision errors and establish the true state of nature for each
decision error. The planning team has determined that the two decision errors arc
(i) deciding that the waste is hazardous when it truly is not, and (ii) deciding that
the waste is not hazardous when it truly is.
The true state of nature for decision error (i) is that the waste is not hazardous.
The true state of nature for decision error (ii) is that the waste is hazardous.
(b) Specify and evaluate the potential consequences of each decision error.
The consequences of deciding that the waste is hazardous when it truly is not
will be that the incinerator company will have to pay more for the disposal of
the fly ash at a RCRA facility than at a sanitary landfill.
EPA QA/G-4
51
September 199*
-------
The consequences of deciding that the waste is not hazardous when it truly is
will be thai the incinerator company will dispose of the waste in a sanitary
landfill which could possibly endanger human health and Ac environment. In
tjjis situation, they may also be liable for future damages and environmental
cleanup costs. Additionally, the reputation of the incinerator company may be
compromised, jeopardizing its future profitability.
(c) Establish which decision error has more severe consequences near the action
level The planning team bias concluded that decision error (u) has the more
severe consequences near the action level since the risk of jeopardizing human
health outweighs the consequences of having to pay more for disposal.
(d) Define the null hypothesis (baseline condition) and the alternative hypothesis and
assign the terms "false positive" and "false negative" to the appropriate decision
error.
The baseline condition or null hypothesis (HJ is "the waste is hazardous."
The alternative hypothesis (HJ is "the waste is not hazardous."
The false positive decision error occurs when the null hypothesis is rejected when
it is true. For this example, the false positive decision error occurs when the
decision maker decides the waste is not hazardous when it truly is hazardous. The
false negative decision error occurs when the null hypothesis is not rejected when
ii is false. For this example, the false negative decision error occurs when the
decision maker decides that the waste is hazardous when it truly is not hazardous.
(3) Specify a range of possible values of the parameter of interest where the consequences
of decision errors are relatively minor (gray region) — The gray region is the area
adjacent to the action level where the planning team feels that the consequences of a
false negative decision error are minimal. To decide'how to set the width of the gray
region, the planning team must decide where the consequences of a false negative
decision error arc minimal. Below the action level, even if the concentration of
cadmium were very close to the action level, the monetary costs of disposing of the
waste at a RCRA facility are the same as if the waste had a much lower concentration
of farfmintw. Clearly anv false negative decision error (to the left of the action level)
wll the incinerator company and their customers to bear the cost of unnecessary
expense (i.e., sending nonhazardous waste to a RCRA facility). The planning team,
however, also realizes that they must define a reasonable gray region that balances the
cost of sampling with risk to human health and the environment and the ability of
measurement instruments to detect differences. Therefore the planning team has
specified a width of 0.25 mg/L for this gray region based on their preferences to detect
decision errors at a concentration of 0.75 mg/L (see Figure B-l).
EPA QA/C-*
52
September IW
-------
(4) Assign probability values to points above and below the action levet thai reflect the
tolerable probability for the occurrence of decision errors — For this example, RCRA
regulations alow a 5% decision error rate ai the action level. The planning team has
set the decision error rale to 5% from 1 mg/L to 1.5 mg/L and 1% from 1.5 mg/L to 2
mg/L as the consequences of health effects from the waste disposed of in the
municipal landfill increase. On the other side of the action level, the planning team
has set the tolerable probability of making a false negative error at 20% when tic true
parameter Is from 0.25 to 0.75 mg/L and 10% when it is below 0.25 mg/L, based on
both experience and an economic analysis that shows that these decision error rates are
reasonable to balance the cost of sampling versus the consequence of sending clean
ash to the RCRA facility (see Figure B-l).
Optimize the Design — select the most resource-effective data collection and analysis design
for generating data that are expected to satisfy the DQOs. Optimizing the design is the one
step of the DQO Process thar will most likely be completed by a statistician or someone who
has data collection design expertise. Using the case study as an example, the following
section has been included to provide the reader with a background on the overall process thai
the statistician might follow to optimize the final data ccllectioa design.
m
© m
*3 <5
£ --g
o»
-------
Overview
Developing a data collection design requires an understanding of the sampled medium
. . * te wss generated in previous DQO steps. The statistician s job is 10
revi=tCih= b^Srcund information. determine the appropriate statistical application to
deauately solve the problem, and develop one or more appropriate data collecuon
Once rs is complete, the statistician will compare the cost and performance of the different
data collection designs. This process can be broken down into five distinct steps.
(1) Review the DQO outputs and easting environmental data.
(2) Develop general data collection design alternatives.
(3) For each data collection design alternative, select the optimal sample size that
satisfies the DQOs.
(4) Select the most resource-effective data, collection design that satisfies all of the
DQOs,
(5) Document the operational details 'and theoretical assumptions of the selected
design in the sampling and analysis plan.
Activities
(1) Review the DQO outputs and existing environmental data — Because the statistician
has participated in the DQO Process for this problem, there is no need to review the
DQO outputs further. The only existing data relevant to this problem arc the pilot
studv data. Based on the plot study, the incineration company has determined that
cach load of ash is fairly homogeneous, and has estimated the standard deviation in
the concentration of cadmium within loads of ash to be 0.6 mg/L.
(2) ' Develop general data collection design alternatives — Generally, the design ^
alternatives are based on a combination of design objectives developed in previous
DOO Process steps and knowledge of statistical parameters about the medium or
contaminant. Below are four examples of possible designs that could apply to the ca*e
study:
- (a) Simple Random SainoBis, - The simplest type of probity sample is A"i™P,c
random sample. With this type of sampling, eveiy possible point in the sampling
medium has an equal chance of being selected. Simple random samples arc used
primarily when the variability of the medium is relatively small and the cost of
analysis is relatively inexpensive. Simple random sampb locaticLS zrt generally
developed through the use of a random number table or through computer
generation of pseudo-random numbers.
EPA QA/G-4
54
September IW
-------
Ill the case of the cadmium-contaminated ash, a fixed number of random grab
samples would be selected and analyzed. Standard lab splits and QC samples
would be according to standard procedures for the RCRA program. Each
sample would be chosen randomly in three dimensions. A Student's t-test is
suggested as a possible method for testing the statistical hypothesis.
(b) Pnmnosite Simple Random Sagjgtiag (composite sampling) — This type of
sampling consists of taking multiple samples, physically combining (compositing)
them, and drawing one or more subsamples for analysis. Composite samples are
taken primarily when an average concentration is sought and there is no need to
delect peak concentrations. By compositing the samples, researchers are able to
sample a larger number of locations than if compositing was not used, while
reducing the cost of analysis by combining several samples.
In the case of the cadmium-contaminated ash, a fixed number of random grab
samples would be taken and composited. The number of grab samples contained
In a composite sample (g) is also fixed. To determine sampling locations within
the composite, a container would be divided into "g" equal-volume strata and
samples would be chosen. randomly within each strata. The use of strata ensure
full coverage of each container. Standard lab splits and QC samples would be
taken according to standard procedures for the RCRA program. A Student s t-test
is suggested as the possible method for testing the statistical hypothesis. ~
(c) Sequential Sampling — Sequential sampling involves, making several rounds of
sampling and analysis. A statistical test is performed after each analysis to arrive
at one of three possible decisions: reject the null hypothesis, accept the null
hypothesis,1 or collect more samples. This strategy is applicable when sampling
and/or analysis costs are high, when information concerning sampling and/or
measurement variability is lacking, when the waste and site characteristics of
interest are stable over the timeframe of the sampling effort, and when the
objective of the sampling is to test a single hypothesis. By taking samples in
sequence, the researcher can hold down the cost of sampling and analysis.
In |jje Casc of the cadmium-contaminated ash. a sequential probability sample
could be performed. The samples in each sampling round would be chosen
randomly in three dimensions. If the decision to stop sampling has not been made
before the number of samples required for the simple random sample are taken,
sampling would stop at this point and the simple random sample test would be
performed. Standard laboratory splits and QC samples would be taken according
to standard procedures for the RCRA program. An approximate ratio test is
'Decide not to reject the null based on tolerable decision error limits.
EPA QA/G-t • 55
September 1994
-------
I
suggested after each round of sampling is complete to decide whether or not to i
conclude thai the waste is hazardous or to continue sampling, j
M) Random Sampling — Stratified sampling involves dividing the study
^^^tolw^ormorenoMverlapping subsets (strata), which cover the enure
' volume to be sampled. These strata should be defined so that physical samples
within a stratum are more similar to each other ton to samples from other strata.
Sampling depth, concentration level, previous cleanup attempts, and confounding
conianiinaiitscaii be used as the basis for crating strata. Once the str» have
been defined, each stratum Is then sampled separately using one of the above
deslens Stratification is often used to ensure that important areas of a sue are •
represented in the sample. In addition, a stratified random sample may provide
more precise estimates of contaminant levels to those obtained from a simple
random sample. Even with imperfect intonation. a stratified sample cm be more
resource-effective.
Since the incineration company has already determined that each load of ash is 1
fairly homogeneous, stratification does not have any advantages over a simple I
random sample. In addition, since the company has decided to test each waste
toad individually before it leaves the facility, stratifying each waste load would be
difficult md unnecessary. Therefore, this data collection design will not be
considered further. . ( ;
(3) - For each data collection design alternative, select the optimal sample size that '
satisfies the DQOs — The formula for determining the sample size (number of
samples to be collected) is chosen based on the hypothesis test and data collection ' |
design- Standard formulas can be found in several references, including:
. Cochran, W. 1977. Sampling Techniques. New York: John Wiley. |
• Desu, M.M., and D. Raghavarao. 1990. Sample Size Methodology. San Diego,
CA: Academic Press.
• Gilbert, Richard O. 1987. Statistical Methods for Environmental Pollution
Monitoring. New York: Van Nostrand Reinhold.
. u.S. Environmental Protection Agency. 1989. Methods for Evaluating the
' Attainment of Cleanup Standards: Volume 1: Soils and Solid Media.
EPA 230/02-89-042, Office of Policy, Planning and Evaluation.
« U.S. Environmental Protection Agency. 1992. Methods for Evaluating the
Attainment of Cleanup Standards: Volume 2: Ground Water.
EPA 230-R-92-014, Office of Policy, Planning and Evaluation.
EFAQA/G-*
Septtfphcr 199*
-------
• U.S. Environmental Protection Agency. 1994. Statistical Methods for
Evaluating the Attainment of Clean-up Standards: Volume 3: Reference-
Based Standards for Soils and Solid Media. EPA 230-R-94-004. Office of
Policy, Planning and Evaiutaion.
These formulas <"af> also be found in many basic statistics textbooks. Different
formulas are necessary for each data collection design, for each parameter, and for
each statistical test These formulas are generally a function of a; p; the detection
difference, A-(delta); and the standard deviation, o. The detection difference. A, is
defined to be the difference between the action level (AL) and the other bound of the
gray region (U); Le., A = AL - U. In this case the standard deviation was derived
from pilot data under approximately the same conditions as expected for the real
facility.
For example, a formula for computing the sample size necessary to meet the DQO
constraints for comparing a mean against a regulatory threshold, when a simple
random sample is selected, is:
A1
where:
# = estimated variance in measurements (from pilot study)
n = number of samples required,
Zp = the p* percentile of the standard normal distribution (torn standard
statistical tables), and
A = U-AL
Simple Random Sample — Using the formula above, it was determined that 37
samples arc necessary to achieve the specified limits on decision errors. This
sampling plan satisfies all the DQOs including budget, schedule, and practical
constraints.
Composite Samnling — To determine sample sizes for a composite sample, it is
necessary to compute the number of composites samples, n; the number of samples, g,
within each composite; and the number of subsamples, m, to be measured for each
composite. Usually m=l; however, since this design is to be used repeatedly, it is
suggested that two subsamples from each composite sample be measured to estimate
composite variability, which can then be used to re-optimize the number of samples m
and g. •
For a composite sample, with random sample locations, it has been determined thai
eight composite samples of eight samples each are sufficient to meet the limits on
decision errors that have been specified. This design is more than sufficient to
H»A QA/G-4
57
September 19**
-------
achieve the specified limiis on decision errors and satisfies all the DQOs including
budget, schedule, and practical constraints.
g«»finff.ntial Sampling — For the purposes of comparing costs, the average number of
samples in a ggwptM sampling design can be estimated, but these estimates are only
averages. The average sample size for concluding that the waste is hazardous is 16
and the average sample size for concluding the waste is not hazardous is 22. The
average giz»» are different because the burden of proof Is placed on disproving the null
hypothesis, thus, more samples on average are required to prove that the alternative
hypothesis (the waste is not hazardous) is true. However, these sample sizes are only
averages. In some cases, fewer samples are necessary; in others, more may be
necessary. This sampling plan satisfies all the DQOs including budget, schedule, and
practical constraints.
Select the most resource-effective data collection design that satisfies the DQOs —
Compare the overall efficiency of each model and choose the one that will solve the
problem most effectively.
Cost Estimates for Each Design
First, the costs for the three designs alternatives will be evaluated:
Simple Random Sampling — A simple random sampling scheme can be implemented
for each load of fly ash by first generating three-dimensional random sampling points.
TMs can most easily be done by using a computer. Samples can then be taken using a
special grab sampler which will be forced into the ash, opened to take the sample,
then closed and removed. The difficulty with this type of sampling scheme is
measuring sampling locations in three dimensions, and it may be difficult to gain
access to the correct sampling locations.
This design meets all of the required limits on decision errors. The cost of this design
is calculated based on the assumed cost of selecting a sample (S10), and the cost of
analyzing a sample ($150). Since 37 samples need to be taken and analyzed, the cost
of this design is:
Costjuj = 37 x $10 + 37 x $150
' = $370 + $5550 = $5920
Composite Samnling — Composite sampling will be performed similarly to simple
random sampling except that after eight random samples are collected (one from each
stratum), they will be combined and homogenized. Two sample aliquots for analysis
will then be drawn from the homogenized mixture. This process will be repeated
eight times.
EPA QA/G-4
58
September IW4
-------
This design meets all of the required limits on decision errors. The cost of this design
Is based on the cost of selecting (SIO) and analyzing (SI50) a sample. Eight samples
will be used to make each composite sample for a sampling cost of $80; two
subsaroples will be analyzed from this composite sample for a cost of $300.
Therefore, each composite sample will cost S380, The total cost of this design is:
Costo = 8 x S3 80 « 53040.
SpfuientiaJ- Sampling — Sequential sampling will be performed similarly to random
sampling. The primary difference is that the ultimate number of samples will be
determined by the results of one or more sampling rounds.
This design has the potential to reduce the number of samples required in the simple
random sampling design and still meet the decision error limits. The average costs of
the two decisions are used below:
The ash is hazardous: 16 x ($160) = 52,560
The ash is non-hazardous: , 22 x ($160) = $3,520
To determine the expected cost, estimate, the number of loads of ash that should be
scot to a RCRA facility versus the number of loads that can be scot to a municipal
facility. Suppose 25% of the loads arc hazardous and should be sent to a RCRA
facility. Then the expected cost (EQ,) of this design should be.
EQj - 0.25 x (cost of sampling when ash is hazardous) + (0,75 x cost of
sampling when ash is non-hazardous)
= 0.25 x ($2,560) + 0.75 x (S3,520) = S 3.280
Selection of a Design
Because the simple random samptmg design requires that many samples be taken and
analyzed, il is inefficient for the goals of this study. Sampling will cost almost as
much to determine whether the waste is hazardous or nonhazardous as it would cost to
send all the waste to a RCRA hazardous waste landfill. Therefore, this decision is not
resource-effective.
The sequential data collection design is more resource-effective than the simple
random sampling design. The potential savings over sending all waste to a RCRA
hazardous waste facility is $6,750 - 53,280 = $3,470. The site owner has expressed
disapproval for this sampling plan because of the time it may take before a decision
can be made. If the ash was not homogeneous within a container, however, tliis data
collection design may be the design of choice.
EPAQA/G-4
59
Sepwmbcr |W
-------
He composite sample design is the best option, li is the most resource-effective
design and requires the least amount of lime to implement In addition, lie use of
strata ensures fill coverage of each container, li is recommended thai each of the
eight composite samples have two subsamples analyzed. In the .future, after sufficient
data have ten collected to estimate the variability within each composite sample, it
may be possible to reduce the number of samples that will be necessary to make a
decision about the waste contents.
(5) Document the operational details and theoretical assumptions of the selected design in
the sampling and analysis plan — A composite sample design should be used to
determine whether each container of ash should be sent to a RCRA landfill or to a
municipal landfill. Eight composite samples, consisting of eight grab samples, should
be taiwn from each container and two subsamples from each composite should be
analyzed at the laboratory. To form the composite samples, the containers will be
divided into eight strata of equal size and one grab sample will be taken randomly
within each stratum and composited. Sample locations will be generated randomly
using computer-generated random numbers. The model assumes that the variability
within a composite sample is negligible. Data from the subsamples can be used to test
fink assumption and make corrections to the model.
¦v
•RCTotti the BOO Process - Evaluation of the Design using the MA Process
For this study, the dala were collected using the composite sampling design. Once the
samples were collected and analyzed, the data were evaluated statistically and scientifically
using the DQA Process to inspect for anomalies, coafiim thai the model assumptions were
correct, select a statistical test, and verify that the test assumptions such as distribution and
independence can be met. For this study, a t-test satisfied the DQOs, and inspection of the
data indicated that there was no reason to believe that the data were not normally distributed
or that there was correlation between data points. It was also verified that the within-
composite variability was negligible.
After three weeks of sampling, approximately 30% of the waste loads leaving the
incinerator were found to have hazardous concentrations of cadmium in the fly ash. The data
collection design was determined to be cost-effective because the combined cost of sampling
and disposal was less than sending all of the waste to a RGRA landfili
EPAQWG-*
September 199*
-------
APPENDIX C
DERIVATION OF SAMPLE SIZE FORMULA FOR TESTING MEAN
OF NORMAL DISTRIBUTION VERSUS AN ACTION LEVEL
This appendix presents a mathematical derivation of the sample size formula used in
ihc DQO example- of Appendix B.
Let X X .X, denote a random sample from a normal distribution with unknown
mean u and known standard deviation c. The decision maker wishes to test the null
hypothesis H»: u = AL versus the alternative HA: M > AL, where AL, the action level, is som
prescribed constant; the false positive (Type I) error rate is a (i.e.. probabtl^ o H.
when u = AL is a); and for some fixed constant U > AL (where U is the other bound of the
gray region), the false negative (Type ID error rale is 0 (i.e„ probability of rejecting H, wbei
p as U is 1-fl). Let X denote the sample mean of the Xs. ll will have a normal distribution
with mean fi and variance oVn. Hence the random variable Z defined by
will have a standard normal distribution (mean 0. variance 1), Let z, denote tic p* percentile
of tie standard normal distribution (available in most statistics books). Recall that the
symmetry of the standard normal distribution implies thai Zp = -Z|^-
Case 1: Standard Deviation Known
The test of Ho versus HA is performed by calculating the test statistic
T _ (x-iOi/iT
(i)
' 0
r =
(2)
0
If T > z,^, the null hypothesis is rejected.
Note thai
T . [(X-nWii-ADlyfc' = z*e(p)
(3)
IPAQA/G-*
61
September I9W
-------
where
m. <4)
Thus T has a normal distribution with mean e(|i) and variance 1, and in particular, e(AL) - 0.
Hence the Type 1 error rale Is
Prirejccting HolHJ - Pt{T>z^AL} = Pr[Z^AL)>zxJ - frfZ>Z,J - «. <5>
Achieving the desired power 1-p when p. = U requires thai,
Prfreject H^'U] « I ~P*
Therefore,
PrlTSz, - fr{Z+e(W * x,.J - M2 * eC™ = P (6)
This implies
z.^-eCO) » V .
or
(U-ALyfn
0
"*.-r
Let A - U-AL, then rearrange terms to obtain
rnmmm
J* " AV" •
or .
. „ (W*,-.)V (7)
' II *"* iniLii.Lm mi
A1
Case 2: Standard Deviation Unknown
If He standard deviation o is unknown, then a lest statistic like (2) is used except thai
tf is replaced by S, an estimate of the standard deviation calculated from the observed Xs.
Such a statistic has a noncentral t distribution rather than a normal distribution, and the n
computed by the above formula will be too small, although for large n (sa| n>40). the
approximation is good. The particular noncentral t distribution involved m the calculation
depends on the sample size* a. Thus, determining the exact minimum n thai, will satisfy the
C*\ Sepnemfeer IW*
epa qa/g-4
-------
Type 1 and Type H error rale conditions requires an iterative approach in which the ^ •
noncentral t probabilities are calculated for various n values until the desired properties arc
achieved With the aid of a computer routine for calculating such probabilities, this is not
difficult; however, a simple and direct approach for approximating n is available This
approach, whose derivation is described in the paragraphs below, leads to the following
approximate but very accurate formula for a:
„ . 1 „ 1 zL, (8)
A 2
In practice, since c is unknown, a prior estimate of it must be used in (8).
The approach is based on the assumption that, for a given constant k. the statistic
X-kS is approximately normal with mean fi-te and variaace (aVnX 1+^/2) (Guenther, 1977
and 1981).
The classical West rejects Ho when T = [(X - AL)/(SAfi)] > D, where the critical
value D is chosen to achieve the desired Type I error rate a. The inequality can be
rearranged as X-kS>AL, where k = D/Vn. Subtracting the mean (assuming Ho) and dividing
by the standard deviation of X-kS on both sides of the inequality leads to
X-kS-iAL-ko) > AL-jAL-ko) = ktfm - (9)
[o/Jn)\ll+k2/2 (olJn\J\+k2a /l~«*V2
By the distributional assumption on X-kS, the left side of (9) is approximately standard
normal when p = AJL, and the condition that the Type I error rate is-a becomes
Pr
Z>kfn^\ ~ a»
i.e., z.
bjntl(ll)
One can show that (11) is equivalent to
l/[l-fcV21 = 1 -zfj2n. (l2)
The condition that the Type H error rate is P (or that power is 1-p) when n = U means that
the event of incorrectly accepting Hq given X-kS ^ AL should have probability p.
Subtracting the mean (U - to) and dividing by the standard deviation of X-kS on both sides
of this inequality yields
SFAQWG-4
63
September l»4
-------
X-kS-jU-ko) AL- = <17>
(U-AL) 1
' Squaring both sides of (17) and solving for n yields equation (3).
Guenther, William C. 1977. Sampling Inspection in Statistical Quality Control Griffin's
Statistical Monographs and Courses. No. 37, London: Charles Griffin.
Guenther, William C. 1981. "Sample Size Formulas for Normal Theory T Tcsl The
American Statistician. Vol. 35, No. 4.
EPAQA/G-*
64
September 1W4
-------
APPENDIX D
GLOSSARY OF TERMS
action level: the numerical value that causes the decision maker to choose one of the
alternative actions (e.g., compliance or noncompliance). It may be a regulatory
threshold standard, such as a Maximum Contaminant Level for drinking water, a risk-
based concentration level; a technological imitation; or a reference-based standard.
[Note: the action level is specified during the planning phase of a data collection
activity; it is "riot calculated from the sampling data.]
alternative hypothesis: See hypothesis.
lias: the systematic or persistent distortion of a measurement process which causes errors in
one direction (le.» the expected sample measurement is different than the sample's *
tnie value).
boundaries: the spatial and temporal conditions and practical constraints under which
environmental data are collected. Boundaries specify the area or volume (spatial
boundary) and the time period (temporal boundary) to which the-decision will apply.
Samples are then collected within these boundaries. *
data collection design: A data collection design specifies the configuration of the
environmental monitoring effort to satisfy the DQOs. It includes the types of samples
or monitoring information to be collected; where, when, and under what conditions
they should be collected; what variables arc to be measured; and the Quality
Assurance and Quality Control (QA/QQ components that ensure acceptable sampling
design error and measurement error to meet the decision error rates specified in the
DQOs. The data collection design is the principal part of the QAPP.
Data Quality Assessment (DQA) Process: a statistical and scientific evaluation of the data
set to assess the validity and performance of the data collection design and statistical
test, and to establish whether a data set is adequate for its intended use.
Data Quality Objectives (DQOs): Qualitative and quantitative statements derived from the
DQO Process that clarify study objectives, define the appropriate type of data, and
specify the tolerable levels of potential decision errors that will be used as the basis
for establishing the quality and quantity of data needed to support decisions.
Data Quality Objectives Process: a Quality Management tool based on the Scientific
Method, developed by the U.S. Environmental Protection Agency to facilitate the
planning of environmental data collection activities. The DQO Process enables
planners to focus their planning efforts by specifying the intended use of the data (the
decision), the decision criteria (action level), and the decision maker's tolerable
decision error rates. The products of the DQO Process are the DQOs..
EPA QA/0-4
65
September 1994
-------
decision error: an error made when drawing an inference from data in. the context of
hotels testing, such that variability or bias in Ac data mislead the decision maker
totew a conclusion that is inconsistent with the true or actual state of the population
under study. See also false negative decision error,, fate positive decision error.
defensible* the ability to withstand any reasonable challenge related to the veracity, integrity,
or quality of the logical, technical, or scientific approach taken in a decision making
process.
false negative decision errors a false negative decision error occurs when the decision
maker does not reject the null hypothesis when the null hypothesis actually is false.
In statistical terminology, a false negative decision error is also called a Type H error.
The measure of the size of the error is expressed as a probability, usually referred to
as "tela CP)"; this probability is also called the complement of power.
false positive decision error: a false positive decision error occurs when a decision maker
rejects the null hypothesis when the null hypothesis actually is true. In statistical
terminology, a false positive decision error is also called a Type I error. The measure
of the size of the error is expressed as a probability, usually referred to as "alpha (a)"
lhe "level of significance," or "size of the critical region.
gray region: a range of values of the population parameter of interest (such as mean
contaminant concentration) where the consequences of making a decision error are
relatively minor. The gray region is bounded on one side by the action level.
hypothesis: a tentative assumption made to draw out and test its logical or empirical
consequences. In hypothesis testing, the hypothesis is labeled "null" or "alternative .
depending on the decision maker's concerns for making a decision error.
limits on decision errors: the tolerable decision error probabilities established by the
decision maker. Potential economic, health, ecological, political, and social
consequences of decision errors should be considered when setting the limits.
mean: (!) a measure of central tendency of the population (population mean), or (ii) the
arithmetic average of a set of values (sample mean).
measurement error: the difference between the true or actual stale and thai which is
reported from measurements,
median: the middle value for an ordered set of n values; represented by the central value
¦when n is odd or by the average of the two most central values when n is even. The
median is the 50th percentile.
medium: a substance (e.g., air, water, soil) which serves as a carrier of the analytes of
interest.
EPA QA/G-4
66
Septentoer 1994
-------
natural variability: the variability that is inherent or natural to the media, objects, or people
being studied.
null hypothesis: See hypothesis.
parameter a numerical descriptive measure of a population.
percentile: the specific value of a distribution thai divides the distribution such that p
percent of the distribution is equal to or below that value. Example for p=95: "The
95th percentile is X" means that 95% of the values in the population (or statistical
sample) are less than or equal to X.
planning team: the group of people that will carry out the DQO Process. Members include
the decision maker (senior manager), representatives of other data users, senior
program and technical staff, someone with statistical expertise, and a QA/QC advisor
(such as a QA Manager).
population: the total collection of objects, media, or people to be studied and from which a
sample is to be drawn.
power function: the probability of rejecting the null hypothesis (HJ over the range of
possible population parameter values. The power function is used to assess the
goodness of a hypothesis test or to compare two competing tests.
quality assurance (QA): an integrated system of management activities involving planning,
quality control, quality assessment, reporting, and quality improvement to ensure that a
product or service (e.g., environmental data) meets defined standards of quality with a
stated level of confidence.
Quality Assurance Project Plan (QAPP): a formal technical document containing the
detailed QA, QC and other technical procedures for assuring the quality of
environmental data prepared for each EPA environmental data collection activity and
approved prior to collecting the data.
quality control (QC): the overall system of technical activities that measures the attributes
and performance of a process, item, or service against defined standards to verify that
they meet the stated requirements established by the customer.
Quality Management Plan (QMP): a formal document describing the management policies,
objectives, principles, organizational authority, responsibilities, accountability, and
implementation protocols of an agency, organization, or laboratory for ensuring qualify
in its products and utility to its users. In EPA, QMPs are submitted to the Quality
Assurance Management Staff (QAMS) for approval.
range: the numerical difference between the minimum and maximum of a set of values.
ffAOA/W
67
September 19*4
-------
'sample: a single item or specimen from a larger whole or group., such as any single sample
of any medium (air, water, soil. etc.).
3sample: a set of individual samples (specimens or readings), drawn from a population,
whose properties arc studied to gain information about the whole.
sampling: the process of obtaMng representative samples and/or measurements of a subset
of a population.
sampling design error: the error due to observing only a limited number of the total
possible values that make up the population being studied. It should be distinguished
from errors due to imperfect selection; bias in response; and errors of observation,
measurement, or recording, etc.
scientific method: the principles and processes regarded as necessary for scientific
investigation, including rules for concept or hypothesis formulation, conduct of
experiments, and validation of hypotheses by analysis of observations.
standard deviation: the square root of the variance.
statistic: a function of the sample measurements; e.g., the sample mean or standard _
deviation.
statistical test: any statistical method that is used to determine which of several hypotheses
are true.
total study error: the combination of sampling design error and measurement error.
true: being in accord with the actual state of affairs.
Type I error: A Type I error occurs when a decision maker rejects the null hypothesis when
it is actually true. See false positive decision error.
' Type II error: A Type D error occurs when the decision maker fails to reject the null
hypothesis when it is actually false. See false negative decision error.
variable: The attribute of the environment that is indetcrminant.
variance: a measure of (i) the variability or dispersion in a population (population variance),
or (ii) the sum of the squared deviations of the measurements about their mean divided
by the degrees of freedom (sample variance).
%
CPA QA/G-4
68
SejwewbcJ' IW*
-------
Appendix A from EPA (1996b)
Geostatistical Sampling and Evaluation Guidance for Soils
and Solid Media, Review Draft
- ft !'* H..." I* t A\k* H( VKU riS4l MASH.!* r
-------
APPENDIX A
The Data Quality Objectives Process
One planning tool available to help in the design and implementation of a sampling and
analysis project is the data quality objective (DQO) process developed by the United States
Environmental Protection Agency's (USEPA's) Quality Assurance Management Staff (QAMS).
DQOs are qualitative and quantitative statements that clarify the study objective, define the type,
quantity, and quality of required data, and specify the tolerable limits on decision errors. DQOs
are used to define the quality control (QQ requirements for da& collection, saraping and analysis,
and data review and evaluation. Hese QC requirements are included in the quality assurance
(QA) objectives for environmental measurements and the DQOs also are incorporated into a
quality assurance project plan (QAPP). DQO development is an ongoing process involving
discussions between management and technical staff. This process is a practical mans for
specifying and ensuring that the requested information is known to be of the "required" quality.
Failure to establish DQOs prior to implementing field and laboratory activities can cause
difficulties for site investigators in the form of inefficiencies, increased costs, or generation of
unusable data. For example, if low-cost analytical techniques will suffice, but higher cost
techniques are selected, time and money are wasted.
Tm DQO Process
A seven-step DQO Process has been developed for uniform and consistent data generation
activities:
Step 1: State the Problem
Step 2: Identify the Decision
Step 3: Identify Inputs to the Decision
Step 4: Define the Study Boundaries
Step 5: Develop a Decision Rule
Oraft
A-l
February 1996
-------
Geosta/istteal Soil Sampfwf Cuidmee—ApBmA_A
Step 6: Specify Tolerable Limits on Decision Errors
Step 7: Optimize the Design for Obtaining Data.
The DQO process is based on the guidance document issued by QAMS as "final in
September 1994. Guidance for the Data Quality Objectives Process (EPA QA/G-4) provides
general guidance to organizations for developing data quality criteria and performance
specifications for decision making. Chapter One of the SW-846 Methods Manual also provides
additional guidance on the QA program in the Office of Solid Waste (OSW).
Step 1: Slate the Problem
He firt step in any dmaon-eiiMtg process is to define the problem thai has resulted in the
inception of the study. A planning team is assembled and is tasked with developing the project-
specific DQOs. The planning team comprises personnel representing all phases of the project and
may include technical project managers, QA/QC managers, data users, and decision makers. The
primary decision maker, or leader, must be identified. Where applicable, field and lab
technicians, chemists, statisticians, and modelers also should be recruited for the planning team.
The responsibilities of each team member should be clearly defined during this initial planning
stage.
A concise description of the problem must be developed during this early stage of DQO
development. Existing information should be summarized, and the need for additional
information should be determined. Performance of literature searches or an evaluation of
historical data or ongoing studies related to the current site can be studied.
Available financial and manpower resources must be identified and project milestones and
deadlines also should be determined, if sufficient information is present
Step 2: Identify the Decision
This step is used to define the decision statement that the study must resolve. The decision
statement is a consolidation of the principal study question and alternative actions. The principal
Dtqfi
A-2 ' , February 1996
-------
GeostMsikai Soil Sampling Guidance—Appendix A
study question identifies the key unknown conditions or unresolved issues that will be used to
reveal the solution to the problem being investigated. Alternative actions, which axe items thai
may be taken to solve the problem based on the outcome or on the decisions arrived at from the
study, also are identified.
Step 3: Identify Inputs to the Decision
Specific information required to resolve the deciaon statement must be identified during this
step in the DQO development process. He selected ia& acquiatiofi approach will lead to the next
set of questions that address the specific types of information needed to support the decision
statement Sources of the necessary information are then developed and can include regulatory
guidance, scientific literature, historical date or past projects that were similar in scope to the
current effort.
A bright-line, defined as the threshold value that provides the criterion for choosing between
alternative actions, needs to be established.
Existing analytical methods are evaluated to determine if the method will perform as
published, or if method modification or method development needs to be included in the study.
Each analyte of interest should have a method detection limit or level of quantitation assigned, as
this performance information is used later in the DQO Process (Steps 5 & 7).
Step 4: Define the Study Boundaries
Two types of boundaries must be defined and quantified: spatial and temporal. Spatial
boundaries define the physical area to be studied and locations to collect samples. Temporal
boundaries describe the timeframe that the study data will represent and when the samples should
be collected. To arrive at these boundaries, the characteristics that define the population must be
identified. For instance, the compounds of interest and the matrix that should be evaluated might
need to be selected to determine if the compounds are present and at what typical concentrations.
Dnjt A-3 • , February 1996
-------
GeostaiMeal Soil SmmpMng Guidance—Appendix A
The spatial boundaries, or the geographic area to be studied, must be specified using some
physical feature or border, such as units of measure. Where possible, the population should be
further segregated into more homogenous subsets, or strata, as a means of reducing variability.
Step 5: Develop a Decision Rule
A statement must be developed that combines the parameters of interest and the action levels
with the DQO outputs already developed. The combination of these three elements forms the
decision rule and summarizes what attributes the deciaon maker wants to study and how the
information will assist in solving the central problem. The four elements that form the decision
rule include: (1) the parameter of interest that describes a characteristic of the statistical
population, (2) the scale of decision making defined in Step 4 when boundaries were defined,
(3) the bright-line (action level or a measurement threshold value), used as a criterion to choose
alternative actions through the use of "if/then" statements, and (4) identifying the alternative
actions, as developed in Step 2.
Step 6: Specify limits on Decision Errors
Pecislon makers are interested in taowkg the true state of some feature of the environment
(e.g., the concentration of the constituent of concern in soil). However, daa generated from a
sampling and analysis program can only be used to estimate this state, and there is a chance that
the data are in error and the correct decision will not be made. This step in the DQO development
process allows the decision makers to specify acceptable or tolerable limits on decision errors.
Hoe are at least two primary reasons why the decision mater might not determine the true
value of a population parameter. First, sampling design error occurs when the sampling design
is unable to capture the complete extent of variability that exists in the true state of the
environment. Second, measurement error, which is a combination of random and systematic
errors, results from various steps in the measurement process including sample collection, sample
handling, sample preparation, sample analysis, data reduction, and data handling. The
combination of sampling design error and measurement error can be viewed as the total study
error and may lead to decision errors.
A-4
February 1994
-------
Geostariftif"ul Soil Sampling Guidance—Appendix A
To estimate the probability of decision errors, the anticipated range of results from the
parameters of interest usually must be determined, perhaps through the use of historical data.
Values between the observed upper and lower bounds or perhaps from a distribution modeled after
the historical data can be used to estimate how likely are the various decision errors that might
occur depending on the hypothesis framework that has been constructed. The statistical
hypotheses tad with any decision criterion consist of a null hypothesis, supposed to
represent the initially assumed condition of the site, and the alternative hypothesis, representing
the condition of the site when the null hypothesis is not true. Often the null hypothesis indicates
the location of the center (in terms of concentration levels) of the hypothesized sampling
distribution, but it can describe other characteristics of the site population (e.g., an upper
percentile). Both the null and alternative hypotheses make statements regarding a characteristic
of the population rather than a characteristic of a sample. The probability of a decision error is
determined by estimating the chance that one of the two hypotheses will be accepted when in fact
the opposite hypothesis is true.
To identify the decision errors and to construct the hypothesis framework, performance of
four steps must be accomplished. (1) Both types of decision errors must be defined, determining
which occurs above the action level and which occurs below the action level. (2) Potential
of each decision error must be specified and the impact of amving at the incorrect
decision considered. The severity of the error may affect economic and social costs or have
ramifications to human health and the environment. One of the two types of errors (e.g., above
or below the action level) often will have a greater impact than the other. (3) The decision maker
should evaluate which scenario results in more serious consequences. (4) The null hypothesis,
or baseline condition, should be defined and the decision as to what constitutes a false-positive or
false-negative result should be answered.- The term false-positive is assigned to the decision error
where the decision maker rejects the null hypothesis when it is tnie. Conversely, a false-negative
is the resulting decision error if the null hypothesis is not rejected when it is false.
Some decision errors may be considered minor and of minimal impact to use of the data.
This "grey region" should be specified as a range of values having little or no adverse
Drqft
A-5
February 1996
-------
Geovtiipiral Soil SonpBme Guidon ' tentMx A
consequences to the project. Use of these grey area regions Is often important as a tool for
building tolerable limits on the probability of making an incorrect decision.
Step 7: Optimize the Design for Obtaining Data
The final step addresses the design of a resource-effective data collection system to satisfy
the DQOs. Verify that the DQO outputs produced in all preceding steps are internally consistent.
The design options should have been developed based on cost benefits, versus achieving DQOs.
General data collection designs can then be developed as either a factorial design, systematic
sampling, composite sampling, or one of the following random sampling designs: simple,
stratified, or sequential.
In general, three statistical expressions need to be selected to optimize the data collection
design. (1) An appropriate method for.testing the statistical hypothesis framework must be
chosen. (2) A statistical model used to compare measured values to the modeled values must be
developed and tested for consistency with the observed date. Once established, the model also on
be used to more thoroughly describe the components of error or bias that may exist in the
measured values. (3) Finally, a cost evaluation of number of samples versus the total cost of
sampling and analysis must be developed. Using these statistical expressions, an optimal sample
size and sampling layout can be chosen to meet the DQOs for each data collection design
alternative.
Quality Assurance Review
Lastly, have the document peer reviewed, preferably by personnel experienced in statistical
datt collection designs. Ensure that all aspects of the project have been documented to minimize
the numbers of assumptions made during performance of the project.
DQO Decision Bwor FfeAsnuunr IMms (DEFT) Software
The two most intensive steps in the DQO Process are Step 6: Specify Tolerable Limits on
Decision Errors, and Step 7: Optimize the Design for Obtaining Data. During Step 7, the entire
set of DQO outputs is incorporated into a sampling design. If the DQO constraints arc not
Drq/t A-6 February 1996
-------
Geesmistkal Spa Sampling Cmdmce^^A^enOxji
feasible, it is necessary to iterate through one or more of the earlier steps of the DQO Process to
identify a sampling design flat will meet the budget and generate adequate data for the decision.
This iteration can be time consuming and costly. EPA developed the DEFT User s Guide and
software (USEPA 1994c) to streamline this iterative process. Users can change DQO constraints
such as limits on decision errors or the 'grey region" and evaluate how these changes affect the
ample ok for several basic samping designs. The output of the DEFT software can be used to
set upper and lower bounds on the sample size (i.e., the appropriate number of observations).
Through this process, the planning team can evaluate whether these constraints are appropriate or
feasible before the sampling and analysis plan is developed.
Ums of the DEFT software aie first prompted to rate" information from the DQO outputs
hayri on a series of entry screens. Specific information requested by the DEFT software includes,
• Parameter of interest
• Minimum and maximum values (range) of the parameter of interest
• Action level (i.e., the bright-line)
• Null and alternative hypothesis
• Bounds of the gray region
• Estimate of the standard deviation
• Cost per sample for sample collection (i.e., field cost per sample)
• Cost per sample for sample analysis (I.e., laboratory cost per sample)
• Probability limits on decision errors for the bounds of the gray region
• Any additional limits on decision errors.
The DEFT software automatically starts with a simple random sampling design, so the
information requested corresponds to this design.
&Ai«E APPLICATION OF THE POT* SOFTWAIE
At a site contaminated with polynuclear aromatic hydrocarbons (PAH) compounds,
contaminated soil has been excavated and placed on a pad. Investigators are interested m
Draft
A-7
February 1996
-------
Geoaatistieal Soil StmpMag GmMmce—Append A
determining whether the mean concentration of benzo(a)pyrene (BAP) exceeds the bright-line
standard of 90 mg/Kg. Hie investigators have decided to use the DEFT software during the DQO
Process to help optimize the study design.
Parameter of Interest
The parameter of interest far this study is the population mean of the concentration of BAP.
Minimum and Maximum Values (Range) of the Parameter of Interest
Based on data generated during a preliminary study of the contaminated soil, the minimum
concentration of BAP was 62 mg/Kg and the maximum was 120 mg/Kg.
Action Level
The action level, or bright-line, for BAP is 90 mg/Kg.
Null and Alternative Hypothesis
H„: mean i. bright-line vs. H„: mean < bright-line.
Bounds of the Gray Region
The gray region is bounded on one side by the bright-line (90 mg/Kg). For "H0:
mean a bright-line vs. H,: mean < bright-line," DEFT sets a default value for the other bound
of the gray region at the midpoint between the minimum concentration (62 mg/Kg) and the bright-
line. In this example, the lower bound of the gray region is 76 mg/Kg.
Estimate of the Standard Deviation
If there is no estimate of the standard deviation available, DEFT calculates a default value
given by:
(Maximum Concentration - Minimum Concentration)^
Draft A-8 February 1996
-------
Gaosmktkai Soil Sampling GuUmg^EE£2**A
In this example, the estimate of the standard deviation Is 9,7.
Cost Per Sample for Sample Collection
The cost of sample collection is approximately S67.00 per sample. This estimate is based
on the following assumptions:
• Two field sampling technicians are required.
• Samplers can collect, prepare, and ship 12 samples per 8-hour day.
• Labor rate is $50.00/hour ('loaded" rate).
Cost Per Sample for Sample Analysis
The cost per soil sample analysis for semi-volatile organic compounds, including BAP, is
approximately $400,00 per sample.
probability Limits on Decision Errors for the Bounds of the Gray Region
For this example, the probability of making a false positive error is set a a ~ .01, and the
probability of making a false negative error is set a P m .05.
Ate the above information is entered into the DEFT software, sampling design and DQO
summary information is provided. For this example, a simple random sampling design would
require n samplesatatotalcostof$5,137.00 (seeattached 'Design/DOO Summary Screen" and
"Decision Performance Goal Diagram Screen with the Performance Curve ).
A q February 1996
Drcjt h'9
-------
Design/DQO Summary Screen
For the sampling Design of: Simple Random Sampling
Total Cost: $5137.00
Laboratory Cost per Samples $400,00
Field Cost per Samples $67.00
Number of Samples: 11
Data Quality Objectives
Action Level: 90.00
Gray Region: 76.00 - 90.00
Null Hypothesis; mean 2 90.00
Standard Deviation (SD): 9.67
Decision Error Limits
cone. prob(error) type
—?— F(-)
76.00 0.0500 F(-)
90.00 0.0100 F(+)
— F(+)
F(+»
-------
GeotfMstfettf Soil Smftolme Guidance ^ppendixj^
m
>
m
™S
fl
TJ ^
O ®
® £
O w
o "§
*S
-Q LL1
CD _
s
2 S
0
0.9 -
62
F(+)
Action Level
90 ¦
Gray Region
73 JO
85.20
96.80
108.4
True Mean Concentration
0.1
- 0.2
120
Draft
A-ll
February 1996
-------
Sag SmtpMag GuUmce-AppenaxA.
PCT.CTA mt GOTOANCE ON DATA QUAUTf OUECTIVB
9
1 USEPA. 1994a. EPA Requirements for Quality Assurance Project Plans for Environmental
Data Operations (Interim Final). EPA QA/R-5. Quality Assurance Management Staff
(QAMS), Washington, DC.
Presents detailed specifications and instructions for the information that must be contained
in a QAPP for environmental data operations performed by or on behalf of USEPA and the
procedures for its review and approval.
2 USEPA, 1994b. Guidance for the Data Qualiiy Objectives Process (Final). EPA QA/G-4.
Quality Assurance Management Staff (QAMS), Washington,, DC.
Offeis general guidance on developing data quality criteria and joforntance specifications
for data operations. The document outlines the seven distinct steps of the DQO Process:
state the problem; identify the decision; identify inputs to the decision; define the decision
boundaries; develop a decision rule; specify limits on decision rule; and optimize the design
for obtaining data. Includes a detailed example and a glossary.
3 USEPA. 1994c. Data Quality Objectives Decision Error Feasibility Trials (DQO/DEFT)
Version 4.0 Software and User's Guide. EPA QA/G-4D (Final). Quality Assurance
Management Staff (QAMS), Washington, DC.
The DEFT software uses the outputs from Steps 1 through 6 of the DQO Process to allow
a decision maker or member of the DQO planning team to quickly generate cost information
about several simple sampling designs based on the DQO constraints.
4 USEPA. 1996. Guidance for Data Quality Assessment (Final). EPA QA/G-9. Quality
Assurance Management Staff (QAMS), Washington, DC.
. The purpose of this guidance is to demonstrate the use of EPA's data quality assessment
(BQA) process in evaluating environmental data sets and to provide some graphical and
statistical tools that are useful in performing DQA.
Draft
A-12
February 1996
-------
ATTACHMENT D
DATA USEABILITY WORKSHEETS
•• M i* ItSkKftr-VISUTtNAl. MASIt* WH> H« fti jw:!-iT \ 2i«p»i> wc
-------
DATA USEABILITY WORKSHEET
Site:
Medium:
Requirement
Comment
Field Sampling
Evaluate field sampler's trip report. Discuss any
sampling problems that affect data useability.
Discuss field conditions that affect data useability.
Discuss changes to approved field sampling
methodology
Are samples representative of receptor exposure for
this medium (e.g. sample depth, grab vs composite,
filtered vs unftltered. low flow, etc.)?
Discuss the effect of field QC results on data
useability.
Summarize the effect of field sampling issues on the
risk assessment, if applicable.
DUWKS 3
1
-------
DATA USEABILITY WORKSHEET
Site:
Medium:
Requirement
Comment
Analytical Techniques
Discuss whether the analytical methods are appropriate
for quantitative risk assessment.
Discuss data useability limitations of non-routine
analytical methods (e.g. immunoassay, low-
concentration, etc.) for use in quantitative risk
assessment.
Were detection limits adequate?
Summarize the effect of analytical technique issues on
the risk assessment, if applicable.
DOWKS 3
-------
DATA USEABILITY WORKSHEET
Site:
Medium:
Requirement
Comment
Data Quality Objectives
Precision - Indicate notable sources of variability in the
data (e.g. similarity between duplicates and between
splits, overall distribution of data, effect of total
number of samples on variability). Discuss how
duplicates were handled.
Accuracy - Indicate any problems associated with
accuracy and notable sources of bias (e.g. problems
with spikes, dilutions, holding times, blank
contamination).
Representativeness - Indicate any problems associated
with data representativeness (e.g.. trip blank or rinsate
blank contamination, COC problems, etc.).
Completeness - Indicate any problems associated with
data completeness (e.g.. incorrect sample analysis,
incomplete sample records, problems with field
procedures, etc.).
Comparability - Indicate any problems associated with
data comparability.
Were the DQOs specified in the QAPP satisfied?
Summarize the effect of DQO issues on the risk
assessment, if applicable.
DUWKS 3
3
-------
DATA USEABILITY WORKSHEET
Site:
Medium:
Data Validation and Interpretation
What are the data validation requirements for this
region?
Whai method or guidance was used to validate the
data?
Was the data validation method consistent with
regional guidance? Discuss any discrepancies.
Were all data qualifiers defined? Discuss those which
were not.
Which qualifiers represent usable data?
Which qualifiers represent unusable data?
How are tentatively identified compounds handled ?
duwks }
4
-------
DATA USEAB1LITY WORKSHEET
Site:
Medium:
Requirement
Comment
Summarize the effect of data validation and
interpretation issues on the risk assessment, if
applicable.
Additional notes:
Note, Hie purpose of this Worksheet is to succinctly summarize the data useabiltty analysis and conclusions. Reference
specific pages in the Risk Assessment text to tun her expand on the information presented here.
duwks i 5
-------
ATTACHMENT E
U.S. ENVIRONMENTAL PROTECTION AGENCY REGION 9
PRELIMINARY REMEDIATION GOALS
TASK*.ftfViSfcr FlNAtMASTER i
-------
ATTACHMENT F
U.S. ENVIRONMENTAL PROTECTION AGENCY REGION 8
SUPERFUND TECHNICAL GUIDANCE NUMBER RA-03
EVALUATING AMD IDENTIFYING CONTAMINANTS OF CONCERN
FOR HUMAN HEALTH
S MP\ W IssM* T ASK* REVISE; fINAL'MASTER WTO- }t«pn»\Me
-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
REGION VIII (8HWM-8M)
§#» 1ith STREET . SUITE 500
DENVER, COLORADO 80202-24S6
Roflien Viil
. •-y "• ;v -*T*f: f". J5-*-'
Hazardous Waste Management Division, Superfund Management Branch, Technical Section
SUMMARY
This regional guidance is intended to clarify the evaluation process for selecting
contaminants of concern (COCs) for the human health risk baseline risk assessment
process, as generally described in EPA's Risk Assessment Guidance for Superfund
RAGS). This guidance sets forth objective criteria (e.g., comparison to background
levels, frequency of detections, essentiality, etc.) and provides explicit
recommendations on measuring attainment for each of these criteria in order to
evaluate whether or not a site-related contaminant should be retained as a COC.
Pag® 1 of 10 Pages
-------
***** U.»- CPA
mm 9n*-09
§«»*»» im*
EVALUATING AND IDENTIFYING CONTAMINANTS Of CONCERN
FOR HUMAN HEALTH
OBJECTIVE
The ©bjectiva of this Regional
Guidance is to outline and describe a
selection process whereby preliminary lists
of potentially site-related contaminants can
be evaluated for elimination or retention as
contaminants of concern (COCs) for the
human health baseline risk assessment.
BACKGROUND
For certain sites, the list of potentially
site-related contaminants and exposure
pathways may be lengthy. Carrying a
large number of contaminants through a
quantitative risk assessment may be
complex, and may consume significant
amounts of time and resources. In these
cases, a selection process should be used
to further reduce the number of
contaminants of potential concern for each
medium to a reasonable and relevant
amount. EPA's Risk Assessment Guidance
for Superfund (RAGS): Part A (EPA,
1989a) describes general qualitative
criteria which should be considered when
evaluating contaminants for either
elimination or retention as contaminants of
concern (COCs) for the • baseline risk
assessment. The purpose of this Regional
Guidance is to present those criteria in a
selection process which can be applied on
a generic basis to US EPA Superfund sites
in Region 8. This Regional Guidance will
also present detailed examples of how
several criteria presented In the upcoming
flow chart can be quantitatively evaluated.
DISCUSSION
EPA's RAGS: Part A (EPA 1989a)
recommends that the following criteria be
evaluated when determining which
chemicals on the initial list of ail potentially
site-related contaminants should be
retained or eliminated as COCs for the
Baseline Risk Assessment:
1. Essential Nutrients
2. Exceedance of background
concentrations
3. Detection frequency
4. Mobility, persistence, and
bioaccumulation
5. Exceedance of ARARs
6. Historical Evidence
7. Concentration and Toxicity
Page 2 of 10 Pages
coo®
89 "WIQ3I3M QNUaHd.lS SB
ski cez coc Ivj m-ti can ts-so/fl
-------
t. U.S.
Figure 1 presents a selection
process which addresses each of the
criteria present in RAGS: Part A (EPA
1989a) and can be used to arrive at a final
list of COCs for the risk assessment
©valuation. This selection process Is
explained below:
1. Is the contaminant an essential
nutrient?
if the contaminant Identified is an -
essential nutrient and is present at low
concentrations (i.e., only siightty elevated
above naturally occurring levels or below
established EPA toxicity values or FDA
recommended nutritive levels), It does not
need to be considered further In the risk
assessment. Examples of EPA toxicity
values which can be used are the slope
factors or Reference Doses listed on EPA's
Integrated Risk information System (IRIS)
Database or Health Effects Assessment
Summary Tables (HEAST). The FDA's
Recommended Daily Allowance (RDA) of
essential dietary minerals and safe
supplemental levels of dietary minerals can
be used as nutritive indexes. Table 1
shows the 'essential elements/nutrients
which can be considered in the COC
selection process and their corresponding
toxicity value or safe nutritive level.
TABLE I
Element/nutrient Dose (mg/kg/day)
Calcium
Phosphorous
Magnesium
Iron
Zinc
Iodine
Copper
Manganese
Fluoride
Sodium
Chromium III
Potassium
Chloride
Selenium
Molybdenum
Cobalt
1.4E + 01 *
1.4E+01#
5.7E + 00*
2.6E-01 *
3.0E-01 I
2.1E-03*
3.7E-02 h
5.0E-03 i
6.0E-02 1
No data
1.0E+00 i
5.7E-01 •
5.1E-01#
S.0E-Q3 I
5.0E-03 I
6.0E-02 e
•FDA RDA of essential minerals or FDA
supplemental dietary mineral levels
1 - IRIS
h « HEAST
9 ¦ EPA provisional toxicity value
2, Does the contaminant exceed
background concentrations?
For the purpose of comparing silo-
related contamination to background levels
of chemicals, EPA's RAGS: Part A (EPA,
1989a) divides background types into
naturally occurring chemicals and
anthropogenic chemicals. Examples of
Page 3 of 10 Pages
too®
HB 1YIQ3K3H GLvUiid.lS 8S
8CZT £82 COC I*¥J It- fl {&» f§/BG"H
-------
*•»« ». W.«. B»A
TamMMiMi
anthropogenic chemicals Include pesticides
from agriculture, toad from auto amissions,
and PAHs from fossils fuel combustion.
This COC selection process will
automatically include comparisons of site-
related contaminants to naturally occurring
chemicals. Inclusion of site comparisons
t° background anthropogenic chemicals
(whether localized or ubiquitous) will be
considered on a site-specific basis.
The USEPA has issued guidance for
ground water detection monitoring
programs being conducted under the
Resource Conservation and Recovery Act
(RCRA). This guidance, entitled
"Statistical Analysis of Ground-Water
Monitoring Data at RCRA Facilities" (EPA,
1989b) provides a conceptual framework
far determining and applying an
appropriate statistical method for
comparison of background and
contaminated groundwater data. This
statistical guidance could also be applied
to soil background comparisons.
The RCRA guidance details two
types of statistical comparisons that can
be made between samples collected from
background and contaminated sites.
These two type of statistical comparisons
are 11} distributional tests, and 12) extreme
value tests. Distributional tests are
•OArMU-03
•mmkwtWK
statistical tests used to determine whether
the central tendencies of two groups of
data are similar. Extreme velues tests are
statistical tests used to compare individual
results (i.e., results from an affected site)
to results from a distribution (e.g., the
distribution of the background data). The
obfecttve of the statistical analysis for the
rink assessment I® to determine If she
concentrations differ significantly from
background concentrations, on the
tVfrfMft. Therefore, distributional tests,
and generally not extreme value tests,
should be chosen for risk analysis.
figure 2 is an example of a flow
¦ chart (based on the RCRA guidance) for
comparing background and site
concentrations using distributional tests,
which depend on the percent of detected
values for each parameter and distribution
of background and site concentrations.
Th# data analysis process was divided in
this way because each statistical method
can handle a certain number of detected
values before the method becomes
ineffective In determining a significant
difference. The risk assessor is not
United, however, to those statistical tests
shown in Figure 2. The choice of
appropriate test should be based on the
distribution of the data, the percent of
non-detects in background and/or site
Page 4 of 10 Pages
500®
ii 1VIQ3K3S
-------
u.s. bp*
Tmtmmrn Sua.en™,
data, the presence of multiple detection
limits, etc.
Caution; Statistical comparisons of data
sets may be inappropriate and the
interpretation of those lasts meaningless
when the number of non-detects are high
(e.g., > 80%) mul the sample sizes are
small (e.g., N < 20). It is recommended
that a statistician be consulted on the
appropriateness of the statistical test{s)
especially for unstable data sets.
At some sites, a concern may exist
for hot spots" or situations where a small
proportion of the site is contaminated
abov® background, yet application of
distributional tests show no difference
between site and background levels of
randomly sampled data. For example,
there may have been too few samples
collected at the site, so that perhaps only
one or two measurements are elevated
ab©¥© background. One method for
dealing with this situation is to compare
each site measurement to a "hot
measurement" concentration value (Gilbert
and Simpson, 1992). This "hot
measurement" value can be a risk based
number, a standard, or some function of
the background data (e.g., upper tolerance
limit}. Generally the hot measurement
value should be selected to identify small
areas that may individually present
excessive health risk beyond that of
average site-wide exposures. If one or
more site measurements equal or exceed
the hot measurement value, the
contaminant can be retained as a COC,
3. Detection Frequency
A contaminant with a detection
frequency of *5% proceeds into the
toxicity concentration screen. A chemical
with <5% detection frequency is further
evaluated with up to four additional
criteria.
4. Persistence, Mobility, and
Bioaccumulation
A chemical Is retained as a COC if It
Is either highly persistent or highly mobile.
Several physico-chemical parameters
describe these processes, including
environmental half-He, water solubility, log
and The log octanot/water
partition coefficient (log is the ratio of
the chemical concentration in octanol to
the concentration In water. A high log
typically greater than 3, indicates
higher concentrations in the octanol rather
than in the water. is an equilibrium
Page 5 of 10 Pages
MS 1YJQ3K3H (K.Ua3
-------
». UM. fPA
constant that measures the partitioning
between organic carbon and water. tcoo is
useful for describing mobility potential
because it correlates bitter with
adsorption to soil arid sediment. A
chemical's mobility is generally
proportional to its water solubility and
inversely proportional to and Km.
Chemicals with log < 2.7 and K.. <
50 ar® considered to be highly mobile,
while chemicals with log > 3 and Km
> 500 generally have low mobility
potential.
In general, chemicals with Log K9W>
3 begin to have a high bioaccumuiatlon
potential. It is immediately obvious that
these criteria would only exclude
chemicals with K^'s of 2.8 and 2.9. For
this reason, it is recommended that the
parameters of bioaccumuiatlon or mobility
Ml be used to exclude contaminants.
Persistence is measured by the number
of days required to reduce a chemical's
concentration by one-half through biotic
and abiotic degradation processes.
Chemicals art considered highly persistent
If their half-lives in water are >90 days,
and not persistant in water with half-lives
< 30 days.
•ONIMS
PARAMETER POTENTIAL FOR ACTION:
> 3 : Bioaccumuiatlon
OR
Km < 2.7 : Mobility
K» < 50 : -
^ Do not use criteria for eliminating
contaminants. -Proceed to Toxicity
Concentration Screen.
*1/2 > 90 : Persistence
Proceed to Toxicity Concentration Screen.
8. Do concentrations exceed Health-
arid Technology-based Numerical
criteria (ARAR's)?
Numerical criteria are federal and
duty-promulgated state environmental and
public health laws, requirements, or
regulations for the protection of human
healtfi from exposure to chemical
contaminants. If the maximum contam-
inant concentration or the 95th percent
upper confidence limit of the mean for
chemical concentrations exceeds hoalth-
aod technology-based criteria, proceed to
the Toxicity Concentration Screen.
Pag® 8 of 10 Pages
.00®
aa 1VI03I3H aoiaad.is w
ski eez coc \vj st-.fi ton fs-eo-n
-------
**«•»«, U.S.CM
Itmmtmt Butmmm "
6. Is there Historical Evidence of the
Compound at the Sits?
Chemicals reliably associated with
site activities baSt«j on historical
Information generally should not be
eliminated from the quantitative risk
assessment.
7. Toxicrty/Conctntratfon Screen
EPA's RAGS: Part A (EPA 1989a)
suggests consideration of a toxicity
concentration screen based on calculating
individual risk factors and eliminating
chemicals .which do not contribute, for
example, more than 1% of the total risk.
If 1 or more chemicals art present at very
Wgh concen»atfonsf this mtthotf may lead
to the elimination of chemicals which do
not contribute much to the overall risk, but
•xcetd health-based towels, none the less.
For this reason, it is recommended that the
toxicity concentration screen be based on
generic Preliminary Remediation Goals
fPRGs) as calculated by RAGS: Part B (EPA
t"1». Regi0ri Ill's Risk-Based
Concentration Tables spreadsheet Is one
such example of screening levels based on
the RAGS: Part B PRG equations. EPA's
Soil Screening levels ISSLsl are another
example, albeit more conservative. Either
1114
the maximum contaminant value or the 95
percent upper confidence limit of the
arithmetic mean can be compared to the
PRG for 8XP°sure to that media. Use of
the litter value is recommended as the
more scientifically rigorous value for use in
these comparisons, if the contaminant
concentration Is lew than the PRG/10 for
non-carcinogens, or less thin the rrq
calculated at i l0- risk ft, carcinogens,
th« contaminant may be excluded as a
COC. f0r non-cardrK®ena, the
comparison value of 0.1 PRG ensures that
any additive adverse effects will still result
In a hazard index of less than one.
RECOMMENDATION
For sites where the preliminary list
of potentially site-related contaminants is
quite lengthy, ft is recommended that the
selection process outlined and described
above be used to evaluate the
contaminants and derive the final list of
COC's which will be carried through the
baseline risk assessment. Use of this
selection process, however, may not be
appropriate for all sites. It takes a fair
amount of time and resources to evaluate
each preliminary contaminant In this
selection process. Therefore, sites with
Pag® 7 of 10 Pages
Mi 1VIQ3K3H QMdM3d IS fU
SFZI m COC Wi OS:H 03u t-6 en it
-------
»¦» t».«. (PA
TnmiHlMBtH
smaller lists of preliminary contaminants
may find It easier to Just to carry all of the
identified contaminants through the
quantitative risk assessment evaluation.
REFERENCES
tot**** i§§4
Manual, Part B: "Development of Risk-
based Preliminary Remediation Goals".
Office of Emergency and Remedial
Response. OSWER Directive 9285.7-01 B.
1. Gilbert, R.O. and Simpson, J.S. (1992J.
Statistical Methods for Evaluating the
Attainment of Cleanup Standards, Volume
3: Reference-Based Standards for Soils
and Solid Media, PNL-7409 Vol. 3, Rev. 1,
Pacific Northwest Laboratory, Richland,
Washington.
2. U.S. Environmental Protection Agency
{EPA). 1983a. Risk Assessment Guidance
for Superfund, Volume 1: Human Health
Evaluation Manual. Office of Emergency
and Remedial Response, EPA/540/1-
89/002.
3. U.S. Environmental Protection Agency
(EPA). 1989b. Statistical Analysis of
Groundwater Monitoring Data at RCRA
Facilities, interim Final Guidance. Office of
Solid Wast®, Waste Management Division.
EPA/530-SW-83-026, April 1989.
4. U.S. Environmental Protection Agency
(EPA). 1931. Risk Assessment Guidance
for Superfund, Human Health Evaluation
Page 8 of 10 Pages
60®®
Hi 1YIQ3K3H ®OJ83«US «H
sezi cbs coc x'vj is:ti aa.it *6'eo u
-------
Figure 1 - Selection Process for COC's
tti
Ym
¥m
Ym
Hi
Km
WMIiMiwidiiiW
Im *1sM&m. Ifflfffinrfrnf mi
if Mini VVBVSINf69l KwwWwf m
Cs«ipCM«lii||lsT
Ortacllon F n^wne^S*
EuhiIW iitliiin!
Dow CMMntraHon EwMd
NwiMrieal Crlara?
(MMRi)
-------
Figure 2 - Decision Tree for Comparisons
of Central Tendency
Canprltoi of Cwttal TandMcy
^ PlfWit ^
afDatacNoM<5Q%oii
BacHpMMferSJI*.
¥n
Km.
WImmm
Rank Sum
TmI
Km
-------
ATTACHMENT G
U.S. ENVIRONMENTAL PROTECTION AGENCY
LAND USE MEMORANDUM
X RtP*R- *.!* TASK* REVISE. FINAL MASltR WltNil-ll
-------
nT
•v.
*V
\ UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
I WASHINGTON, DC 20460
^,y 2 5 :;-35
CF?IC£ cs
sono waste ande^ct.-.c*
SESPONSc
OSWER Directive So. 9355,7-04
MfORMPW
SUBJECT: Land Use in the CEfiCLA Remedy Selection Process
FROM: Elliott P. Laws~Zf/ Qj(
Asa is cant Adraind^Bjflrttoc^ X
TO: Director, Waste Management Division
Regions I, IV, v, VII
Director, Emergency and Remedial Response Division
Region II
Director, Hazardous Waste Management Division
Regions III, VI, VIII, IX
Director, Hazardous Waste Division,
Region X
Director, Environmental Services Division
Regions I, VI, VII
EUESSJIS:
This directive presents additional information for
considering land use in making remedy selection decisions under
the Comprehensive Environmental Response, Compensation, and
Liability Act (CERCLA) at National Priorities List (NPL) sites.
The U.S. Environmental Protection Agency {EPA) believes that
early community involvement, with a particular focus on the
community's desired future uses of property associated with the
CERCLA site, should result in a more democratic decisionmaking
process; greater community support for remedies selected as a
result of this process; and more expedited, cost-effective .
cleanups.
The major points of this directive are:
• Discussions with local land use planning authorities,
appropriate officials, and the public, as appropriate,
should be conducted as early as possible in the scoping
phase of the Remedial Investigation/Feasibility Study
(RI/FS). This will assist SPA in understanding the
^•cyemtrRccycJab*
P'mto w»»n jov>Ci"(Ni 70 sac*'
somami m tetvt SC% ''J»f
-------
reasonably anticipated future uses of the land en which
the Superfunci sits is located;
• If the site is located in a community that is likely to
have environmental justice concerns, extra efforts
should be made to reach out to and consult with
segments of the community that are not necessarily
reached by conventional communication vehicles or
through local officials and planning commissions,
• Remedial action objectives developed during the RI/FS
should reflect the reasonably anticipated future land
use or uses;
• Future land use assumptions allow'the baseline risk
assessment and the feasibility study to be focused on
developing practicable and cost effective remedial
alternatives. These alternatives should lead to site
activities which are consistent with the reasonably
anticipated future land use. However# there may be
reasons to analyze implications associated with
additional land uses;
• Land uses that will be available following completion
of remedial action are determined as part of the remedy
selection process. During this process, the goal of
realizing reasonably anticipated future land uses is
considered along with other factors. Any combination
of unrestricted uses, restricted uses, or use for long-
term waste management may result.
Discussions with local land use authorities and other
locally affected parties to make assumptions about future land
use are also appropriate in the RCRA context. EPA recognizes
that RCRA facilities typically are industrial properties that are
actively managed, rather than, the abandoned, sites that are often
addressed under CERCLA. Therefore, consideration of non-
residential uses is especially likely to be appropriate for RCRA
facility cleanups. Decisions regarding future land use that are
made as part of RCRA corrective actions raise particular issues
for RCRA (e.g., timing, property transfers, and the viability of
long-term permit or other controls) in ensuring protection of
human health and the environment. EPA intends to address the
issue of future land use as it relates specifically to RCRA
facility cleanups in subsequent guidance and/or rulemakings.
This guidance is also relevant for Federal Facility sites.
Land use assumptions at sites that are undergoing base closure
may be different than at sites where a Federal agency will be
maintaining control of the facility. Most land management agency
sites will remain in Federal ownership after remedial actions.
In these cases, Forest Land Management Plana and other resource
-------
manacement guidelines may help develop reasonable assumptions
about future uses of the land* Ac all such sices, however, zhz.s
document^ can focus the land use consideration toward appropriate
options.
Background:
Reasonably anticipated future use of the land at NPL sites
is an important consideration in determining Che appropriate
extent of remediation. Future use of the land will affect the
types of exposures and the frequency of exposures that may occur
to any residual contamination remaining on the site, which in
turn affects the nature of the remedy chosen. On the other hand,
the alternatives selected through the National Oil and Hazardous
Substance Contingency Plan (NCP) [55 Fed. Reg. 8666, March 8,
1990] process for CERCLA remedy selection determine the extent to
which hazardous constituents remain at the site, and therefore
affect subsequent available land and ground water uses.
The NCP preamble specifically discusses land use assumptions
regarding the baseline risk assessment, The baseline risk
assessment provides the basis for taking a. remedial'action at a
Superfund site and supports the development of remedial action
objectives. Land use assumptions affect the exposure pathways
that are evaluated in the baseline risk assessment. Current land
use is critical in determining whether there is a current risk
associated with a Superfund site, and future land use is
important in estimating potential future threats. The results of
the risk assessment aid in determining the degree of remediation
necessary to ensure long-term protection at NPL sites.
EPA has been- criticized for too often assuming that future
use will be residential. In many cases, residential use is the
least restricted land use and where human activities axe
associated with the greatest potential for exposures. This
directive is intended to facilitate future remedial decisions at
NPL sites by outlining a public process and sources of
information which should be considered in developing reasonable
assumptions regarding future land use.
This directive expands on discussions; provided is the
preamble to the National Oil and Hazardous Substance Contingency
Plan (NCP) ; "Risk Assessment Guidance for Superfund Vol. I, Human
Health Evaluation Manual" (Part AJ (EPA/540/1-89/002, Dec. 1983}?
"Guidance for Conducting Remedial Investigations and Feasibility
Studies Under CERCLA" (OSWER Directive 9355.3-01, Oct. 1988); and
Federal agency responsibility under CERCLA 120 (h) (31,
which relates to additional clean up which may be required to
allow for unrestricted use of the property, is not addressed in
this guidance.
-------
4
"Rcle of the Baseline Risk Assessment in Superfund Remedy
Selection Decisions" {OSWER Directive 9355.0-30, April 22, 1991! .
This land use directive may have the most relevance in
situations where surface soil is the primary exposure pathway.
Generally, where soil contamination is impacting ground water,
protection of the ground water may drive soil cleanup levels.
Consideration of future ground water use for CERCLA sites is not
addressed in this document. There are separate expectations
established for ground water in the NCP rule section 300.430
(a}(l)(iii)(F) that "EPA expects to return usable ground waters
to -their beneficial uses wherever practicable, within a timeframe
that is reasonable given the particular circumstances of the
site."
efei&stiza
This directive has two primary objectives. First, this
directive promotes early discussions with local land use planning
authorities, local officials, and the public regarding reasonably
anticipated future uses of the property on which an NPL site is
located. Second, this directive promotes the use of that
information to formulate realistic assumptions regarding" future
land use and clarifies how these assumptions fit in and influence
the baseline risk assessment, the development of alternatives,
and the CERCLA remedy selection process.
The approach in this guidance is meant to be considered at
current and future sites in the RI/FS pipeline, to the extent
possible. This directive is not intended to suggest that
previous remedy selection decisions should be re-opened.
In order to ensure use of realistic assumptions regarding
future land uses at a site, El&_atoiM-di^aafljgaaQnably
attfctoCLJU&adu.lQgal-affislalaj—and-tlte_BabllgJ_aa_aBDEgBgiafce^_aa
should gain an understanding of the reasonably anticipated future
land uses at a particular Superfund site to perform the risk
assessment and select the appropriate . remedy.
A visual inspection of the site -and its surrounding area is
a good starting point in developing assumptions regarding future
land use. Discussions with the local land use authorities and
appropriate officials should follow. Discussions with the public
can be accomplished through a public meeting and/or other means.
By developing realistic assumptions based on information gathered
from these sources early in the El/FS process, EPA may develop
-------
remedial alternatives that are consistent with the anticipated
future use.
The development of assumptions regarding the reasonably
anticipated future land use should not become an extensive,
independent research project.. Site managers should use existing
information to the extent possible, much of which will be
available from local land use planning authorities. Sources and
types of information that may aid EPA in determining the
reasonably anticipated future land use include, but are not
limited to:
• Current land use
• Zoning laws
• Zoning maps,
• Comprehensive community master plans
• Population growth patterns and projections (e.g.#
Bureau of Census projections!
• Accessibility of-site to existing infrastructure (e.g.,
transportation and public utilities)
• Institutional controls '-urrently in place
• site location in relation to urban, residential,
commercial, industrial, agricultural and recreational
areas
• Federal/State land use designation (Federal/State
control over designated lands range from established
uses for the general public, such as national parks or
State recreational areas, to governmental facilities
providing extensive site access restrictions, such as
Department of Defense facilities
• Historical or recent development patterns
• Cultural factors (e.g., historical sites, Native
American religious site*}
• Natural resources information
• Potential vulnerability of ground water to contaminants
that might migrate from soil
• Environmental justice issues
• Location of on-site or nearby wetlands
¦ • Proximity of site to a floodplain
• Proximity of site to critical habitats of endangered or
threatened species
• Geographic and geologic information
• Location of Wellhead Protection areas, recharge areas,
and other areas identified in a State's Comprehensive
Ground-water Protection Program .
These types of information should be considered when
developing the assumptions about future land use. Interaction
with the public, which includes all stakeholders affected by the
site, should serve to increase the certainty in the assumptions
made regarding future land use at an NPL site and increase the
-------
o
confidence sx?ec:acicns about ancicipatec future land use are, ir.
face, reasonable.
For example# fucure industrial land use is likely to be a
reasonable assumption where a site is currently used for
industrial purposes, is located in an area where the surroundings
are zoned for industrial use, and the comprehensive plan predicts
the site will continue to be used, fox* industrial purposes.
fMimmitv laTOlyaaaBt.
NPL sites are located, in divers© areas of the country, with,
great variability in land use planning practices,. For some NPL
sites, the future land use of a site may have been, carefully
considered through local, public# participatory, planning
processes, such as zoning hearings, master plan approvals or
other vehicles. When this is the case, local residents around,
the Superfund site are likely to demonstrate substantial
agreement with the local land use planning authority on the
future use of the property. Where there is substantxal agreement
among local residents and land use planning agencies, owners and
developers, EPA can rely with a great deal of certainty on the
future land use already anticipated for the site. For other NPL
sites, however, the absence or nature of a local planning process
may yield considerably less certainty about what assumptions
regarding future use are reasonable. In some instances the local
residents near the Superfund. site may feci disenfranchised from
the local land use planning and development process. This may be
an especially important issue where there are concerns- regarding
environmental justice in the neighborhood around the NPL site.
Consistent with the principle of fairness, EPA should make an
extra effort to reach out to the local community to establish
appropriate future land use assumptions at such sites.
Future iand^jiag_^aaimBfclgna-allggJAs..teafigliJna--CL3l£
aaffiRaaBianfc-aBd^hg-ffl*aihi^ifi3LJ£MdaJa-fflfflML.giiJ&fi-flggBij£gaBSB&
of ,Pgac&i^abig_and^fig&iflffgg&lgaLjangflU^Jl&figȣJUaflx-lsflfliBg
to fits activities,..Hhigh ,liC6 MtlBliBtCBt ^Itll She r6ft5QtHblY
enticigated future land ugg.
The baseline risk assessment generally needs only to
consider the reasonably anticipated future land use; however, it
may be valuable to evaluate risks associated with other land
uses. The NCP preamble (55 Fed. Reg. 0710) states that in the
baseline risk assessment,, more than one future land use
assumption, may ba considered when decision makers wish, to
understand the implications of unexpected exposures. Especially
where there is some uncertainty regarding the anticipated future
land use, it may be useful to compare the potential risks
associated with, several land use scenarios to estimate the impact
-------
cn human health and the environment shcuia cr.e land use
unexpectedly change. The magnitude of such potentia. i.T.pacts may
be an important consideration in determining whether and how
institutional controls should be used co restrict future uses.
If the baseline risk assessment evaluates a future use under
which exposure is limited, it will not serve the traditional
role, evaluating a "no action" scenario. A remedy, i.e.
institutional controls to limit future exposure, will be required
to protect human health and the environment. In addition to
analyzing human health exposure scenarios associated with certain
land uses, ecological exposures nay also need to be considered.
Remedial action objectives provide the foundation upon which
remedial cleanup alternatives are developed. In general*
remedial action objectives ahPUld dSYglPPtti in ^9 ,
dfigelflB-al£fi£na£A3Ea«-£bafi-jggiild^ag±M3Eft-glgaauB-lg3EBla,Aaagstfl^sa
with thft_rgaaflnabl3g-an£ifiiBa£giMfMfiMEB.JLanfl*Jtfg-ggBC-&a;:fflUsa^£
the site an possible. CPA recognizes, however, that achieving
either the reasonably anticipated land use, or the land use
preferred by the community, may not be practicable across the
entire site, or in some cases, at all. For example, as RI/FS
data become available, they may indicate that the remedial
alternatives under consideration for achieving a level of cleanup
consistent with the reasonably anticipated future land use are
not cost-effective nor practicable. If this is the case, the
remedial action objective may be revised which may result in
different, more reasonable land use(s).
EPA's remedy selection expectations described in section
300 .430 (a) (1) (iii) of the NCP should also be considered when
developing remedial action, objectives. Where practicable, EPA
expects to treat principal threats, to use engineering controls
such as containment for low-level, threats, to use institutional
controls to supplement engineering controls, to consider the use
of innovative technology, and to return usable ground waters to
beneficial uses to protect human health and the environment.
(Some types of applicable or relevant and appropriate
requirements (ARARs) define protective cleanup levels which may,
in turn, influence post-remediation land use potential.)
In cases where the future land use is relatively certain,
the remedial action objective generally should reflect this land
use. Generally, it need not include alternative land use
scenarios unless, as discussed above, it is impracticable to
provide a protective remedy that allows for that use. A landfill
site is an example where it is highly likely that the future land
use will remain unchanged (i.e., long-term waste management
area) , given the NCP's expectation that treatment of high volumes
of waste generally will be impracticable and the fact that EPA's
presumptive remedy for landfills is containment. In such a case,
-------
3
a remedial action objective could be established with a very hicr.
degree of certainty to reflect the reasonably anticipated future
land use.
In cases where the reasonably anticipated future land use is
highly uncertain, a range of the reasonably likely future land
uses should be considered in developing remedial action
objectives. These likely future land uses can be,reflected by
developing a range of remedial alternatives that will achieve
different land use potentials. The remedy selection process will
determine which alternative ia moac appropriate for the sice and,
consequently, the land use Is) available following remediation,.
As discussed in "Role of the Baseline Risk Assessment: in
Superfund Remedy Selection Decisions" (OSWER Directive 9355.0-30,
April 22, 1991), EPA has -established a risk range for carcinogens
within which EPA strives to manage site risks. EPA recognizes
chat a specific cleanup level within the acceptable risk range
may be associated with more than, one land use (e.g., an
industrial cleanup to 10 may also allow for residential use at
'a 10 risk level.) It is not EPA's intent that the risk range
be partitioned into risk standards based solely on categories of
land use {e.g., with residential cleanups at the 10 level and
industrial cleanups at the 10" risk level.) Rather, the risk
range provides the necessary flexibility to address the technical
and cost limitations, and the performance and risk uncertainties
inherent in all waste remediation efforts.
Land Use Considerations in tfdy Select
As a result of the comparative analysis of alternatives with
respect to EPA's nine evaluation criteria, EPA selects a site-
specific remedy. The remedy determines the cleanup levels, the
volume of contaminated material to be treated, and the volume of
contaminated material to be contained. Consequently, the remedy
selection decision determines the size of the area that can be
returned to productive use and the particular types of uses that
will be possible following remediation. -
The volume and concentration of contaminants left on-site,
and thus the degree of residual risk at a site, will affect
future land use. For example, a remedial alternative may include
leaving in place contaminants in soil at concentrations
protective for industrial exposures, but not protective for
residential exposures. In this case, institutional controls
should be used to ensure that industrial use of the land is
maintained and to prevent risks from residential exposures.
Conversely, a remedial alternative may result in no waste left in
place and allow for unrestricted use (e.g., residential use) .
-------
avaults of Remedy Selection Process
Several potential land use situations could result from
E?A"s remedy selection decision. They are:
• The remedy achieves cleanup levels that allow the
entire site to be available for the reasonably
anticipated future land use in the baseline risk^
assessment (or, where future land use is uncertain, all
uses that could reasonably be anticipated).
• The remedy achieves cleanup levels that allow most, but
not all, of the site to be available for the reasonably
anticipated future land use. For example, in order to
be cost effective and practicable, the remedy may
require creation of a long-term waste management area
for containment of treatment residuals or low-level
waste on a small portion of the site. The cleanup
levels in this portion of the site might allow for a
more restricted laud use.
• The remedy achieves cleanup levels that require a more
restricted land use than the reasonably anticipated
future land use for the encirs site. This situation
occurs when no remedial alternative that is cost-
effective or practicable will achieve the cleanup
levels consistent with the reasonably anticipated
future land use. The site may still be used for
productive purposes, tat the use would be more
restricted than the reasonably anticipated future land
use. Furthermore, the more restricted use could be a
long-term waste management area over all or a portion
of the site.
if any remedial alternative developed during the PS will
require a restricted land use in order to be protective, it is
essential that the alternative include components that will
ensure that it remain protective. In particular, institutional
controls will generally have to be included in the alternative^to
prevent an unanticipated change in land use that could result in
unacceptable exposures to residual contamination, or, at a
minimum, alert future users to the residual risks and monitor for
any changes in use. In such cases, institutional controls will
play a key role in ensuring long-term protectiveness and should
be evaluated and implemented with the same degree of care as is
given to other elements of the remedy. In developing remedial
alternatives that include institutional controls, EPA should
determine: the type of institutional control to be used, the
existence of the authority to implement the institutional
control, and the appropriate entity's resolve and ability to
-------
10
implement the institutional control. to alternative may
anticipate two or more options for establishing institutional
controls, but should fully evaluate all such options. A variety
of institutional controls may be used such as deed restrictions
and deed notices, and adoption of land use controls by a local
government. These controls either prohibit certain kinds of site
uses or, at a minimum, notify potential owners or land users of
the presence of hazardous substances remaining on site at levels
that are not protective for all uses. Where exposure must be
limited to assure protectiveness, a deed notice alone generally
will not provide a sufficiently protective remedy. While the ROD
need not always specify the precise type of control to be
imposed, sufficient analysis should be shown in the PS and ROD to
support a conclusion that effective implementation of
institutional controls can reasonably be expected.
Suppose, for example, that a selected remedy will be
protective for industrial land use and low levels of hazardous
substances will remain on site. An industry may still be able to
operate its business with the selected remedy in place.
Institutional controls, however, Generally will need to be
established to ensure the.land is not used for other, less
restricted purposes, such as residential use, or to alert
potential buyers of any remaining contai&ination.
Where waste is left on-site at levels that would require
limited use and restricted exposure, EPA will conduct reviews at
least every five years to monitor the site for any changes. Such
reviews should analyze the implementation and effectiveness of
institutional controls with the same degree of care as other
parts of the remedy. Should land use change, it will be
necessary to evaluate the implications of that change for the
selected remedy, and whether the remedy remains protective.
EPA's role in any subsequent additional cleanup will be
determined on a site-specific basis, if landowners or others
decide at a future date to change the land use in such a way that
makes further cleanup .necessary to ensure protectiveness, CERCLA
does not prevent them from conducting such a cleanup as^long as
protectiveness of the remedy is not compromised. (EPA may invoke
CERCLA section 122(e)(6), if necessary, to prevent actions that
are inconsistent with the original remedy.) In general, EPA
. would not expect to become involved actively in the conduct or
oversight of such cleanups« EPA, however, retains its authority
to take further response action where necessary to ensure
protectiveness.
6
-------
Further Information
If you have any questions concerning this directive, please
call Sherri Clark at 703-603-9043,
NOTICEi The policies sec out in this memorandum are intended
solely as guidance. They are not intended, nor can they be
relied upon, to create any rights enforceable by any party in
litigation with the United States. EPA officials may decide to
follow the guidance provided in this memorandum, or to act at
variance with the guidance, based on an analysis of specific site
circumstances. Remedy selection decisions are made and justified
on a case-specific basis. The Agency also reserves the right co
change this guidance at any time without public notice.
-------
ATTACHMENT H
SOIL SCREENING GUIDANCE
CHEMK tOPERTIES TABLE C
\ RfP* REVlStlFWJKLMASTER WpJ*|*|»inmiwro«ftiw2lMfct
-------
Attachment C
Chemical Properties for SSL Development
-------
Attachment C
Chemical Properties
This attachment provides the chemical properties necessary to calculate inhalation and migration to
ground water SSLs (see Section 2,5.2) for 110 chemicals commonly found at Superfumd sites. The
Technical Background Document for Soil Screening Guidance describes the derivation and sources
for these property values.
• Table C-I provides soil organic carbon - water partition coefficients (K^), air and water
diffusivities (Di a and D,tW). water solubilities (S). and dimensionless Henry's law constants
(H'>.
Table C-2 provides pH-specific Koc values for organic contaminants thai ionize under natural
pH conditions. Site-specific soil pH measurements (see Section 2.3.5) can be used to select
appropriate Koc values for these chemicals. Where site-specific soil pH values are not
available, values corresponding to a pH or 6 J should be used (note that the K
-------
Table C-1. Chemical-Specific Properties used in SSL Calculations
Koc
D,..
Di,w
S
H'
CAS No.
Compound
(L/kg)
(cm2/s)
(cm2/»)
(mg/L)
(dimensionlessf
83-32-9
Acenaphthene
7.08E+03
4.21E-02
7.89E-06
4.24E+00
6.38E-03
87-64-1
Acetone
5.75E-01
1.24E-01
1.14E-05
1.00E+06
1.53E-03
309-00-2
Aldrin
2.45E+06
1.32E-02
4.86E-06
1.80E-01
6.97E-03
120-12-7
Anthracene
2.95E+04
3.24E-02
7.74E-06
4.34E-02
2.67E-03
56-55-3
Benz(a)anthracene
3.98E+05
5.10E-02
9.00E-06
9.40E-03
1.37E-04
71-43-2
Benzene
5.B9E+01
8.80E-02
9.80E-06
1.75E+03
2.28E-01
205-99-2
Benzo
-------
Table C-1 (continued)
^oc
D,,«
D|,w
S
H*
CAS No.
npound
(L/kg)
(cm2/t)
(mg/L)
(dimenslontess)
51-28-5
2,4-Pinitrophenol
1.00E-02
2.73E-02
nwe&oeT
"~2,79E+03
1 82E-05
121-14-2
2,4-Pinitrototuene
9.55E+01
2,.03E-<31
7.06E-06
2.70E+02
3 80E-06
006-20-2
2,6-Dinitrololuerie
6.92E+01
3.27E-02
7.26E-06
1J2E+02
3 06E-05
117-84-0
Di-n-octyl phthalate
8.32E+07
1.51E-02
3.58E-06
2.00E-02
2.74E-03
115-29-?
Endosuifan
2.14E+03
1.15E-02
4.55E-06
5.10E-01
4.59E-04
72-20-8
Endrin
1.23E+04
1.25E-02
4.74E-06
2.50E-01
3.08E-04
100-41-4
Ethylbenzene
3.63E+02
7.50E-02
7.80E-06
1.69E+02
3.23E-01
206-44-0
Ftuoranlhene
1.07E+05
3.02E-02
6.35E-06
2.06E-01
6 60E-04
86-73-7
Fluorene
1.38E+04
3.63E-02
7.88E-06
1J8E+00
2.61 E-03
76-44»8
Heptachlor
1.41E+06
1.12E-02
5.69E-06
1J0E-01
4.47E-02
1024-57-3
Heptachlor epoxide
8.32E+04
1.32E-02
4.23E-06
2.00E-01
3.90E-04
118-74-1
Hexachlorobenzene
5.50E+04
5.42E-Q2
5.91 E-06
6.20E+00
5.41 E-02
87-68-3
Hexachloro-1,3-buladlene
5.37E+04
5.61 E-02
6.16E-06
3.23E+00
3.34E-01
319-84-8
a-HCH (a-8HG)
1.23E+03
1.42E-02
7.34E-06
2.00E+00
4.35E-04
319-85-7
BrHCH (&-BHC)
1.26E+03
1.42E-02
7.34E-06
240E-01
3 05E-05
58-89-3
Y-HCH (Lindane)
1.07E+03
1.42E-02
7.34E-06
8.80E+00
5.74E-04
77-47-4
Hexachlorocyclopentadiene
2.0QE+05
1,81 E-02
7.21 E-06
1.80E+00
M1E+00
87-72-1
Hexachloroethane
1.78E+03
2.50E-03
6.80E-06
5.00E+01
1 59E-01
193-39-5
lndeno(1,2.3-ccOpyrene
347E+06
1.90E-02
5.66E-06
2.20E-05
6.56E-05
78-59-1
isophorone
4.88E+01
6.23E-02
8.76E-06
1.20E+04
2.72E-04
7439-97-6
Mercury
_
3.07E-02
6 30E-06
—
4.67E-01
72-43-5
Methoxychter
9.77E+04
1 56E-02
4.46E-06
4.50E-02
6.48E-04
74-83-9
Methyl bromide
1.05E+01
7.28 E-02
1.21E-05
1 52E+04
2.56E-01
75-09-2
Methylene chloride
1J7E+01
1.01E-01
1.17E-05
1J0E+04
8.98E-02
95-48-7
2-Methytphenot
9.12E+01
7.40E-02
8.30E-06
2 60E+04
4.92E-05
91-20-3
Naphthalene
2.00E+03
5.90E-02
7.50E-06
3.10E+01
1.98E-02
98-35-3
Nitrobenzene
8.46E+01
7.60E-02
8.60E-06
2.09E+03
9.84E-04
06-30-8
W-Nitrosodiphenylamine
1.29E+03
3 12E-02
6.35E-06
3.51 E+01
2.05E-04
821-64-7
J^Nitresodi-n-prapyiarnine
2.40E+01
5.45E-02
8.17E-06
9.89E+03
9.23E-05
1336-38-3
FCBs
3.09E+05
—
—
7.00E-01
—
87-86-5
Pertachlorophenol
5 92E+Q2
5.60E-02
6.10E-06
1.9SE+03
1 00E-O6
108-35-2
Phenol
288E+01
8.20E-02
9.10E-06
8.28E+04
1.63E-05
129-00-0
Pyrene
105E+05
2.72E-02
7.24E-06
1.35E-01
4.51 E-04
100-42-5
Styrene
7.76E+02
7.10E-02
8.00E-06
3.10E+02
1.13E-01
79-34-5
1,1,2,2-T etracftioroelhans
9.33E+01
7.10E-02
7.90E-06
2.97E+03
1.41 E-02
127-18-4
T effachloroelhyferte
1 55E+02
7 20E-02
8.20E-06
2.00E+02
7 54E-01
108-88-3
Toluene
1.82E+02
8.70E-02
8.60E-06
5.26E+Q2
2 72E-01
8001-35-2
Toxaphene
2.57E+05
1.16E-02
4.34E-06
740E-01
2.46E-04
120-82-1
1,2,4-T rtchlorobenzene
1.78E+03
3.00E-02
823E-06
3.00E+02
5.82E-02
71-55-8
1,1,1-T richloroethane
1 10E+02
7 80E-02
8 80E-06
1.33E+03
7 05E-01
79-00-5
1»1,2-T richloroethane
5.01 E+01
7J0E-02
8.80E-06
4.42E+03
3 74E-02
79-01-6
T richtoroethytene
1.66E+02
7.90E-02
9.10E-06
1.10E+03
4.22E-01
95-35-4
2,4,5-T richlorophenol
1 60E+03
2.91 E-02
7.03E-06
1.20E+03
1 78E-04
88-06-2
2,4.6-T richlorophenol
3.81 E+02
3 18E-02
6.2SE-06
8.00E+02
3 19E-04
C-3
-------
Table C-1 (continued)
^OC
Dlt,
D|.w
S
H*
CAS No.
Compound
(L/kg)
(cmJ/s)
(cmZ/s)
Img'Li
(dimensioriless)
108-05-4
Vinyl acetate
5 25E+00
8.50E-02
&20E-06
2.00E+04
2 10E-02
75-01-4
Vinyl chloride
1.B6E+01
1.06E-01
1.23E-06
2.76E+03
1.11E+00
108-38-3
m-Xytene
4.07E+02
7.00E-02
7.806-06
1.61E+02
3 01E-01
95-47-6
o-Xytene
3.83E+02
8.70E-02
1.00E-05
1.78E+02
2.13E-01
106-42-3
p-Xytene
3.89E+02
7.69E-02
8.44E-06
1.S5E+02
3.14E-01
Koe = Soil organic cartoorwtoater partition coefficient
Di.o * Difteivity in air (28 *C),
D1W = Dtffuwvity in water (25 •€)
S = Solubility in water (20-26 *C).
H' = Dimensiontess Henry's law constant {HlC[alm-m3/(T»IJ*41H25«C),
K
-------
Table C-2. Koc Values for Ionizing Organics as a Function of pH
PH
4.9
5.0
5.1
5.2
5.3
5.4
5.5
5.6
5.7
5.8
5.9
8.0
6.1
6.2
6.3
6.4
6.5
66
6.7
6.8
6.9
7.0
7.1
7.2
7.3
74
7.5
7.6
7.7
78
7.9
80
Benzoic
Acid
5.54E+00
4.64E+00
3J8E+00
3.25E+00
2.72E+00
2 29E+00
1.84E+00
1.85E+00
1.42E+00
1.24E+00
1.09E+00
9.69E-01
8.75E-01
7.99E-01
7.36E-01
6.89E-01
6 51E-01
6.20E-01
5.35E-01
5.76E-01
5 60E-01
5.47E-01
5.38E-01
5 32E-01
5.25E-01
5.19E-01
5.16E-01
5 13E-01
5.09E-01
5 06E-01
5.06E-01
5.06E-01
2-
Chloro-
phcnol
3.98E+02
3.98E+02
3.98E+02
3.98E+02
3.98E+02
3.98E+02
3.97E+02
3.97E+02
3.97E+02
3 97E+02
3.97E+02
3.96E+02
3.96E+02
3.96E+02
3.95E+02
3.94E+02
3.93E+02
3 92E+02
3.90E+02
3.88E+02
3.88E+02
3J3E+02
3 79E+02
3.75E+02
3.69E+02
3.82E+02
3.54E+02
3.44E+02
3.33E+02
3.19E+02
3.04E+02
2 88E+02
2,4-
2,4-Dtehtoro-Olnitro-
phenol phenol
PtBtacfior®-
ptwnol
2,3,4,5-
Tatrachloro-
phenol
2A4.6-
Telrachloro-
phenol
2,4,5-Trichloro-
phertoi
2,4,6-
Trtcbtaf©-
pnenot
1.59E+02 2.94E-02
1.59E+02 2 55E-02
1.59E+02 2.23E-02
1.59E+02 1.98E-02
1.53E+02 1.78E-02
1.58E+02 1.62E-02
1 58E+02 1 50E-02
1.58E+02 1.40E-02
1.58E+02 1.32E-02
1.58E+02 1.25E-02
1.57E+02 1.20E-02
1.57E+02 1.16E-02
1.57E+02 1.13E-02
1.56E+02 1.10E-02
1.55E+02 1.08E-02
154E+02 1 06E-02
1.53E+02 1 05E-02
1.52E+02 1 04E-02
1.50E+02 1.03E-02
1.47E+02 1.02E-02
1.4SE+02 1.02E-02
1.41E+02 1.Q2E-02
1.38E+02 1.02E-02
1.33E+02 1.01E-02
1.28E+02 1.01E-02
1.21 E+02 1.01E-02
1.14E+02 1.01E-02
1.07E+02 1.01E-02
9.84E+01 1 00E-02
8.97E+01 1.006-02
8.07E+01 1.00E-02
7.17E+01 1.00E-02
9.05E+03
1.73E+04
4.45E+03
2.37E+03
1.04E+03
7.96E+03
1.72E+04
4.I5E+03
2.36E+03
1.03E+03
8.93E+03
1.7QE+04
3.83E+03
2.36E+03
1 02E+03
5.97E+03
1.67E+04
3 49E+03
2.35E+03
1.01 E+03
5.10E+O3
1.65E+04
3.14E+03
2.34E+03
9.99E+02
4.32E+03
1.61E+04
2.79E+03
2.33E+03
9.82E+02
3.65E+03
1.57E+04
2.45E+03
2.32E+03
9.62E+02
3.07E+03
1.52E+04
2.13E+03
2,31 E+03
9.3SE+02
2.58E+03
147E+04
1.83E+03
2 29E+03
9.10E+02
2.18E+03
1.40E+04
1.56E+03
2.27E+03
8.77E+02
1.84E+03
1.32E+04
1 32E+03
2 24E+03
8 39E+02
1.56E+03
1.24E+04
1.11 E+03
2.21 E+03
7.96E+02
1.33E+03
1.15E+04
9.27E+02
2.17E+03
7.48E+02
1.15E+03
1.05E+04
7.75E+02
2.12E+03
6.97E+02
9 98E+02
9.51 E+03
6.47E+02
2.06E+03
8.44 E+02
8.77E+02
8.48E+03
5.42E+02
1.99E+03
5.89E+02
7.81 E+02
7.47E+03
4.55E+02
1.91 E+03
5.33E+02
7.03E+02
6.49E+03
3.84E+02
1.82E+03
4.80E+02
6.40E+02
5.53E+03
3.27E+02
1.71 E+03
4.29E+02
5.92E+02
4.74E+03
2.80E+02
1.60E+03
3,81 E+02
5.52E+02
3.99E+03
242E+02
1.47E+03
3.38E+02
5.21 E+02
3.33E+03
2.13E+02
1.34E+03
3 00E+02
4.98E+02
2.76E+03
1.88 E+02
1,21 E+03
2.67E+02
4.76E+02
2.28E+03
1 69E+02
1.07E+03
2.39E+02
4,81 E+02
1.87E+03
1 53E+02
9.43E+02
2.15E+02
4.47E+02
1.53E+03
1.41 E+02
8.19E+02
1.95E+02
4.37E+02
1.25E+03
1,316+02
7.03E+02
1 78E+02
4.29E+02
1.Q2E+03
1.23E+02
5.99E+02
1.64E+02
4.23E+02
8.31 E+02
1.17E+02
5.07E+02
1J3E+02
4.18E+02
8.79E+02
1.13E+02
4,261+02
1.44 E+02
4.14E+02
5.56E+02
1.08E+02
3.57E+02
1 37E+02
4.10E+02
4.58E+02
1.05E+Q2 ,
2.98E+02
1.31 E+02
C-5
-------
Table C-3, Physical Slate of Organic SSL Chemicals
Compounds liquid at m&lmiwmtoim ' ' ' Compoundssolid at soil temperatures
CAS No.
Chemical
Melting
Point {C)
CAS No. Chemical
Melting
Poin
Acetone
-94,8
83-32-9 Acenaphthene
93 4
71-43-2
Benzene
5.5
309-00-2 Aldrtn
104
117-81-7
Bis(2-ethylhexfl)phthalate
-55
120-12-7 Anthracene
215
111-44-4
Bis(2-chloroethyl)ether
-S1.9
56-55-3 Benz(a)arthracene
84
75-27-4
iromodichloromethane
-57
50-32-8 Benzo(a)pyrene
176.5
75-25-2
Bromoform
a
205-99-2 Benzo(b)fluorar»thene
188
71-38-3
Butanoi
-89.8
207-08-9 Benzo(lc)flyoranthene
217
85-68-7
Butyl benzyl phthalate
-35
85-85-0 Benzoic acid
122.4
* 75-15-0
Carbon disulfide
-115
86-74-8 Carbazole
246.2
56-23-5
Carbon tetrachloride
-23
67-74-9 Chlordane
106
108-90-7
Chtorobenzene
-45.2
106-47-8 p-Chloroaniline
72.5
124-48-1
Chlorodibromomethane
-20
218-01-9 Chrysene
258.2
67-66-3
Chloroform
-63.8
72-54-8 ODD
109.5
95-57-8
2-Chlorophenol
9.8
72-55-9 DDE
89
B4-74-2
Di-n-butyl phthalate
-35
50-29-3 DDT
108.5
95-50-1
1,2-Dichlorobenzene
-16.7
53-70-3 Dibenzo(a,ft)anthracerie
269.5
75-34-3
1,1-DichIoroethane
-96.9
106-46-7 1,4-Dichlorobenzene
52.7
107-06-2
1,2-DichIoroelhane
-35.5
31-34-1 3,3-Dichlorobenzidine
132.5
75-35-4
1,1-DichloroethyIene
-122.5
120-83-2 2,4-Dichlorophenol
45
156-59-2
cis-1,2-DicNoroethylene
-80
60-57-1 Dieldrin
175.5
158-60-5
(rans-1,2-Dichlofoethylene
-49.8
105-67-9 2,4-Dimethyt phenol
24.5
78-87-5
1,2-DichIoropropane
-70
51-28-5 2,4-Dinitrophenol
115-116
542-75-6
1,3-Dichloropropene
NA
121-14-2 2.4-Dinitrotoluene
71
84-86-2
Diethyl phthalate
-40.5
506-20-2 2,8-Dinitrotoluene
68
117-84-0
Di-n-octyl phthalate
-30
72-20-8 Endrin
200
100-41-4
Ethylbenzene
-94,9
206-44-0 Fluoranthene
107.8
87-68-3
Hexach!oro-1,3-butadiene
-21
86-73-7 Fluorene
114.8
77-47-4
Hexachtorocyclopentadiene
-9
76-44-8 Heptaehtor
95.5
78-59-1
isophorone
-8.1
1024-57-3 Heptachlor epoxide
160
74-83-9
Methyl bromide
-93.7
118-74-1 Hexachlorobenzeme
231.8
75-09-2
Methylene chloride
-95.1
313-84-8 ct-HCH (ot-BHC)
180
98-95-3
Nitrobenzene
5.7
319-85-7 BrHCH (&-BHC)
315
100-42-5
Styrene
-31
58-89-9 y-hCH (Lindane)
112.5
79-34-5
1,1,2,2-Tetrachloroelharie
-43.8
67-72-1 Hexachloroetharte
187
127-18-4
T etrachloroethy lene
-22.3
193-39-5 lr»deno(1,2,3-ctf)pyrene
181.5
106-88-3
Toluene
-94.9
72-43-5 Methoxychlor
87
120-82-1
1,2,4-T richtorobenzene
17
95-48-7 2-Methytphenoi
29.8
71-55-6
1,1»1 -T richloroethane
-30.4
821-64-7 M-Nitrosodi-n-propylamirie
NA
79-00-5
1.1,2-T richloroetha ne
-38.8
86-30-6 W-Ntfrosodiphenylamine
86.5
73-01-6
Trichtoroethylene
-84.7
91-20-3 Naphthalene
80.2
108-05-4
Vinyl acetate
-93.2
87-88-5 Pentaehlorophenol
174
75-01-4
Vinyl chloride
-153.7
108-95-2 Phenol
40.9
108-38-3
m-Xylerte
-47.8
129-00-0 Pyrene
151.2
95-47-8
o-Xyierte
-25.2
8001-35-2 Toxaphene
65-90
106-42-3
p-X ylene
13.2
95-95-4 2,4,5-Trichlorophenol
89
88-06-2 2,4.6-Tnchlorophenol
115-29-7 Endosuttfan
69
108
NA * Not available.
C-6
-------
Table €-4. Metal Kd Values (L/kg) as a Function of pH*
pH
As
B«
Be *
Cd
Cr (+3)
4,9
2.5E+01
1 JE+01
2.3E+01
1.5E+01
1.2E+03
5.0
2.5E+Q1
1.2E+01
2.8E+01
1 7E+01
1.9E+03
5,1
2.5E+01
1.4E+01
2.8E+01
1 9E+01
3.0E+03
5.2
2.6E+01
1.5E+01
3.1E+01
2 JE+01
4.9E+03
5.3
2.6E+01
1 JE+01
3.5E+01
2.3E+01
8.1E+03
5,4
2.8E+01
1 JE+01
3.8E+01
2.5E+01
1.3E+04
5.5
2.8E+01
2.1 £+01
4.2E+01
2.7E+01
2.1E+04
5.8
2,8E+§1
2.2E+01
4.7E+01
2.9E+01
3.5E+04
5.7
2.7E+01
2.4E+01
5.3E+01
S.1E+01
5.5E+04
S.B
2.7E+01
2.8E+D1
6 0E+01
3.3E+01
B.7E+04
5.9
2.7E+01
2.8E+01
6.9E+01
3.5E+01
1.3E+05
6.0
2.7E+01
3.0E+01
8.2E+0I
3.7E+01
2.0E+05
6.1
2.7E+01
3.1E+01
9.9E+01
4.0E+01
3.0E+05
8.2
2.8E+01
3.3E+01
1.2E+02
4.2E+01
4.2E+05
8.3
2.8E+01
3.5E+01
1 JE+02
4 4E+01
5.8E+05
6.4
2.8E+01
3 JE+01
2.1E+02
4.BE+01
7.7E+05
6.8
2.8E+01
3.7E+01
2,JE+02
5.2E+01
9.9E+05
6.8
2.BE+01
3.9E+01
3.9E+02
5.7E+01
1.2E+08
8.7
2.9E+01
4.0E+01
5.5E+02
6.4E+01
1.5E+06
S.B
2.9E+01
4 JE+01
7.9E+02
7.5E+01
1.8E+06
e.i
2JE+01
4.2E+01
1.1E+03
i JE+01
2.1E+08
7.0
2 JE+01
4.2E+01
1.7E+03
1 JE+02 ,
2.5E+06
7.1
2J6+01
4.3E+01
2.5E+03
1.5E+02
2.BE+06
7,2
3.0E+01
4.4E+01
3.8E+03
2.0E+02
3.1E+08
7.3
3.0E+01
4.4E+01
5.7E+03
2.8E+02
3.4E+06
7,4
3.0E+01
4.5E+01
8.8E+03
4.0E+02
3.7E+08
7.5
3.0E+01
4.6E+01
1.3E+04
5.9E+02
3.9E+06
7.8
3.1E+01
4.8E+01
2.0E+04
8.7E+02
4.1E+08
7.7
3 JE+01
4.7E+01
3.0E+04
1.3E+03
4.2E+06
7.8
3 JE+01
4.9E+01
4.6E+04
1.9E+03
4.3E+06
7.9
3 JE+01
5.0E+01
6.9E+04
2.9E+03
4.3E+08
B.O
3 JE+01
5.2E+01
1.0E+05
4.3E+03
4.3E+06
Cr (+6)
Mfl
Ni
Afl
Se
Tl
Zn
3 JE+01
3 JE+01
3.0E+01
2.9E+01
2.BE+01
2.7E+01
2.7E+01
2.6E+01
2.5E+01
2.5E+01
2.4E+01
2.3E+01
2.3E+01
2.2E+01
2.2E+01
2 JE+01
2.0E+01
2.0E+01
1JE+01
1.9E+01
1.BE+01
1 JE+01
1.7E+01
1.7E+01
1.6E+01
1.6E+01
1 JE+01
1.SE+01
1.5E+01
1.4B+01
1 4E+01
1.4E+01
4.0E-02
6.0E-02
9.0E02
1.4E-01
2.0E-01
3.0E0!
4.6E-01
6.9E-01
1.0E+00
1.6E+00
23E+00
3.5E+00
5.1E+00
7.5E+00
1 1E+01
1.6E+01
2.2E+01
3.0E+01.
4.0E+01
S.2E+01
6.8E+01
B.2E+01
9.9E+01
1.2E+02
1.3E+02
1.5E+02
1.6E+02
1 7E+02
1BE+02
1.9E+02
1.9E+02
2.0E+02
1.6E+01
1 JE+01
2.0E+01
2.2E+01
2.4E+01
2.6E+01
2.8E+01
3.0E+01
3.2E+01
3.4E+01
3.6E+01
3.8E+01
4.0E+01
4.2E+01
4.5E+01
4.7E+01
8 .Of rt1
5.4E+01
5.8E+Q1
6.5E+01
7.4E+01
8.8E+01
1 JE+02
1.4E*02
1JE+02
2.5E+02
3.5E+02
4 JE+02
7.0E+02
9.9E+02
1.4E+03
1.9E+03
non pH-dapandant inorganic Kd valuas tor antimony, cyanide, and vanadium are 45, 9,9, and 1,000 respectively.
1.0E-01
1.3E-01
1.6E-01
2.1E-01
2.6E-01 •
3.3E-01
4.2E-01
5.3E-0I
6.7E-01
8.4E-01
1.1E+00
1.3E+00
1.7E+00
2.1E+00
2.7E+00
3.4E+00
4.2E+00
5.3E+00
6.6E+00
0.3E+OO
1.0E+01
1.3E+01
1.6E+01
2.0E+01
2.5E+01
3.1E+01
3.9E+01
4 8E+01
5 JE+01
7.3E+0T
B.iE+01
1JE+02
1.BE+01
1 JE+01
1.6E+01
1.5E+01
1.4E+01
1.3E+01
1.2E+01
1.1E+01
1.1E+01
9.8E+00
9.2E+00
B.6E+00
8.0E+00
7.5E+00
7.0E+00
8.5E+00
B.1E+00
5.7E+00
5.3E+00
S JE+00
4.7E+00
43E+00
4.1E+00
3.8E+00
3.5E+00
3 3E+00
3.1E+00
2 9E+00
2.7E+00
2 5E+00
2 4E+00
2 2E+00
4.4E+01
4.5E+01
4.6E+01
4.7E+01
4.8Ef01
5.0E+01
5.1E+01
5.2E+01
5.4E+01
S.iE+01
8.8E+01
5.8E+01
S.iE+01
6.1E+01
6 2E+01
6.4E+01
6.6E+01
S.7E+01
6.9E+01
7 JE+01
7.3E+01
7 4E+01
7 6E+01
7 8E+01
8.0E+01
8.2E+01
8.5E+01
8.7E+0I
8 9E»01
91E+01
9.4E+01
3 8E+01
1.6E+01
1.BE+01
1.9E+01
2.1E+01
2.3E+01
2.5E+01
2.6E+01
2.BE+01
3.0E+0I
3.2E+01
3.4E+01
3.6E+01
3 JE+01
4 2E+01
4.4E+01
4 7E+01
5 JE+01
5.4E+01
5 JE+01
6 2E+01
6 BE+01
7.SE+01
8.3E+01
9.5E+01
1 JE+02
1 3E+02
1 6E+02
1 9E+02
2 4E+02
31E+02
4 0E+02
5 3E+02
-------
ATTACHMENT I
REGION 10 EPA MEMORANDUM ON INORGANIC ARSENIC
Hi ('«, H « 5* MSM fcl VlSi; ft*
-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
REGION 10
1200 Sixth Avenue
Seattle, Washington 98101
March 26, 1997
OEA095
MEMORANDUM
SUBJECT: Inorganic Arsenic in Fish
FROM I Dana Davoli
TO; Karen Keeley
After talking with Roseanne about this, we have decided to
assume that 10% of the arsenic in seafood is inorganic. If use of
this value results in arsenic in seafood (e.g., fish, shellfish,
seaweed) driving a remediation, we would take a closer look at
the site-specific data {e.g., types and amounts of species
consumed, speciated arsenic analyses in seafoods).
The reasons we chose 10V are listed below:
CD A review of the literature (see Attachment 1) shows that the
range of,arsenic in aquatic species ranges from below detection
to as high as about 3.0%.
(2) Although most values are well below 10%, we chose 10% due to
the uncertainties listed below:
(i) Inorganic arsenic data are missing on many species,
including ones that might be expected to have high levels of
inorganic arsenic (e.g., seaweed, bottomfish, shellfish,
crustaceans).
(ii) Although it has always been assumed that methylation of
inorganic arsenic to form dimethyl arsenic (DMA) results in
Reply To
Attn Of:
Pttfissd on R0Cfci«t Paper
-------
detoxification, more recent data have shown that DMA may be a
probable human carcinogen {see Attachment 2}. Data that we have
from a Superfund site in Washington show that the DMA levels are
much higher than the inorganic levels {see attached report). We
have no way of including this potential risk in our risk
assessments because the Agency has not yet developed a potency
factor for DMA,
-------
ATTACHMENT J
WORLD HEALTH ORGANIZATION ABSTRACT
ON DIOXIN TOXIC EQUIVALENCY FACTORS
- hi i-k a * r \n:* pi vise n\*i. mastih *r
-------
WHO Toxic Equivalency Factors (TEFs) for dioxin-like compounds
for humans and wildlife
Introduction
Polychlorinated dibenzodioxins (PCDDs), polychlorinatcd dibenzofurans (PCDFs) and
poiychlorisated biphenyis (PCBs) ate persistent - and toxic - OTrtronmental chemicals. They
enter the food chain and accumulate in fatty tissues of humans and other humans. Several
representatives of these groups of chemicals have been shown to cause similar toxic effects as
2J.7.8-TCDD, the most toxic dioxin. Their health effects include dermal toxicity,
immunotoxicity, reproductive effects and teratogenicity, endocrine effects and
carcinogenicity. These chemicals are also found in human milk, which raises serious health
concerns.
As a result of their different chemical properties, the relative concentrations of the
various PCDD, PCDF and PCB compounds vary from sample to sample. They also differ
front the mixtures originally released into the environment. TMs complex situation hampers
the evaluation of the health risk for humans and for the environment and the establishment of
regulatory control of exposure to mixtures of these compounds.
TEF concept
The concept of toxic equivalency factors (TEFs) has been developed to deal with thi s
problem. The TEF concept is based on the evidence thai dioxin-like compounds share a
common mechanism of action - binding to the Ah-receptor. By applying this TEF concept,
the toxicity of the different compounds relative to that of23,7,8-TCDD is determined on the
basis of in vivo and in vitoo data.
Several TEF schemes have been developed for PCDDs and PCDFs and for dioxin-like
PCBs. Recognizing the need for a harmonized approach in setting internationally agreed
TEFs, the WHO European Centre for Environment and Health aid the International
Programme on Chemical Safety (IPCS) have initiated a programme to derive consensus TEFs
for compounds with dioxin-like activity for assessing the impact of these compounds on
human health. The first Consultation on the Derivation of TEFs for Dioxin-like PCBs was
convened in December 1993. For this meeting data on the relative toxicity of dioxin-like
PCBs for mammalian specie were collected and criteria for deriving TEFs were established.
These data were altered into a database set 151 by the Karoimska Institute in
Stockholm, Sweden, and evaluated with respect to the applicability for the derivation of
TEFs. As a result of this process, consensus TEFs for human intake were derived for 13
different PCBs (1). In addition, it became apparent that the database should be extended to
include PCDDs and PCDFs, as well as data on the relative toxicity of dioxin-like compounds
for wildlife.
O
-------
THU 17:08 FAX 2187205538
mid-cont ecol div duuth
Risks for wildlife?
One of the open questions was whether or not the TEFs derived for human risk assessment
purposes were appropriate for estimating the risk for wildlife, or whether a separate set of
TEFs for wildlife should be developed. This was discussed at an initial WHO consultation on
the derivation of TEFs for wildlife for PCBi, PCDDs, PCDFs and other dioxin-like
compounds, convened in August 1996. This WHO meeting identified the type of dam
necessary for the derivation of TEFs For wildlife, and defined a workplan and the way data
should be collected. It decided to combine the efforts to derive TEFs for wildlife with the
update of existing TEFs for human risk assessment Furthermore, it recommended that TEFs
for human health and wildlife be harmonized to the extent possible.
Derivation of TEFs
Many scientific articles on PCDDs, PCDFs and PCBs were analysed, of which 185 fulfilled
the selection criteria. Based on the information available in these articles about 1600 sets of
information have been inserted into the database. Following collection of all available
information, a WHO meeting on the derivation of toxic equivalency factors (TEFs) for PCBs,
PCDDs PCDFs and other dioxin-like compounds for humans and wildlife was held at the
^tute of Environmental Medicine of the Karolinska Institute in June 1997. The meeting
evaluated the information in the database and discussed several general issues related to the
TEF concept.
The term TEF was defined to bean order of magnitude estimate of the toxicity of a
compound relative to the toxicity ofTCDD that is derived using careful scientific judgement
after considering all available data. The relative potency of a compound obiainedina single
in vivo or in vitro study will be referred to as a relative potency (REP) value. TEFs, in
combination with chemical residue data, can be used to calculate toxic equivalent (TEQ)
concentrations in various media, including animal tissues, soil, sediment and water. TEQ
concentrations in samples containing PCDDs, PCDFs and PCBs are calculated using the
following equation:
TEQ - ([PCDDi * TEFjlxj) + ([FCDFj * TEPjJn) + ([PCBt * TEPjJn)
Substantial evidence indicated that the TEF approach is equally valid for human risk
assessment as for wildlife, although wildlife risk assessments usually attempt to estimate
population-level effects (unlike traditional human risk assessments, which focus on protecting
individuals) because effects on populations are of greater ecological relevance than are effects
on individuals. The criteria used for including a compound in a wildlife TEF scheme are the
same as those used for human TEFs. Compounds must;
« show a structural relationship to the PCDDs and PCDFs
» bind to the Ah receptor
• elicit dioxin-specific biochemical and toxic responses
• be persistent and^ccumulate in the food chain.
A i 1-- In ¦ft-*-* «ikitr«»4tnf in rraH»r<» or hsVP
-------
UftMU4/4M K.UO fAA
MIH-UJM fcLUi- l/i * uuiAUtt
uy 4
New WHO TEF scheme
Based on the available information .both the previously established TEFs for PCDDs and
PCDFs (2) and the WHO TEFs forPCBs (1) for human risk assessment were re-evahuted.
for revision of the existing TEFs for PCDDs, PCDFs and PCBs It was agreed by the working
group that if the available information was considered insufficient to warrant a change, the
existing value would be adopted.
In deriving TEFs for wild mammals it was concluded that there was insufficient
evidence to discriminate between laboratory and wild mammalian species, and it was therefore
decided that the TEFs for human risk assessment based on laboratory animals would be equally
applicable for wild mammal species.
The relative potency factors were primarily taken from in mm toxicity data, which were
given more weight than In vitro and/or QSAR data. In vivo toxicity data were prioritized
according to the ranking scheme chronic > subchronic > subacute > acute. In the final TEF
selection different Ah-receptor specific endpoints were also ranked according to toxic >
biochemical (e.g. enzyme induction) response.
Also for the derivation of TEFs for PCDDs, PCDFs, and coplanar and mono-ortho
PCBs for fish, and birds a tieredapp roach was followed that gives a higher weight to overt
toxicity in vitro studies than to biochemical effects. An even lower weight was given to
biochemical effects (enzyme induction) in vitro. In case none of these data, were available an
estimate was made based on quantitative structure-activity relationships (QSAR).
When comparing the final TEF value across differmt taxa the wotting group tried to
harmonize the TEFs to the extent possible, as this would have a clear advantage from a risk
assessment and management perspective. However total synchronisation of TEFs between
mammals, birds and fish was considered to be not feasible in case of obvious indications of
orders of a magnitude difference between the taxa.
In line with the already existing TEF values new TEFs were rounded to a value of
either lor 5, irrespective of the order of magnitude difference with the reference compound,
TCDD. It is important to point out that in this rounding procedure a conservative approach
has been chosen to provide optimal protection of humans and wildlife.
1. Ahlborg et al.(1994), Toxic equivalency factors for dioxin-like PCBs: Report on a WHO-
ECEH and IPCS consultation, December 1993. Chemosphere, Vol 28, No. 6, 1049-1067.
2. NATO (1983), International toxicity equivalency factor (I-TEF) method of risk assessment
for complex mixtures of dtoxins and related compounds. Report No. 176, Brussels, Belgium.
For more information please contact F.XJRolaf van Leeuwen, WHO European Centre for
Environment and Health, P.O. Box 10, NL-3730 AA De Bilt, tel. 31 30 2295 307, fax 31 30
2294 252, email t^@who.nl
-------
08/04/87 TOT 17:88 FAI 2X8720553#
WD-CONT ECOL DIV DULUTH
0005
Table I. WHO-TEFs for humans, mammals, fish and birds
CONGENER
TOXIC EQUTTALSfCY FACTOR (TB5)
HUMANS/
MAMMALS
FISH =
BIRDS «
Zj,7,8-TCDD
1.2.3.7.8-PeCDD
i,2,3,4,7,S-HxCDD
t»2J,6»7J-H*CDD
l,Z,3J,l,9-HxCBD
1,2,3,4,6,7,8-HpCDD
OCDD
I
t
0,1
0.1
0,1
0.0 i
0.0001
1
I
0.5
0.01
0.01 «
0.001
1
1 f
0.05 r
0.01 r
0,1 t
<0.001 r
2,3,7,S-TCDF
0.1
0.05
t
r
!.2J,7,S-PeCDF
0.05
0.05
0.1
f
2,3,4,7,8-PeCDF
0.5
05
1
r
1,2.3,4,7,8-HxCDF
0.1
0.1
0.1
tj
1,2,3,6,7.8-HxCDF
0.1
0.1 e
0.1
cj
UJ,7,g,9-HxCDF
0.1 »
0.1 *•«
0.1
c
2,3.4,6.7.8-HxCDF
0.1 *
0.1 =
0.1
e
1,2 J ,4,6,7,8-BpCPF
0.01 •
0.01 k
0.01
b
1,2,3.4.7.8.9-HpCDF
0.01 »
0.01 M
0.01
h
OCDf
0.0001 *
0.0001 fc.»
O.OOOI
b
3,4,4',5-TCB (81)'
0.0001 *>.«•«
0.0005
0.1
€
JXIC-ICB (77)
0.0001
0.0001
0.05
3,3',4,4',5-PeCB (126)
0.1
0.005
0.1
3,3,,4,#,5,S*-HxCl (169)
0.01
0.00005
0.001
W.4.4'-PeCB (105)
0.0001
<0.000005
0.0001
2,3,4,4',5-PeCB (114)
0.0005
<0.000005 k
0.0001
I
2,3',4,4',5-PcCB (It8)
0.0001
<0.000005
0.00001
2',3,4,4',5-PeCB (123)
0,0001 w»
<0.000005 h
0.00001
s
2,J,3',4,4',5-H*CB (156)
0.0005 b*
<0.000005
0.0001
2f3t3,A4,»5*.ffeCB (15?)
0.0005
<0.000005 fci
O.OOOI
2,3',4.4',5,5'-HxCB (167)
0,00001
<0.000005 *
§.00001
g
IJ^VVAS'-HpCB (189)
0.0001 3*
<0.000005
0,00001
6
indieat bo TEF became of lick of dau
a) Itmkcd iiu ttt
b) tcnicwnl timiljfity
c) QSAR n»4el!iog prediction from CVPIA iniwefion (monk«y. pig, chicken, or fiih}
d) m «e« Can from IWJ review
e) m viiro CYPIA induction
0 in >t*o CYP IA induction after m 0*9 expOlwe
J) QSAR modelling prediction from clu> jpcdfio TEFj
o
-------
ATTACHMENT K
EPA REGION 1 RISK UPDATES,
REVISED MANGANESE REFERENCE DOSE
* T*SM WISE: FIN 41 .MASTER WPO m-nrnm^
-------
&EPA
Region 1, New England
RISK
Number 4
R,
kJSK UPDATES is a periodic
bulletin prepared by EPA - Region I,
New Englanrt risk assessors to provide
Information on new regional guidance
Risk Updates is distributed to contract ore
supporting Syperfund and RCRA.
regulators, and interested parties. Risk
assessment questions may be directed
to the following EPA scientists (area
code 817):
Regional Risk Assessment Contact
Ann-Marie Burke 223-5S28
Superfund
Human Health Risk Assessment
Ann-Marie Burice 223-6528
Sarah levinson 573-9614
Margaret McDonough 573-5714
Jayne Michaud 223-SS83
Ecological Risk Assessment
Susan Svirsky 573-9849
Patti Tyler 860-4342
RCRA Corrective Action
Mary Ballew
573-5718
Stephanie Can-
223-5593
Air Modeling
Brian Hennessey
565-3572
Combustion Risk Issues
Jui-Yu Hsieh
565-3501
Comparative Risk
Katrina Kipp
5(55-3520
Cost Benefit Analysis
Ronnie Levin
585-3351
Drinking Water
Maureen McClelland
565-3543
Air Risk Issues
Jerri Weiss
565-3448
ORO Technical Liaison
Ruth ileyter
573-5792
EPA Region I, New England receives
additional ecological technical support
from Ken Finkelstein (223-5537) of the
National Oceanic Atmospheric
Administration (NOAA), and US Fish &
Wildlife (Steve Mierzykowski 207/827-
5338. Ken Murmey 803/225-1411, and
Tim Prior 401/364-3124).
Contents
EPA Ecological Guidance
Update™ Page 1
Superfund Risk Assessment
Reform Initiatives. Page 2
Risk Characterization
Update. Page 3
| Mercury Updata Page 3
EPA Finalizes Soil Screening
Guidance Page 3
EPA's Proposed Cancer Guidance
and Implementation Plaru„Pa§«s 4
Ground Water Use and Value
Guidance. .....Page S
Lead Risk at CERCLA Sites and
RCRA Correction Action
Facilities. Page 6
Revised Manganese Reference
Dose.— Page 8
New Cancer Slope Factors for
PCB's. Page 8
Editors
Stephanie Carr ana Jayne Michaud
Layout Glona Hums
I
-------
Revised
Manganese
Reference Dose
The manganese reference
dose (RID) in the IRIS data
base was revised in
November, 1995, This
revision results in a lower risk
(and thus, higher cleanup
level) for drinking water
compared to the previous RfD.
The IRIS RfD of 1.4E-1
mg/kg/day is for the total oral
intake of manganese. As
stated in the IRIS file, it is
recommended that a modifying
factor of 3 be applied to the
RfD for non-dietary exposures.
Background
Prior to November, 1995 the
IRIS data base provided two
references doses for
manganese, one for food and
one for water, The food RID
was based on dietary intake of
manganese. The water RfD
was based on a study of
humans who had ingested
drinking water containing
elevated levels of manganese
as well as on assumptions
regarding differences in
absorption of manganese in
food as opposed to water.
The drinking water RfD was
withdrawn from IRIS in
November, 1335 because of
concerns about the validity of
the human exposure study and
because new information
indicated that the disparity
between absorption of
manganese from food as
opposed to water was
overestimated.
New Approach
The revised RfD for
manganese is for the total oral
intake of manganese. This
value is 0,14 mg/kg/day and is
derived as follows:
• 10 mg/day of manganese
may be consumed without
adverse effects (the
"critical dose"). This value
comes from several dietary
studies.
average adult body
weight = 70 kg
Therefore, the RfD =
10 ma/dav "0.14 mg/kg/day
70 kg
Soil Exposure
A modifying factor of 3 may be
appropriate for assessing risks
via exposure to soils if neonates
fa child 12 months or younger)
are a potentially exposed
population. For most RCRA arid
Superfund risk assessments
neonates are unikeiy to be
exposed to significant amounts
of soils. Therefore, a modifying
factor of 1 is appropriate.
Assuming exposure to a young
child under a residential
scenario,, a hazard index of 1 for
manganese in soil would
correspond to,—^a soil
concentration o07s§O^ig/kg,
wattm by Mtrgaret McDonough
A modifying factor of 3 is
recommended in IRIS when
assessing exposure from
drinking water.
Drinking Water Exposures
The average dietary
manganese content of the
U.S. population, 5 mg/day, is
subtracted from the "critical
dose" of 10 mg/day:
10 mg/day - 5mg/day =
' S mg/day
Apply modifying factor of 3 per
IRIS recommendation;
SuSSiM'
3
1,87 mg/day
Compute RfD:
1.67 mg/day = ,024 mg/kg day
70 kg
The Hazard Index (HI) for
chinking water is calculated as
follows (using a simplified
equation):
Concentrationfmg/L}' 2lrtefsAfoy
0.024 mg/kg/day * 70 kg
A HI of 1 corresponds to a
concentration of 840 ug/L.
8
-------
ATTACHMENT L
AN APPROACH FOR DETERMINING TOXICITY VALUES
FOR DERMAL EXPOSURE,
OAK RIDGE NATIONAL LABORATORY INTERNAL PAPER
t V*I. VMnYIR 'APf> .
-------
ATTACHMENT L
AN APPROACH FOR DETERMINING TOXICITY VALUES
FOR DERMAL EXPOSURE,
OAK RIDGE NATIONAL LABORATORY INTERNAL PAPER
V RFPtJU'* T\SM HCMSEi FtNAL MAStER Wpfr Hl-ft
-------
AN APPROACH FOR DETERMINING TOXICITY VALUES FOE DERMAL EXPOSURE.
C. B. Bast and H. T. Borges. Biomedical and Environmental Information Analysis. Health
Sciences Research Division. Oak Ridge National Laboratory', Oak Ridge, TN.
"Managed by Martin Marietta Energy Systems, Inc., for the U.S. Department of Energy under
contract No. DE-A005-840R2I400.
-------
ABSTRACT
Oral-toxicity data are available for many chemicals allowing for the calculation of Oral Reference
Daces (RfD) for nndiiogeiiic effects and slops factor* (%") for carcinogenic effects. In contrast,
derma) toxicity data for touf-tem atpam are not arable for most chanfcafc wbkb precludes
caJculalion of tanil RID> and slope imMm. HeaNi risb &®n» dennal q»
-------
INTRODUCTION
Oral toxicity
-------
METHOD
1, Mfmttefatinm nf Pnbnntial 4emM Effects. A first step in denial risk assessment is to review dermal
toxicity of tic compound and determine if it causes point of entej effects (direct skin effects). For
wattiwifc strong adds awl muse direct sMa destruction, and metcuiy, chromium, and lead can
cause *Kn irritation al «efai«efj low concenteaioos. Ewen. If the amount of chemical Mwotwi in
rfermal closure k Miai compared to lie amount inhaW or ingested, dennal tooric% may be
important If mite sfcai eflfceft esnt. Abo,, when aqppiyioqg the following risk equations. It is important
#
to evaluate the risk value obtained in nefenoaoe to the contribution of contact site toxicity.
2 Identification of Gastrointestinal Absorption Facton. Absorption data were obtained through on-
line literature seareies of National library of Medicine (NLM) databases and from hardcopy sources.
Seoonclafj sources such * monographs, surveys, review articles, and criteria documents were mm! to
obtain gwtrointolmal absorption data when possible. However, if no absorption data were present
in the seooniaiy sources or if the absorjJtiQO data woe not dearly presented, primary publications
were consulted.
Gasteolntesiitat (GI) ahsoiptioi values calculated and reported in the literature wet© utflfaed
when the methods employed appeared to be scientifically found. In caw where no absorption
facton were reported or where the factors appeared to be ieiwei by inappropriate methods,
attempts mete made to estimate abmption facton from the published absorption and ocretin data.
In these instancy absorption mm estimated by adding amounts of omprnni seoowefei in aO
reported oqjan* (aduding the luminal cooteote of the stomach and intestines) at necropsy and/or
from amounts irawiwi from lamping of plasma, saliva, urine, awl taeatk Jto many cases it wis
possible to estimate only a tower limit of absorption since it could not be determined if OTmponnd
detected fa the feces wm actually absorbed, In tine ma, abmptnii m leportoi as "greater thai
or equal to" a gjneo value.
-------
M some eves qualitative desoiptioiK of gastrointestinal abKxptioii, such, * absorbed"
or poodf abeorttcd** arc available. When comparing quaitaiwe mi quantitative GI afasuption data,
lie tew rare, little, sparse, and tow" tend to refer to absorptions between 1-20%, "readily and
rapidy to ihsorplioiii between 20-90%, and Vefl and almost complete" between 70-100%.
terms such as fead% and rapMfy may refer to rate rather than amount of absorption. It k difficult to
dmse « qualitative absorption ranking system bated on the to* mi mmmaiy found fa the
He feltawng ipl® is suggested wfa „ quntiwiw data but qntfufe. data are
present:
Negligibly Absorbed; 80%
If no quriufe or qualitative data «rc avribbfe fa specific cbcnucals, an attempt was
made to estimate absoipioii fiicton hj sbuctunl mak®-.
3. S^chon of Gasavauemnal Absorption Faaocfor Use in Dermal RukAwsmtm. The following
scheme was developed to select OI absorption [acton for toe in dermal risk assessment:
For chemicals with (pintititwc absorption H»t»»
^ ^ ^ ^ ^ ^ taanan »W" (UA
Human. > Non-human Primate, Pfe > Rat, Guinea Tig > Mouse, Rabbit
2) Sefcrt the mart eoMewatwe ataxpfion value fount -r-m m ¦ i_ _• v ^
¦ppopkte ipecw. (When the only value
to" ifwaitaijtim, the value ftself was selected.) peater or equal
fot diamciii with only qualitative absorption
1) UtHne the qualitative system previously presented.
^tSXtSPrDP,kte •te0rPd°n ^ °f the
-------
4, Cmwmiimfimm jUmmatimxt to Absorbed Dose. GI afamptioit fisdon for 135 dmmcals of
intetest at DOF* Oak RMge Jtearwtioii wax MeniBed in the Mteralmc and placed into a dermal
risk «fatah»w» Hie CAS nunitws of the* dmicA were compared electronically with the CAS
numtes of a dmtabase containing the RfD sod Slope Factom of ewer 600 hundred chemicals. When
an cm* match w* found, the KJED and Slope Factor wm Inserted Into the dermal risk database and
converted to toxicity values bawd an aborted imed 6* dennal ev»re * fetows (US. EPA, 1992;
US. EPA, 1969):
MD.^ = PfT> . . .x ABS-.
= %^«^/ABSai
-------
RESULTS
Using this oietiwfcicw. MD» ranging from 3JOOE4B to 4JOE+0O mgAgtiay and slope factors
ranging bom 7MEM to 43©B+®2 (mg/kg/day)1 hme been ifawwl for over 60 chemicals. The
computer-generated table provides m consistent mathematical metW for ofculatiiig dermal toxicity
values and is presented below.
-------
US.-EPA. 1®. R»fc Assessment Guidance for Supcrfiiad, Volume L Human Health Evaluation
Manual (Part A). Interim RnaL Office of Emageney and jRjciiicscliflil IBLcspcMiisc^ AAfsshinjE^to^D^ DCZ1*
December, 1969*
U.S.EPA. 1992. Dermal Exposure Assessment: Principles and Applications. Interim Report. Office
of Health 'gnd Pnvwnnwwitil Washington, DC, Januaiy, 1992. HPA/6OW6-91/011B.
"Rgfemtco from wbMt gastrointestinal absorption *w obtamed am EMed at the md of the
cmnpmio'igmBttted laMe.
-------
DERMAL RISK VALUES DERIVED BY CALCULATION
FROM GASTROINTESTINAL (Gl) ABSORPTION DATA IN ALPHABETICAL ORDER
Chamlcai
CAS Nombw
Gl
Absorption
Factor l*l*
Oi
Absorption
Riltnno*
Oral RfO |mg/kg/dayl
Chronic Bub chronic
Oral Slop*
Factor
(mg/kg/day) 1
Dermal RfO tmg/kg/dayl
Chronic Subchronic
D«m«l Slop#
Factor
ting/kg/dayl'*
Actnftphthen«
000083 32 9
31
2
b
8 OOE 02
8 OOE 01 °
NA
1.88E 02
1 88E 01
NA
Acenaphthytar>«
000208 98 8
31
2
NA
NA
NA
NA
NA
NA
Acelom
000087 64 1
83
3
1 OOE 01 b
1.00E + 00C
NA
8 30E 02
8 30E 01
NA
Akhin
000309-00-2
50
65
3.00C 06b
3 OOE OS C
1.70E + 01 b
1.50E-05
1 60E 06
3.40E ~ 01
Aluminum
007429-90-5
10
4.5
NA
NA
NA
NA
NA
NA
Anthracene
000120-1 2-7
78
6
3 00E-01 b
3.OOE * 00 C
NA
2.28E 01
2.28E tOO
NA
Antimony (metallic)
007440 38-0
2.0
7
4.00€ 04b
4 OOE 04 C
NA
8.OOE-08
8 OOE 08
NA
Aroclor 1010
011674-11-2
90
46
7 00E 06b
NA
NA
S.3OE-0S
NA
NA
Aroclor 1 234.
011097-69-1
90
48
2.00E-05 b _
5 OOE 06 C
NA
1 BOE 06
4.50E-05
NA
Aroclor 1 260
Ot 1096 82 5
80
46
NA
NA
NA
NA
NA
NA
Araanic Still
NA
80
20
NA
NA
NA
NA
NA
NA
Araanlc. Inorganic
007440 38 2
<33>
8
3.00E-O4 b
3 00EO4 C
NA
1.23E-04
1 23E-04
NA
Barium
007440-39-3
7.0
9
7.00C-02 b
7.00E-02 C
NA
4 90E 03
4 90C 03
NA
Baiulalanttuacana
000058 55 3
31
2
NA
NA
NA
NA
NA
000071 -43-2
97
10
NA
NA
NA
NA
2 99E-0J
Bant ana Haxachlorida
NA
•7
10
NA
NA
NA
NA
NA
NA
Bant ana. EthyMimethyl
NA
9?
10
NA
NA
NA
NA
NA
" NA
Baruana. Ethylmathyl
NA
97
10
NA
NA
NA
• NA
NA
NA
Baruana. Mathylpropanyl
NA
97
10
NA
NA
NA
NA
NA
NA
Baruana. Mathylpropyl
NA
97
10
NA
NA
NA
NA
NA
NA
Baruana. Trtmalhyl
025551-13-1
97
10
NA
NA
NA
NA
NA
MA
Benzidine
OO0O92-87-5
80
11
3.00E-03 b
3 OOE 03 C
2.30E »02 b
2 40E 03
7 40E 03
2 OBE « 02
Banzolalpyrana
000050-32-8
31
2
NA
NA
7 .30E ~ 00 b
NA
NA
2 J5E t 01
Beniolbllluoranthana
000205-99-2
31
2
NA
NA
NA
NA
NA
NA
Banio(a.h.ilp«rylan«
000191-24-2
31
2
NA
NA
NA
NA
NA
NA
Baniolkl'luorantHene
000207 08 9
31
2
NA
NA
NA
NA
NA
NA
OI/OS/85
The dermal values in this labia have no! undergone poor review
f'sg* 1 I
-------
¦ DERMAL RISK VALUES DERIVED BY CALCULATION
FROM GASTROINTESTINAL (Gl) ABSORPTION DATA IN ALPHABETICAL ORDER
Gl
Abiorption
Gt
Absorption
Oral RID (mg/kg/dayl
Oral Slop#
Factor
Dermal RfO (mg/kg/day)
Dermal Slop*
factor
Chemical
CAS Number
Factor 1*1®
Reference
Chronic
Subchronlc
ImB/kg/dayl
1
Chronic
Subchronlc
ImgAit/dayl'
Benzoic Acid
000086 85 0
100
12
b
4.00E+00
4.OOE -1-00 C
NA
.4.OOE+ 00
4.00E+00
NA
Benzyl Alcohol
000100-51-6
60
12
3 OOE 01 °
1 OOE ~ 00 C
NA
1906 01
6 80E01
NA
BarylHym
007440-41-?
1.0
13. 14
5.OOE 03
5 OOE 03 C
b
4.30E+00
S.MJE-OS
5 OOE-06
4.306 * 02
Bta( 2-athythexyOphthalale •,
000117 81 7
19
IB
2.00E-02
2 OOE 02
1.406-02 b
3.006-03
3.80E 03
T 37E02
Boron And Borates Only
007440-42-8
90
75
1 OOE 02
• OOE 02 C
NA
8.106 02
8.106-02
NA
Bcomodtchloromelharw
000075 27 4
98
87
2.00E 02
2 OOE 02 C
6.206 02 b
1 96E 02
1 >6E 02
8 33E 02
Bromoforrn
000076 25 2
60
69
2 OOE 02
2 OOE 01 C
7.90E 03b
1.206 02
1.20E-01
1 32C 02
ButafKHie-2. 4 cMoro 4.4 diflimro
NA
80
1
NA
NA
NA
NA
NA
NA
Butyl B»niyl Phthlate
000085 68 7
61
78
2 OOE 01
2.00C + 00 C
NA
1 22E 01
1 22E tOO
NA
Cadmttim |OI®tl
007440-43-9
(B>
18. 17. IB. 19
b
1 OOE 03
NA
NA
1 OOE 05
NA
NA
Cidmkffn IWtlvr!
007440-43-9
1.0
16. 17. 18. 19
5 OOE 04
NA
NA
6.006 06
NA
NA
Cerbazole
00006&74 8
70 "
78
NA
NA
2.00E 02
NA
NA
2.B66 02
Carbon Dfauiikto
000075 15 0
63
20
1 OOE-OI
1 OOE 01 C
NA
6.30E 02
. 8 30E 02
NA
Carbon Tetrachloride
000068 23-5
65
21
1.00E-04
7.00C-03
b
1.30E-01
3S
4.55E-04
4.5SE-03
2 . OOF 01
Chlordana
000057-74-9
50
22
b
e.ooE-os
6 OOE 05
i.aoE+oo b
3 OOE-05
3.006-05
2 80E +00
Chtoro benzene
oooioe-*o-7
31
23
2 OOE 02
2 OOE 01
NA
6 206 03
6.20E-02
NA
Chloroform
000)67-66-3
20
24
1.OOE 02
1.OOE 02
6.10E-03 b
2 OOE 03
2.00E 03
3 05E 02
Chromium {till ftnsoiubta Salt*)
01B065-83-1
0.50 -
74
1.OOE+00
1 .OOE+ 00 €
NA
5. OOE 03
5 OOE 03
NA
Chromium (VII
018540-29-9
2.0
25
5.00E 03 '
2 OOE 02
NA
1 OOE 04
4 OOE 04
NA
Chromium Salt*
NA
2.0
74
NA
NA
NA
NA
NA
NA
Chryaene
00021B01 9
31
2
NA
NA
NA
NA
NA
NA
Cobalt
007440 484
• BO
26
NA
NA
NA
NA
NA
MA ¦
Copper
007440-60-B
30
27
NA
NA
NA
NA
NA
NA
CtBiol. p-
000106 44 6
65
72
5 OOE 03C
5 OOf 03 °
NA
3 25E-03
3 25E 03
NA
Cyanide (CN 1
000057-12-5
17
37
2 OOE 02 b
2 OOE 02 C
NA
3 40E 03
3 401 03
NA
ODD
000072 54 B
70
28
NA
NA
2 40E 01
NA
NA
3 4.1F 01
02/09/36
The dermal values
table have not undergone peer review
Puge 7 2
-------
DERMAL RISK VALUES DERIVED BY CALCULATION
FROM GASTROINTESTINAL (Gl) ABSORPTION DATA IN ALPHABETICAL ORDER
01
Absorption
Gl
Absorption
Oral RID Img/kg/dayt
Oral Slops
Factor
Ditmil RID (mg/kg/day)
Darmal Slop*
Fact of
Chemical
CAS Number
factor (%(•
Rafaranca
Chronic
Subchronic
(mg/kg/day)'
Chronic
Sub chronic
(mg/kg/day)'
DDE
000072 66 S
70
28
NA
NA
b
3.40E-O1
NA
NA
4 68E-01
DOT
000050 29 3
70
28
5 OOE 04 b
5 OOE 04 C
3.40E-01 b
3.50E-04
3 50E 04
4.80E-OI
Dibanila.hlanthiacana
000053 70 3
31
2
NA
NA
NA
NA
NA
NA
Dibrainochtoromathane
0001 24-48-1
60
89
2.00E 02 b
2.00E 01 C
8 40E 02 b
1.20E 02
1 20E 01
1 40E 01
Dibutyl Phthalata
000064 74 2
100
29
b
1 OOE 01
1 OOE+ 00 C
NA
1. OOE 01
NA
NA
Dichtorobanzena. 1.4
000106-46-7
SO
82
NA
NA
2.40E 02
NA
NA
2 87E-02
Dkhtorodifluoromolbane
000075-71-8
23
12
2 OOE 01 C
9 OOE 01 C
NA
4 60E 02
2 07E 01
NA
Dichloraathana. 1.1-
000075-34-3
100
BO
1 OOE 01 6
1 OOE 4 00 C
NA
1.OOE-01
NA
NA
Ofchloroathana. 1.2-
000107 06-2
100
30
NA
NA
9.10E-02 b
NA
NA
9 10E 02
Dichkwoathylana. 1.1-
000075-35-4
100
31
9 OOE 03 b
• OOE 03 C
8 OOE 01 b
9 OOE 03
9 OOE 03
8 OOE 01
Dichlan>phaiiol. 2.4-
000120 83 2
82
81
3 00C 03b
3 OOE 03 C
NA
2 46E 03
1 4BE 03
NA
DtcMoroprepana, 1.2-
000078 87 5
74
88
NA
b
3 OOf. 04
NA
6 BOE 02
NA
NA
9 19E-02
DicMoropfopana. 1.3-
000542 75 6
66
73
3 OOE 03 C
1.80E 01
1.60E + 01 b
1 65E 04
1.65E 03
3 27E-01
OtoWrin
000060 57 1
50
65
6.00E-05 b
5 OOE 05 C
2 50E OS
2 50E 05
3 20E 4 01
Dlilhyl Phthalaia
000084-68-2
90
79
b
8.O0E-0I
b.ooe+oo c
NA
7 20E 01
7 20f »00
NA
Dimalhylph'thalata
000131-11-3
90
7»
NA
NA
NA
NA
NA
NA
DMlra-o craiol. 4.6-
000534-52-1
100
12
NA
NA
NA
NA
NA
NA
Dtnttrobatuana. 1.2-
000528 29 0
93
32
4.00E-04 °
4 OOE-03 C
NA
b.d
6 80E 01
b.d
6 BOE 01
3 72E 04
3 J2E-03
NA
Dtnitrolotuane, 2.4-
000121 14 2
85
33
2.00€ 03b •
2 OOE-03 C
1 70E 03
1 /OE 03
8 OOE-01
Dioilrotoluene. 2.6-
000606-20-2
85
33
1.00E 03C
1 OOE 02 C
B.50E-04
8 50E-03
8 OOE 01
Endrin
000012-20-B
2.0
77
3.00E-04
3 OOE 04 C
NA
6. OOE-06
6 OOE-00
NA
Ethylbanzana
000100-4 1 4
97
10
1 OOE 01
NA
NA
9 70E 02
NA
NA
Fluor anlhana
000206-44-0
31
2
4 OOE 02
4 OOE 01 C
NA
1 24E 02
1 24f 01
NA
Fluorida
007782 41-4
97
34
6 OOE 02
6 OOE 02 C
NA
5 B2E 02
5 B2F 01
NA
Maplachtor
000076 44 B
72
35
5 OOE 04
5 TOE-04 °
4 BOE * OO h
3 60E 04
3 60E 04
6 25E < 00
Haptachfor t povide
001024 57 3
72
35
b
1.30E OS
t 30E 05 C
9 «OE »(X) h
9 If if 06
i 1M on
1 71% <01
Ths dermoi values in this labia have not underflow ponr rewiaw P»o® 1 3
-------
DERMAL RISK VALUES DERIVED BY CALCULATION
FROM GASTROINTESTINAL {Gil ABSORPTION DATA IN ALPHABETICAL ORDER
<31
Absorption
Qi
Absorption
Oral RID (mg/kg/ilay|
Oral Slope
Factor
Dermal RID (mgfltg/dayl
Dermal Slop*
Factor
Chamlcai
CAS Number
Factor |%|«
Raleranca
Chronic
Subchronic
Img/kg/dayl 1
Chronic
Subchronic
Img/kg/dayl"'
Hsxachtorocyclohexane. Alpha
000319 64 6
97
70
NA
NA
b
6.3OE+00
NA
NA
8 49E +00
Haxachtorocyclohexane. Bat®-
000319 85 7 ¦
91
70
NA
NA
I.BOE-fOO1*
NA
NA
1.9BE+00
Hexachlorocyclohaxane. Gammi-
000068 89 9
97
10
3.00E-04 b
3.00E-03 C
1.30E + 00
2.91E-04
2 91E 03
1 34E +00
Kssmone, 2-
000591-78 0
.88
36
NA
NA
NA
NA
NA
NA
Indenol 1,2.3-cdlpyrene
000193-39-6
31
1
NA
NA
NA
NA
NA
NA
lion
007439 8V 0
15
26
NA
NA
NA
NA
NA
NA
laopropanol
000067 03 0
100
12
NA
NA
NA
NA
NA
NA
lead And Compounds
007439-92-1
15
26
NA
NA
NA
NA
NA
NA
LUhlum
0O7439 93-2 '
80
20
NA
NA
NA
NA
NA
NA
bl»f
007439-95-4
20
26
NA
NA
NA
NA
NA
NA
Manganeae (Diet)
007439 90 5
4.0
38
1.40E-01 b
1 40E 01 6
NA
S.60E 03
6 60t 03
NA
Martgeneee (Water!
007439-98-8
4.0
38
5.006 03 b
5 OOE 03 C
NA
2.00E-O4
2 OOE 04
NA
Mercury. Inorganic
007439 97 6
0.01
20. 39
3.00E-04C
3 OOE 04 C
NA
3 OOE 00
3 OOE OB
NA
Methyl Ethyl Ketone
000078 S3 3
BO
1
6.00E-01 b
2.OOE + 00 C
NA
4 BOE 01
1.60E +00
NA
Methyl Mercury
022967 92 0
90
26, 39. 40
3.O0E-O4 b
3 OOE 04 C
NA
2.70C-04
7 70f 04
NA
Methylene Chloride
000075-09 2
95
41
• O.OOt 02b
O.OOE 02 C
7 60E-03 b
S.70E 02
5 70E 02
7 89E 03
Molybdenum
007439 98 7
38
42
5.00C-03 b
5 OOE 03 °
NA
1 90C 03
1 90E 03
NA
Naphthalene
000091 20 3
80
43
NA
NA
NA
NA
NA
NA
Naphthalene, t Methyl
000090 12 0
BO
43
NA
NA
NA
NA
NA
NA
Naphthalene. 2-Methyl
000091 57 6
80
43
NA
2.00E-02 b
NA
NA
NA
NA
NA
Nickel Soluble Sails
007440 02-0
27
44
2.00E 02 °
NA
5 40E 03
5 40E 03
NA
Nitrobenzene
000098 95 -3
87
10
5.00E-04 b
6 OOE 03 C
NA
4.BSE 04
4 B5E-03
NA
NHrophenol. 4-
000100 02 7
100
12
NA
NA
NA
NA
NA
NA
NMroao «H-N propylamine. N-
000621 64 7
28
45
NA
NA
7 . OOE ~ 00 b
NA
NA
2 eoE + 01
Nttroaodipbenylamtne. N-
000086 30 6
25
45
NA
NA
4 90E 03 b
NA
NA
1 96E 02
Octyi Phthalate. d< N
000117-B4-0
90
78
2 OOE 02C
2.0OE-02 C
NA
1 SOE 02
1 BOE 0?
NA
02/09/95
Th« ctarmal values • . able have no! undergone peer review
P*d* 7.4
-------
DERMAL RISK VALUES DERIVED BY CALCULATION
FROM GASTROINTESTINAL (Gt) ABSORPTION DATA IN ALPHABETICAL ORDER
Ql
Ab«ct»ption
Factor 1*1 ¦
G(
Absorption
R#fWMC*
Oral RID (mfi/fcg/day)
Oral Stop*
Factor
Darmat RID {mo/kg/dayl
Deranai Slop*
Factor
Chemical
CAS Numbar
Chronic
Subchronlc
(mgAa/day)''
Chronic
Subchrontc
tmeAgMay}
P«nt»chtofophanol
000087 86 5
100
71
b
3 006 02
3.006 02 °
b
1.20E-01
3.006 02
3 006 02
1 206 01
Penlyl Alcohol. N •
000071 41 0
50
12
NA
NA
NA
NA
NA
NA
Pb#n»rttbr«n»
000065 01 a
73
2
NA
b
6 006 01
NA
NA
NA
NA
NA
Phenol
000106 96 2
80
12
8.006 01 °
NA
S.40E-OI
5 40F 01
NA
Potybrormnsted Biphanyii
059536-65-1
S3
64
7.006-06°
7 006 05 °
e.90E+C»
8.516 06
8.516-05
9 5 76 « 00
Polfchk>rln»t«d Biphenyl*
001338-36-3
90
46
NA
NA
7.70E 4-00
NA
NA
8 566 f 00
Pyf»n»
000129-00-0
31
2
3.00E 02b
3.006 01 C
NA
9 306 03
9 306 02
NA
Scknloui Acid
007783-00-B
87
41
5 006 03 b
5 006 03 °
NA
4 366 03
4 35f 03
NA
Salanita
014124 07 5
70
48
NA
b
6.006 03
NA
NA
NA
NA
NA
Selenium
007182-49-2
44
47
5.006 03 °
NA
2.206 03
2 206 03
NA
SHvar
007440-22-4
18
49
5.006-03*
5.006 03 C
NA
9 006 04
9.OOE 04
NA
Sulfata
014008-79 8
20
50
NA
NA
NA
NA
NA
NA
Tetrachlofoethane. 1.1,2.2-
000079-34-5
70
61
NA
NA
2.006-01 b
NA
NA
2.866-01
T «tr»chteroBth*l«ne
000127 IB 4
100
83
1.006-02 b
1.006 01 C
NA
1.006 02
1 006 01
NA
Thulium
007440-28-0
IB
62
NA
NA
NA
NA
NA
NA
Tfmrfum
007440- 29-1
1.0
53
NA
NA
NA
NA
NA
NA
Tin
007440-31-5
10
28
8.006 01C
8 006 01 C
NA
8 006 02
6 OOF 02
NA
Titanium
007440 32 6
3.0
28
NA
2 006 01 b
NA
NA
NA
NA
NA
Toluene
000108 88 3
BO
54
2 006 + 00 C
NA
1 006 01
1 60E * 00
NA
Trichloroathene, 1,1.1-
000071-55-6
¦0
55
NA
NA
NA
b
NA
NA
NA
Trichloroelhane. 1.1.2-
000079 00 5
01
66
4 006 03b
4 006 02 C
5.70E-02
3 24E 03
3 246 02
1 04E 02
T richloroathylena
000079 01-6
15
57
NA
NA
NA
NA
NA
NA
Trichlorolluoromethane
O00075 69 4
23
12
3 006 01 b
7.006 01 °
NA
6 90E 07
1 61E 01
NA
Uranium
007440-61-1
86
28
NA
NA
NA
NA
NA
NA
UrmHjm. Soluble SalU
NA
85
20
3 006 03
NA
NA
J 55E 03
NA
NA
Vanadium. Metallic
007440-62-2
1.0
58. 59
7 006 03 C
7 006 01 °
MA
7 OOE 05
7 001 05
NA
Ilia dermal vetoes in (his table havo not undergo™ !>«"' rovmw Pege ' '
-------
^SSSSSSSSST.SStn^ „„„„
Chemical
Vinyl Acaltii
Vinyl Chlorida
Xfton#. Mintut*
Zinc IMelaHicI
Zirconium
CAS Number
000108-05-4
000075 01 4
TO 1330-20-7
007440-66-8
007440-87-7
Gi
Absorption
Factor |*|«
Ql
Abcorptkwi
R«far«nc«
Oral RfO Img/ltg/dayl
Subchronlc
Chronic
Oral Slop#
Factor
(mg/hg/dayl't
Dermal RID Img/kg/dayI
Subchronic
b ®' *opr,'on leclore obtained from'literalur, by BEIA .Off
c SoU,C": Hi.k information Syatam (ni5|
d Envt,oom*m'' A"re,¦summ^ t— «ast» ,9B3
"* " D<™lrotoiuen» mixture, 2 4-/2 8-" In «ic m ...
.4/2.6 ,n«.S. The v®tue li baaad on m »,udy uiing technic®! grade DNT
Dermal Slope
Factor
fmg/kg/day)1
t.OOE + OO
I.OOC+MJ
8.50E-01
8 50€ 01
1.90E + 00
2.O0E + OO
!.90f+00
1.84E400
3 00E 01
3.00C-O1
8.00E-02
eoofoi
'09/95
The dermal values in thi: ,
nave nol undergone poor rovia
-------
1
2
3
4
5
6
?
8
9
10
11
12
13
14
15
18
17
18
19
20
21
Gi ABSORPTION REFERENCES
Reference
ATSDR (Agency lor Toxic Subtleness and Diseas# Registry!. 1992. Toxicotogical Profile lor 2-Butonono. ATSDRAJ.S. Public Health Service
Rahman, A., J.A. Barrowman and A. Rahimtula. 1986. The influonco of bila on the bioavailability of potyrtucloar aromatic hydrocarbons from the rat Intestine. Can. J. Physiol.
Pharmacol. 64:1214-1218
ATSDR (Agortcy for Toxic Substances and Disease Registry!. 1992. Tonicoloflical Profile for Acolon®. ATSDR/U.S. Public Haatth Sarvica
Venugopal, B. and T.D. Luckey. 1978. Metal Toxicity in Mammals 2. Chemical Toxicity of Metals and Metalloids. Plenum Press, pp. 104-112
Kortu*. J. 196?. The carbohydrate metabolism accompanying intoxication by aluminum salts in the rat. Expariantia 23:912-913
Rahman',' A., J.A. Barrowman and A. Rahimtula. 1988. The influence of bila on the bioavailability of potynocleor aromatic hydrocarbons from the rat intestine. Cart. J. Physiol.
Pharmacol. 64:1214-1218
Gerber. G.B.. J. Maes and B. Eyfcens. 1982. Transfer of antimony and arsenic to the developing organism. Arch. Toxicol. 49:159-168
Battfey. F.R., O'Shea, J.A. 1975. The absorption of arsenic and its relation to carcinoma. Br. J. Dermatology. 32:563. (Cited in Hindmarsh and McCurdy. 19861
ATSDR (Agency for Toxic Substances and Disease Registry). 1992. Toxicotogical ProfKe for Barium. ATSDR/U.S. Public Health Service
Sabourin, P.J., B.T. Chen, G. Lucior, et al. 1987. Effects of dote on the absorption and excretion of |14C(benzene administered orally or by inhalation in rets nnd mien. Toxicol.
Appl. Pharmacol. 8.7:325-336
ATSDR I Agency for Toxic Substances and Disease Registry). 1993. Toxicologic al Profile for Benzidine. ATSDR/U.S. Public Health Service
H5DB (Hazardous Stibstencvs Data Bank!. 1994. MEDLARS Ontina Information Retrieval System, National library of Medicine. Retrieved March, 1934
Furchner, J.E., C.R. Richmond and J.E. London. 1973. Comparative metabolism of radionuclides in mammals: VII. Retention of beryllium in the mouse, rat. monkey and dog.
Health Phys. 24:292-300
Reeves, A.L. 1965. The absorption of berytiium from the gastrointestinal tract. Arch. Environ. Health 11:209-214
Teirlynck, O.A. end J. Betpaire. 1985. Disposition of orolly administered «Ji(2-ethylh«xyl|ph!hal8te and monol2-ethylhexyl)phthalete in the rat. Arch. Toxicol. 57(41:226-230
Foulkos, E.G. 1988. Absorption of cadmium. In: Handbook of Experimental Pharmacology. Vol. SO, E.C. foulkas, ed„ Springer-Verlag, Berlin, pp. 75-100
Nordberg, G.F., T. Kjeilstrom and M. Nordberg. 1985. Kinetics and ntetaboiism. In: Cadmium and Health: A Toxicotogical and Epidemiological Apprntsal. Vol. I. Exposure,
Dose, and Metabolism. L. Friberg, E.G. Elinder, T. Kjeilstrom and G.F. Nordberg, ads. CRG Press. Boca Raton, Fl, pp. 103-178
McLettari. J.S., P R. Flanagan, M.J. Chamberlain and L.S. Vaiberg. 1978. Measurement of dietary cadmium absorption in humans. J. Toxicol. Environ. Health 4:131 138
Shaikh. Z.A. end J.C. Smith. 1980. Metabolism of orally ingested cadmium in humans. In: Mechanisms of Toxicity and Hazard Evaluation, B. Holmstead, et a!., ods.
Elsevier/North Holland, Amsterdam, pp. 569-574
ATSDR (Agency lor Toxic Substances and Disease Registry). 1992. Toxicologies! Profile for Carbon Disulfide. ATSDR/U.S. Public Health Service
U.S. EPA, 1989. Updated Health Effects Assessment for Carbon Tetrachloride. Prepared for the Office Emergency and Remedial Response, Washington. DC. Environmental
Criteria and Assessment Office, Cincinnati, OH
ATSDR (Agency for Toxic Substances and Diseas® Registry). 1992. Toxicotogical Profile for CHordane. ATSDR/U.S. Public Health Service
Page 9.1
-------
Gi ABSORPTION REFERENCES
Rafaranca
Number Rafarcnce |
23 A1SDR {Agency lor Toxic Substances and Disease Registry). 1990. Toxicologtcal Profile for Chlorobanzane. ATSDR/U.S. Public Health Service
24 Brown, D.M., P.F. Langaly, 0. Smith, at ai. 1974. Metabolism of Chloroform. I. Th® metabolism of |14C|-chloro(orm by different spocies. Xnnobiotica 4:151-163
25 U.S. EPA. 1984. Health Effect* Assessment lor Chromium. Final Report. Environmental Criteria and Assessment Olfice. Research Triangle Park, NC. EPA 600/8-83-014F.
PP8S-115905
26 Goyer, R.A. 1991. Toxic Effect* o( Metals. In: Casarett end DouM's Toxicology: The Baste Science of Poisons, fourth edition. M. 0. Ammdvr, J. Ootid, and C D. Klaassen,
edi. pp. 623-680
27 Venugopal, B. and T.O. Luckey. 1978. Metal Toxicity in Mammals - 2. Chemical Toxicity of Metals and Metalloids. Plenum Press, pp. 104-112.
28 ATSDR f Agency for Toxic Substances and Disease Registry). 1392. Toxicotogical Profile for p.p'-DDT, DDE, DDD. ATSDR/U.S. Public Health Service
29 ATSDR (Agency for Toxic Substances and Disease Registry). 1990. Toxicological Profile lor Di-n-Butylphtbalate. ATSDR/U.S. Public Health Service
30 ATSDR (Agency for Toxic Substances and Disease Registry!. 1989. Toxicologic^! Profile for 1,2-Dichioraathano. ATSDR/U.S. Public Health Service
31 ATSDR (Agency for Toxic Substances and Disease Registry!. 1989. ToMicologieal Profile for 1,1 -Dichloroatbeno. ATSDR/U.S. Public Health Sarvice .
32 Parka, D. 1961. Studies in detoxification. V. The metabolism of m-dtnitrof 14C|benzene in the rabbit. Biochem. J. 78:262-271
33 U.S. EPA. 1986. Health ami Environmental Effects Profile for Dinitrotoiuone. Office of Health and Environmental Assessment, Environmental Criteria ami Assessment Olfice.
Cincinnati, OH. ECAO- CIN-P183
34 ATSDR (Agency for Toxic Substances end Disease Registry!. 1991. Toxicologtcal Profile for Fluoride, Hydrogen Fluoride, and Fluorine. ATSDR/U.S. Public Health Service
35 ATSDR (Agency for Toxic Substances and Disease Registry). 1993. Toxicologies! Profile for Heptachlor/Heptachlor Epoxide. ATSDR/U.S. Public Health Service
36 ATSDR (Agency for Toxic Substances and Disease Registry). 1992. Toxicologies! Profile for 2-Hexanono. ATSDR/U.S. Public Health Service
37 U.S. EPA. 1984. Health Effects Assessment for Cyanides. Prepared by the Environmental Criteria and Assessment Office, Cincinnati, OH for the Emergency and Remedial
Response Offica, Washington, DC. EPA/540/1 -86-011
38 U.S. EPA. 1990. Integrated Risk Information System (IRIS). Health Risk Assessment for Manganese. Online. (Verification dale 6/21/30). Office of Health arid Environmental
Assessment, Environmental Criteria and Assessment Offica, Cincinnati, OH
33 ATSDR (Agency lor Toxic Substances and Disease Registry). 1989. Toxicologicel ProfilA for Mercury. ATSDR/U.S. Public Health Service
40 Miettinen, J.K. 1973. Absorption and elimination of dietary (Hg+ + I and methylmercury in man. In: Mercury, Mercurials, and Mercaptans. M.W. Miller end T.W. Clakrson, Eds.
Springliefd, il pp 233-243
41 Angeio, M.J., A.B. Pritchard, D.R. Hawkins, et al. 1986. The pharmacokinetics of dichlorometbene. I. Disposition in B6C3F1 mice following intravenous and ore! administration.
Food Chem. Toxicol. 24:965-974
42 Friberg, i. and J. Loner. 1986. Moiybenum. In: Handbook on the Toxicology of Metals. 2nd ed. I. Friberg. G.F. Nordberg and V.B. Vouk, ads. Elsevier/North Holland Biomedical
Press, New York. pp. 446-461
43 ATSDR (Agency for Toxic Substances and Disease Registry). 1993. Toxicol ogical Profile for Naphthalene. ATSDR/U.S. Public Hoalth Service
02/03/35
\ Pnga S.2
-------
44
45
46
47
48
43
50
51
52
53
54
55
56
57
58
59
60
61
62
83
64
65
Gi ABSORPTION REFERENCES
Reference
ATSDR (Agency (or Toxic Substances and Disease Registry). 1993. Toxicological Prolila lor Nickel. ATSDR/U.S. Public Health Seivice
ATSDR {Agency lor Toxic Substances and Disease Registry). 1989. Toxicological Prolile for N Nitrosodi n propylamine. ATSDR/U.S. Public Health Service
ATSDR (Agency lor Toxic Substances and Disease Registry). 1989. Toxicological Profile for PolycMorinaled Biphanyis: Aroctor 1260, 1254, 1248, 1242, 1232, 1221, and
1016. ATSDR/U.S. Public Health Service
Madlnsky. M.A., R.G. Cuddihy, R.O. McCleHan. 1981. Systemic absorption of sdenious acid and elemental selenium aerosols in rats. J. Toxicol. Environ. Health 8:917-928.
Thomson C.D. end R.D.H. Stewart. 1974. The metabolism of |75Se| in young women. Br. J. Nutr. 32:47-57
East B.W., K. Boddy, E.D. Williams, et al. 1980. Silver retention, total body silver and tissue silver concentrations in argyria associated with exposure to en anti-smoking remedy
containing silver acetate. Clin. Exp. Dermatol. 5:305-311
World Health Organization (WHO). 1984. Guidelines for Drinking Water Quality, Vol. 2. Health Criteria and Other Supporting Information
ATSDR (Agency lor Toxic Substances and Disease Registry). 1S89. Toxicological Profile for 1,1,2.2-Tatrachloroethana. ATSDR/U.S. Public Health Service
ATSDR (Agency for Toxic Substances ami Disease Registry). 1992. Toxicological Profile for Thallium. ATSDR/U.S. Public Health Service
ATSDR (Agency for Toxic Substances and Disease Registry). 1990. Toxicological Profile for Thorium. ATSDR/U.S. Public Health Service
ATSDR (Agency for Toxic Substances and Disease Registry). 1989. Toxicological Profile for Toluene. ATSDR/U.S. Public Health Service
ATSDR (Agency for Toxic Substances and Disease Registry). 1993. Toxicological Profile for 1,1,1-Trichloroethane. ATSDR/U.S. Public Health Service
ATSDR (Agency for Toxic Substances and Disease Registry). 1989. Toxicological Profile for 1.1.2 TricMoroethane. ATSDR/U.S. Public Health Set vice
Daniel, J.W. 1963. The metabolism of 36CJ-labeHed trichk>roethylene and tatrachiorathytone in the rat. Biochem. Pharmacol. 12:735-802
Curran, G.L. and R E. Burch. 1967. Biological and health effects of vanadium. In: Proceedings of the Conference on Trace Substances in Environmental Health. D O. Hemphill
Ed. University ol Missouri, Columbia, MO, pp. 98-104
1CRP (International Commission on Radiological Protection). 1960. Report of Committee (I on Permissible Dose for Internal Radiation. Recommendations of the International
Commission on Radiological Protection. ICRP Publ. No. 2. Pergamon Press, Oxford, England, p.163
ATSDR (Agency for Toxic Substances and Disease Registry). 1992. Toxicological Proliln for Vinyl Acetate. ATSDR/U.S. Public Health Service
ATSDR (Agency for Toxic Substances and Disease Registry). 1993. Toxicological Profiln'for Total Xylenes. ATSDR/U.S. Public Health Service
U.S. EPA. 1984. Health Eflects Assessment for Zinc and Compounds. U.S. Environmental Protection Agency, Office of Research and Development. Washington, DC.
EPA/540/1 86-048
ATSDR (Agency lor Toxic Substances and Disease Registry). 1993. Toxicological Profile for TetrachloroathylanB. ATSDR/U.S. Public Health Service
ATSDR (Agency for Toxic Substances end Disease Registry). 1993. Toxicological Profile for Polybrominaled biphanyis (PBBsl. ATSDR/U.S. Public Health Sorvicn
ATSDR (Agency for Toxic Substances and Disease Registry). 1993. Toxicological Profile lor AkJrin/Dieldrin. ATSDR/U.S. Public Health Service
ATSDR [Agency lor Toxic Substances and Disease Registry). 1983. Toxicological Profit# lor Vinyl Chloride. ATSDR/U.S. Public Health Service
ATSDR (Agency for Toxic Substances and Disease Registry). 1989. Toxicological Profile for Bromodichloromethane. ATSDR/U.S. Public Health Service
Pb{j« 9.3
-------
68
69
70
71
72
73
74
75
76
77
78
73
80
81
82
Gi ABSORPTION REFERENCES
IUftr«ic«
ATSDR (Agency for Toxic Substances and Disease Registry). 1989. Toxicological Profile for 1,2-Dichtoropfopane. ATSDR/U.S. Public Health Service
ATSDR (Agency lor Toxic Substance# and Disease Registry). 1990. Toxic oiogical Profile for Bromoform/ChlorocHbromomethane. ATSDR/U.S. Public Health Service
ATSDR (Agency (or Toxic Substances and Disease Registry). 1994. Toxicological Profile lor Alpha-, Beta-, Gamma-, and Delta- Hexachlorocyclohexane. ATSDR/U.S. Public
Health Service
ATSDR (Agency for Toxic Substances and Disease Registry). 1994. Toxicologies! Profile for Pentachlorophanol. ATSDR/U.S. Public Health Service
ATSDR (Agency for Toxic Substances and Disease Registry). 1992. Toxicological Profile for Cresols. ATSDR/U.S. Public Health Service
ATSDR (Agency for Toxic Substances and Disease Regisfryl. 1992. Toxicologic ai Profile for 1,3-Dichloroproper»«. ATSDR/U.S. Public Health Service
ATSDR (Agency for Toxic Substances awl Disease Registryl. 1993. Toxic oiogical Profile for Chromium. ATSDR/U.S. Public Health Service
U.S. EPA. 1980. Health and Environmental Effects Document for Boron and Boron Compounds. Prepared for the Office of Solid Waste and Emergency Response, Washington,
DC by the Environmental Criteria and Assessment Office, Cincinnati, OH. ECA0-CIN-G100
U.S. EPA. 1986. Health and Environmental Effects Profile for Carbamle. Prepared for the Office of Solid Waste and Emergency Response, Washington, DC by the Environmental
Criteria and Assessment Office, Cincinnati, OH. ECAO-CIN-P157
U.S. EPA. 1987. Health Effects Assessment for Endrki. Prepared for the Office of Solid Waste and Emergency Response, Washington, DC by the Environmental Criteria and
Assessment Office, Cincinnati, OH. ECAO-CIN-HOS9
U.S. EPA. 1989. Health and Environmental Effects Document for Butyl benzyl phthaiate. Prepared for the Office of Solid Waste and Emergency Response, Washington, DC by
the Environmental Criteria and Assessment Office, Cincinnati, OH. ECA0-CIN-G081
U.S. EPA. 1987. Health and Environmental Effects Profile for Phthalic acid alkyf, aryl, and aikyl/aryl esters. Prepared for the Office of Solid Waste and Emergency Response,
Washington, DC by lhe Environmental Criteria and Assessment Office, Cincinnati, QH. ECAO-CINP188
U.S. EPA. 1985. Health and Environmental Effects Document for Dichloroethenes. Prepared for the Office of Solid Waste and Emergency Response, Washington, DC by the
Environmental Criteria and Assessment Office, Cincinnati, OH. ECAO-CIN-P139
U.S. EPA. 1987. Health and Environmental Effects Profile for Chlorinated phsnols. Prepared for the Office of Solid Waste and Emergency Response. Washington, DC by the
Environmental Criteria and Assessment Office, Cincinnati, OH. ECA6-CIN-G013
U.S. EPA. 1987. Health Effects Assessment for Dichlorobemanes. Prepared for the Office of Solid Waste and Emergency Response, Washington, DC by the Environmental
Criteria and Assessment Office. Cincinnati, OH. ECAO-CIN-H079
f'ngn 9.4
-------
ATTACHMENT M
EPA GUIDANCE FOR CALCULATING EXPOSURE POINT CONCENTRATIONS
s FINAL MASTER WPD Jmituj-r
-------
LJnfttd Slaws
Environmental PfWecoon
Agency
OfSca of SoW Waste and
Emeigeney Response
Washington, D.C. 20460
PubltcaDon 92BS 7-061
May 1992
Supplemental Guidance to
RAGS: Calculating the
Concentration Term
Office of Emergency and Remedied Responee
HaMrtous Site Evaluation Divwon, OS-230
intermittent Bulletin
Velum® 1 Number 1
The overarching mandate of the Comprehensive Environmental Response, Compensation, and Liability
Act (CERCLA) is to pro tea human health and the environment from current and potential threats posed by
uncontrolled releases of hazardous substances. To help meet this mandate, the U.S. Environmental Protection
Agency's (EPA's) Office of Emergency and Remedial Response has developed a human health risk assessment
process as pan of its remedial response program. His process is described in Risk Assessment Guidance for
Superfund: Volume I — Human HmMh Evaluation Mmmai (RAGS/HHEM). Part A of RAGSifOIEM
addresses the baseline risk assessment, and inscribes a general approach for estimating exposure to individuals
from hazardous substance releases at Superfund sites.
This bulletin. opiates the concentration term in the acposureHntalte equation to remedial project
managers (RPMs), risk assessors, statisticians, and other personnel. This bulletin presents the general intake
equation as presented in RAG5/HHEM Part A, discusses basic coicepts ooiceiuiflg the concentration term,
describe generally how to calculate the concentration tenn, presents camples to illustrate several important
points, and, lastly, identifies where to get additional help.
THE CONCENTRATION TERM
How is the conceatwflto# tenon used?
RAGS/HHEM Pan A presents the
Superfund risk assessment process in four "steps*:
(1) data coieetiOB and evaluation; (2) acposuje
assessment; (3) toodcity assessment; and (4) risk
characterization. The concentration term Is
calculated for use in the exposure assessment step.
Highlight 1 presents the general equation
Superfund wo for calculating otposwe, and
illustrates that the concentration term (C) is one
of several parameter needed to estimate
contaminant intake for an individual. .
For Superfund assessments, the
- concentration term (C) in the intake equation is
an estimate of the arithmetic average concentration
for a contaminant based on a set of site sampling
cstimatini the true average concentration at a site.
the arithmetic mean should be used for this
variable. The 95 percent UCL provides reasonable
confidence that the true site average will not be
MndgreM ifwatftfi
Why o* an average wioe for the «mcente»ilon
term?
As estimate of average conceotfatioa is used
beause:
Supplemental €kMu%et: m MAGS to a Mlctts acriei oa risk wafflni of Snpofuatf litm. Ttae WUetnB aem aa wppteaeats to
Kak AtsammM GuMtmet for Supofimd: Vekmtel—MmmmMmMtEmhmiamMmmai. The intomitw pM«tert ¦ iateoded as
I«jdan.eE[ to EPA aad otko* pwmisl eaptoycm It doca so* tommms rutewUag by lie Agcaqr, nd saw doc Be reBed on to
txe»te i substantive or pfoeaiufml right cafonable by aay other poaoe. The Govobbmi mwf take action that is it wum with
tbe*e buUetiaa.
-------
Highlight I
GENERAL EQUATION FOE ESTIMATING EXPOSURE
TO A SITE CONTAMINANT
j r CRxEFD I
I = Cx———— x —
BW AT
where:
I =» intake (i.e.. He quantitative measure of exposure in RAGS/HHEM)
C = contaminant concentration
CR = contact (intake) rate
EFD =* exposure frequency and duration
BW =t body weight
AT = averaging time
(1) carcinogenic and chronic noncarcinogenic
• toxicity criteria1 are based on lifetime
average exposures; and
(2) average concentration is most
representative of the concentration that
would be contacted at a site over time.
For example, if you assume that an exposed
individual moves randomly across an otposutre
area, then the spatially averaged soil concentration
can be used to estimate the true average
concentration contacted over time. In this
example, the average concentration contacted over
lime would equal the spatially averaged
concentration met the exposure area, WMIe an
individual may not actually exhibit a truly random
pattern of movement across an exposure area, the
assumption of equal time spent ia different para
of the area is a simple but reasonable approach.
When should an ¦wage coxaartrmUoa be used?
The two types of exposure estimates now
being required for Superfund risk assessments, a
reasonable maximum exposure (RME) and an
average, should both use an average concentration.
To be protective, the overall estimate of intake
{see Highlight 1) used as a basis for action at
When acute toxicity is of most concern, a long*
term average concentration generally should not be
used for risk assessment purposes, as the focus
should be to estimate short-term, peak
concentrations.
Superfund sites should be an estimate in the high
end of the intake/dose distribution. One high-end
option is the RME used in the Superftmd
program. He MMDS, which is defined as the
highest exposure that could reasonably be expected
to occur for a given exposure pathway at a site, is
intended to account for both uncertainty in the
contaminant concentration and variability in
exposure parameters (e.g., exposure frequency,
averaging time). For comparative purposes.
Agency guidance (US, EPA, Guidance on Risk
Characteriztaion for Bisk Managers and Risk
Assessors, February 26,1992) states that an average
estimate of exposure also should be presented in
risk assessments. For decision-making purposes in
the Superfund prapam, however, RME is used to
estimate risfc*
Why am an estimate of the arithmetic mean
rather than the geomefcfc mean?
The choice of the arithmetic mean
concentration as the appropriate measure for
estimating exposure derives from the need to
estimate an individual's long-term average
exposure. Most Agency health criteria are based
on the long-term average daily dose, which is
simply the sum of all daily doses divided by the
total number of days in the averaging period. Has
is the definition of an arithmetic mean. The
For additional information on RME, see
RAGS/HHEM Part A and the National Oil and
Hazardous Substances Pollution Contingency Plan
(NCP), 55 Federal Register 8710, March 8, 1990.
-------
arithmetic mean is appropriate regardless of the
pattern of daily exposures over lime or the type of
statistical distribution thai might best describe the
sampling data. The geometric mean of a set of
sampling results, however, bears no logical
connection 10 the cumulative intake thai would
result from long-term contact with site
contaminants, and it may differ appreciably from —
and be much lower than — the arithmetic mean.
Although the geometric mean is a convenient
parameter for describing central tendencies of
lopormat distributions, it is not an appropriate
basis for estimating the concentration term used in
Superfund exposure assessments. The following
simple example may help clarify the difference
between the arithmetic and geometric mean when
used for an exposure assessment:
Assume the daily exposure for a trespasser
subject to random exposure at a site is 1.0,
0.01, 1.0, 0.01, 1.0, 0.01, 1.0, and 0.01
units/day over an 8-day period. Given
these values, the cumulative exposure is
simply their summation, or 4.04 units.
Dividing this by 8 days of exposure results
in an arithmetic mean of 0.505 units/day.
This is the value we would want to use in
a risk assessment for this individual, not
the geometric mean of 0.1 units/day.
Viewed another way, multiplication of the
geometric mean by the number of days
equals 0J units, considerably lower than
the known cumulative exposure of 4.04
units.
UCL AS AN ESTIMATE OF THE
AVERAGE CONCENTRATION
What Is a 95 percent UCL?
The 95 percent UCL of a mean is defined
as a value that, when calculated repeatedly for
randomly drawn subsets of site data, equals or
exceeds the tree mean 95 percent of the time.
Although the 95 percent UCL of the mean
provides a conservative estimate of the average (or
mean) concentration, it should not be confused
with a 95th percentile of site concentration data (as
shown in Highlight 2).
Why use the UCL as the average concentration?
Statistical confidence limits are the classical
tool for addressing uncertainties of a distribution
average. The 95 percent UCL of the arithmetic
mean concentration is used as the average
concentration because it is not possible to know
the true mean. The 95 percent UCL therefore
accounts for uncertainties due to limited sampling
data at Superfund sites. As sampling data become
less limited at a site, uncertainties decrease, the
UCL moves closer to the true mean, and exposure
evaluations using either the mean or the UCL
produce similar results. This concept is illustrated
in Highlight 2.
Should a value other than the 95 percent UCL be
used for the concentration?
A value other than the 95 percent UCL
can be used provided the risk assessor can
document that high coverage of the true
population mean occurs (i.e., the value equals or
exceeds the true population mean with high
probability). For exposure areas with limited
amounts of data or extreme variability in measured
or modeled data, the UCL can be greater than the
highest measured or modeled concentration. In
these cases, MaddfflgiilJata cannot practicably be
obtained, the highest measured or modeled value
could be used as the concentration term. Note,
however, that the true mean still may be higher
(i.e., the 95 percent UCL
indicates a higher mean is possible), especially if
the most contaminated portion of the site has not
been sampled.
CALCULATING THE UCL
Btaw many samples are necessary to calculate the
95 percent UCL?
Sampling data from Superfund sites have
shown that data sets with fewer than 10 samples
per exposure area provide poor estimates of the
mean concentration (i.e., there is a large difference
between the sample mean and the 95 percent
UCL), while data sets with 10 to 20 samples per
exposure area provide somewhat better estimates
of the mean, and data sets with 20 to 30 samples
provide fairly consistent estimates of the mean
(ie., the 95 percent UCL is close to the sample
mean). Remember that, in general, the UCL
approaches the true mean as more samples are
included in the calculation.
Should the data be transformed?
EPA's experience shows that most large or
"complete" environmental contaminant data sets
-------
Highlight 2
COMPARISON OF UCL AND W* PERCENTILE
Upptr CoaMmm
Um* (UCL)
of DM ll«m
Conewritratfen
As sample size increases, the UCL of the mean moves closer to tie true mean, wfcle the 95*
percentile of tie distribution remains at the upper end of tie distribution.
front soil sampling to lognormally distributed
rather thai normally distributed (see Hgligfcis 3
and 4 for illustrations of lognormal and normal
distributions). In most cases, it is reasonable
to assume that Superfund soil sampling data are
tepermaUf distributed. Beomse transformation is
a necessary step in ealcutaimg the UCL of the
arithmetic mean for a lognormal distribution, the
data should be transformed by using the natural
logarithm function (it, calculate ln(x), where x is
the value from the data set). However, In
where there is a question about the distribution of
the data set, a statistical test should be used to
identify the best distributional assumption for the
data set The W-test (Gilbert 1987) is one
statistical method that on be used to determine if
a data set is consistent with a normal or lognormal
distribution. In aU cases, it is valuable to plot the
data to better understand the contaminant
distribution at the site.
How do you calculate the UCL for a lognormal
distribution?
To calculate the 95 percent UCL of the
arithmetic mean for a lognormally distributed data
set, fiist tnnsfofm the dan using the natural
logarithm function as iHscnssei previously (ie.,
calculate ln(x)). After transforming the data,
determine the 95 percent UCL for the data set by
completing the following four steps:
(1) Calculate the arithmetic mean of the
transformed data (which is also the log of
the geometric mean);
(2) Calculate the standard deviation of the
uansfonied data;
(3) Determine the H-statistic (c.g.» see Gilbert
1987); mi
(4) Calculate the UCL using the equation
shown in HWiiwa 5.
How do pw calculate the UCL for a normal
distribution?
IUjaitfetigUgLsHBponsthe assumption
tha|jhejja|ajgLgjOTmallv distributed, calculate
the 95 percent UCL by completing-the following
four steps:
-------
Highlight 3
EXAMPLE OF A LOGNORMAL DISTRIBUTION
40
i
s
B%h%bl 4
IXMffLE OF A NORMAL DISTRIBUTION
40
Coneantratton
-------
Highlight 5
CALCDL4TOIG THE UCL OF THE ARTTTOIETIC fcffiAN
FOR A LOGNORMAL DISTRIBUTION
where:
UCL ~jWv/^T)
UCL
e
I
s
H
E
upper confidence limit
constant (base of the natural log, eqoal to 2.718)
mean of tie transformed data
standard deviation of the transformed data
H-statfetic (e.f.» foil table published to Gilbert 1987)
number of samples
_ _ Highlight <
CALCULATING THE UCL OF THE AKHBMEIIC MEAN FOR A NORMAL DISTRIBUTION
UCL*x + t(sfJn)
where;
%
UCL -
x »
S s
t m
n s
upper confidence limit
¦earn of the untransformed data
standard deviation of the untransformed data
Student-! statistic (e.g„ from table published in Gilbert 1987)
number of samples
———______________
(1) Calculate the arithmetic mean of the
untransformed data;
(2} .Calculate the standard deviation of the
untranslbRned data;
(3) Determine lie one-tailed Statistic (e.t
see Gilbert 1987); and
(4) Calculate the UCL using the equation
presented la Highlight g.
Use caution when applying normal distribution
calculations if there is a possibility that heavily
contaminated portions of lie site have not been
adequately sampled, to such cues, a UCL ftom
normal distribution calculations could fall below
the true mean, even if a limited data set at a site
appears normally distributed.
EXAMtttS
The examples shown in Highlights 7 and I
address the exposure scenario where an individual
at a Superfttnd site has equal opportunity to
contact soil in any sector of the mwnrmfiai^ mm
ewer time. Even though the examples address oily
soil exposures, the UCL approach is applicable to
al exposure pathways. Guidance and examples for
ether exposure pathways wM be presented in
forthcoming bulletins.
Highlight 7 presents a simple data set and
provides a stepwise demonstration of transforming
the data — assuming a lognormal distribution —
aad calculating the UCL Highlight 8 uses the
sane data set to show the difference between the
UCLj that would result from ass timing' normal and
lognormal distribution of the data. These
-------
examples demonstrate the importance of using the
correct assumptions.
WHERE CAN i GET MORE HELP?
Additional information on Superfund's
policy and approach 10 calculating lie
concentration tern and estimating oposures at
waste sites can be obtained in:
• U.S. EPA. Risk Assessment Guidance
for Superfiuid: Volume I — Human
Health Evaluation Manual (Part A),
EPA/540/1-89/002, December 1989.
• U.S. EPA, Guidance for Data
Useability m Risk Assessment,
EPA/540/G-90/008 (OSWER
Directive 9285.7-05), October 1990.
• U.S. EPA, Risk Assessment GrnMace
for Supofmd (Part A —Bmetme Msk
Assessment) Supplemental Guidancel
Standard Exposure Foam, OSWER
Directive 9285.MB, May 1991.
Useful statistical guidance can be found in many
standard textbooks, including:
• Gilbert, R.O., Statistical Methods for
Environmental Pollution Monttomg,
Van Nostrand Eemhold, New York,
New York. 1987.
Questions or comments concerning the
concentration term can be directed to:
• Toxics Integration Branch
Office of Emergency and Remedial
Response
401 M Street SW
Washington, DC 20460
Phone: 202-260-9486
EPA staff can obtain additional copies of this
bulletin by calling EPA's Center for Environmental
Research Information at FTS 684-7562 (513-569-
7652). Others can obtain copies by contacting
NHS at 800-336-4700 (703-457-4650 in the
Washington, DC area).
Flrat-Clasa Mail
Postage and F«ta Paid
P»m# No. G-3S
SEPA
United Stataa
Environmental Protection
Aganqr (OS»J
Washington, DC 20460
Official Bualnaaa
Panalty for Privata Uaa
S300
-------
Highlight 7
EXAMPLE OF DATA TRANSFORMATION AND CALCULATION OF UCL
Hfe example show the calculation of a 95 percent UCL of the arithmetic mean
concentration for chromium in soil at a Superfund site. This example a applicable nniv t0 a
sceqano >n which a sp^tjaUy raqdotq exposure pattern is assume^ The concentrations of chromium
obtained from random sampling in soil at this site (in me/kg) are 10,13, 20 36 41 59 67 nn i in
136. 140, 160, 200, 230. and .300. Usfcg , or 502 mgkg.
- , Highlight g
OMPAJUNG UCL8 OF TBI AMnBMEIlC MEAN ASSUMING DiTFEwrNT DISTRIBUTIONS
,K. t._to ^ the data presented in Highlight 7 are used to demonstrate the difference in
the UCL list is seen if the normal distribution approach were inappropriately applied to this data
sci (i.e., if, hi this example, a normal distribution is assumed),
ASSUMED DISTRIBUTION: Normal Lognonnal
TEST STATISTIC Student-t H-statistic
95 PERCENT UCL (mg/kg): 325
-------
Environmental Prelection
Agency
Emergency Response
Washington, D.C. 20460
December 1933
Supplemental Guidance to RAGS:
Estimating Risk from Groundwater
Contamination
Office of Emergency and Remedial Response
Hazardous Site Evaluation Division
Intermittent Bulletin
Volume X Number X
The Toxics Integration Branch and Regional risk
assessors have formed a Total Quality Management
(TQM) Quality Action Team (QAT), known as the
Concentration Term workgroup, to address the broad
goal of improving the qualify of data used in baseline
risk assessments.
For this fact sheet, the Concentration Term
workgroup consulted with representatives of the
Groundwater Forum to address the risk assessment
challenges posed by groundwater, in particular.
Calculating Risk
Risk ai Superfund sites generally is calculated by
comparing estimates of human exposure with Agency-
verified toxicity criteria. Exposure is calculated by
combining concentration with other parameters, such
as the contact rate, exposure frequency and duration,
and body weight. Current Agency guidance (U.S.
EPA, Guidance on Risk Characterization for Risk
Managers and Risk Assessors* February 26, 1993)
requires risk assessments to present multiple descrip-
tors of risk, including estimates for both average and
high-end exposure scenarios. The Superfund site
manager uses the calculated risk value to help deter-
mine the need for and extent of contaminant cleanup.
Since risk and exposure are linearly related, the
pollutant's concentration has a significant influence on
the risk analysis, and, consequently, the remedial
decision at a given site. The calculation of the concen-
tration term is crucial; miscalculation could result in a
false estimate of risk and, ultimately, result in inappro-
priate cleanup decisions and misdirected Superfund
actions.
The Superfund program uses a reasonable maxi-
mum exposure (RME) or high-end risk calculation as
the basis for remedial decisions. RME is intended to
estimate a conservative case that is both protective of
human health and the environment, while remaining
within the range of potential exposure levels.
Because groundwater is a very complex and
dynamic medium with characteristics that can change
seasonally, it is likely that concentration of a given
contaminant in each well will vary over lime. There-
fore, the concentration term is best described by an
arithmetic average, regardless of whether the overall
exposure estimate is high-end or average. Time and
resource considerations generally preclude collecting
enough data to calculate a true average; therefore,
Superfund has relied on an upper-confidence limit on
the arithmetic mean (UCLW) to represent the average
concentration.
! he Challenge of Groundwater
Risk Assessment
When determining the need for action, there are
both policy and technical issues that set groundwater
apart from other media, such as soil. EPA's policy is to
consider the maximum beneficial use of groundwater
and to protect it against future contamination. The
National Oil and Hazardous Substances Pollution
Contingency Plan (NCP) (U.S. EPA Publication
9200.2-14, January 1992) states that groundwater is an
inherently valuable natural resource to be protected and
restored where necessary and practical, as groundwater
that is not currently used may be a drinking water
supply in the future. An example of this practice ts
where a deeper, uncontaminated aquifer is hydrauli-
cally connected to a shallow, contaminated aquifer.
(continued on p.2)
Paoe '
-------
I continued from p. I)
Although the shallow aquifer may not
currently be a dnnking wafer resource.
EPA may choose to remediate it 10
protect the deeper aquifer. In addi-
tion, few states have designated
aquifers as unpotable, resulting in
most aquifers being considered
drinking water sources thai must be
addressed In the risk assessment
These policies are sometimes
at odds with Superfund's attempts
10 reasonably assess potential risks
to human health. Risk assessment
should be based upon the likelihood
that a person will be continuously
exposed lo the contaminants present
at the site over lime. In a true
assessment of risk, the usability of
the aquifer must be considered.
This includes such factors as the
quality of the water (pH, redox
potential, salinity, etc.), the size of
the aquifer, the hydraulic character-
istics, the community's water
needs, and the availability of other
drinking water sources.
Technical issues center around
the characteristics of groundwater
and make estimating long-term
exposures particulary difficult.
Most groundwater plumes move
over time. The rate at which the
plume moves, both horizontally and
vertically, can greatly affect the
concentration of contaminants at
the same well. Seasonal variations
in precipitation can cause Jow-to-
high shifts in the groundwater table,
flushing some contaminants out of
the sample area.
The complexity of ground-
water as a medium has a very
definite impact on the ability to
calculate a reliable concentration
term. Because the concentration
term is key to determining risk, it is
imperative that the risk assessor has
enough information to properly
calculate the concentration term.
Analysis has shown that as the
•mber of samples increases, the
or More Information
Additional information on Superfurxfs policy and
approach to calculating risk at groundwater sites can be
obtained in:
• U.S. EPA, Risk Assessment Guidance for Superfund
(RAGS): Volume (—Human Health Evaluation
Manual (Part A), EPA/540/1-89/002. December
1989.
• U.S. EPA, Supplemental Guidance to RAGS: Calculating the Concentra-
tion Term, Publication 9285.7-081, March 1992.
• U.S. EPA, Guidance for Data Usability in Risk Assessment, EPA/540/G-
9G/OOB (OSWER Directive 9285.70S), October 1990.
« U.S. EPA, Guidance on Bisk Characterization for Risk Managers and
Ftisk*A$S8S$ots, jnemorandum from F. Henry Habicht II lo Assistant and
fe&niaif 26,1993. (Available tot the Office of.
. the Administrator.) • ¦
' . • U.Sf-EPA, Guidance Document lor Providing Alternate Water SuppSes,
: . EPA/54Q/G-87/006 {OSWER Directive 9355.3-03), February 1988.
(Available from ttte Superfund Document Center at 202/260-9760.)
• U.S. EPA, Natioaal 09 'and Hazardous Substances PoBuOon Cmtin-
gency Plan (The NCP), Publication 92002-14, January 1992.
degree of uncertainty and inherent
conservatism is reduced. Prelimi-
nary results with soil analyses have
shown that data from 10 to 20
samples per exposure area can support
the calculation of a UCL95 that is
reasonably close to the true mean.
Risk assessors have found that
groundwater pollutant concentra-
tion data collected during the
remedial investigation often are
insufficient to support a statistically
meaningful average. For ground-
water. the exposure area is difficult
to define, and due to the expense
and labor required to install
monitoring wells, adequate data
may not be available for risk
analysis use. If the available data
cannot support statistical calcula-
tion of a pollutant's average
concentration, the risk assessor is
forced lo calculate risk values
from a single concentration
measurement, usually relying on a
maximum value. This approach
provides very low scientific
confidence, as a single measure-
ment cannot represent the con-
tamination present in the entire
plume. Thus, the risk assessors
and site managers must reach a
compromise between the desire
for the optimum amount of data
and the cost of installing arid
sampling wells.
-------
E2B.
281
Q
argeting
Groundwater Risk
Members of the Groundwa-
ter Forum provided the following
description of a groundwater site
investigation, first, they stated
that if is common practice in the
initial phases of a groundwater
site investigation to install be-
tween five to six wells across the
site, targeting source areas and
potential downgradient migration.
Second, around wells with high
hits, one or two additional wells
may be installed to further define
the "center of the plume." Fi- .
nally, once the "center" has been
located, future efforts focus on
defining the extent of contamina-
tion downgradient of the "center."
The number of sampling rounds
varies from site to site, but four
quarters* worth of data will
provide a very good picture of the
Influence of seasonal changes on
the level of the water table.
Although other exposure
estimates are made, the NCP
directs that the risk assessment
focus on estimating an RME.
Therefore, it is appropriate for the
samples. The
primary purpose
of calculating the
UCL,5 is to ensure
that the true mean of the
entire site would not be under-
estimated, as is common with
limited data sets. However, in
this case, we arc targeting data
from the more highly contami-
nated area of the plume, and it is
unlikely that the site-wide average
will be underestimated. Thus, for
the concentration term in ground-
water risk assessments, it is
sufficient to take the simple
arithmetic average of sample data
obtained from two to three wells
I in the "center" of the plume,
j Again, to account for the impact
i of seasonal variations, data from
1 at least two quarters is required,
i and data from four quarters is
I preferred.
This guidance is most appli-
cable to sites where groundwater
is not currently used for drinking
water. For residential wells that
are currently in use, action may be
taken where "Removal Action ~ .
Levels" are exceeded. This action
can be taken based on one round
of sampling with confirmation
analyses.
stated above, the assessor may
have data from only two or three
wells, and calculation of a mean-
ingful UCLW requires 10 to 20
-------
issues to
Consider Jointly at Each Site
Our discussions with the Groundwater Forum produced the
following list of issues thai Regional risk assessors and
hydrogcologists should explore together on a site-specific basis.
Plumes move at different rates,
but few are static. This fad
underscores the indefensibility
of using one data point to
represent a long-term exposure ¦"
point concentration.
Also, speed of plume movement can affect duration of exposure lo
contaminants, leading to either acute or chronic exposure.
(55y\ Speed of
Plume Movement
. Hydraulic Forces
^3 in'the'.Aquifer
Contaminants in a plume are
subject to a variety of forces
thai can retard migration or
attenuate lie concentration
overtime. Even quarterly »
sample data over a year's time represents only a "snapshot" of
contaminant levels that may not be representative of the true long-
term exposure point concentration.
Width/Size of
Capture Zone
The true exposure area for
groundwater is the area
"captured" by a residential
pumping well. The area
defined as the "center" of the
plume may be different.
(jffif Type of Well
alii Rrodmtai, ttwmnpai. o> warmotma
Residential wells usually
pump on the order of 1 to 5
gallons per minute, and data
obtained from monitoring
wells may not reflect the type
of exposure in a residential
setting. A more representative
estimate of exposure point
concentration may be
achieved by modeling the
impact of a pumping well in
the "center" of the plume.
Municipal wells or well fields
can differ substantially in
construction and pumping
capacity from monitoring
wells.
Page*
-------
ATTACHMENT N
A BIBLIOGRAPHY RELATED TO ECOLOGICAL RISK ASSESSMENT
x J«I r-\ J ft|:VlSIS:?f?!MAl Af ASTER WP|> it 'Jim t *«^ni UC
-------
This bibliography is based on a limited literature search and review and is not a complete list of documents
related to ecological risk assessment.
U.S. ENVIRONMENTAL PROTECTION AGENCY GUIDANCE
U.S. Environmental Protection Agency. 1994. Ecological Risk Assessment Guidance for Superfund;
Process for Designing and Conducting Ecological Risk Assessments. Environmental Response Team.
Edison, New Jersey. September.
This proposed agency-wide guidance document describes an accepted process for designing and
conducting ecological risk assessments under the Superfund program, including methods for calculating
risk-based cleanup levels. The final version of this guidance document will supersede Risk Assessment
Guidance for Superfund, Volume II, Environmental Evaluation Manual, which still can be used as a basic
tutorial on Comprehensive Environmental Response, Compensation, and Liability Act (CERCLA)
ecological risk assessment.
EPA. 1993 and 1994. A Review of Ecological Assessment Case Studies From a Risk Assessment
Perspective. Volumes 1 and 2. Risk Assessment Forum. Washington, B.C.
EPA scientists present a cross section of ecological assessment case studies. The case studies present a
variety of work scopes, ecosystems, ecological endpoints, chemical and nonchemical stressors, and
programmatic requirements within EPA. The approaches used in the case studies are generally consistent
with some, but not all, of the principles in the Framework for Ecological Risk Assessment (EPA 1992).
While these case studies are useful examples of the "state-of-the-practice," they should not be regarded
models to be followed.
EPA. 1992. Framework for Ecological Risk Assessment EPA/630/R-92/001. Risk Assessment
Forum. Washington, B.C. February.
This reference should be used in conjunction with other technical guidelines since it is only a conceptual
framework and is not considered a "stand alone" guidance document. The framework provides a format
that facilitates consistent ecological risk assessment formulation at regulated facilities, including Resource
Conservation and Recovery Act (RCRA) hazardous waste facilities.
EPA. 1989. The Nature and Extent of Ecological Risks at Superfund Sites and RCRA Facilities.
EPA-230-03-89-043. Office of Policy Analysis. Washington, B.C. June.
This report presents the results of a study of ecological risks posed by Superfund sites and RCRA facilities.
The report includes discussions of methods used in the identification of sites and review of reports, the
nature and extent of ecological threats, and summaries of key findings.
EPA. 1989. Risk Assessment Guidance for Superfund, Volume II, Environmental Evaluation
Manual. Interim Final. EPA/540/1-89/001. Office of Solid Waste and Emergency Response.
Washington, B.C. March.
Risk Assessment Guidance for Superfund (RAGS) Volume II provides conceptual guidance in planning
studies to evaluate the ecology of a site, including RCRA facilities. The draft Ecological Risk Assessment
Guidance for Superfund supersedes RAGS Volume II as guidance on how to design and conduct
Nj-1
S REP* RIIJOIH TASKS.REVISE!-F!NAL\MASTER WPD l<|-RllX)|Xl>7(NMW24"??n Jtipm-uc 1 * 1
-------
ecological risk assessments. RAGS Volume II does include useful information on the regulatory and
statutory bases of ecological risk assessment, basic ecological concepts, and other background information
that are not presented in RAGS Volume II.
EPA. 1989. Ecological Assessments of Hazardous Waste Sites: A Field and Laboratory Reference.
EPA/600/3-89/013. Office of Research and Development. Washington, D.C. March.
This document provides introductory discussions on various techniques in ecological risk assessment that
may be appropriate for RCRA facilities.
The following ECO Updates are ecological risk assessment bulletins intermittently issued by the EPA that
supplement RAGS Volume II. These bulletins are used to provide technical information that pertains to
various aspects of ecological risk assessment. The bulletins do not constitute rule making by the EPA.
There are currently three volumes of ECO Updates containing specific topics as follows:
Volume 1:
EPA. 1991. The Role of BTAGs in Ecological Assessment. Volume 1. Number 1. Office of Solid
Waste and Emergency Response. Washington, D.C. September.
This bulletin summarizes the Biological Technical Assistance Group (BTAG) structure and function in the
CERCLA process. It explains how the BTAG can assist project managers in evaluating ecological risks.
EPA. 1991. Ecological Assessment of Superfund Sites: An Overview. Volume 1. Number 2. Office
of Solid Waste and Emergency Response. Washington, B.C. December,
This bulletin provides an updated framework for ecological assessment in the Superfund (or CERCLA)
program. It describes ecological assessment components and how they fit into the remedial investigation
and feasibility study (RI/FS) process.
EPA. 1992. The Role of Natural Resource Trustees in the Superfund Process. Volume 1. Number 3.
Office of Solid Waste and Emergency Response. Washington, D.C. March.
This bulletin facilitates the working relationship between project managers involved in site cleanup and
natural resource trustees. It also helps ensure compliance with applicable or relevant and appropriate
requirements (ARAR) and increases understanding of trustee issues relevant to the CERCLA process.
EPA. 1992. Developing a Work Scope for Ecological Assessments. 1992. Volume 1. Number 4.
Office of Solid Waste and Emergency Response. Washington, D.C. May.
This bulletin helps project managers involved in site cleanup to plan and manage ecological assessments as
part of the RI/FS process under CERCLA. It includes information on project scoping, preparing
statements of work, and work plan development.
EPA. 1992. Briefing the BTAG: Initial Description of Setting, History, and Ecology of a Site.
Volume 1. Number 5. Office of Solid Waste and Emergency Response. Washington, D.C. August.
REPt RI'K'IK 1 ASKK REVISE! FINAL MASTER WPD'<>f|.RIMIIIUI7l«r.lltf24/'J7.'1 311pm uac N"2
-------
This bulletin focuses on the first opportunity (usually in the early Rf planning stage) that a project manager
has for conferring with the BTAG about possible ecological effects at a site. Pertinent information to
present to the BTAG includes the site's setting and history, constituents expected, and ecological
characteristics.
Volume 2:
EPA. 1994. Using Toxicity Tests in Ecological Risk Assessment. Volume 2. Number!. Office of
Solid Waste and Emergency Response. Washington, D.C. September.
This bulletin presents measurement endpoints in toxicity testing, elements in a toxicity assessment, and
general guidelines for selecting toxicity tests. These tests may help to determine whether concentrations of
COPECs detected in site media are high enough to cause adverse effects in organisms; demonstrate
whether constituents are bioavailable; evaluate the aggregate toxic effects of all hazardous constituents in a
medium; and evaluate the toxicity of substances whose biological effects are not well understood.
EPA. 1994. Catalogue of Standard Toxicity Tests for Ecological Risk Assessment. 1994. Volume 2.
Number 2. Office of Solid Waste and Emergency Response. Washington, D.C. September.
This bulletin consists of a list of standardized aquatic, sediment, terrestrial, and microbial toxicity tests
used at CERCLA sites. It indicates source documents that more fully describe test protocols.
EPA. 1994. Field Studies for Ecological Risk Assessment. Volume 2. Number 3. Office of Solid
Waste and Emergency Response. Washington, D.C. September.
This bulletin addresses ecological field studies as part of the ecological risk assessment that occur in the
area of ecological concern at a site. It covers important considerations such as the organisms to be
evaluated in the field study and elements in the design of a field study. A catalogue of field methods is
also presented.
EPA. 1994. Selecting and Using Reference Information in Superfund Ecological Risk Assessments.
Volume 2. Number 4. Office of Solid Waste and Emergency Response. Washington, B.C.
September.
This bulletin summarizes methods for identifying and using reference information sources such as existing
relevant data, mathematical models, and new data collected from unimpacted or reference sites.
Volume 3:
EPA. 1996. Ecological Significance and Selection of Candidate Assessment Endpoints. Volume 3.
Number 1. Office of Solid Waste and Emergency Response. Washington, D.C. January.
This bulletin helps project managers identify ecological elements at a site, estimate their relative value and
significance in the ecosystem, and identify potential assessment endpoints.
EPA. 1996. Ecotox Thresholds. Volume 3. Number 2. Office of Solid Waste and Emergency
Response. Washington, D.C. January.
N-1
S REPA RIINUXTASKX-REVISE2-JF1NAL MASTER WPDt|<|-R|iMHXti7
-------
This bulletin describes how ecotoxicological values are to be used for screening purposes in the Superfund
ecological risk assessment process. It summarizes the methodologies used to calculate ecotox thresholds
for each medium and discusses limitations of using ecotox thresholds.
REGION-SPECIFIC U.S. ENVIRONMENTAL PROTECTION AGENCY GUIDANCE
Region 1
EPA. 1989. Supplemental Risk Assessment Guidance for the Superfund Program - Part 2: Guidance
for Ecological Risk Assessments. EPA/901/5-69/01. Risk Assessment Work Group. Boston,
Massachusetts.
This manual reviews elements of the ecological risk assessment process and the types of data needed to
evaluate risks to ecological receptors.
Region 3
EPA. 1994. Interim Ecological Risk Assessment Guidelines. 9107-4431. Hazardous Waste
Management Division. Philadelphia, Pennsylvania.
This draft document briefly summarizes three levels of ecological risk assessment: the screening level, the
semiquantitative level, and the quantitative level.
Region 4
EPA. 1995. Supplemental Guidance to RAGS: Region 4 Bulletins - Ecological Risk Assessment.
Draft. Office of Health Assessment. Atlanta, Georgia. November.
This draft document includes bulletins that discuss preliminary risk evaluation or screening-level
ecological risk assessment, ecological screening values used to identify COPECs in the screening-level
ecological risk assessment, assessment and measurement endpoint selection based on the results of the
screening-level ecological risk assessment, and inclusion of natural resource trustees in the CERCLA
process.
Region 5
EPA. 1992. Regional Guidance for Conducting Ecological Assessments. Draft final. Chicago,
Illinois. April.
This reference supplements existing CERCLA guidelines; it outlines a framework for conducting
ecological assessments at Superfund sites.
EPA. 1994. Ecological Risk Assessment Guidance for RCRA Corrective Action. Interim draft.
Waste Management Division. Chicago, Illinois. July.
This is a draft document that discusses ecological risk assessment and how to effectively integrate risk
assessment and corrective action processes.
. REP* MUM* TASKS REVISE!/INA1.1M ASTER WPD>15I.Rlflf>i»IMM
-------
Region 7
EPA. 1995. Assessing Ecological Risk at RCRA Hazardous Waste Treatment, Storage, and Disposal
Facilities: A Project Manager's Guide. Kansas City, Kansas. June.
This guide is design to help project managers and risk managers that are unfamiliar with ecological
concepts used in ecological risk assessment. The focus is on evaluating ecological risks posed by RCRA
treatment, storage, and disposal (TSD) facilities in Region 7. The guide relies primarily on CERCLA
guidance to provide the framework for ecological risk assessment at RCRA TSD facilities in Region 7.
Region 8
EPA. 1994. Operation of the Ecological Technical Assistance Group (ETAG) for EPA Region VIII
Ecological Risk Assessments. ER-01. Hazardous Waste Management Division. September.
This technical guidance describes the goals of the ETAG (or BTAG) for CERCLA ecological risk
assessments in Region 8. ETAGs and BTAGs help EPA achieve better and more consistent ecological risk
assessments at CERCLA sites.
Region 10
EPA. 1995. Supplemental Risk Assessment Guidance for Superfund. Draft. Seattle, Washington.
December.
This draft guidance incorporates CERCLA human health and ecological risk assessment guidance. It
summarizes important concepts for the agency-wide guidance, highlights steps of the CERCLA RI/FS
process in which risk assessors need to be involved, and identifies specific deliverables required by EPA
Region 10 during the development of baseline risk assessments.
DEPARTMENT OF ENERGY GUIDANCE
U.S. Department of Energy (DOE). 1996. Screening Benchmarks for Ecological Risk Assessment.
Version 1.5 (computer database program). Oak Ridge National Laboratory. Oak Midge,
Tennessee. January.
Note: Screening benchmarks approved for use at some DOE facilities are formatted in a computer
database format. The database provides screening values for aquatic biota, wildlife, terrestrial
plants, sediment-associated organisms, and soil and litter organisms. The sources and derivation
of these screening values are presented in the following DOE documents.
S uter II, G.W. and J.B. Mabrey. 1994. Toxicological Benchmarks for Screening of Potential
Contaminants of Concern for Effects on Aquatic Biota on Oak Ridge Reservation: 1994 Revision.
ES/ER/TM-96. Oak Ridge National Laboratory. Oak Ridge, Tennessee.
Opresko, D.M., B.E. Sample, and G.W. Suter II. 1995. Toxicological Benchmarks for Wildlife: 1995
Revision. ES/ER/TM-86/R2. Oak Ridge National Laboratory. Oak Ridge, Tennessee.
S REPA.R MkMS TASKK REVISE! FINAL MASTER WPD' f 51-R |
-------
Will, M.E. and G.W. Suter II. 1995. Toxicological Benchmarks for Screening Potential Contaminants of
Concern for Effects on Terrestrial Plants: 1995 Revision. ES/ER/TM-8S-R2. Oak Ridge National
Laboratory. Oak Ridge, Tennessee.
Suter II, G.W., and R.N. Hull. 1994. Toxicological Benchmarks for Screening Contaminants of Potential
Concern for Effects on Sediment-Associated Biota: J994 Revision. ES/ER/TM-95/R1. Oak Ridge
National Laboratory. Oak Ridge, Tennessee.
Will, M.E. and G.W. Suter II. 1995. Toxicological Benchmarks for Potential Contaminants of Concern
for Effects on Soil and Litter Invertebrates and Heterotrophic Process. ES/ER/TM-126/R1. Oak Ridge
National Laboratory. Oak Ridge, Tennessee.
Barnthouse, L.W., et al. 1992. Survey of Ecological Risk Assessment at DOE Facilities. NTIS
Accession Number: DE93OO0972/XAB. Oak Ridge National Laboratory. Oak Midge, Tennessee.
This document is a survey of ecological risk assessment procedures at a subset of major DOE facilities.
The survey identifies ecological risk assessment approaches used by DOE and its contractors. It
documents lessons learned with these approaches. The survey identifies new technical developments and
approaches that may apply to DOE facilities. The report also identifies major data needs, data resources,
and methodological deficiencies.
Suter II, G.W. 1994. Approach and Strategy for Performing Ecological Risk Assessments for the U.S.
Department of Energy's Oak Ridge Reservation. ES/ER/TM-33/R1. Oak Ridge National
Laboratory. Oak Ridge, Tennessee. August.
This document includes guidelines on developing conceptual models for the ecological risk assessment
process, selecting assessment and measurement endpoints, specific data requirements, and risk
characterization.
DEPARTMENT OF DEFENSE GUIDANCE
LaPoint, T.W., M. Simini, J.D. Fiorian, Jr., and R.S. WentseL 1995. Procedural Guidelines for
Ecological Risk Assessments at U.S. Army Sites - Volume 2: Research and Biomonitoring Methods for
the Characterization of Ecological Effects. Report Number: ERDEC-TR-221. Aberdeen Proving
Ground, Maryland. February.
Volume 2 contains information about more than 100 environmental models and test methods used in
ecological risk assessment. The .methodologies are designed to assist risk assessors in selecting appropriate
models and tests that are relevant to ecological hypotheses and goals of RIs and FSs.
Wentsel, R.S., T.W. LaPoint, M. Simini, D. Ludwig, and L. Brewer. 1994. Procedural Guidelines
for Ecological Risk Assessments at U.S. Army Sites - Volume 1. Report Number: ERDEC-TR-221.
Aberdeen Proving Ground, Maryland. December.
Volume 1 provides ecological risk assessment guidance for U.S. Army National Priority sites and sites
listed under the Base Realignment and Closure program. This report provides an enhanced understanding
of CERCLA guidance, cost-effective and tiered procedures, and a conceptual framework to standardize
•» RFPA RiiKMX TASKKVREV1SE2'JF]NAJ,\M ASTER WPD' 151-RHVJ|Xil70m|lltt4/V7n
-------
ecological risk assessments at U.S. Army facilities. The conceptual framework is based on EPA's
agency-wide framework.
NATIONAL OCEANIC and ATMOSPHERIC ADMINISTRATION GUIDANCE
Long, E.R. and L.G. Morgan. 1990. The Potential for Biological Effects of Sediment-Sorbed
Contaminants Tested in the National Status and Trends Program. Technical memorandum NOS
OMA 52. March.
This report assesses the potential for adverse biological effects through exposure of biota to hazardous
constituents in sediments sampled and analyzed under the National Status and Trends Program.
Guidelines for assessing potential effects are included and were developed from data assembled for a
variety of approaches and many geographic areas.
Long, E.R. 1992. Ranges in chemical concentrations in Sediments Associated with Adverse Biological
Effects. Marine Pollution Bulletin 24. Volume 1. Pages 38-45.
Data derived from many geographic regions, methods, and approaches are evaluated in this paper to
identify the ranges in chemical concentrations associated with adverse biological effects. Data from three
basic approaches to determining health-based criteria were evaluated: the equilibrium partitioning
approach, the spiked-sediment bioassay approach, and various methods of evaluating biological and
chemical data collected during field surveys.
Long, E.R. et al. 1995. "Incidence of Adverse Biological Effects Within Ranges of Chemical
Concentrations in Marine and Estuarine Sediments." Environmental Management. 19(1): 81-97.
This paper presents effects range-low (ER-L) and effects range-median (ER-M) guideline values for
selected chemicals, based on biological and chemical data compiled from numerous modeling, laboratory,
and field studies performed in marine and estuarine sediments. The incidence of adverse effects was
quantified within 3 separate concentration ranges for the selected chemicals.
AMERICAN SOCIETY for TESTING and MATERIALS GUIDANCE
American Society for Testing and Materials (ASTM). 1996. Standard Guide for Selecting and Using
Ecological Endpoints for Contaminated Sites. Draft. ASTM subcommittee E-47.13. January.
This guide presents an approach to identifying, selecting, and using assessment and measurement
endpoints that may be affected by direct or indirect chemical and nonchemical stressors associated with
wastes and contaminated media at specific sites under current and future land uses. NOTE: According to
ASTM, this document is a draft and not an ASTM standard. As such, it has not been approved by ASTM.
The final approved version of this draft may or may not include the same information as it appears in this
draft. For more information on this draft, contact ASTM at 100 Barr Harbor Drive, West Conshohocken,
Pennsylvania 19428.
ASTM. 1995. Standard Guide for Developing Conceptual Site Models for Contaminated Sites.
STP E-1689. ASTM. Philadelphia, Pennsylvania.
N--7
S REPA RltatlXTASM REV!SE2-FINAL\MASTER WPD:nhRI(KilXii7iNlv|li/24fMl 1 2Upin.»e 1 ^ '
-------
This guide assists in the development of conceptual site models used to integrate technical information,
support sampling design elements such as identifying data needs and data collection activities, and
evaluate risk to human health and the environment posed by a contaminated site.
Gorsuch, J.W., J. Dwyer, C. Ingersoll, and T.W. LaPoint, editors. 1993. Environmental Toxicology
and Risk Assessment - Volume 2. ASTM STP 1216. ISBN 0-8031-1485-0. Philadelphia,
Pennsylvania.
This document presents 48 papers on new research techniques, findings concerning various environmental
stressors, and the application of techniques and processes of environmental assessment. Specific topics
include aquatic toxicology and the use of experimental ecosystems, plants for toxicity assessments, and
sediment toxicology.
Hughes, J.S., G.R. Biddinger, and E. Mones, editors. 1995. Environmental Toxicology and Risk
Assessment - Volume 3. ASTM STP 1218. ISBN 0-8031-1485-0. Philadelphia, Pennsylvania.
This text provides a comprehensive overview of the current status of ecological risk assessment and
suggested advances. Also, 22 papers are presented on topics such as models in ecological risk assessment,
ecotoxicology and the measurement of ecological effects at various sites, fate and effects of chemicals, and
the development and refinement of new methods to evaluate exposure and toxicity.
Land is, W.G., J.S. Hughes, and M.A. Lewis, editors. 1993. Environmental Toxicology and Risk
Assessment. ASTM STP 1179. ISBN 0-8931-1860-0. ASTM. Philadelphia, Pennsylvania.
This document presents 28 papers addressing such topics as evaluating ecological impacts at the
population and community levels, biomarkers, and marine toxicity test methods and methods development.
Other topics include evaluating regulatory concerns, basic research, risk and hazard assessment, and
methods development in environmental toxicology.
CANADIAN GUIDANCE
Canadian Council of Ministers of the Environment (CCME). 1996. A Framework for Ecological
Risk Assessment-General Guidance. National Contaminated Sites Remediation Program. March.
This document provides general guidance for using the framework for ecological risk assessment at
contaminated sites in Canada. Topics include planning ecological risk assessments, screening-level
assessments, and preliminary and detailed quantitative ecological risk assessments.
CCME. 1996. A Protocol for the Derivation of Environmental and Human Health Soil Quality
Guidelines. National Contaminated Sites Remediation Program. March.
This document includes rationale and guidance for developing environmental and human health soil
quality guidelines for contaminated sites in Canada. Topics include derivation of environmental soil
quality guidelines, relevant endpoints for deriving soil quality guidelines, potential ecological receptors and
exposure pathways of soil contamination, and uncertainties in guidelines derivation.
CCME. 1995. Protocol for the Derivation of Canadian Sediment Quality Guidelines for the Protection
of Aquatic Life. CCME EPC-98E. Environment Canada. Ottawa, Canada. March.
S REPA RKKIIX TASkK REV1SE2 FINAL MASTER WPD »5l-RK»MK«l?i*l'10/24^7 1
N-8
-------
These guidelines can be used to assess sediment quality, help set goals for sediment quality that will
sustain aquatic system health for the long term, and develop site-specific objectives. The document
outlines procedures for deriving scientifically defensible sediment quality guidelines for the protection of
aquatic life. The document also discusses the use of sediment quality guidelines as benchmarks, the
National Status and Trends Program (NSTP) approach, the spiked-sediment toxicity approach, and
derivation of safety factors.
Gaudet, C. 1994. A Framework for Ecological Risk Assessment at Contaminated Sites in Canada:
Review and Recommendations. Scientific Series Number 199. Environment Canada. Ottawa,
Canada.
This document presents the ecological risk assessment framework for the National Contaminated Sites
Remediation Program. The framework includes a tiered approach to ecological risk assessment. The
document discusses ecological risk assessment components such as problem definition, exposure
assessment, receptor characterization, hazard assessment, and risk characterization.
Jaagumagi, R. 1993. Development of the Ontario Provincial Sediment Quality Guidelines for Arsenic,
Cadmium, Chromium, Copper, Iron, Lead, Manganese, Mercury, Nickel, and Zinc. ISBN
0-7729-9249-5. Ontario Ministry of tie Environment. Canada. August.
This document describes the derivation of the metals guidelines and summarizes data used to derive the
values listed in the guidelines. The document also summarizes properties and fate of the metals, describes
the forms in which metals can exist in sediments, and provides details of the calculations used to arrive at
the sediment quality guidelines.
Jaagumagi, R. 1994. Development of the Ontario Provincial Sediment Quality Guidelines for
Poly cyclic Aromatic Hydrocarbons. ISBN 0-7778-1710-1. Ontario Ministry of the Environment.
Canada. January.
This document describes the derivation of the guidelines for 12 individual polycyclic aromatic
hydrocarbons (PAH) as well as total PAH and summarizes data used to derive these values. The document
also summarizes the fate of PAHs in sediments, and provides details of the calculations of the sediment
quality guidelines.
Keddy, C., J.C. Greene, and M.A. Bunnell. 1994. A Review of Whole Organism Bioassays for
Assessing the Quality of Soil, Freshwater Sediment, and Freshwater in Canada. Scientific Series
Number 198. Ecosystem Conservation Directorate. Ontario, Canada.
This report addresses application of recommended bioassays to site assessment and remediation. The
report identifies potentially suitable test methods, assesses their applicability, and recommends tests for
soil, freshwater sediment, and fresh water. The report also evaluates the future for hazardous constituent
assessment using biological organisms, including alternative test endpoints, in situ tests, and assessment
beyond whole organisms and fresh water, such as methods for assessing impacts to microbial processes
and multispecies testing.
Persaud, D., R. Jaagumagi, and A. Hayton. 1993. Guidelines for the Protection and Management of
Aquatic Sediment Quality in Ontario. ISBN 0-7729-9248-7. Ontario Minister of the Environment.
Canada. August.
N-9
5 REPA RI'tllK TASKK REV1SE2 .FINAL MASTER WPD Ht-RhKllX>l7m M> 24 'iT't 2Upnt*.ic 1 * 7
-------
The purpose of the sediment quality guidelines is to protect aquatic systems by setting safe concentrations
for metals, nutrients, and organic compounds. The guidelines help decision makers with sediment issues,
including determining which sediments are contaminated and how to effectively manage the problem. The
guidelines establish three levels of effect: no effect level, lowest effect level, and severe effect level.
ADDITIONALGMBAHCE
Kartell, S.M., R.H. Gardner, and R.V. O'Neill. 1992. Ecological Risk Estimation. ISBN
0-873711637. Lewis Publishers. Chelsea, Michigan.
This publication presents an approach to estimating ecological risks using laboratory toxicity data to
predict ecological consequences of toxic chemicals. The text includes discussions of the following
subjects: toxicological and ecological data for risk analysis, modeling aquatic ecosystems, modeling
sublethal toxic effects, predicting risks, evaluating predictive methodology, and comparisons of predicted
and measured effects.
Burmaster, D.E. and P.D. Anderson. 1994. "Principles of Good Practice for the Use of Monte
Carlo Techniques in Human Health and Ecological Risk Assessment." Risk Analysis. Volume 14.
Pages 477-481.
This paper proposes 14 principles of good practice in performing and reviewing probabilistic or Monte
Carlo risk assessments of toxic chemicals in the environment. Monte Carlo risk assessments that follow
these principles will be easier to understand, will explicitly distinguish assumptions from data, and will
consider and quantify effects that could otherwise lead to misinterpretation of the results.
Calabrese, E.J. and L.A. Baldwin. 1993. Performing Ecological Risk Assessments. Lewis
Publishers. Chelsea, Michigan.
This text presents an extensive compilation that addresses components of an ecological risk assessment,
including environmental fate modeling and pharmacokinetic factors, uncertainty factors, deriving
chemical-specific and species-specific maximum acceptable tissue concentrations, and sediment quality
criteria.
Cairns, J., R.R. Niederlehner, and D.R. Orvos, editors. 1992. Predicting Ecosystem Risk. Volume
XX. Princeton Scientific Publishing Company. Princeton, New Jersey.
This text includes a series of papers on topics such as predicting ecological risks posed by changes in
hydrological regime, forest management, genetically engineered microorganisms and products, highways,
and radioactive materials. The use of experimental stream mesocosms in assessing risks is also discussed.
Cardwell, R., et al. 1991. Aquatic Risk: An Assessment Report. Evaluation of the Protocols for
Aquatic Ecological Risk Assessment. Water Environment Research Foundation Order Number:
D0005. Alexandria, Virginia.
This report evaluates aquatic and ecological risk models. Additional research is identified to help develop
a conceptual framework to integrate, organize, and validate aquatic ecological risk assessment protocols,
and to apply these protocols to water quality issues.
S REPA.RIiki|H TASKS REV35E2\FINAL'MASTER WFD' I5i-RE(XliMt7
-------
Clifford, P.A., D.E. Barchers, D.F. Ludwig, R.L. Sielken, J.S. Klingensmith, R.V. Graham, and M.I.
Bantom. 1995. "An Approach to Quantifying Spatial Components of Exposure for Ecological Risk
Assessment." Environmental Toxicology and Chemistry. Volume 14. Pages 895-906.
This paper presents an approach to quantifying spatial components of exposure using the Geographic
Information System (GIS). G1S is used to estimate spatially weighted exposure concentrations within an
organism's foraging or exposure ranges. GIS is also used for comparing exposure concentrations to
benchmark concentrations and presenting site-specific results in a three-dimensional format to effectively
present site-specific quantified ecological risks and to provide an effective risk management
decision-making tool.
International Atomic Energy Agency. 1992. Effects of Ionizing Radiation on Plants and Animals at
Levels Implied by Current Radiation Protection Standards. Technical Report Series Number 332.
ISBN 92-0-100992-5. Vienna, Austria.
This report addresses potential effects on plant and animal populations through chronic releases of
radionuclides. The report includes reviews of available information on the effects of ionizing radiation on
natural organisms and determines the doses above which there are adverse effects to plants and animals.
The report also establishes whether on not plant and animal populations are adequately protected under
radiation protections standards for humans.
Landis, W.G., G.B. Matthews, R.A. Matthews, and A. Sergeant. 1994. "Application of Multivariate
Techniques to Endpoint Determination, Selection, and Evaluation in Ecological Risk Assessment."
Environmental Toxicology and Chemistry. Volume 13. Pages 1917 - 1927.
The paper reviews the role of ecological endpoints and introduces and discusses the ramifications of
multivariate analysis on the assessment of risk to ecological systems. Three methods are discussed in the
paper: the mean strain measurement, the state-space analysis, and the nonmetric clustering method.
MacDonald, D.D. 1994. Approach to the Assessment of Sediment Quality in Florida Coastal Waters,
Volume 1 - Development and Evaluation of Sediment Quality Assessment Guidelines. MacDonald
Environmental Sciences, Ltd. British Columbia, Canada. November.
This report recommends a scientifically defensible framework for assessing the biological significance of
sediment-associated hazardous constituents. Numerical sediment quality assessment guidelines (SQAG)
provide the basis for assessing potential effects of sediment-associated constituents. The report reviews a
variety of approaches and recommends an integrated strategy and relevant assessment tools. The SQAGs
are derived from a variety of sediment quality data and are based on a weight-of-evidence approach that
links constituent concentrations with adverse biological effects.
MacDonald, D.D. 1994. Approach to the Assessment of Sediment Quality in Florida Coastal Waters,
Volume 2 - Application of the Sediment Quality Assessment Guidelines. MacDonald Environmental
Sciences, Ltd. British Columbia, Canada. November.
This document assists potential users in applying SQAGs and other relevant sediment quality assessment
tools. The report lists applications of SQAGs that are considered inappropriate and presents a framework
for assessing the significance of sediment-associated constituents.
N-l 1
S REPA RHHUK TASKS REMSE2.FISALMASTER WPD1l*l.RH«lWi7lli>\|ii/24«7/l jDpmase 1 ^ 1 1
-------
Macintosh, D.L., G.W. Suter II, and F.O. Hoffman. 1994. "Use of Probabilistic Exposure Models
in Ecological Risk Assessments of Contaminated Sites." Risk Analysis. Volume 14. Pages 405 - 419.
This paper discusses the use stochastic food web models in ecological risk assessment, particularly in
estimating exposure to endpoint species as well as subsequent effects and determining cleanup levels by
estimating concentrations in environmental media that will not cause significant adverse effects in endpoint
species.
Maughan, J.T. 1993. Ecological Assessment of Hazardous Waste Sites. ISBN 0-442-01091-5. Van
Nostrand Reinhold. New York, New York.
Essential technical and regulatory information necessary to plan, prepare, and implement an ecological risk
assessment is presented in this text. The ecological risk assessment process is examined along with
techniques for evaluating three important components of ecological risk assessments: terrestrial pathways
of constituents, sediment quality and contamination, and toxicity testing.
McCarthy, J.F. and L.R. Shugart, editors. 1990. Biomarkers of Environmental Contamination.
ISBN 0-87371-284-6. Lewis Publishers. Boca Raton, Florida.
This text provides and introduction and review of the research on biological markers in plants as well as
animals and provides an approach to evaluating ecological and health effects of environmental
contamination. The focus is on the development, application, and validation of biological markers as
indicators of exposure to toxic chemicals or as predictors of the averse consequences of that exposure.
Peterle, TJf. 1991. Wildlife Toxicology. Van Nostrand Reinhold. New York, New York.
This text provides information on environmental pollution as it affects wildlife. The text covers relevant
laws and regulations, materials found in environmental pollutants, transport and distribution in natural
systems, accumulation in organisms, lethal and chronic effects on organisms, and effects on ecosystems.
Suter II, G.W. 1993. Ecological Risk Assessment. ISBN 0-87371-875-5. Lewis Publishers. Chelsea,
Michigan.
This text emphasizes risks to aquatic systems because of the preponderance of data and modeling
techniques for aquatic systems. There are also more ecological risk assessments that address aquatic rather
than terrestrial systems. Predictive risk assessments are the main focus because the risk assessment
paradigm is based on predictive assessments. A chapter in the text is dedicated to retrospective
assessments. Chemical, physical, and biological stressors are discussed in the text. Effects of exposure to
chemicals and fate and transport of chemicals are discussed. This text also refers to other literature
concerning ecological risk assessment.
Suter II, G.W. 1990. "Endpoints for Regional Ecological Risk Assessments." Environmental
Management. Volume 14. Pages 19 - 23.
This article distinguishes between assessment and measurement endpoints in terms of their roles in
ecological risk assessment. Topics include endpoint selection criteria and regional ecosystem effects.
7fXn|IM24/V7/1 2i)pmvsae
N-12
-------
ATTACHMENT O
EPA GUIDANCE AND POLICY FOR PROBABILISTIC RISK ASSESSMENT
-------
I j r%j nwu'
win., j
POLICY FOR USE OF PROBABILISTIC ANALYSIS IN
RISK ASSESSMENT
at the U.S. Environmental Protection Agency
May 15, 1997
Guiding Principles for Monte Carlo Analysts (EPA/630./R-97/001)
INTRODUCTION
The importance of adequately characterizing variability and uncertainty in risk assessments has been
emphasized in several science and policy documents. These include the 1992 U.S. Environmental
Protection Agency (EPA) Exposure Assessment Guidelines, the 1992 EPA Risk Assessment Council (RAC)
Guidance, the 1995 EPA Policy for Risk Characterization, the EPA Proposed Guidelines for Ecological Risk
Assessment, the EPA Region 3 Technical Guidance Manual on Risk Assessment, the EPA Region 8
Superfund Technical Guidance, the 1994 National Academy of Sciences "Science and Judgment in Risk
Assessment," and the report by the Commission on Risk Assessment and Risk Management. As part of the
implementation of the recommendations contained in these reports, the Agency is issuing guidance on the
appropriate use of an application for analyzing variability and uncertainty in Agency risk assessments.
This policy and the guiding principles attached are designed to support the use of various techniques for
characterizing variability and uncertainty. Further, the policy defines a set of Conditions for Acceptance.
These conditions are important for ensuring good scientific practice in quantifying uncertainty and variability.
In accordance with EPA's 1995 Policy for Risk Characterization, this policy also emphasizes the importance
of clarity, transparency, reasonableness, and consistency in risk assessments.
There are a variety of different methods for characterizing uncertainty and variability. These methods cover
a broad range of complexity from the simple comparison of discrete points to probabilistic techniques like
Monte Carlo analysis. Recently, interest in using Monte Carlo analysis for risk assessment has increased.
This method has the advantage of allowing the analyst to account for relationships between input variables
and of providing the flexibility to investigate the effects of different modeling assumptions. Experience has
shown that to benefit fully from the advantages of such probabilistic techniques as Monte Carlo analysis,
certain standards of practice are to be observed. The Agency is issuing, therefore, this policy statement and
associated guiding principles. While Monte Carlo analysis is the most frequently encountered probabilistic
tool for analyzing variability and uncertainty in risk assessments, the intent of this policy is not to indicate
that Monte Carlo analysis is the only acceptable approach for Agency risk assessments. The spirit of this
policy and the Conditions for Acceptance described herein are equally applicable to other methods for
analyzing variability and uncertainty.
POLICY STATEMENT
05/27/97 03:56:03
-------
EPA/ORD/NCEA • Policy For Use Of Probabilistic Analysis in Risk Assessment
http://www.epa.gov/nceaxncpolicy_htm
It is the policy of the U.S. Environmental Protection Agency that such probabilistic analysis techniques as
Monte Carlo analysis, given adequate supporting data and credible assumptions, can be viable statistical
tools for analyzing variability and uncertainty in risk assessments. As such, and provided that the conditions
described below are met, risk assessments using Monte Carlo analysis or other probabilistic techniques will
be evaluated and utilized in a manner that is consistent with other risk assessments submitted to the Agency
for review or consideration. It is not the intent of this policy to recommend that probabilistic analysts be
conducted for all risk assessments supporting risk management decisions. Such analysis should be a part of
a tiered approach to risk assessment that progresses from simpler (e.g., deterministic) to more complex
(e.g., probabilistic) analyses as the risk management situation requires. Use of Monte Carlo or other such
techniques in risk assessments shall not be cause, perse, for rejection of the risk assessment by the
Agency. For human health risk assessments, the application of Monte Carlo and other probabilistic
techniques has been limited to exposure assessments in the majority of cases. The current policy,
Conditions for Acceptance and associated guiding principles are not intended to apply to dose response
evaluations for human health risk assessment until this application of probabilistic analysis has been studied
further. In the case of ecological risk assessment, however, this policy applies to all aspects including
stressor and dose-response assessment.
CONDITIONS FOR ACCEPTANCE
When risk assessments using probabilistic analysis techniques (including Monte Carlo analysis) are submitted
to the Agency for review and evaluation, the following conditions are to be satisfied to ensure high quality
science. These conditions, related to the good scientific practices of transparency, reproducibility, and the
use of sound methods, are summarized here and explained more fully in the Attachment, "Guiding Principles
for Monte Carlo Analysis."
1. The purpose and scope of the assessment should be clearly articulated in a "problem formulation"
section that includes a full discussion of any highly exposed or highly susceptible subpopulations
evaluated (e.g., children, the elderly). The questions the assessment attempts to answer are to be
discussed and the assessment endpoints are to be well defined.
2. The methods used for the analysis (including all models used, all data upon which the assessment is
based, and all assumptions that have a significant impact upon the results) are to be documented and
easily located in the report. This documentation is to include a discussion of the degree to which the
data used are representative of the population under study. Also, this documentation is to include the
names of the models and software used to generate the analysis, Sufficient information is to be
provided to allow the results of the analysis to be independently reproduced.
3. The results of sensitivity analyses are to be presented and discussed in the report. Probabilistic
techniques should be applied to the compounds, pathways, and factors of importance to the
assessment, as determined by sensitivity analyses or other basic requirements of the assessment.
4. The presence or absence of moderate to strong correlations or dependencies between the input
variables is to be discussed and accounted for in the analysis, along with the effects these have on
the output distribution.
5. Information for each input and output distribution is to be provided in the report. This includes tabular
and graphical representations of the distributions (e.g., probability density function and cumulative
distribution function plots) that indicate the location of any point estimates of interest (e.g., mean,
median, 95th percentile). The selection of distributions is to be explained and justified. For both the
input and output distributions, variability and uncertainty are to be differentiated where possible.
6. The numerical stability of the central tendency and the higher end (i.e.. tail) of the output distributions
are to be presented and discussed.
7. Calculations of exposures and risks using deterministic (e.g., point estimate) methods are to be
reported if possible. Providing these values will allow comparisons between the probabilistic analysis
and past or screening level risk assessments. Further, deterministic estimates may be used to
answer scenario specific questions and to facilitate risk communication. When comparisons are
2 of 4
05/27'97 09 56 03
-------
UTSUf • runcy rui useui n uuauuisuu ns sestjf sis in i\ian njjcjji i
answer scenario specific questions and to facilitate risk communication, When comparisons are
made, it is important to explain the similarities and differences in the underlying data, assumptions,
and models.
8. Since fixed exposure assumptions (e.g., exposure duration, body weight) are sometimes embedded in
the toxicity metrics (e.g., Reference Doses, Reference Concentrations, unit cancer risk factors), the
exposure estimates from the probabilistic output distribution are to be aligned with the toxicity metric.
LEGAL EFFECT
This policy and associated guidance on probabilistic analysis techniques do not establish or affect legal
rights or obligations. Rather, they confirm the Agency position that probabilistic techniques can be viable
statistical tools for analyzing variability and uncertainty in some risk assessments. Further, they outline
relevant Conditions for Acceptance and identify factors Agency staff should consider in implementing the
policy.
The policy and associated guidance do not stand alone; nor do they establish a binding norm that is finally
determinative of the issues addressed. Except where otherwise provided by law, the Agency's decision on
conducting a risk assessment in any particular case is within the Agency's discretion. Variations in the
application of the policy and associated guidance, therefore, are not a legitimate basis for delaying action on
Agency decisions.
IMPLEMENTATION
Assistant Administrators and Regional Administrators are responsible for implementation of this policy within
their organizational units. The implementation strategy is divided into immediate and follow-up activities.
Immediate Activities
To assist EPA program and regional offices with this implementation, initial guidance on the use of one
probabilistic analysis tool, Monte Carlo analysis, is provided in the Attachment, "Guiding Principles for Monte
Carlo Analysis" (EPA/630/R-97/001), The focus of this guidance is on Monte Carlo analysis because it is
the most frequently encountered technique in human health risk assessments. Additional information may be
found in the "Summary Report for the Workshop on Monte Carlo Analysis" (EPA/630/R-96/010). This report
summarizes discussions held during the May 1996 Risk Assessment Forum sponsored workshop that
involved leading experts in Monte Carlo analysis.
Follow-Up Activities
To prepare for the use and evaluation of probabilistic analysis methods, including Monte Carlo analysis,
within the next year, EPA's Risk Assessment Forum (RAF) will develop illustrative case studies for use as
guidance and training tools. Further, the RAF will organize workshops or colloquia to facilitate the
development of distributions for selected exposure factors. EPA's National Center for Environmental
Assessment (NCEA) will develop an Agency training course on probabilistic analysis methods, including
Monte Carlo analysis for both risk assessors and risk managers which will become available during Fiscal
Year (FY) 1997 or FY 1998. Also. NCEA will develop detailed technical guidance for the quantitative analysis
of variability and uncertainty.
In the longer term, various Regions, Programs and the Office of Research and Development (ORD) may need
to modify existing or develop new guidelines or models to facilitate use of such techniques as Monte Carlo
analysis. Also, the NCEA will revise or update the Exposure Factors Handbook to include distributional
3 ol 4
05/27/97 03:56:03
-------
EPA/ORD/NCEA • Policy For Use Of Probabilistic Analysis in Risk Assessment
http://www.epa.gov/ncea/mcpolicy.htm
information. ORD's National Exposure Research Laboratory
(NERL) has formed a modeling group that may provide assessment and analysis advice to Program and
Regional Offices. The issue of using probabilistic techniques, including Monte Carlo analysis in the dose
response portion of human health risk assessments requires further study. NCEA will conduct research in
this area and additional guidance will be provided if necessary.
Fred Hansen
Deputy Administrator
Cwawiewts
CWfD ' SPA
Home f aqe rtortePagc Mo*te
acca#
5§iiM„£siiiEii§iits I sifumsli. 1
11 .Page
Last Revised: May 21, 1937
URL: http://www epa.gov/ncea/mcpoiicy.htm
4 of 4
05/27/97 09 56 04
-------
EP A/630/R-97/001
March 1997
Guiding Principles for Monte Carlo
Analysis
Technical Panel
Office of Prevention, Pesticides, and Toxic Substances
Michael Firestone (Chair) Penelope Fenner-Crisp
Office of Policy, Planning, and Evaluation
Timothy Barry
Office of Solid Waste and Emergency Response
David Bennett Steven Chang
Office of Research and Development
Michael Callahan
Regional Offices
AnneMarie Burke (Region I) Jayne Michaud (Region I)
Marian Olsen (Region II) Patricia Cirone (Region X)
Science Advisory Board Staff
Donald Barnes
Risk Assessment Forum Staff
William P. Wood Steven M. Knott
Risk Assessment Forum
U S. Environmental Protection Agency
Washington, DC 20460
-------
DISCLAIMER
This document has been reviewed in accordance with U.S. Environmental Protection
Agency policy and approved for publication. Mention of trade names or commercial products
does not constitute endorsement or recommendation for use
ii
-------
TABLE OF CONTENTS
Preface iv
Introduction 1
Fundamental Goals and Challenges 3
When a Monte Carlo Analysis Might Add Value to a Quantitative Risk Assessment 5
Key Terms and Their Definitions 6
Preliminary Issues and Considerations 9
Defining the Assessment Questions 9
Selection and Development of the Conceptual and Mathematical Models , 10
Selection and Evaluation of Available Data 10
Guiding Principles for Monte Carlo Analysis 11
Selecting Input Data and Distributions for Use in Monte Carlo Analysis . 11
Evaluating Variability and Uncertainty 15
Presenting the Results of a Monte Carlo Analysis ' 17
Appendix Probability Distribution Selection Issues 22
References Cited in Text . . . 29
iii
-------
PREFACE
The U.S. Environmental Protection Agency (EPA) Risk Assessment Forum was
established to promote scientific consensus on risk assessment issues and to ensure that this
consensus is incorporated into appropriate risk assessment guidance To accomplish this, the Risk
Assessment Forum assembles experts throughout EPA in a formal process to study and report on
these issues from an Agency-wide perspective. For major risk assessment activities, the Risk
Assessment Forum has established Technical Panels to conduct scientific reviews and analyses.
Members are chosen to assure that necessary technical expertise is available
This report is part of a continuing effort to develop guidance covering the use of
probabilistic techniques in Agency risk assessments. This report draws heavily on the
recommendations from a May 1996 workshop organized by the Risk Assessment Forum that
convened experts and practitioners in the use of Monte Carlo analysis, internal as well as external
to EPA, to discuss the issues and advance the development of guiding principles concerning how
to prepare or review an assessment based on use of Monte Carlo analysis. The conclusions and
recommendations that emerged from these discussions are summarized in the report "Summary
Report for the Workshop on Monte Carlo Analysis" (EPA/630/R-96/010) Subsequent to the
workshop, the Risk Assessment Forum organized a Technical Panel to consider the workshop
recommendations and to develop an initial set of principles to guide Agency risk assessors in the
use of probabilistic analysis tools including Monte Carlo analysis It is anticipated that there will
be need for further expansion and revision of these guiding principles as Agency risk assessors
gain experience in their application.
IV
-------
Introduction
The importance of adequately characterizing variability and uncertainty in fate, transport,
exposure, and dose-response assessments for human health and ecological risk assessments has
been emphasized in several U.S. Environmental Protection Agency (EPA) documents and
activities. These include:
• the 1986 Risk Assessment Guidelines;
• the 1992 Risk Assessment Council (RAC) Guidance (the Habicht memorandum);
• the 1992 Exposure Assessment Guidelines; and
• the 1995 Policy for Risk Characterization (the Browner memorandum).
As a follow up to these activities EPA is issuing this policy and preliminary guidance on
using probabilistic analysis The policy documents the EPA's position "that such probabilistic
analysis techniques as Monte Carlo analysis, given adequate supporting data and credible
assumptions, can be viable statistical tools for analyzing variability and uncertainty in risk
assessments." The policy establishes conditions that are to be satisfied by risk assessments that
use probabilistic techniques. These conditions relate to the good scientific practices of clarity,
consistency, transparency, reproducibility, and the use of sound methods.
The EPA policy lists the following conditions for an acceptable risk assessment that uses
probabilistic analysis techniques. These conditions were derived from principles that are
presented Iater in this document and its Appendix. Therefore, after each condition, the relevant
principles are noted
I. The purpose and scope of the assessment should be clearly articulated in a "problem
formulation" section that includes a full discussion of any highly exposed or highly
susceptible subpopulations evaluated (e.g., children, the elderly, etc.). The questions
the assessment attempts to answer are to be discussed and the assessment endpoints
are to be well defined
2 The methods used for the analysis (including all models used, all data upon which the
assessment is based, and all assumptions that have a significant impact upon the
results) are to be documented and easily located in the report. This documentation is
1
-------
to include a discussion of the degree to which the data used are representative of the
population under study. Also, this documentation is to include the names of the
models and software used to generate the analysis Sufficient information is to be
provided to allow the results of the analysis to be independently reproduced.
(Principles 4, 5, 6, and 11)
3. The results of sensitivity analyses are to be presented and discussed in the report
Probabilistic techniques should be applied to the compounds, pathways, and factors of
importance to the assessment, as determined by sensitivity analyses or other basic
requirements of the assessment. (Principles 1 and 2)
4. The presence or absence of moderate to strong correlations or dependencies between
the input variables is to be discussed and accounted for in the analysis, along with the
effects these have on the output distribution. (Principles 1 and 14)
5. Information for each input and output distribution is to be provided in the report This
includes tabular and graphical representations of the distributions (e.g., probability
density function and cumulative distribution function plots) that indicate the location
of any point estimates of interest (e.g., mean, median, 95lh percentile) The selection
¦ of distributions is to be explained and justified For both the input and output
distributions, variability and uncertainty are to be differentiated where possible
(Principles 3, 7, 8, 10, 12, and 13)
6. The numerical stability of the central tendency and the higher end (i.e., tail) of the
output distributions are to be presented and discussed. (Principle 9)
7. Calculations of exposures and risks using deterministic (e.g., point estimate) methods
are to be reported if possible. Providing these values will allow comparisons between
the probabilistic analysis and past or screening level risk assessments Further,
deterministic estimates may be used to answer scenario specific questions and to
facilitate risk communication. When comparisons are made, it is important to explain
the similarities and differences in the underlying data, assumptions, and models
(Principle 15).
o
-------
8. Since fixed exposure assumptions (e.g., exposure duration, body weight) are
sometimes embedded in the toxicity metrics (e.g., Reference Doses, Reference
Concentrations, unit cancer risk factors), the exposure estimates from the probabilistic
output distribution are to be aligned with the toxicity metric.
The following sections present a general framework and broad set of principles important
for ensuring good scientific practices in the use of Monte Carlo analysis (a frequently encountered
tool for evaluating uncertainty and variability). Many of the principles apply generally to the
various techniques for conducting quantitative analyses of variability and uncertainty; however,
the focus of the following principles is on Monte Carlo analysis. EPA recognizes that quantitative
risk assessment methods and quantitative variability and uncertainty analysis are undergoing rapid
development. These guiding principles are intended to serve as a minimum set of principles and
are not intended to constrain or prevent the use of new or innovative improvements where
scientifically defensible.
Fundamental Goals and Challenges
In the context of this policy, the basic goal of a Monte Carlo analysis is to chatacterize,
quantitatively, the uncertainty and variability in estimates of exposure or risk. A secondary goal is
to identify key sources of variability and uncertainty and to quantify the relative contribution of
these sources to the overall variance and range of model results.
Consistent with EPA principles and policies, an analysis of variability and uncertainty
should provide its audience with clear and concise information on the variability in individual
exposures and risks; it should provide information on population risk (extent of harm in the
exposed population), it should provide information on the distribution of exposures and risks to
highly exposed or highly susceptible populations; it should describe qualitatively and
quantitatively the scientific uncertainty in the models applied, the data utilized, and the specific
risk estimates that are used
Ultimately, the most important aspect of a quantitative variability and uncertainty analysis
may well be the process of interaction between the risk assessor, risk manager and other
interested parties that makes risk assessment into a dynamic rather than a static process.
Questions for the risk assessor and risk manager to consider at the initiation of a quantitative
variability and uncertainty analysis include:
J
-------
• Will the quantitative analysis of uncertainty and variability improve the risk
assessment?
• What are the major sources of variability and uncertainty? How will variability
and uncertainty be kept separate in the analysis 9
• Are there time and resources to complete a complex analysis9
• Does the project warrant this level of effort?
• Will a quantitative estimate of uncertainty improve the decision? How will the
regulatory decision be affected by this variability and uncertainty analysis 9
• What types of skills and experience are needed to perform the analysis9
• Have the weaknesses and strengths of the methods been evaluated?
• How will the variability and uncertainty analysis he communicated to the public
and decision makers?
One of the most important challenges facing the risk assessor is to communicate,
effectively, the insights an analysis of variability and uncertainty provides. It is important for the
risk assessor to remember that insights will generally be qualitative in nature even though the
models they derive from are quantitative. Insights can include:
• An appreciation of the overall degree of variability and uncertainty and the
confidence that can be placed in the analysis and its findings.
• An understanding of the key sources of variability and key sources of uncertainty •
and their impacts on the analysis.
• An understanding of the critical assumptions and their importance to the analysis
and findings.
• An understanding of the unimportant assumptions and why they are unimportant'.
• An understanding of the extent to which plausible alternative assumptions or
models could affect any conclusions
• An understanding of key scientific controversies related to the assessment and a
sense of what difference they might make regarding the conclusions.
4
-------
The risk assessor should strive to present quantitative results in a manner that will clearly
communicate the information they contain.
When a Monte Carlo Analysis Might Add Value to a
Quantitative Risk Assessment
Not every assessment requires or warrants a quantitative characterization of variability and
uncertainty. For example, it may be unnecessary to perform a Monte Carlo analysis when
screening calculations show exposures or risks to be clearly below levels of concern (and the
screening technique is known to significantly over-estimate exposure). As another example, it
may be unnecessary to perform a Monte Carlo analysis when the costs of remediation are low.
On the other hand, there may be a number of situations in which a Monte Carlo analysis
may be useful. For example, a Monte Carlo analysis may be useful when screening calculations
using conservative point estimates fall above the levels of concern. Other situations could include
when it is necessary to disclose the degree of bias associated with point estimates of exposure;
when it is necessary to rank exposures, exposure pathways, sites or contaminants; when the cost
of regulatory or remedial action is high and the exposures are marginal; or when the consequences
of simplistic exposure estimates are unacceptable.
Often, a "tiered approach" may be helpful in deciding whether or not a Monte Carlo
analysis can add value to the assessment and decision. In a tiered approach, one begins with a
fairly simple screening level model and progresses to more sophisticated and realistic (and usually
more complex) models only as warranted by the findings and value added to the decision.
Throughout each of the steps in a tiered approach, soliciting input from each of the interested
parties is recommended. Ultimately, whether or not a Monte Carlo analysis should be conducted
is a matter of judgment, based on consideration of the intended use, the importance of the
exposure assessment and the value and insights it provides to the risk assessor, risk manager, and
other affected individuals or groups.
5
-------
Key Terms and Their Definitions
The following section presents definitions for a number of key terms which are used
throughout this document.
Bayesian
The Bayesian or subjective view is that the probability of an event is the degree of belief
that a person has, given some state of knowledge, that the event will occur. In the classical or
frequentist view, the probability of an event is the frequency with which an event occurs given a
long sequence of identical and independent trials. In exposure assessment situations, directly
representative and complete data sets are rarely available, inferences in these situations are
inherently subjective. The decision as to the appropriateness of either approach (Bayesian or
Classical) is based on the available data and the extent of subjectivity deemed appropriate.
Correlation, Correlation Analysis
Correlation analysis is an investigation of the measure of statistical association among
random variables based on samples. Widely used measures include the linear correlation
coefficient (also called the product-moment correlation coefficient or Pearson's correlation
coefficient), and such non-parametric measures as Spearman rank-order correlation coefficient,
and Kendall's tan When the data are nonlinear, non-parametric correlation is generally
considered to be more robust than linear correlation.
Cumulative Distribution Function (CDF)
The CDF is alternatively referred to in the literature as the distribution function,
cumulative frequency function, or the cumulative probability function. The cumulative
distribution function, F(x), expresses the probability the random variable X assumes a value less
than or equal to some value x, F(x) = Prob (X ^ x). For continuous random variables, the
cumulative distribution function is obtained from the probability density function by integration, or
by summation in the case of discrete random variables.
Latin Hypercube Sampling
In Monte Carlo analysis, one of two sampling schemes are generally employed simple
random sampling or Latin Hypercube sampling Latin hypercube sampling may be viewed as a
stratified sampling scheme designed to ensure that the upper or lower ends of the distributions
6
-------
used in the analysis are well represented. Latin hypercube sampling is considered to be more
efficient than simple random sampling, that is, it requires fewer simulations to produce the same
level of precision. Latin hypercube sampling is generally recommended over simple random
sampling when the model is complex or when time and resource constraints are an issue.
Monte Carlo Analysis, Monte Carlo Simulation
Monte Carlo Analysis is a computer-based method of analysis developed in the 1940's that
uses statistical sampling techniques in obtaining a probabilistic approximation to the solution of a
mathematical equation or model.
Parameter
Two distinct, but often confusing, definitions for parameter are used. In the first usage
(preferred), parameter refers to the constants characterizing the probability density function or
cumulative distribution function of a random variable. For example, if the random variable W is
known to be normally distributed with mean \x and standard deviation a, the characterizing
constants \x and a are called parameters. In the second usage, parameter is defined as the
constants and independent variables which define a mathematical equation or model. For
example, in the equation Z = aX + PY, the independent variables (X,Y) and the constants (a,p)
are all parameters.
Probability Density Function (PDF)
The PDF is alternatively referred to in the literature as the probability function or the
frequency function. For continuous random variables, that is, the random variables which can
assume any value within some defined range (either finite or infinite), the probability density
function expresses the probability that the random variable falls within some very small interval.
For discrete random variables, that is, random variables which can only assume certain isolated or
fixed values, the term probability mass function (PMF) is preferred over the term probability
density function. PMF expresses the probability that the random variable takes on a specific
value
Random Variable
A random variable is a quantity which can take on any number of values but whose exact
value cannot be known before a direct observation is made. For example, the outcome of the toss
-------
of a pair of dice is a random variable, as is the height or weight of a person selected at random
from the New York City phone book.
Representativeness
Representativeness is the degree to which a sample is characteristic of the population for
which the samples are being used to make inferences.
Sensitivity, Sensitivity Analysis
Sensitivity generally refers to the variation in output of a mathematical model with respect
to changes in the values of the model's input. A sensitivity analysis attempts to provide a ranking
of the model's input assumptions with respect to their contribution to model output variability or
uncertainty. The difficulty of a sensitivity analysis increases when the underlying model is
nonlinear, nonmonotonic or when the input parameters range over several orders of magnitude
Many measures of sensitivity have been proposed. For example, the partial rank correlation
coefficient and standardized rank regression coefficient have been found to be useful. Scatter
plots of the output against each of the model inputs can be a very effective tool for identifying
sensitivities, especially when the relationships are nonlinear For simple models or for screening
purposes, the sensitivity index can be helpful
In a broader sense, sensitivity can refer to how conclusions may change if models, data, or
assessment assumptions are changed.
Simulation
In the context of Monte Carlo analysis, simulation is the process of approximating the
output of a model through repetitive random application of a model's algorithm
8
-------
Uncertainty
Uncertainty refers to lack of knowledge about specific factors, parameters, or models.
For example, we mav be uncertain about the mean concentration of a specific pollutant at a
contaminated site or we may be uncertain about a specific measure of uptake (e.g., 95th percentile
fish consumption rate among all adult males in the United States). Uncertainty includes parameter
uncertainly (measurement errors, sampling errors, systematic errors), model uncertainty
(uncertainty due to necessary simplification of real-world processes, mis-specification of the
model structure, model misuse, use of inappropriate surrogate variables), and scenario
uncertainty (descriptive errors, aggregation errors, errors in professional judgment, incomplete
analysis).
Variability
Variability refers to observed differences attributable to true heterogeneity or diversity in a
population or exposure parameter. Sources of variability are the result of natural random
processes and stem from environmental, lifestyle, and genetic differences among humans.
Examples include human physiological variation (e.g., natural variation in bodyweight, height,
breathing rates, drinking water intake rates), weather variability, variation in soil types and
differences in contaminant concentrations in the environment. Variability is usually not reducible
by further measurement or study (but can be better characterized).
Preliminary Issues and Considerations
Defining the Assessment Questions
The critical first step in any exposure assessment is to develop a clear and unambiguous
statement of the purpose and scope of the assessment. A clear understanding of the purpose will
help to define and bound the analysis. Generally, the exposure assessment should be made as
simple as possible while still including all important sources of risk. Finding the optimum match
between the sophistication of the analysis and the assessment problem may be best achieved using
a "tiered approach" to the analysis, that is, starting as simply as possible and sequentially
employing increasingly sophisticated analyses, but only as warranted by the value added to the
analysis and decision process
9
-------
Selection and Development of
the Conceptual and
Mathematical Models
To help identify and select plausible
models, the risk assessor should develop
selection criteria tailored to each assessment
question. The application of these criteria
may dictate that different models be used for
different subpopulations under study (e.g.,
highly exposed individuals vs. the general
population) In developing these criteria, the
risk assessor should consider all significant
assumptions, be explicit about the
uncertainties, including technical and
scientific uncertainties about specific
quantities, modeling uncertainties,
uncertainties about functional forms, and
should identify significant scientific issues about which there is uncertainty
At any step in the analysis, the risk assessor should be aware of the manner in which
alternative selections might influence the conclusions reached.
Selection and Evaluation of Available Data
After the assessment questions have been defined and conceptual models have been
developed, it is necessary to compile and evaluate existing data (e.g., site specific or surrogate
data) on variables important to the assessment. It is important to evaluate data quality and the
extent to which the data are representative of the population under study.
10
Some Considerations in the Selection of Models
. appropriateness of the model's assumptions vis-a-vis
the analysis objectives
. compatibility of the model input/output and linkages to
other models used in the analysis
. the theoretical basis for the model
. level of aggregation, spatial and temporal scales
. resolution limits
. sensitivity to input variability and input uncertainty
. reliability of the model and code, including peer review
of the theory and computer code
. verification studies, relevant field tests
. degree of acceptance by the user community
. friendliness, speed and accuracy
. staff and computer resources required
-------
Guiding Principles for Monte Carlo Analysis
This section presents a discussion of principles of good practice for Monte Carlo
simulation as it may be applied to environmental assessments. It is not intended to serve as
detailed technical guidance on how to conduct or evaluate an analysis of variability and
uncertainty.
Selecting Input Data and Distributions for Use in Monte Carlo
Analysis
l. Conduct preliminary sensitivity analyses or numerical experiments to identify model
structures, exposure pathways, and model input assumptions and parameters that make
important contributions to the assessment endpoint and its overall variability and/or
uncertainty.
The capabilities of current desktop computers allow for a number of "what if scenarios to
be examined to provide insight into the effects on the analysis of selecting a particular model,
including or excluding specific exposure pathways, and making certain assumptions with respect
to model input parameters. The output of an analysis may be sensitive to the structure of the
exposure model Alternative plausible models should be examined to determine if structural
differences have important effects on the output distribution (in both the region of central
tendency and in the tails).
v Numerical experiments or sensitivity analysis also should be used to identify exposure
pathways that contribute significantly to or even dominate total exposure. Resources might be
saved by excluding unimportant exposure pathways (e.g., those that do not contribute appreciably
to the total exposure) from full probabilistic analyses or from further analyses altogether. For
important pathways, the model input parameters that contribute the most to overall variability and
uncertainty should be identified. Again, unimportant parameters may be excluded from full
probabilistic treatment. For important parameters, empirical distributions or parametric
distributions may be used. Once again, numerical experiments should be conducted to determine
the sensitivity of the output to different assumptions with respect to the distributional forms of
the input parameters. Identifying important pathways and parameters where assumptions about
distributional form contribute significantly to overall uncertainty may aid in focusing data
gathering efforts
-------
Dependencies or correlations between model parameters also may have a significant
influence on the outcome of the analysis. The sensitivity of the analysis to various assumptions
about known or suspected dependencies should be examined Those dependencies or correlations
identified as having a significant effect must be accounted for in later analyses.
Conducting a systematic sensitivity study may not be a trivial undertaking, involving
significant effort on the part of the risk assessor. Risk assessors should exercise great care not to
prematurely or unjustifiably eliminate pathways or parameters from full probabilistic treatment
Any parameter or pathway eliminated from full probabilistic treatment should be identified and the
reasons for its elimination thoroughly discussed.
2. Restrict the use of probabilistic assessment to significant pathways and parameters.
Although specifying distributions for all or most variables in a Monte Carlo analysis is
useful for exploring and characterizing the full range of variability and uncertainty, it is often
unnecessary and not cost effective. If a systematic preliminary sensitivity analysis (that includes
examining the effects of various assumptions about distributions) was undertaken and
documented, and exposure pathways and parameters that contribute little to the assessment
endpoint and its overall uncertainty and variability were identified, the risk assessor may simplify
the Monte Carlo analysis by focusing on those pathways and parameters identified as significant.
From a computational standpoint, a Monte Carlo analysis can include a mix of point estimates and
distributions for the input parameters to the exposure model. However, the risk assessor and risk
manager should continually review the basis for "fixing" certain parameters as point values to
avoid the perception that these are indeed constants that are not subject to change. '
3. Use data to inform the choice of input distributions for model parameters .
The choice of input distribution should always be based on all information (both
qualitative and quantitative) available for a parameter. In selecting a distributional form, the risk
assessor should consider the quality of the information in the database and ask a series of
questions including (but not limited to):
• Is there any mechanistic basis for choosing a distributional family?
• Is the shape of the distribution likely to be dictated by physical or biological
properties or other mechanisms?
• Is the variable discrete or continuous *
-------
• What are the bounds of the variable 7
• Is the distribution skewed or symmetric7
• If the distribution is thought to be skewed, in which direction ?
• What other aspects of the shape of the distribution are known?
When data for an important parameter are limited, it may be useful to define plausible
alternative scenarios to incorporate some information on the impact of that variable in the overall
assessment (as done in the sensitivity analysis). In doing this, the risk assessor should select the
widest distributional family consistent with the state of knowledge and should, for important
parameters, test the sensitivity of the findings and conclusions to changes in distributional shape.
4. Surrogate data can be used to develop distributions when they can be appropriately
justified.
The risk assessor should always seek representative data of the highest quality available.
However, the question of how representative the available data are is often a serious issue. Many
times, the available data do not represent conditions (e.g., temporal and spatial scales) in the
population being assessed. The assessor should identify and evaluate the factors that introduce
uncertainty into the assessment. In particular, attention should be given to potential biases that
may exist in surrogate data and their implications for the representativeness of the fitted
distributions.
When alternative surrogate data sets are available, care must be taken when selecting or
combining sets. The risk assessor should use accepted statistical practices and techniques when
combining data, consulting with the appropriate experts as needed.
Whenever possible, collect site or case specific data (even in limited quantities) to help
justify the use of the distribution based on surrogate data. The use of surrogate data to develop
distributions can be made more defensible when case-specific data are obtained to check the
reasonableness of the distribution.
5. When obtaining empirical data to develop input distributions for exposure model
parameters, the basic tenets of environmental sampling should be followed. Further,
-------
particular attention should be given to the quality of information at the tails of the
distribution.
As a general rule, the development of data for use in distributions should be carried out
using the basic principles employed for exposure assessments. For example,
• Receptor-based sampling in which data are obtained on the receptor or on tin-
exposure fields relative to the receptor;
• Sampling at appropriate spatial or temporal scales using an appropriate
stratified random sampling methodology;
• Using two-stage sampling to determine and evaluate the degree of error,
statistical power, and subsequent sampling needs; and
• Establishing data quality objectives.
In addition, the quality of information at the tails of input distributions often is not as good
as the central values. The assessor should pay particular attention to this issue when devising data
collection strategies.
6. Depending on the objeetives of the assessment, expert1 judgment can be included either
within the computational analysis by developing distributions using various methods or
by using judgments to select and separately analyze alternate, but plausible, scenarios.
When expert judgment is employed, the analyst should be very explicit about its use.
' Expert judgment is used, to some extent, throughout all exposure assessments. However,
debatable issues arise when applying expert opinions to input distributions for Monte Carlo
analyses. Using expert judgment to derive a distribution for an input parameter can reflect bounds
on the state of knowledge and provide insights into the overall uncertainty. This may be
particularly useful during the sensitivity analysis to help identify important variables for which
additional data may be needed. However, distributions based exclusively or primarily on expert
judgment reflect the opinion of individuals or groups and, therefore, may be subject to
considerable bias. Further, without explicit documentation of the use of expert opinions, the
1 According to NCRP (1996). an expert has (1) training and experience in the subject area resulting in
superior knowledge in the field, (2) access to relevant information. (3) an ability to process and effectivelv use the
information, and (4) is recognized by his or her peers or those conducting the studv as qualified to provide ludgments
about assumptions, models, and model parameters at the level of detail required
14
-------
distributions based on these judgments might be erroneously viewed as equivalent to those based
on hard data When distributions based on expert judgement have an appreciable effect on the
outcome of an analysis, it is critical to highlight this in the uncertainty characterization.
Evaluating Variability and Uncertainty
7. The concepts of variability and uncertainty are distinct. They can be tracked and
evaluated separately during an analysis, or they can be analyzed within the same
computational framework. Separating variability and uncertainty is necessary to
provide greater accountability and transparency. The decision about how to track
them separately must be made on a case-by-case basis for each variable.
Variability represents the true heterogeneity or diversity inherent in a well-characterized
population. As such, it is not reducible through further study. Uncertainty represents a lack of
knowledge about the population. It is sometimes reducible through further study. Therefore,
separating variability and uncertainty during the analysis is necessary to identify parameters for
which additional data are needed. There can be uncertainty about the variability within a
population For example, if only a subset of the population is measured or if the population is
otherwise under-sampled, the resulting measure of variability may differ from the true population
variability. This situation may also indicate the need for additional data collection.
8. There are methodological differences regarding how variability and uncertainty are
addressed in a Monte Carlo analysis.
There are formal approaches for distinguishing between and evaluating variability and
uncertainty When deciding on methods for evaluating variability and uncertainty, the assessor
should consider the following issues.
• Variability depends on the averaging time, averaging space, or other dimensions
in which the data are aggregated.
• Standard data analysis lends to understate uncertainty by focusing solely on
random error within a data set. Conversely, standard data analysis tends to
overstate variability by implicitly including measurement errors.
• I 'arums types of model errors can represent important sources of uncertainty.
Allernative conceptual or mathematical models are a potentially important source
of uncertainty. A major threat to the accuracy of a variability analysis is a lack of
representativeness of the data.
15
-------
9. Methods should investigate the numerical stability of the moments and the tails of the
distributions.
For the purposes of these principles, numerical stability refers to observed numerical
changes in the characteristics (i.e., mean, variance, percentiles) of the Monte Carlo simulation
output distribution as the number of simulations increases. Depending on the algebraic structure
of the model and the exact distributional forms used to characterize the input parameters, some
outputs will stabilize quickly, that is, the output mean and variance tend to reach more or less
constant values after relatively few sampling iterations and exhibit only relatively minor
fluctuations as the number of simulations increases. On the other hand, some model outputs may
take longer to stabilize. The risk assessor should take care to be aware of these behaviors. Risk
assessors should always use more simulations than they think necessary. Ideally, Monte Carlo
simulations should be repeated using several non-overlapping subsequences to check for stability
and repeatability. Random number seeds should always be recorded. In cases where the tails of
the output distribution do not stabilize, the assessor should consider the quality of information in
the tails of the input distributions. Typically, the analyst has the least information about the input
tails. This suggest two points.
• Data gathering efforts should be structured to provide adequate coverage at the
tails of the input distributions.
• The assessment should include a narrative and qualitative discussion of the
quality of information at the tails of the input distributions.
10. There are limits to the assessor's ability to account for and characterize all sources of
uncertainty. The analyst should identify areas of uncertainty and include them in the
analysis, either quantitatively or qualitatively.
Accounting for the important sources of uncertainty should be a key objective in Monte
Carlo analysis. However, it is not possible to characterize all the uncertainties associated with the
models and data. The analyst should attempt to identify the full range of types of uncertainty
impinging on an analysis and clearly disclose what set of uncertainties the analysis attempts to
represent and what it does not. Qualitative evaluations of uncertainty including relative ranking of
the sources of uncertainty may be an acceptable approach to uncertainty evaluation, especially
when objective quantitative measures are not available. Bayesian methods may sometimes be
16
-------
useful for incorporating subjective informatioa into variability and uncertainty analyses in a
manner that is consistent with distinguishing variability from uncertainty.
Presenting the Results of a Monte Carlo Analysis
11. Provide a complete and thorough description of the exposure model and its equations
(including a discussion of the limitations of the methods and the results).
Consistent with the Exposure Assessment Guidelines, Model Selection Guidance, and
other relevant Agency guidance, provide a detailed discussion of the exposure model(s) and
pathways selected to address specific assessment endpoints. Show all the formulas used. Define
all terms. Provide complete references If external modeling was necessary (e.g., fate and
transport modeling used to provide estimates of the distribution of environmental concentrations),
identify the model (including version) and its input parameters. Qualitatively describe the major
advantages and limitations of the models used.
The objectives are transparency and reproducibility - to provide a complete enough
description so that the assessment might be independently duplicated and verified.
12. Provide detailed information on the input distributions selected. This information
should identify whether the input represents largely variability, largely uncertainty,
or some combination of both. Further, information on goodness-of-fit statistics
should be discussed.
It is important to document thoroughly and convey critical data and methods that provide
an important context for understanding and interpreting the results of the assessment. This
detailed information should distinguish between variability and uncertainty and should include
graphs and charts to visually convey written information.
The probability density function (PDF) and cumulative distribution function (CDF) graphs
provide different, but equally important insights. A plot of a PDF shows possible values of a
random variable on the horizontal axis and their respective probabilities (technically, their
densities) on the vertical axis. This plot is useful for displaying:
• fhe relative probability of values;
• the most likely values (e.g., modes);
• the shape of the distribution (e.g., skewness, kurtosis); and
I
17
-------
• small changes in probability density.
A plot of the cumulative distribution function shows the probability that the value of a random
variable is less than a specific value. These plots are good for displaying:
• fractiles, including the median:
• probability intervals, including confidence intervals;
• stochastic dominance; and
• mixed, continuous, and discrete distributions.
Goodness-of-fit tests are formal statistical tests of the hypothesis that a specific set of
sampled observations are an independent sample from the assumed distribution Common tests
include the chi-square test, the Kolmogorov-Smirnov test, and the Anderson-Darling test.
Goodness-of-fit tests for normality and lognormality include Lilliefors' test, the Shapiro-Wilks'
test, and D'Agostino's test.
Risk assessors should never depend solely on the results of goodness-of-fit tests to select
the analytic form for a distribution. Goodness-of-fit tests have low discriminatory power and are
generally best for rejecting poor distribution fits rather than for identifying good fits. For small to
medium sample sizes, goodness-of-fit tests are not very sensitive to small differences between the
observed and fitted distributions. On the other hand, for large data sets, even small and
unimportant differences between the observed and fitted distributions may lead to rejection of the
null hypothesis. For small to medium sample sizes, goodness-of-fit tests should best be viewed as
a systematic approach to detecting gross differences The risk assessor should never let
differences in goodness-of-fit test results be the sole factor for determining the analytic form of a
distribution.
Graphical methods for assessing fit provide visual comparisons between the experimental
data and the fitted distribution. Despite the fact that they are non-quantitative, graphical methods
often can be most persuasive in supporting the selection of a particular distribution or in rejecting
the fit of a distribution. This persuasive power derives from the inherent weaknesses in numerical
goodness-of-fit tests. Such graphical methods as probability-probability (P-P) and quantile-
quantile (Q-Q) plots can provide clear and intuitive indications of goodness-of-fit.
18
-------
Having selected and justified the selection of specific distributions, the assessor should
provide plots of both the PDF and CDF, with one above the other on the same page and using
identical horizontal scales. The location of the mean should be clearly indicated on both curves
[See Figure 1 ]. These graphs should be accompanied by a summary table of the relevant data
13. Provide detailed information and graphs for each output distribution.
In a fashion similar to that for the input distributions, the risk assessor should provide
plots of both the PDF and CDF for each output distribution, with one above the other on the
same page, using identical horizontal scales. The location of the mean should clearly be indicated
on both curves. Graphs should be accompanied by a summary table of the relevant data.
14. Discuss the presence or absence of dependencies and correlations.
Covariance among the input variables can significantly affect the analysis output. It is
important to consider covariance among the model's most sensitive variables. It is particularly
important to consider covariance when the focus of the analysis is on the high end (i.e., upper
end) of the distribution.
When covariance among specific parameters is suspected but cannot be determined due to
lack of data, the sensitivity of the findings to a range of different assumed dependencies should be
evaluated and reported.
15. Calculate and present point estimates.
Traditional deterministic (point) estimates should be calculated using established
protocols Clearly identify the mathematical model used as well as the values used for each input
parameter in this calculation. Indicate in the discussion (and graphically) where the point estimate
falls on the distribution generated by the Monte Carlo analysis. Discuss the model and parameter
assumptions that have the most influence on the point estimate's position in the distribution. The
most important issue in comparing point estimates and Monte Carlo results is whether the data
and exposure methods employed in the two are comparable. Usually, when a major difference
between point estimates and Monte Carlo results is observed, there has been a fundamental
change in data or methods. Comparisons need to call attention to such differences and determine
their impact
In some cases, additional point estimates could be calculated to address specific risk
management questions or to meet the information needs of the audience for the assessment. Point
estimates can often assist in communicating assessment results to certain groups by providing a
-------
scenario-based perspective. For example, if point estimates are prepared for scenarios with which
the audience can identify, the significance of presented distributions may become clearer This
may also be a way to help the audience identify important risks.
16. A tiered presentation style, in which briefing materials are assembled at various levels
of detail, may be helpful. Presentations should be tailored to address the questions
and information needs of the audience.
Entirely different types of reports are needed for scientific and nonscientific audiences
Scientists generally will want more detail than non-scientists Risk managers may need more
detail than the public. Reports for the scientific community are usually very detailed. Descriptive,
less detailed summary presentations and key statistics with their uncertainty intervals (e.g., box
and whisker plots) are generally more appropriate for non-scientists.
To handle the different levels of sophistication and detail needed for different audiences, it
may be useful to design a presentation in a tiered format where the level of detail increases with
each successive tier. For example, the first tier could be a one-page summary that might include a
graph or other numerical presentation as well as a couple of paragraphs outlining what was done
This tier alone might be sufficient for some audiences. The next tier could be an executive
summary, and the third tier could be a full detailed report. For further information consult Bloom
etal., 1993.
Graphical techniques can play an indispensable role in communicating the findings from a
Monte Carlo analysis. It is important that the risk assessor select a clear and uncluttered graphical
style in an easily understood format. Equally important is deciding which information to display
Displaying too much data or inappropriate data will weaken the effectiveness of the effort.
Having decided which information to display, the risk assessor should carefully tailor a graphical
presentation to the informational needs and sophistication of specific audiences. The performance
of a graphical display of quantitative information depends on the information the risk assessor is
trying to convey to the audience and on how well the graph is constructed (Cleveland, 1994). The
following are some recommendations that may prove useful for effective graphic presentation
Avoid excessively complicated graphs. Keep graphs intended for a glance (e.g.,
overhead or slide presentations) relatively simple and uncluttered. Graphs
intended for publication can include more complexity
Avoid pie charts, perspective charts (3-dimensional bar and pie charts, ribbon
charts), pseudo-perspective charts (2-dimensional bar or line charts).
20
-------
Color and shading can create visual biases and are very difficult to use effectively.
Use color or shading only when necessary and then, only very carefully Consult
references on the use of color and shading in graphics.
When possible in publications and reports, graphs should be accompanied by a
table of the relevant data.
If probability density or cumulative probability plots are presented, present both,
with one above the other on the same page, with identical horizontal scales and
with the location of the mean clearly indicated on both curves with a solid point.
• Do not depend on the audience to correctly interpret any visual display of data.
Always provide a narrative in the report interpreting the important aspects of the
graph.
Descriptive statistics and box plots generally serve the less technically-oriented
audience well. Probability density and cumulative probability plots are generally
more meaningful to risk assessors and uncertainty analysts.
21
-------
Appendix: Probability Distribution Selection Issues
Surrogate Data, Fitting Distributions, Default Distributions
Subjective Distributions
Identification of relevant and valid data to represent an exposure variable is prerequisite to
selecting a probability distribution However, often the data available are not a direct measure of
the exposure variable of interest. The risk assessor is often faced with using data taken in spatial
or temporal scales that are significantly different from the scale of the problem under
consideration. The question becomes whether or not or how to use marginally representative or
surrogate data to represent a particular exposure variable. While there can be no hard and fast
rules on how to make that judgment, there are a number of questions risk assessors need to ask
when the surrogate data are the Only data available. .
Is there Prior Knowledge about Mechanisms? Ideally, the selection of candidate probability
distributions should be based on consideration of the underlying physical processes or mechanisms
thought to be key in giving rise to the observed variability. For example, if the exposure variable
is the result of the product of a large number of other random variables, it would make sense to
select a lognormal distribution for testing. As another example, the exponential distribution
would be a reasonable candidate if the stochastic variable represents a process akin to inter-arrival
times of events that occur at a constant rate. As a final example, a gamma distribution would be a
reasonable candidate if the random variable of interest was the sum of independent exponential
random variables.
Threshold Question - Are the surrogate data of acceptable quality and representativeness to
support reliable exposure estimates?
What uncertainties and biases are likely to be introduced by using surrogate data? For
example, if the data have been collected in a different geographic region, the contribution of
factors such as soil type, rainfall, ambient temperature, growing season, natural sources of
exposure, population density, and local industry may have a significant effect on the exposure
concentrations and activity patterns. If the data are collected from volunteers or from hot spots,
they will probably not represent the distribution of values in the population of interest. Each
difference between the survey data and the population being assessed should be noted. The
effects of these differences on the desired distribution should be discussed if possible.
How are the biases likely to affect the analysis and can the biases be corrected? The risk
assessor may be able to state with a high degree of certainty that the available data over-estimates
or under-estimates the parameter of interest Use of ambient air data on arsenic collected near
smelters will almost certainly over-estimate average arsenic exposures in the United States.
However, the smelter data can probably be used to produce an estimate of inhalation exposures
that falls within the high end. In other cases, the assessor may be unsure how unrepresentative
data will affect the estimate as in the case when data collected by a particular State are used in a
22
-------
national assessment. In most cases, correction of suspected biases will be difficult or not possible.
If only hot spot data are available for example, only bounding or high end estimates may be
possible Unsupported assumptions about biases should be avoided. Information regarding the
direction and extent of biases should be included in the uncertainty analysis.
How should any uncertainty introduced by the surrogate data he represented?
In identifying plausible distributions to represent variability, the risk assessor should examine
the following characteristics of the variable:
1. Nature of the variable.
Can the variable only take on discrete values (e.g., either on or off; either heads or tails) or is
the variable continuous over some range (e.g., pollutant concentration; body weight, drinking
water consumption rate)? Is the variable correlated with or dependent on another variable?
2. Hounds of the variable.
What is the physical or plausible range of the variable (e.g., takes on only positive values,
bounded by the interval [a,b]). Are physical measurements of the variable censored due to limits
of detection or some aspect of the experimental design9
3. Symmetry of the Distribution.
Is distribution of the variable known to be or thought to be skewed or symmetric? If the
distribution is thought to be skewed, in which direction? What other aspects of the shape of the
distribution are known9 Is the shape of the distribution likely to be dictated by physical/biological
properties (e.g., logistic growth rates) or other mechanisms?
4. Summary Statistics.
Summary statistics can sometimes be useful in discriminating among candidate distributions.
For example, frequently the range of the variable can be used to eliminate inappropriate
distributions; it would not be reasonable to select a lognormal distribution for an absorption
coefficient since the range of the lognormal distribution is (0,«) while the range of the absorption
coefficient is (0,1). If the coefficient of variation is near 1.0, then an exponential distribution
might be appropriate. Information on skewness can also be useful. For symmetric distributions,
skewness = 0; for distributions skewed to the right, skewness > 0; for distributions skewed to the
left, skewness < 0.
5. Graphical Methods to Explore the Data.
The risk assessor can often gain important insights by using a number of simple graphical
techniques to explore the data prior to numerical analysis. A wide variety of graphical methods
have been developed to aid in this exploration including frequency histograms for continuous
distributions, stem and leaf plots, dot plots, line plots for discrete distributions, box and whisker
plots, scatter plots, star representations, glyphs, Chemoff faces, etc. [Tukey (1977); Conover
(1980); du Toit el a/. (1986), Morgan and Henrion, (1990)]. These graphical methods are all
23
-------
intended to permit visual inspection of the density function corresponding to the distribution of
the data. They can assist the assessor in examining the data for skewness, behavior in the tails,
rounding biases, presence of multi-modal behavior, and data outliers.
Frequency histograms can be compared to the fundamental shapes associated with standard
analytic distributions (e.g., normal, lognormal, gamma, Weibull). Law and Kelton (1991) and
Evans et al. (1993) have prepared a useful set of figures which plot many of the standard analytic
distributions for a range of parameter values. Frequency histograms should be plotted on both
¦ linear and logarithmic scales and plotted over a range of frequency bin widths (class intervals) to
avoid too much jaggedness or too much smoothing (i.e., too little or too much data aggregation)
The data can be sorted and plotted on probability paper to check for normality (or log-normality)
Most of the statistical packages available for personal computers include histogram and
probability plotting features, as do most of the spreadsheet programs. Some statistical packages
include stem and leaf, and box and whisker plotting features.
After having explored the above characteristics of the variable, the risk assessor has three
basic techniques for representing the data in the analysis. In the first method, the assessor can
attempt to fit a theoretical or parametric distribution to the data using standard statistical
techniques. As a second option, the assessor can use the data to define an empirical distribution
function (EDF). Finally, the assessor can use the data directly in the analysis utilizing random
resampling techniques (i.e., bootstrapping). Each of these three techniques has its own benefits.
However, there is no consensus among researchers (authors) as to which method is generally
superior. For example, Law and Kelton (1991) observe that EDFs may contain irregularities,
especially when the data are limited and that when an EDF is used in the typical manner, values
outside the range of the observed data cannot be generated. Consequently, when the data are
representative of the exposure variable and the fit is good, some prefer to use parametric
distributions. On the other hand, some authors prefer EDFs (Bratley, Fox and Schrage, 1987)
arguing that the smoothing which necessarily takes place in the fitting process distorts real
information. In addition, when data are limited, accurate estimation of the upper end (tail) is
difficult. Ultimately, the technique selected will be a matter of the risk assessor's comfort with the
techniques and the quality and quantity of the data under evaluation.
The following discussion focuses primarily on parametric techniques. For a discussion of the
other methods, the reader is referred to Efron and Tibshirani (1993), Law & Kelton (1991), and
Bratley et al (1987).
Having selected parametric distributions, it is necessary to estimate numerical values for the
intrinsic parameters which characterize each of the analytic distributions and assess the quality of
the resulting fit.
Parameter Estimation. Parameter estimation is generally accomplished using conventional
statistical methods, the most popular of which include the method of maximum likelihood.
method of least squares, and the method of moments. See Johnson and Kotz (1970), Law and
24
-------
Kelton (1991). Kendall and Stewart (1979), Evans et al. (1993), Ang and Tang (1975),
Gilbert (1987), and Meyer (1975).
Assessing the Representativeness of the Fitted Distribution. Having estimated the
parameters of the candidate distributions, it is necessary to evaluate the "quality of the fit"
and, if more than one distribution was selected, to select the "best" distribution from among
the candidates. Unfortunately, there is no single, unambiguous measure of what constitutes
best fit. Ultimately, the risk assessor must judge whether or not the fit is acceptable.
Graphical Methods for Assessing Fit Graphical methods provide visual comparisons
between the experimental data and the fitted distribution. Despite the fact that they are non-
quantitative, graphical methods often can be most persuasive in supporting the selection of a
particular distribution or in rejecting the fit of a distribution. This persuasive power derives
from the inherent weaknesses in numerical goodness-of-fit tests. Commonly used graphical
methods include: frequency comparisons which compare a histogram of the experimental data
with the density function of the fitted data; probability plots compare the observed cumulative
density function with the fitted cumulative density function. Probability plots are often based
on graphical transformations such that the plotted cumulative density function results in a
straight line; probability-probability plots (P-P plots) compare the observed probability with
the fitted probability. P-P plots tend to emphasize differences in the middle of the predicted
and observed cumulative distributions, quantile-quantile plots (Q-Q plots) graph the ith-
q nan tile of the fitted distribution against the ith quantile data. Q-Q plots tend to emphasize
differences in the tails of the fitted and observed cumulative distributions; and box plots
compare a box plot of the observed data with a box plot of the fitted distribution.
(7oodness-of-Fit Tests. Goodness-of-fit tests are formal statistical tests of the hypothesis that
the set of sampled observations are an independent sample from the assumed distribution.
The null hvpothesis is that the randomly sampled set of observations are independent,
identically distributed random variables with distribution function F. Commonly used
goodness-of-fit tests include the chi-square test, Kolmogorov-Smirnov test, and Anderson-
Darling test. The chi-square test is based on the difference between the square of the
observed and expected frequencies. It is highly dependent on the width and number of
intervals chosen and is considered to have low power. It is best used to reject poor fits. The
Kolmogorov-Smirnov Test is a non-parametric test based on the maximum absolute
difference between the theoretical and sample Cumulative Distribution Functions (CDFs).
The Kolmogorov-Smirnov test is most sensitive around the median and less sensitive in the
tails and is best at detecting shifts in the empirical CDF relative to the known CDF It is less
proficient at detecting spread but is considered to be more powerful than the chi-square test.
The Anderson-Darling test is designed to test goodness-of-fit in the tails of a Probability
Density Function (PDF) based on a weighted-average of the squared difference between the
observed and expected cumulative densities.
25
-------
Care must be taken not to over-interpret or over-rely on the findings of goodness-of-fit tests
It is far too tempting to use the power and speed of computers to run goodness-of-fit tests
against a generous list of candidate distributions, pick the distribution with the "best"
goodness-of-fit statistic, and claim that the distribution that fit "best" was not rejected at some
specific level of significance. This practice is statistically incorrect and should be avoided
[Bratley et at, 1987, page 134], Goodness-of-fit tests have notoriously low power and are
generally best for rejecting poor distribution fits rather than for identifying good fits For
small to medium sample sizes, goodness-of-fit tests are not very sensitive to small differences
between the observed and fitted distributions. On the other hand, for large data sets, even
minute differences between the observed and fitted distributions may lead to rejection of the
null hypothesis. For small to medium sample sizes, goodness-of-fit tests should best be
viewed as a systematic approach to detecting gross differences.
Tests of Choice for Normality and Lognormality. Several tests for normality (and
lognormaiity when log-transformed data are used) which are considered more powerful than
either the chi-square or Komolgarov-Smirnoff (K-S) tests have been developed: Lilliefors'
test which is based on the K-S test but with "normalized" data values, Shapiro-Wilks test (for
sample sizes £ 50), and D'Agostino's test (for sample sizes * 50). The Shapiro-Wilks and
D'Agostino tests are the tests of choice when testing for normality or lognormality
If the data are not well-fit by a theoretical distribution, the risk assessor should consider the
Empirical Distribution Function or bootstrapping techniques mentioned above.
For those situations in which the data are not adequately representative of the exposure
variable or where the quality or quantity of the data are questionable the following approaches
may be considered.
Distributions Based on Surrogate Data. Production of an exposure assessment often
requires that dozens of factors be evaluated, including exposure concentrations, intake rates,
exposure times, and frequencies. A combination of monitoring, survey, and experimental
data, fate and transport modeling, and professional judgment is used to evaluate these factors.
Often the only available data are not completely representative of the population being
assessed. Some examples are the use of activity pattern data collected in one geographic
region to evaluate the duration of activities at a Superfund site in another region; use of
national intake data on consumption of a particular food item to estimate regional intake, and
use of data collected from volunteers to represent the general population.
In each such case, the question of whether to use the unrepresentative data to estimate the
distribution of a variable should be carefully evaluated. Considerations include how to express
the possible bias and uncertainty introduced by the unrepresentativeness of the data and
alternatives to using the data In these situations, the risk assessor should carefully evaluate
the basis of the distribution (e g , data used, method) before choosing a particular surrogate or
before picking among alternative distributions for the same exposure parameter The
26
-------
following table indicates exposure parameters for which surrogate distributions may be
reasonable and useful.
Table 1 Examples of exposure parameters for which
distributions based on surrogate data might be reasonable
Receptor Physiological
Parameters
body weight
height
total skin surface area
exposed skin - hands, forearms, head, upper
body
Behavioral
Receptor
Time-Activity
Patterns
residency periods - age, residency type
weekly work hours
time since last job change
showering duration
Receptor
Contact Rates
soil ingestion rates
soil adherence
food ingestion - vegetables, freshwater finftsh,
saltwater finfish, shellfish, beef
water intake - total water, tapwater
inhalation rates
Rough Characterizations of Ranges and Distributional Forms. In the absence of
acceptable representative data or if the study is to be used primarily for screening, crude
characterizations of the ranges and distributions of the exposure variable may be adequate.
For example, physical plausibility arguments may be used to establish ranges for the
parameters. Then, assuming such distributions as the uniform, log-uniform, triangular and
log-triangular distributions can be helpful in establishing which input variables have the
greatest influence on the output variable However, the risk assessor should be aware that
there is some controversy concerning the use of these types of distributions in the absence of
data Generally, the range of the model output is more dependant on the ranges of the input
variables than it is on the actual shapes of the input distributions. Therefore, the risk assessor
should be careful to avoid assigning overly-restrictive ranges or unreasonably large ranges to
variables. Distributional assumptions can have a large influence on the shapes of the output
distribution. When the shape of the output distribution must be estimated accurately, care and
attention should be devoted to developing the input distributions.
Distributions Based on Expert Judgment One method that has seen increasing usage in
environmental risk assessment is the method of subjective probabilities in which an expert or
experts are asked to estimate various behaviors and likelihoods regarding specific model
variables or scenarios. Expert elicitation is divided into two categories: (1) informal
elicitation, and (2) formal elicitation. Informal elicitation methods include self assessment,
brainstorming, causal elicitation (without structured efforts to control biases), and taped
group discussions between the project staff and selected experts.
27
-------
Formal elicitation methods generally follow the steps identified by the U.S. Nuclear
Regulatory Commission (USNRC, 1989; Oritz, 1991; also see Morgan and Henrion. 1990;
IAEA, 1989; Helton, 1993, Taylor and Burmaster, 1993) and are considerably more elaborate
and expensive than informal methods.
28
-------
References Cited in Text
A. H-S. Ang and W. H. Tang, Probability Concepts in Engineering Planning and Design,
Volume /, Basic Principles, John Wiley & Sons, Inc., New York (1975).
D L Bloom, et al., Communicating Risk to Senior EPA Policy Makers: A Focus Group
Study, U.S. EPA Office of Air Quality Planning and Standards (1993).
P. Bratley, B L. Fox, L. E. Schrage, A Guide to Simulation, Springer-Verlag, New York
(1987).
W S. Cleveland, The Elements of Graphing Data, revised edition, Hobart Press,
Summit, New Jersey (1994).
W. J. Conover, Practical Nonparametric Statistics, John Wiley & Sons, Inc., New York
(1980).
5 H C du Toit, A. G W Steyn, R.H. Stumpf, Graphical Exploratory Data Analysis,
Springer-Verlag, New York (1986).
B Efron and R Tibshirani, An introduction to the bootstrap. Chapman & Hall, New York
(1993).
M Evans, N. Hastings, and B Peacock, Statistical Distributions, John Wiley & Sons, New
York (1993).
R O Gilbert, Statistical Methods for Environmental Pollution Monitoring, Van Nostrand
Reinhold, New York (1987).
J. C Helton, "Uncertainty and Sensitivity Analysis Techniques for Use In Performance
Assessment for Radioactive Waste Disposal," Reliability Engineering and System Safety,
Vol. 42, pages 327-367 (1993).
IAEA, Safety Series 100, Evaluating the Reliability of Predictions Made Using
Environmental Transfer Models, International Atomic Energy Agency, Vienna, Austria
(1989)
N L Johnson and S. Kotz, Continuous Univariate Distributions, volumes 1 & 2, John Wiley
6 Sons, Inc , New York (1970).
M Kendall and A Stuart, The Advanced Theory of Statistics, Volume I - Distribution
Theorv. Volume II - Inference and Relationship, Macmillan Publishing Co , Inc , New York
(1979)
29
-------
A. M. Law and W. D. Kelton, Simulation Modeling & Analysis, McGraw-Hill, Inc., (1991)
S. L. Meyer, Data Analysis for Scientists and Engineers, John Wilev & Sons, Inc.. New York
(1975).
M. G. Morgan and M. Henri on, Uncertainty A guide to Dealing with Uncertainty in
Quantitative Risk and Policy Analysis, Cambridge University Press, New York (1990).
NCRP Commentary No. 14, "A Guide for Uncertainty Analysis in Dose and Risk Assessments
Related to Environmental Contamination," National Committee on Radiation Programs,
Scientific Committee 64-17, Washington, D C. (May, 1996).
N. R. Oritz, M. A. Wheeler, R. L. Keeney, S. Hora, M. A. Meyer, and R. L. Keeney, "Use of
Expert Judgment in NUREG-1150, Nuclear Engineering and Design, 126:313-331 (1991)
A. C Taylor and D E. Burmaster, "Using Objective and Subjective Information to Generate
Distributions for Probabilistic Exposure Assessment," U.S. Environmental Protection Agency,
draft report (1993).
J. W. Tukey, Exploratory Data Analysis, Addison-Wesley, Boston (1977).
USNRC, Severe Accident Risks: An Assessment for Five U.S. Nuclear power Plants (second
peer review draft), U.S. Nuclear Regulatory Commission, Washington, D C. (1989).
30
-------
References for Further Reading
B F Baird, Managerial Decisions Under Uncertainty, John Wiley and Sons, Inc., New
York (1989).
D E Burmaster and P. D Anderson, "Principles of Good Practice for the Use of Monte
Carlo Techniques in Human Health and Ecological Risk Assessments," Risk Analysis, Vol.
14(4), pages 477-482 (August, 1994).
R Clemen, Making Hard Decisions, Duxbury Press (1990).
D C Cox and P. Bay butt, "Methods for Uncertainty Analysis: A Comparative Survey," Risk
Analysis, Vol. 1 (4), 251-258 (1981).
R D'Agostino and M A Stephens (eds), Goodness-of-Fit Techniques, Marcel Dekker, Inc.,
New York (1986).
L Devroye, Non-Uniform Random Deviate Generation, Springer-Verlag, (1986).
D M. Hamby, "A Review of Techniques for Parameter Sensitivity Analysis of Environmental
Models," Environmental Monitoring and Assessment, Vol. 32, 135-154 (1994).
D. B Hertz, and H. Thomas, Risk Analysis and Its Applications, John Wiley and Sons, New
York (1983).
D B Hertz, and H. Thomas, Practical Risk Analysis - An Approach Through Case Studies,
John Wiley and Sons, New York (1984).
F O Hoffman and J. S. Hammonds, An Introductory Guide to Uncertainty Analysis in
Environmental and Health Risk Assessment, ES/ER/TM-35, Martin Marietta (1992).
F 0 Hoffman and J. S. Hammonds, "Propagation of Uncertainty in Risk Assessments: The
Need to Distinguish Between Uncertainty Due to Lack of Knowledge and Uncertainty Due to
Variability," Risk Analysis, Vol. 14 (5), 707-712 (1994).
R L I man and J. C Helton, "An Investigation of Uncertainty and Sensitivity Analysis
Techniques for Computer Models," Risk Analysis, Vol. 8(1), pages 71-90 (1988)
R L I man and W J Conover, "A Distribution-Free Approach to Inducing Rank Correlation
Among Input Variables," Commun. Statistics, Communications and Computation, 11, 311-
331 (1982).
-------
R L. Iman, J. M. Davenport, and D. K. Zeigler, "Latin Hypercube Sampling (A Program
Users Guide)," Technical Report SAND 79:1473, Sandia Laboratories, Albuquerque (1980)
M. E Johnson, Multivariate Statistical Simulation, John Wiley & Sons. Inc., New York
(1987).
N. L. Johnson, S. Kotz and A. W. Kemp, Univariate Discrete Distributions, John Wiley &
Sons, Inc., New York (1992).
R. LePage and L. Billard, Exploring the Limits of Bootstrap, Wiley, New York (1992).
J. Lipton, et al. "Short Communication: Selecting Input Distributions for Use in Monte Carlo
Analysis," Regulatory Toxicology and Pharmacology, 21, 192-198 (1995).
W. J. Kennedy, Jr. and J. E Gentle, Statistical Computing, Marcel Dekker, Inc , New York
(1980).
T E. McKone and K. T Bogen, "Uncertainties in Health Risk Assessment: An Integrated
Case Based on Tetrachloroethylene in California Groundwater," Regulatory Toxicology and
Pharmacology, 15, 86-103 (1992).
R. E. Megill (Editor), Evaluating and Managing Risk, Penn Well Books, Tulsa, Oklahoma
(1985).
R. E. Megill, An Introduction to Risk Analysis, end Ed. Penn Well Books, Tulsa, Oklahoma
(1985).
Palisade Corporation, Risk Analysis and Simulation Add-In for Microsoft Excel or Lotus 1-2-
3. Windows Version Release 3.0 User's Guide, Palisade Corporation, Newfield, New York
(1994).
W. H. Press, B. P. Flannery, S. A. Teulolsky, and W. T. Vetterling, Numerical Recipes in
Pascal, the Art of Scientific Computing, Cambridge University Press (1989).
W. H Press, S. A. Teulolsky, W. T. Vetterling, and B P. Flannery, Numerical Recipes in
FORTRAN: the Art of Scientific Computing, Cambridge University Press (1992).
W. H. Press, S. A. Teulolsky, W. T. Vetterling, and B P. Flannery, Numerical Recipes in C:
the Art of Scientific Computing, Cambridge University Press (1992)
T. Read and N. Cressie, Goodness-of-fit Statistics for Discrete Multivariate Data, Springer-
Verlag, New York (1988).
32
-------
Y K Rohatgi, Statistical Inference, John Wiley & Sons, New York (1984).
R. Y. Rubinstein. Simulation and the Monte Carlo Method, John Wiley and Sons, New York
(1981)
L Sachs. Applied Statistics - A Handbook of Techniques, Spring-Verlag, New York (1984).
A Saltelii and J Marivort, "Non-parametric Statistics in Sensitivity Analysis for Model
Output: A Comparison of Selected Techniques," Reliability Engineering and System Safety,
Vol. 28, 229-253 (1990).
H Schneider, Truncated and Censored Distributions form Normal Populations, Marcel
Dekker, Inc., New York (1986)
F A. Seiler and J. L Alvarez, "On the Selection of Distributions for Stochastic Variables,"
Risk Analysis, Vol 16 (1), 5-18 (1996).
F A. Seiler, "Error Propagation for Large Errors," Risk Analysis, Vol 7 (4), 509-518 (1987).
W Slob, "Uncertainty Analysis in Multiplicative Models," Risk Analysis, Vol. 14 (4), 571-576
(1994).
A. E. Smith, P B Ryan, J. S. Evans, "The Effect of Neglecting Correlations When
Propagating Uncertainty and Estimating the Population Distribution of Risk," Risk Analysis,
Vol 12 (4), 467-474 (1992).
U.S. Environmental Protection Agency, Guidelines for Carcinogenic Risk Assessment,
Federal Register 51(185), 33992-34003 (May 29, 1992).
U.S. Environmental Protection Agency, Source Assessment: Analysis of Uncertainty -
Principles and Applications, EPA/600/2-79-004 (August, 1978)
U.S. Environmental Protection Agency, Guidelines for Exposure Assessment, Federal
Register 57(104), 22888-22938 (May 29, 1992).
U S Environmental Protection Agency, Summary Report for the Workshop on Monte Carlo
Analysis. EPA/630/R-96/010 (September, 1996).
-------
figure la. Example Monte Carlo Estimate of the PDF for Lifetime Cancer Risk
0. 02s
I - ; : Pt>i t«
a. 020
|| ,1 : Est i mte
1
Man i J K. /
So. 018
Rsk v Iff n ; /
it
II ¦ 1 /
8
\j" | f
i»-o«
i*\. ! \
M
p V ^
0. MS
J i \
m mM
. \ ; V. ,
1. 06-08 1. OB-07 1. OE-m f. OS-OS 1. 0B-O4 1. OB- 03
tJ fail am Oncsr Fisk
Figure lb: Example Monte Carlo Estimate of the CDF for Lifetime Cancer Risk
o. m
o. m
--a m
a 30
o. 20
ft 10
e. m
1. or oa f, as- or 1. oe- m 1. oe- os t, ee- m 1, oe- 03
Ufwtfmm Oncer W sk
34
-------
Figure 2: Example Box and Whiskers Plot of the Distribution of Lifetime Cancer Risk
1B04
-
£ 1BOS
•firt jxreanllfi
!~ffr0 8
7fth fmemtll*
35th jMrcantlla
Sfft narcantlli
1B-67
o Wnt SI I mf •• Aten ffslr
>3
-------
ATTACHMENT P
DEALING WITH DATA BELOW DETECTION LIMITS,
QUALITY ASSURANCE COURSE MODULE 492,
IPA NATIONAL CENTER FOR ENVIRONMENTAL
RESEARCH AND QUALITY ASSURANCE
S IIEP4 Siwi»'-Ti»»;8'«IWIErHHM,aMSTa WPPllli.»HMi«ra»IHBll.«lf» ipMBE
-------
492: g with Data
"lelow De.-stion
492 -10SS 1
-------
Monitoring Well Data for Gladstone
Recycling Corporation
¦ Arsenic in ground water
- Method 606a (1987) with a detection limit of 2 jug/liter
- intensive monitoring (almost daily) for February 1994
4.9
<2
<2
3.5
4.2
4.4
<2
2.2
2.9
2.7
3.2
2.2
2.1
2.9
4.3
3.1
2.9
<2
4.9
3.3
2.6
5.1
2.1
<2
5.3
¦ 5 out of 25 readings were "less than detection"
¦ Problem is to estimate the mean level of arsenic
192-1095 2
-------
Strategies for Estimating Mean
¦ Strategy 1: Throw them away, ignore them
¦ Strategy 2: Make them all zero
- Strategy 3: Set them all at the detection limit
¦ Strategy 4: Use a statistical approach
¦ Strategy 5: Set them all at some value (e.g., DL/2)
(Note that strategies 2 and 3 are subsets of this strategy)
492-1095 3
-------
Consequences of the 3 Easy
Strategies
Strategy 1: Ignore them (invisible)
- Overestimates true mean (non-detects are small)
- Underestimates the variability (data set more compact)
Strategy 2: Make them all zero
- Underestimates true mean (non-detects probably
greater than zero)
- Overestimates the variability (increases spread in data)
Strategy 3: Set them all at the detection limit
- Overestimates true mean (non-detects are small)
- Underestimates the variability (reduces spread in data)
;432 -1095 4
-------
Comparison of the 3 Easy Strategies
Estimated Estimated
Mean Variance
1. Ignore them 3.440 1.143
2. Make them 0 2.752 2.877
3. Set to DL 3.152 1.250
¦ Real values for "<2" are 1.9, 0.9,1.5,1.7,1.8
¦ True Mean is 3.0G4,
¦ True Variance is 1.520
492-1095 5
-------
Recommendations
Strategy 3: If you want to "nail them"
Strategy 2: If you want to conceal things
Strategy 1: Provided your boss does not catch you
wasting resources!
-------
Strategies 4 and 5
If there is a reasonable amount of data available:
Approximate
Percentage
of Non-Detects
Analysis Method
< 15 %
Replace non-defects with DL/2f DLr or a very small number
15% - 50%
Use a trimmed mean, Cohen's adjustment, or the
Winsorized mean and standard deviation
50% - 90%
Use a test for proportions*
>S0%
Use a Poisson Approach*
* Consult a statistician!
492-1095?
-------
Case (i): Less than 15%"" an-Detects
¦ Set all non-defects equal to DL/2
- Resulting bias in estimating mean and variance quite
small
¦ In the Arsenic Example:
~ Adjusted Mean = 2.952 (True Mean = 3.064)
~ Adjusted Variance = 1.897 (True Variance = 1.520)
. V.492-1095 8
-------
Case (if): Between 15-50%
Mon-Detects
Cohen's Method
- Mends the Data
Trimmed Mean
- Discards the Data
Winsorized Mean and Standard Deviation
- Substitutes the Data
492-1095 9
-------
Cohen's Method
¦ "Mends" the data essentially by estimation; estimates
what values should have been found
¦ Uses information from the values above the detection
limit together with assumptions about the distribution
of the data
¦ Calculates mean and variance from the values above
the Detection Limit, then adjusts the mean down and
the variance up
¦ Requires special tables
32-109510
-------
Why Cohen's Method Works
X
/
/
X
\
/
\
\
Normal
Distribution
*
/
/
/
/
\
\
\
\
Cohen's method uses Normality and the principle of
Maximum Likelihood to estimate what impact the below
Detection Limit values (dotted) would have on the
estimates calculated from the above Detection Limit
-------
How Cohen's Method Works
The data look like:
x x x x x | xxxxxxxxxx (n) Total number
(n-m) DL (m) observations of observations
observations above DL
below DL
2
1) Calculate the mean (Xd) and variance (sd ) of the m
observations above the DL.
32-1095 12
-------
How Cohen's Method Works
Continued
2) Then Adjusted X = Xd - %(Xd - DL)
Adjusted s2 = sj + A, - DL)2
where X is calculated from special tables.
In the Arsenic Example:
Adjusted Mean = 2.998 (True = 3.064)
Adjusted Variance = 1.780 (true = 1.520)
492-109513
-------
The Trimmed Mean
¦ Discards data below the DL arid an equivalent amount
of the largest values above the DL.
« The percentage of data discarded below the DL defines
the trimmed mean percentage. For example, if 10% of
the data are below the detection level (so 20% of the
data are discarded), this method will compute a 10%
trimmed mean.
¦ Although good for estimating the mean, it does not help
with the variance.
) 92 -1095 14
-------
How Trimming Works
1) Find the percentage (p%) of readings below the DL.
(In the example, this was 5 out of 25 = 20%).
2) Determine the equivalent highest above the DL and
discard these values. (In the example, 5.3, 5.1, 4.9, 4.9,
4.4 are discarded.)
3) Calculate the mean of the remaining values. This is
the p% trimmed mean.
In the Arsenic Example:
20% Trimmed Mean is 2.947 (True = 3.064)
Trimmed Variance, incidently, is 0.476 (True = 1.520)
492-109515
-------
Winsorized Mean and Variance
Similar to trimming but instead of throwing data
away, the ones below the DL are replaced by the
value closest to the DL, and the equivalent number of
highest values are replace by an equivalent high
value.
,92- lies 18
-------
How the Winsorized Mean Works
1) Rank the data from smallest to largest.
In the example,
>2, >2, >2, >2, >2 | 2.1, 2.1, 2.2,..., 4.3, 4.4, 4.9, 4.9, 5.1, 5.3
2) Replace those below the DL by the first value greater than
the DL. Replace the same number of the largest values by
the next closest value.
In the example, 2.1 is the first value greater than the DL so
the 5 non-detects are replaced with 2.1.
The five largest values are 4.4,4.9, 4.9, 5.1, and 5.3. These
values are replaced by 4.3.
492-109517
-------
How the Winsorized Mean Works
Continued
3) Calculate sample mean, Xw, and sample variance, s2,
as usual. Then the winsorized mean is X, and the
winsorized variances2 is w
2 r n- 1 2 2
sw = [ ] S
2m - n- 1
In the Arsenic Example:
Winsorized Mean = 3.048 (True = 3.064)
Winsorized Variance = 2.344 (True = 1.520)
92-1095 18
-------
Comparison of Methods
True
1
-------
But Why Does it Matter?
Data are likely to be used in a statistical test to determine
compliance with a standard or evidence of
contamination.
For example, standard Student's t-test:
X - u
t =
\
n
where ji comes from the null hypothesis (e.g., the
regulatory standard)
92-1095 20
-------
Bias
¦ If less than detection values are replaced by something
else (i.e., some number) and X^,and sjLare calculated,
then
true mean (X) = Xmw + bias (Xnew)
2 2 2
true variance (s ) = sBew + bias (sBew)
> It is impossible to eliminate these biases
¦ Statistical methods (Cohen or Winsorization) reduce
the bias considerably where as substitution methods
(1/a DL, etc.) do not.
492-1095 21
-------
What's the Practical Effect?
Since is impossible to make the biases cancel, the simple
t-test becomes:
t =
[ ^ new (*„) ] - II
\
[ sL, + bias (sL) ]
n
does not have the intended false positive rate (a)
does not have the intended false negative rate ((5)
does not have the intended statistical power
,32 - 1095 22
} *
-------
Hypothesis Testing and Bias
It is difficult to predict what the true false positive
rate and false negative rate will be as both biases are
functions of the DL, amount of data, amount of data
below DL, and the relative standard deviation
(coefficient of variation).
-------
What if the Data are Not Normally
Distributed?
If approximately Lognormal, transform and use Cohen's
method
If roughly Lognormal shaped and only a few data below
DL, substitute Vz DL
Trimming and Winsorization perform poorly unless data
are approximately symmetric
92- 1095 24
-------
Lognormal Example
-10 Carbon Monoxide readings in a furnace room
collected a random intervals over a 24-hour period.
- Data from smallest to largest are:
1.53, 2.20, 2.72, 3.16, 3.76, 4.15, 4.43, 7.81, 8.42, 20.76
- Data are lognormally distributed
¦ Let DL = 2.00, so 9 readings are above DL
492-1095 25
-------
Lognormal Example - Continued
* True Mean = 5.89 *
* True Variance = 32.33 *
Winsorized: Replace < DL with 2.20, Replace 20.76 with 8.42
Winsorized Mean = 4.72
Winsorized Variance = 6.45
Substitution: Replace < DL with % DL (1,00)
Substitution Mean = 5.84
Substitution Variance = 32.891
£2-1095 26
-------
Conclusions
Note: All percentages are approximate.
NC\ / Are data \ YES
—
-------
|