United States Office of Radiation 9355.4-16
Environmental Protection and Indoor Air EPA/540-R-00-006
Agency Office of Solid Waste and PB2000 963306
Emergency Response October 2000
Washington, DC 20460
ORIA/Superfund
&EPA
Soil Screening Guidance
for Radionuclides:
Technical Background
Document
-------
Publication 9355.4-16
October 2000
Soil Screening Guidance for
Radionuclides:
Technical Background Document
Office of Radiation and Indoor Air
Office of Emergency and Remedial Response
U.S. Environmental Protection Agency
Washington, DC 20460
-------
DISCLAIMER
Notice: The Soil Screening Guidance for Radionuclides is based on policies set out in the Preamble to the Final Rule
of the National Oil and Hazardous Substances Pollution Contingency Plan (NCP), which was published on March 8,
1990 (55 Federal Register 8666).
This guidance document sets forth recommended approaches based on EPA's best thinking to date with respect to soil
screening for radionuclides. Alternative approaches for screening radionuclides in soil may be found to be more
appropriate at specific sites (e.g., where site circumstances do not match the underlying assumptions, conditions, and
models of the guidance). The decision whether to use an alternative approach and a description of any such approach
should be placed in the Administrative Record for the site.
The policies set out in both the Soil Screening Guidance for Radionuclides: User's Guide and the supporting Soil
Screening Guidance for Radionuclides: Technical Background Document are intended solely as guidance to the U.S.
Environmental Protection Agency (EPA) 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
with the United States government. EPA officials may decide to follow the guidance provided in this document, or 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.
-------
TABLE OF CONTENTS
Section Page
Disclaimer ii
List of Tables vi
List of Figures viii
List of Highlights viii
Preface ix
Acknowledgments x
Parti: INTRODUCTION
1.1 Background 1-2
1.2 Purpose of SSLs 1-2
1.3 Scope of Soil Screening Guidance for Radionuclides 1-4
1.3.1 Exposure Pathways 1-4
1.3.2 Exposure Assumptions 1-5
1.3.3 RiskLevel 1-6
1.3.4 SSL Model Assumptions 1-6
1.4 Organization of the Document 1-7
Part 2: DEVELOPMENT OF PATHWAY-SPECIFIC SOIL SCREENING LEVELS
2.1 Human Health Basis 2-2
2.1.1 Additive Risk 2-11
2.1.2 Toxicokinetics of Radionuclides 2-11
2.2 Direct Ingestion of Soil 2-11
2.3 Inhalation of Fugitive Dusts 2-13
2.3.1 Screening Level Equation for Inhalation of Fugitive Dusts 2-13
2.3.2 Dispersion Model 2-13
2.3.3 Paniculate Emission Factor 2-16
2.4 External Exposure from Radionuclides in Soil 2-17
2.4.1 Gamma Shielding Factor - GSF (Unitless) 2-18
2.5 Ingestion of Homegrown Produce 2-19
2.6 Migration to Ground Water 2-20
2.6.1 Development of Soil/Water Partition Equation 2-21
2.6.2 Inorganics (Metals, including radionuclides)-Partition Theory 2-24
2.6.3 Assumptions for Soil/Water Partition Theory 2-24
2.6.4 Dilution/Attenuation Factor Development 2-25
2.6.5 Default Dilution-Attenuation Factor 2-30
2.7 Mass-Limit Model Development 2-31
2.7.1 Mass Balance Issues 2-31
2.7.2 Migration to Ground Water Mass-Limit Model 2-31
-------
TABLE OF CONTENTS (continued)
Section Page
PART 3: UNSATURATED ZONE MODELS FOR RADIONUCLIDE
FATE AND TRANSPORT
3.1 Considerations for Unsaturated Zone Model Selection 3-9
3.2. Model Applicability to SSLs for Radionuclides 3-9
PART 4: MEASURING CONTAMINANT CONCENTRATIONS IN SOIL
4.1 Sampling Surface Soils - Radionuclide Not Present in Background 4-2
4.1.1 State the Problem 4-2
4.1.2 Identify the Decision 4-2
4.1.3 Identify Inputs to the Decision 4-4
4.1.4 Define the Study Boundaries 4-4
4.1.5 Develop a Decision Rule 4-6
4.1.6 Specify Limits on Decision Errors for the Max Test 4-6
4.1.7 Optimize the Design for the Max Test 4-8
4.1.8 Using the DQA Process: Analyzing Max Test Data 4-16
4.1.9 Specify Limits on Decision Errors for Sign Test 4-18
4.1.10 Optimize the Designforthe Sign Test 4-19
4.1.11 Using the DQA Process: Analyzing Sign Test Data 4-22
4.1.12 Special Considerations for Multiple Radionuclides 4-27
4.1.13 Quality Assurance/Quality Control 4-27
4.1.14 Final Analysis 4-27
4.1.15 Reporting 4-27
4.2 Sampling Subsurface Soils - Radionuclide Not Present in Background 4-27
4.2.1 State the Problem 4-29
4.2.2 Identify the Decision 4-29
4.2.3 Identify Inputs to the Decision 4-29
4.2.4 Define the Study Boundaries 4-32
4.2.5 Develop a Decision Rule 4-32
4.2.6 Specify Limits on Decision Errors 4-32
4.2.7 Optimize the Design 4-33
4.2.8 Analyzing the Data 4-33
4.2.9 Reporting 4-34
IV
-------
TABLE OF CONTENTS (continued)
Section Page
4.3 Basis for the Surface Sampling Strategies: Technical Analysis Performed 4-34
4.3.1 1994 Draft Guidance Sampling Strategy 4-35
4.3.2 Test of Proportion Exceeding a Threshold 4-36
4.3.3 Relative Performance of Land, Max, and Chen Tests 4-38
4.3.4 Treatment of Observations Below the Limit of Quantitation 4-44
4.3.5 Multiple Hypothesis Testing Considerations 4-44
4.3.6 Investigation of Compositing Within EA Sectors 4-46
PARTS: RADIONUCLIDE-SPECIFIC PARAMETERS
5.1 Area Correction Factor - ACF (unitless) 5-1
5.2 Soil -Water Distribution Coefficients (Kd) for Radionuclides 5-2
PART 6: REFERENCES
References R-l
APPENDICES
A Generic SSLs A-l
B Sample Measurement Units, Activity and Mass B-l
-------
LIST OF TABLES
Table 2.1. Radionuclides Included in Generic Soil Screening Analysis 2-4
Table 2.2. Radionuclide Cancer Morbidity - Slope Factors 2-7
Table 2.3. Radionuclide Drinking Water MCLs 2-9
Table 2.4. Q/C Values by Source Area, City, and Climatic Zone 2-15
Table 3.1. Input Parameters Required for HYDRUS 3-2
Table 3.2. Input Parameters Required for MULTIMED 3-3
Table 3.3. Input Parameters Required for FECTUZ Module within EPACMTP 3-6
Table 3.4. Input Parameters Required for CHAIN Model 3-7
Table 3.5. Input Parameters Required for CHAIN-2D 3-8
Table 3.6. Summary Comparisons of the Vadose Zone Model for Radionuclides in the SSL Process 3-14
Table3.7. Summary of the Use of the Unsaturated Zone Model for Radionuclides in the SSL Process ... 3-16
Table 3.8. Soil Hydraulic Properties at the Las Graces Trench Site for SSG Model Evaluation Study
(Modified from Wierenga et al., 1991) 3-17
Table 3.9. Soluble Transport Properties at the Las Graces Trench Site for SSG Model Evaluation Study
(Modified fromPorro and Wierenga, 1993) 3-18
Table 3.10. Characteristics of the Las Graces Trench Site for SSG Model Evaluation Study
(Modified from Gee et al., 1994) 3-19
Table 3.11. Base Values of Input Parameters for the Unsaturated Zone Models 3-20
Table 3.12. Input Parameters Used for the Dilution Factor Calculation and Transport Simulation of
Radionuclide in the Saturated Zone using FECTUZ/EPACMTP Model 3-21
Table 4.1. Sampling Soil Screening DQOs for Surface Soils 4-3
Table 4.2. Sampling Soil Screening DQOs for Surface Soils under the Max Test 4-9
Table4.3. Probability of Decision Error at 0.5 SSLand2 SSL Using Max Test 4-13
Table 4.4. Values of Sign P for Given Values of the Relative Shift, A/o 4-20
Table 4.5. Percentiles Represented by Selected Values of a and B 4-21
Table 4.6. Values of N for a Given Relative Shift (A/o) for a = 0.05 and B = 0.20 4-22
Table 4.7. Critical Values kslgn for the Sign Test Statistic S+ 4-25
Table 4.8. Soil Screening DQOs for Subsurface Soils 4-28
Table 4.9. Comparison of Error Rates for Max Test, Chen Test (at .20 and . 10 Significance Levels), and
Original Land Test, Using 8 Composites of 6 Samples Each, for Gamma Contamination Data . 4-41
Table 4.10. Error Rates of Max Test and Chen Test at .2 (C20) and . 1 (CIO) Significance Level for
CV= 2, 2.5, 3, 3.5 4-43
Table4.11. Probability of "Walking Away" from an EA When Comparing Two Chemicals to SSLs 4-45
Table 4.12. Means and Cvs for Dioxin Concentrations for 7 Piazza Road Exposure Areas 4-47
Table 5.1. Recommended Area Correction Factors as Function of Source Area 5-2
Table 5.2. Area Correction Factors as Function of Source Area for Selected Radionuclides Calculated Using
Microshield 5-2
Table 5.3. Default Kd Values for Selected Elements, based on EPA (1999) 5-4
vi
-------
Table 5.4. Kd values for cesium as a function of cation exchange capacity or clay content low mica content
soils 5-4
Table 5.5. Kd values for cesium as a function of cation exchange capacity or clay content high mica content
soils 5.5
Table 5.6. Kd values for plutonium as a function of soluble carbonate and soil clay values 5.5
Table 5.7. Kd values for strontium as a function of CEC, clay content, and pH 5.5
Table 5.8. Kd values for thorium as a function of pH and dissolved thorium concentrations 5.6
Table 5.9. Kd values for uranium as a function of pH 5.6
VII
-------
LIST OF FIGURES
Figure 1. Conceptual Risk Management Spectrum for Contaminated Soil 1-3
Figure 2. The Data Quality Objectives process 4-2
Figure 3.1. Base case simulations of "TC breakthrough (through the unsaturated zone) with three distribution
coefficients using the CHAIN, HYDRUS, and CHAIN 2D models 3-22
Figure 3.2. Base case simulations of "TC breakthrough (through the unsaturated zone) with three distribution
coefficients using the FECTUZ, and MULTIMED_DP models 3-23
Figure 3.3. Base case simulations of the daughter product - "Ru of "TC breakthrough (through the unsaturated
zone) with a distribution coefficient of 0.007 mL/g using FECTUZ, and CHAIN models 3-24
Figure 3.4. Base case prediction of 99TC concentrations at the receptor well by employing dilution factor
method and using saturated zone model (SZM) - FECTUZ. The numbers in the parentheses are
ground water velocity. The solid lines are used to represent the MCL (5.3 x 10-5 ml/L) Level . 3-25
Figure 4. Design performance goal diagram 4-10
Figure 5. Systemic (square grid points) sample with systemic compositing scheme (6 composite samples
consisting of 4 specimens) 4-12
Figure 6. Systemic (square grid points) sample with random compositing scheme (6 composite samples
consisting of 4 specimens) 4-14
Figure 7. Stratified random sample with random compositing scheme (6 composite samples consisting of 4
specimens) 4-15
LIST OF HIGHLIGHTS
Highlight 1. Key Attributes of the Soil Screening Guidance for Radionuclides 1-4
Highlight 2. Simplifying Assumptions for the Migration to Ground Water Pathway 2-21
Highlight 3. Procedure for Compositing of Specimens from a Grid Sample Using a Systematic Scheme
(Figure 6) 4-12
Highlight 4. Procedure for Compositing of Specimens from a Grid Sample Using a Random Scheme
(Figure 7) 4-14
Highlight 5. Procedure for Compositing of Specimens from a Stratified Random Sample Using a Random
Scheme (Figure 8) 4-15
Highlight 6. Directions for Data Quality Assessment for the Max Test 4-18
Highlight 7. Directions for the Sign Test 4-25
VIM
-------
PREFACE
This document provides the technical background for the development of methodologies described in the Soil
Screening Guidance for Radionuclides: User's Guide (EPA/540-R-00-007), along with additional information
useful for soil screening. Together, these documents define the framework and methodology for developing Soil
Screening Levels (SSLs) for radionuclides commonly found at Superfund sites. This document is similar to a
previous guidance document issued by EPA entitled Soil Screening Guidance: Technical Background Document
(EPA/540/R-95/128), which contains information regarding the soil screening of chemicals commonly found at
Superfund sites. This guidance document intended to be consistent with the 1996 guidance document except where
it was necessarry to be different based on a technical difference posed by radionuclides.
IX
-------
ACKNOWLEDGEMENTS
Early drafts of this technical background document were prepared by Sandy Cohen and Associates (SC&A) under EPA
Contract 68D70073 for the Office of Radiation and Indoor Air (ORIA), and the Office of Emergency and Remedial
Response (OERR) U.S. Environmental Protection Agency (EPA). Phil Newkirk of EPA, the EPA Work Assignment
Manager for this effort, guided the effort. Phil Newkirk and Ron Wilhelm of ORIA, and Stuart Walker of OERR, are
the principal EPA authors of the document, with significant contributions from Ken Lovelace and Janine Dinan of
OERR. John Mauro is the SC&A work assignment Leader and principal SC&A author of the document. Phil Newkirk
and Mike Boyd of ORIA conducted the analysis of area correction factors for direct gamma radiation. David Jewett
of the Office of Research and Development (ORD) conducted the analysis of more complicated groundwater models.
The authors would like to thank all EPA reviewers whose careful review and thoughtful comment greatly contributed
to the quality of this document.
-------
Parti: INTRODUCTION
This document provides the technical background for the Soil Screening Guidance for Radionuclides. The Soil
Screening Guidance for Radionuclides is a tool that the U.S. Environmental Protection Agency (EPA) developed to
help standardize and accelerate the evaluation and cleanup of soils contaminated with radioactive materials at sites
on the National Priorities List (NPL) with anticipated future residential land use scenarios.1 This guidance provides
a methodology for environmental science/engineering professionals to calculate risk-based, site-specific, soil
screening levels (SSLs), for radioactive contaminants in soil that may be used to identify areas needing further
investigation at NPL sites.
SSLs are not national cleanup standards. SSLs alone do not trigger the need for response actions or define
"unacceptable" levels of radionuclides in soil. "Screening," for the purposes of this guidance, refers to the process
of identifying and defining areas, radionuclides, and conditions at a particular site that do not require further Federal
attention. Generally, at sites where radionuclide concentrations fall below SSLs, no further action or study is
warranted underthe Comprehensive Environmental Response, Compensation, and Liability Act (CERCLA). (Some
States have developed screening numbers or methodologies that may be more stringent than SSLs; therefore further
study may be warranted under State programs.) Where radionuclide concentrations equal or exceed the SSLs, further
study or investigation, but not necessarily cleanup, is warranted.
EPA's Soil Screening Guidance: User's Guide (EPA. 1996a) and the Soil Screening Guidance: Technical Background
Document (EPA 1996b) apply the SSL framework to NPL sites with hazardous organic and inorganic soil
contaminants. They do not address sites with radioactive contaminants. These documents provide standardized
exposure equations for deriving generic and site-specific SSLs for chemicals under a residential land use setting,
assuming three soil exposure pathways—soil ingestion, inhalation of volatiles and fugitive dusts, and ingestion of
contaminated ground water. Chemical- specific SSLs are based on a target risk of one-in-a-million (10~6) for
carcinogens, a hazard quotient of 1 for noncarcinogens, or, for the ground water migration pathway, a nonzero
maximum contaminant level goal (MCLG), maximum contaminant level (MCL), or a risk-based level. For each
contaminant, the lowest pathway-specific SSL is selected as the appropriate screening level. The guidance provides
default, generic SSLs for 110 chemicals based on parameter values consistent with Superfund's concept of reasonable
maximum exposure (RME). The guidance also presents a tiered approach for evaluating contaminant migration to
ground water, methods for deriving site-specific dilution/attenuation factors (DAFs) and particulate emission factors
(PEFs), and guidelines for measuring and comparing contaminant soil concentrations with SSLs.
A soil screening framework for radionuclides has been developed that is consistent and compatible with the SSL
framework for chemicals. Since radiologically contaminated sites are a subset of hazardous waste sites, much of the
analysis conducted for the two 1996 SSL documents was considered valid for this effort to address radioactive
contaminants. Radionuclide SSLs are based on a target risk of one-in-a-million (10~6), or, for the ground water
migration pathway, a maximum contaminant level (MCL). These SSLs, in activity units of picocuries per gram of
soil (pCi/g), are derived from equations combining exposure information assumptions with EPA radiotoxicity data.
The Soil Screening Guidance provides a framework for screening soils contaminated with radionuclides that
encompasses both simple and more detailed approaches for calculating site-specific SSLs, and generic SSLs for use
where site-specific data are limited. The Soil Screening Guidance for Radionuclides: User's Guide (U.S. EPA, 2000)
1. Note that the Superfund program defines "soil" as having a particle size under 2 millimeters, while the RCRA program
allows for particles under 9 millimeters in size.
1-1
-------
focuses on the application of the simple site-specific approach by providing a step-by-step methodology to calculate
site-specific SSLs and plan the sampling necessary to apply them. This Technical Background Document describes
the development and technical basis of the methodology presented in the User's Guide. It includes detailed modeling
approaches for developing screening levels that can take into account more complex site conditions than the simple
site-specific methodology emphasized in the User's Guide. It also provides generic SSLs for the most common
contaminants found at NPL sites.
1.1 Background
The Soil Screening Guidance for chemicals was the result of technical analyses and coordination with numerous
stakeholders. The effort began in 1991 when the EPA Administrator charged the Office of Solid Waste and
Emergency Response (OSWER) with conducting a 30-day study to outline options for accelerating the rate of
cleanups at NPL sites. One of the specific proposals of the study was for OSWER to "examine the means to develop
standards or guidelines for contaminated soils." From 1991 to 1995, several drafts of the guidance and the
accompanying technical background document received widespread reviews both within and outside EPA. In the
Spring of 1995, final drafts were released for public comment and external scientific peer review. Many reviewers'
comments contributed significantly to the development of this flexible tool that uses site-specific data in a
methodology that can be applied consistently across the nation.
These SSL guidance documents specifically excluded radionuclides. ORIA developed a soil-screening guidance
document for radionuclides to supplement 1996 guidance. The SSG for radionuclides guidance and TBD are
intended to be consistent with the original chemical SSG documents, except where there is a technical difference
between chemicals and radionuclides. Almost all of the new information contained in the radionuclide SSG guidance
and TBD have been previously released for public comment and/or externally scientific peer reviewed as part of other
guidance/rulemaking efforts.
Since radionuclides are considered a hazardous substance under section 101(14)(E) of CERCLA, the Agency has the
authority and responsibility to oversee the cleanup of Superfund sites contaminated with radionuclides. This
guidance will assist OSC's and RPM's in making decisions at these sites.
1.2 Purpose of SSLs
In identifying and managing risks at sites, EPA considers a spectrum of radionuclide concentrations. The level of
concern associated with those concentrations depends on the likelihood of exposure to radioactive soil contamination
at levels of potential concern to human health. Figure 1 illustrates the spectrum of soil contamination encountered
at Superfund sites and the conceptual range of risk management. At one end are levels of contamination that clearly
warrant a response action; at the other end are levels that warrant no further study under CERCLA. Appropriate
cleanup goals for a particular site may fall anywhere within this range depending on site-specific conditions.
Screening levels identify the lower bound of the spectrum—levels below which no further study is warranted under
CERCLA, provided conditions associated with the SSLs are met.
1-2
-------
No further study
warranted under
CERCLA
Site-specific
cleanup
goal/level
Response
action clearly
warranted
"Zero"
concentration
Screening
level
Response
level
Very high
concentration
Figure 1. Conceptual Risk Management Spectrum for Contaminated Soil
Although the application of SSLs during site investigations is not mandatory at sites being addressed by CERCLA
or RCRA, EPA recommends the use of SSLs as a tool to facilitate prompt identification of radionuclides and
exposure areas of concern. EPA developed the Soil Screening Guidance for Radionuclides to be consistent with and
to enhance the current Superfund investigation process and anticipates it will be used primarily during the early
stages of a remedial investigation (RI) atNPL sites. It does not replace the Remedial Investigation/Feasibility Study
(RI/FS) or risk assessment, but use of screening levels can focus the RI and risk assessment on aspects of the site that
are more likely to be a concern under CERCLA. By screening out areas of sites, potential radionuclides of concern,
or exposure pathways from further investigation, site managers and technical experts can limit the scope of the
remedial investigation or risk assessment. SSLs can save resources by helping to determine which areas do not
require additional Federal attention early in the process. Furthermore, data gathered during the soil screening process
can be used in later Superfund phases, such as the baseline risk assessment, feasibility study, treatability study, and
remedial design. This guidance may also be appropriate for use by the removal program when demarcation of soils
above residential risk-based numbers coincides with the purpose and scope of the removal action. EPA created the
Soil Screening Guidance for Radionuclides to be consistent with and to enhance current Superfund processes.
The purpose of the process presented in this guidance is to develop and apply simple, site-specific soil screening
levels. This approach is likely to be most useful where it is difficult to determine whether areas of soil are
contaminated to an extent that warrants further investigation or response (e.g., whether areas of soil at an NPL site
require further investigation under CERCLA through an RI/FS). The screening levels have been developed assuming
future residential land use assumptions and related exposure scenarios. Although some of the models and methods
presented in this guidance could be modified to address exposures under other land uses, EPA has not yet
standardized assumptions for those other uses. Using this guidance for sites where residential land use assumptions
do not apply could result in overly conservative screening levels. However, EPA recognizes that some parties
responsible for sites with non-residential land use might still benefit from using SSLs as a tool to conduct
conservative initial screening.
EPA created the Soil Screening Guidance for Radionuclides: User's Guide (U.S. EPA, 1999) to be easy to use: it
provides a simple step-by-step methodology for calculating SSLs that are specific to the user's site. Applying site-
specific screening levels involves developing a conceptual site model (CSM), collecting a few easily obtained site-
specific soil parameters (such as the dry bulk density and percent soil moisture), and sampling soil to measure
radionuclide levels in surface and subsurface soils. Often, much of the information needed to develop the CSM can
be derived from previous site investigations (e.g., the preliminary assessment/site inspection [PA/SI]) and, if properly
planned, SSL sampling can be accomplished in one mobilization.
SSLs can be used as Preliminary Remediation Goals (PRGs) provided appropriate conditions are met (i.e., conditions
found at a specific site are similar to conditions assumed in developing the SSLs). The concept of calculating risk-
based soil levels for use as PRGs (or "draft" cleanup levels) was introduced in the Risk Assessment Guidance for
Superfund (RAGS), Volume I, Human Health Evaluation Manual (HHEM), Part B (U.S. EPA, 1991b). PRGs are
risk-based values that provide a reference point for establishing site-specific cleanup levels. The models, equations,
1-3
-------
and assumptions presented in the Soil Screening Guidance for Radionuclides and described herein supersede those
described in RAGS HHEM, Part B, for residential soils. In addition, this guidance presents methodologies to
address the leaching of radionuclides through soil to an underlying potable aquifer. This pathway should also
be addressed in the development of PRGs.
EPA emphasizes that SSLs are not cleanup standards. SSLs should not be used as site-specific cleanup levels unless
a site-specific nine-criteria evaluation using SSLs as PRGs for soils indicates that a selected remedy achieving the
SSLs is protective, compliant with applicable or relevant and appropriate requirements (ARARs), and appropriately
balances the other criteria, including cost. PRGs may then be converted into final cleanup levels based on the nine-
criteria analysis described in the National Contingency Plan (NCP; Section 300.430 (3)(2)(A)). The directive entitled
Role of the Baseline Risk Assessment in Superfund Remedy Selection Decisions (U.S. EPA, 1991c) discusses the
modification of PRGs to generate cleanup levels.
The generic SSLs provided in Appendix A are calculated from the same equations used in the simple site-specific
methodology, but are based on a number of default assumptions chosen to be protective of human health for most
site conditions. Generic SSLs can be used in place of site-specific screening levels; however, they are expected to
be generally more conservative than site-specific levels. The site manager should weigh the cost of collecting the
data necessary to develop site-specific SSLs with the potential for deriving a higher SSL that provides an appropriate
level of protection.
1.3 Scope of Soil Screening Guidance for Radionuclides
The Soil Screening Guidance for Radionuclides incorporates readily obtainable site data into simple, standardized
equations to derive site-specific screening levels for selected radionuclides and exposure pathways. Key attributes
of the Soil Screening Guidance for Radionuclides are given in Highlight 1.
Highlight 1: Key Attributes of the Soil Screening Guidance for Radionuclides
Standardized equations are presented to address human exposure pathways in a residential
setting consistent with Superfund's concept of "Reasonable Maximum Exposure" (RME).
Source size (area and depth) can be considered on a site-specific basis.
Parameters are identified for which site-specific information is needed to develop site-specific
SSLs.
Default parameter values are provided to calculate generic SSLs where site-specific information is
not available.
SSLs are generally based on a 10"6 lifetime cancer risk. SSLs for migration to ground water are
based on maximum contaminant levels (MCLs).
Radiation risk coefficients used to calculate SSLs represent the average risk per unit exposure to
members of a population exposed throughout life to a constant concentration of a radionuclide in
a specific environmental medium. They assume no radioactive decay.
1.3.1 Exposure Pathways. In a residential setting, potential pathways of exposure to radionuclides in soil
included in this guidance are as follows (see Figure 2):
• Direct ingestion of soil
Inhalation of fugitive dusts
External radiation exposure from photon-emitting radionuclides in soil
1-4
-------
• Ingestion of homegrown produce that has been contaminated via plant uptake
• Ingestion of contaminated ground water caused by migration of radionuclides through soil to an underlying
potable aquifer
The Soil Screening Guidance for Radionuclides addresses each of these pathways to the greatest extent practical.
The mode of exposure to radionuclides is different than that of chemicals. This renders some chemical pathways
inapplicable to radionuclides (e.g., inhalation of volatiles, dermal absorption) while adding other pathways unique
to radiation (e.g., external exposure to photons emitted by radionuclides). The radiological pathways listed above
represent the most likely exposure mechanisms for individuals in a residential setting. The external exposure
pathway is, for most radionuclides, the dominant exposure and typically represents the most significant risk. For
some radionuclides, the ingestion of contaminated produce and drinking water constitute the most likely exposure
pathways provided that these items are obtained from onsite sources. The inhalation of fugitive dust pathway is
included in the analysis; however, it is of significance for only a very few radionuclides. All of these pathways have
generally accepted radiological risk methods, models, and assumptions that lend themselves to a standardized
approach.
The Soil Screening Guidance for Radionuclides addresses the human exposure pathways listed previously and
will be appropriate for most residential settings. The presence of additional pathways or unusual site
conditions does not preclude the use of SSLs in areas of the site that are currently residential or likely to be
residential in the future. However, the risks associated with these additional pathways or conditions (e.g., fish
consumption, raising of livestock for meat or milk consumption, fugitive dusts caused by heavy truck traffic
on unpaved roads) should be considered in the remedial investigation/feasibility study (RI/FS) to determine
whether SSLs are adequately protective.
An ecological assessment should also be performed as part of the RI/FS to evaluate potential risks to ecological
receptors.
The Soil Screening Guidance for Radionuclides should not be used for areas with chemical contaminants.
1.3.2 Exposure Assumptions. SSLs are risk-based concentrations derived from equations combining
exposure assumptions with EPA radiotoxicity data. The models and assumptions used to calculate SSLs were
developed to be consistent with Superfund's concept of "reasonable maximum exposure" (RME) in the residential
setting. The Superfund program's method to estimate the RME for chronic exposures on a site-specific basis is to
combine an average exposure point concentration with reasonably conservative values for intake and duration in the
exposure calculations (U.S. EPA, 1989b; U.S. EPA, 1991a). The default intake and duration assumptions presented
in U.S. EPA (1991a) were chosen to represent individuals living in a small town or other nontransient community.
(Exposure to members of a more transient community is assumed to be shorter and thus associated with lower risk.)
Exposure point concentrations are either measured at the site (e.g., ground water concentrations at a receptor well)
or estimated using exposure models with site-specific model inputs. An average concentration term is used in most
assessments where the focus is on estimating long-term, chronic exposures. Where the potential for acute toxicity
is of concern, exposure estimates based on maximum concentrations may be more appropriate.
The resulting site-specific estimate of RME is then compared with a radionuclide-specific toxicity criterion. EPA
recommends using criteria from the Health Effects Assessment Summary Tables (HEAST) which may be found on
the internet at the following address: http://www.epa.gov/superfund/programs/risk/calctool.htm. although values from
other sources may be used in appropriate cases.
SSLs are concentrations of radionuclides in soil that are designed to be protective of exposures in a residential
setting. A site-specific risk assessment is an evaluation of the risk posed by exposure to site radionuclides in various
1-5
-------
media. To calculate SSLs, the exposure equations and pathway models are run in reverse to backcalculate an
"acceptable level" of a radionuclide in soil corresponding to a specific level of risk.
1.3.3 Risk Level. For each pathway, radiotoxicity criteria are used to define an acceptable level of radionuclides
in soil, based on a one-in-a-million (10~6) individual excess lifetime cancer risk. SSLs are backcalculated for
migration to ground water pathway using ground water concentration limits [maximum contaminant levels (MCLs)].
The potential for additive effects has not been "built in" to the SSLs through apportionment. While the pathways
included in the analysis are considered to represent those a residential setting, SSLs are not calculated for a specific
scenario (i.e., SSLs are not summed over a set of pathways). EPA believes that setting a 10~6 risk level for individual
radionuclides and pathways will generally lead to cumulative risks within the risk range (10~4 to 10~6) for the
combinations of radionuclides typically found at Superfund sites.
1.3.4 SSL Model Assumptions. The models used to calculate SSLs were designed for use at an early stage
of site investigation when site information may be limited. Because of this constraint, they incorporate a number
of simplifying assumptions.
The models assume that the source is infinite (steady overtime). Although the assumption is highly conservative,
a finite source model cannot be applied unless there are accurate data regarding source size and volume. EPA
believes it to be unlikely that such data will be available from the limited subsurface sampling that is done to apply
SSLs. However, EPA also recognizes that infinite source (i.e., steady-state) models can violate mass balance (i.e.,
can release more contaminants than are present) for certain contaminants and site conditions (e.g., small sources).
To address this problem, this guidance includes simple models that provide a mass-based limit for the inhalation and
migration to ground water SSLs (see Section 2.6). A site-specific estimate of source depth and area are required
to calculate SSLs using these alternative models.
The infinite source (i.e., steady-state) assumption leads to several other simplifying assumptions. Fractionation of
contaminant mass between the inhalation and migration to ground water pathways cannot be addressed with infinite
source (i.e., steady-state) models. For the migration to ground water pathway, an infinite source (i.e., steady-state)
overrides adsorption in the unsaturated zone or in the aquifer. The models also assume that contamination is evenly
distributed throughout the source (i.e., homogeneous) and that no biological or chemical degradation occurs in the
soil or in the aquifer. Again, models capable of addressing heterogeneities or degradation processes require
collection of site-specific data that is well beyond the scope of the Soil Screening Guidance for Radionuclides.
Although the Soil Screening Guidance for Radionuclides encourages the use of site-specific data to calculate SSLs,
conservative default parameters are provided for use where site-specific data are not available. These defaults are
described in Part 2 of this document. Appendix A provides an example set of "generic" SSLs for 60 radionuclides
that are calculated using these defaults. Because they are designed to be protective of most site conditions across
the nation, they are conservative.
A default 0.5 acre source area is used to calculate the generic SSLs. A 30 acre source size was used in the December
1994 Soil Screening Guidance for chemicals. EPA received an overwhelming number of comments that suggest that
most contaminated soil sources addressed under the Superfund program are 0.5 acres or smaller. Because of the
infinite source (i.e., steady-state) assumption, generic SSLs based on a 0.5 acre source size can be protective of larger
sources as well (see Appendix A). However, this hypothesis should be examined on a case-by-case basis before
applying the generic SSLs to sources larger than 0.5 acre.
1-6
-------
1.4 Organization of the Document
Part 2 of this document describes the development of the simple equations used to calculate SSLs. It describes and
supports the assumptions behind these equations and presents the results of analyses conducted to develop the SSL
methodology. Some of the more sensitive parameters are identified for which site-specific data are likely to have
a significant impact. Default values are provided along with their sources and limitations.
Part 3 presents information on other, more complex models that can be used to calculate SSLs when more extensive
site data are available or can be obtained. Some of these models can consider a finite source and fractionation
between exposure pathways. They also can model more complex site conditions than the simple SSL equations,
including conditions that can lead to higher, yet still protective, SSLs (e.g., thick unsaturated zones, biological and
chemical degradation, layered soils).
Part 4 provides the technical background for the development of the soil sampling design methodology for SSL
application. It addresses methods for surface soil, including a test based on a maximum soil composite sample, the
Max test, and the Sign test, which allows decision errors to be controlled. Lastly, Part 4 provides simulation results
that measure the performance of these methods and sample size tables for different contaminant distributions and
compositing schemes. Step-by-step guidance is provided for developing sample designs using each statistical
procedure.
Part 5 describes the selection and development of the radionuclide properties used to calculate SSLs.
1-7
-------
Part 2: DEVELOPMENT OF PATHWAY-SPECIFIC
SOIL SCREENING LEVELS
This part of the Technical Background Document for Radionuclides describes the methods used to calculate SSLs
for residential exposure pathways, along with their technical basis and limitations associated with their use. Simple,
standardized equations have been developed for the more common radiological exposure pathways at Superfund
sites:
• Direct ingestion of soil (Section 2.2)
• Inhalation of fugitive dust (Section 2.3)
• External radiation exposure from photon-emitting radionuclides in soil (Section 2.4)
• Ingestion of homegrown produce that has been contaminated via plant uptake (Section 2.5)
• Ingestion of contaminated ground water caused by migration of radionuclides through soil to an
underlying potable aquifer (Section 2.6).
The equations were developed under the following constraints:
• They should be consistent with current Superfund risk assessment methodologies and guidance.
• To be appropriate for early-stage application, they should be simple and easy to apply.
• They should allow the use of site-specific data where they are readily available or can be easily
obtained.
• The process of developing and applying S SLs should generate information that can be used and built
upon as a site evaluation progresses.
The equations include easily obtained site-specific input parameters. Conservative default values have been
developed for use where site-specific data are not available. Generic SSLs, calculated for 60 radionuclides using
these default values, are presented in Appendix A. The generic SSLs are conservative, since the default values are
designed to be protective at most sites across the country.
The migration to ground waterpathway equation assumes an infinite source (e.g., source remains constant overtime).
As pointed out by several commenters to the December 1994 draft Soil Screening Guidance for chemicals (U.S. EPA,
1994h), SSLs developed using these models may violate mass-balance for certain contaminants and site conditions
(e.g., small sources). To address this concern, EPA has incorporated a simple mass-limit model for this pathway
assuming that the entire volume of contamination leaches over the duration of exposure and that the level of
contaminant at the receptor does not exceed the health-based limit (Section 2.7). Because they require a site-
specific estimate of source depth, these alternative models cannot be used to calculate generic SSLs.
2-1
-------
The mode of exposure to radionuclides is different than that of chemicals. This renders some chemical pathways
inapplicable to radionuclides (e.g., inhalation of volatiles, dermal absorption) while adding other pathways unique
to radiation (e.g., external exposure to photons emitted by radionuclides). The radiological pathways listed above
represent the most likely exposure mechanisms for individuals in a residential setting. For soils under the residential
land use assumption, the external exposure pathway will typically be the dominant exposure pathway for most
radionuclides (e.g., 60Co, 137Cs, 226Ra, 238U). For a good number of radionuclides (e.g., 14C, 63Ni, 90Sr, 99Tc), the plant
ingestion pathway often dominates, although not to the extent that the external exposure pathway does. The soil
ingestion pathway also plays a dominant role for some radionuclides of interest (i.e., 1291,232Th, 239Pu, 241Am). In the
majority of cases, the inhalation of fugitive dust pathway and the migration to groundwater pathway play an
insignificant role.
In addition to the more common pathways of exposure in a residential setting, concerns have been raised regarding
the potential for migration of radon from subsurface soils into basements. The dominant factor in indoor radon levels
is home construction practices and the extent to which these practices employ radon-resistant techniques. Homes
built atop soil with identical levels of radium can have orders of magnitude differences in indoor radon levels
depending on the extent to which radon-resistanttechniques are used. As a naturally-occurring radionuclide, radium
is present in all soils. Reducing the radium content in the soil may not result in any reduction in indoor radon levels.
However, taking simple and inexpensive steps in home construction will ensure that radon levels in homes are kept
below Applicable or Relevant and Appropriate Requirements (ARARs) levels. For existing homes with elevated
levels of radon, a variety of methods can be used to reduce radon concentrations to ARAR levels. Discussion of
radon mitigation standards may be found in several EPA publications, including Radon Mitigation Standards, EPA
402-R-93-078 (U.S. EPA, 1994i). Also note that potential ARARs exist for indoor radon under 192.12(b)(l) and
192.41(b). For further guidance on using these ARARs, see the August 1997 memorandum from Stephen Luftig
(OERR) and Larry Weinstock (ORIA) titled "Establishment of cleanup levels for CERCLA sites with radioactive
contamination," OSWER Directive 9200.4-18, (U.S. EPA, 1997a).
Section 2.1 describes the human health basis of the Soil Screening Guidance for Radionuclides and provides the
human toxicity and health benchmarks necessary to calculate SSLs. The selection and development of the
radiological properties required to calculate SSLs are described in Part 5 of this document.
2.1 Human Health Basis
The 60 radionuclides for which generic SSLs have been calculated are listed in Table 2.1. Principal radionuclides
are radionuclides with half-lives greater than six months. The decay products of any principal radionuclide down
to, but not including, the next principal radionuclide in its decay chain are called associated radionuclides and consist
of radionuclides with half-lives less than six months. It is assumed that a principal radionuclide is in secular
equilibrium with its associated radionuclides at the point of exposure. This assumption is reasonable because it
usually takes about three years or longer to clean up a site. Associated decay chains are indicated, as well as principal
radionuclide half-life and the terminal nuclide or radionuclide (i.e., the principal radionuclide or stable nuclide that
terminates an associated decay chain).
Tables 2.2 and 2.3 list the regulatory and human health benchmarks necessary to calculate SSLs for 60 radionuclides.
For all pathways other than migration to ground water, these benchmarks are based on the latest available
radionuclide cancer slope factors. For migration to ground water, drinking water MCLs are shown.
The human health benchmarks in Table 2.2 were obtained from Health Risks from Low-Level Environmental
Exposure to Radionuclides, Federal Guidance Report No. 13, Part I - Interim Version, EPA 402-R-97-014 (U.S. EPA
1998a) (also know as FGR 13). When the slope factors are not available in FGR 13, data are taken from Health
Effects Assessment Summary Tables (HEAST): Annual Update, FY-1995, (U.S. EPA, 1995a) (also known as the
2-2
-------
HEAST report). Slope factors are listed for each route of intake for principal radionuclides in units of picocuries
(pCi).1 Radionuclides are presented alphabetically by element and atomic weight.
MCLs in Table 2.3 were obtained from Drinking Water Regulations and Health Advisories (U.S. EPA, 1995a). The
current MCLs for radionuclides were enacted in 1976 and are set at 4 mrem/yr for the sum of the doses from beta
particles and photon emitters, 15 pCi/L for gross alpha particle activity (including Ra-226, but excluding uranium
and radon), and 5 pCi/L combined for Ra-226 and Ra-228. In July 1991, EPA proposed to revise the MCLs for Ra-
226 and Ra-228 to 20 pCi/L for each, change the methodology used for determining a 4 mrem/yr dose for the sum
of the doses from beta particles and photon emitters, alter the definition of alpha particle activity to exclude Ra-226,
and establishing new MCLs of 300 pCi/L for Rn-222 and 20 (jg/L (30 pCi/L) for uranium (56 FR 33050). EPA is
under Court Order to either finalize the 1991 proposal for radionuclides (except for radon), or to ratify existing
standards by November 2000. On April 21,2000 EPA solicited comment in a Notice of Data Availability (NODA)
on three options for a uranium MCL: 1) 20 ug/1 and 20 pCi/1 as a preferred option, 2) 40 ug/1 and 40 pCi/1, and 3)
80 ug/1 and 80 pCi/1 (65 FR 21576). In this NODA, EPA indicated that changes would not be made to the existing
MCLs for radium, alpha particle activity, and beta particles and photon emitters. The 1996 Amendments to the Safe
Drinking Water Act (SDWA) require EPA to propose a MCL for radon by August 1999, and to finalize the MCL
by August 2000. To comply with the requirements of the amended SDWA, on August 6, 1997, EPA withdrew its
1991 proposal for Rn-222 (62 FR 42221). EPA issued a new proposal for Rn-222 on November 2, 1999 (64 FR
59246). EPA proposed an MCL of 300 pCi/1 with an alternative MCL of 4,000 pCi/1 if a state or local indoor radon
mitigation program was established.
References for each table are updated regularly. Prior to calculating SSLs, the values in Tables 2.2 and 2.3 should
be checked against the most recent version of these sources to ensure that they are up-to-date.
Selected radionuclides and radioactive decay chain products are designated with the suffix "+D" (e.g., U-238+D, Ra-
226+D, Cs-137+D) to indicate that cancer risk estimates for these radionuclides include the contributions from their
short-lived decay products, assuming equal activity concentrations (i.e., secular equilibrium) with the principal or
parent nuclide in the environment.2 In most cases, site-specific analytical data should be used to establish the actual
degree of equilibrium between each parent radionuclide and its decay products in each media sampled. However,
in the absence of empirical data, the "+D" values for radionuclides should be used unless there are compelling
reasons not to. Note that there may be circumstances, such as long disposal times or technologically enhanced
concentrations of naturally occurring radionuclides, that may necessitate the combination of the risks of a parent
radionuclide and its decay products over several contiguous subchains. For example, Ra-226 soil analyses at a site
might show that all radium decay products are present in secular equilibrium down to stable Pb-206. In this case,
Ra-226 risk calculations should be based on the ingestion, inhalation and external exposure slope factors for the Ra-
226+D subchain. plus the ingestion, inhalation and external exposure factors for the Pb-210+D subchain. For actual
sites, users should consult with a health physicist or radiochemist (1) to evaluate the site-specific analytical data to
determine the degree of equilibrium between parent radionuclides and decay members of contiguous decay chains
and (2) to assist in the combination of appropriate slope factor values.
1 Slope factors are reported in the customary units of risk per picocurie (1 pCi = 10"12 curies (Ci) = 3.7xlO"2
nuclear transformations per second) for consistency with the system used for radionuclides in the IRIS database. If
required, slope factors can be converted into the International System (SI) units of becquerels (1 Bq = 1 nuclear
transformation per second) by multiplying each inhalation, ingestion, or external exposure value by 27.03. Users can
calculate cancer risks using slope factors expressed in either customary units or SI units with equivalent results,
provided that they also use air, water and soil concentration values in the same system of units.
2 There is one exception to the assumption of secular equilibrium. For the inhalation slope factor for Rn-
222+D reported in the table, ORIA assumes a 50% equilibrium value for radon decay products (Po-218, Pb-214, Bi-214
andPo-214) in air.
20
-3
-------
Table 2.1. Radionuclides Included in Generic Soil Screening Analysis
Principal Radionuclide*
Nuclide
Ac-227+D
Ag-108m
Ag-110m
Am-241
Am-243+D
Bi-207
C-14
Cd-109
Ce-144+D
CI-36
Cm-243
Cm-244
Co-57
Co-60
Cs-134
Cs-135
Cs-137+D
Eu-152
Eu-154
Eu-155
Fe-55
Gd-153
H-3
1-129
K-40
Mn-54
Half-life (yr)
22
127
0.7
432
7,400
38
5,730
1.3
0.8
3.0E+5
28
18
0.7
5
2
3E+6
30
13
8
5
3
0.7
12
1.6E+7
1.3E+9
0.9
Associated Decay Chainb
[Th-227 (98.6%, 19 d)
Fr-223(1.4%, 22 min)]
Ra-223(11 d)
Rn-219(4s)
Po-215(2ms)
Pb-211 (36 min)
Bi-211 (2 min)
[TI-207 (99.7%, 5 min)
Po-211 (0.3%, 0.5s)]
-
Np-239 (2 d)
-
-
-
[Pr-244(9%, 17 min)
pr-244m (2%, 7 min)]
-
-
-
-
-
-
-
Ba-137m(95%, 3 min)
-
-
-
-
-
-
-
-
-
Terminal Nuclide or
Radionuclide0
Nuclide
Pb-207
Pd-108(91%)
[Cd-108(98%}
Ag-108(9%)
Pd-108(2%)]
Cs-110(99%)
[Cd-11 0(99.7%)
Ag-110(1%)
Pd-110(0.3%)]
Np-237
Pu-239
Pb-207
N-14
Ag-109
Nd-144
S-36
Am-243 (0.2%)e
Pu-240
Fe-57
Ni-60
Ba-134(~100%)
Ba-135
Ba-137
Sm-152(72%)
Gd-152(28%)
Gd-154(~100%)
Gd-155
Mn-55
Eu-153
He-3
Xe-129
Ca-40 (89%)
Ar-40(11%)
Cr-54
Half-life (yr)
stable
stable
stable
2 min
stable
stable
stable
25s
stable
2.1E+6
2.4E+4
stable
stable
stable
stable
stable
7,400
6,600
stable
stable
stable
stable
stable
stable
1.1E+14
stable
stable
stable
stable
stable
stable
stable
stable
2-4
-------
Table 2.1. Radionuclides Included in Generic Soil Screening Analysis
Principal Radionuclide"
Nuclide
Na-22
Nb-94
Ni-59
Ni-63
Np-237+D
Pa-231
Pb-210+D
Pm-147
Pu-238
Pu-239
Pu-240
Pu-241
Pu-242
Pu-244+D
Ra-226+D
Ra-228+D
Ru-106+D
Sb-125+D
Sm-147
Sm-151
Sr-90+D
Tc-99
Th-228+D
Th-229+D
Th-230
Th-232
Half-life (yr)
3
2.0E+4
7.5E+4
100
2.1E+6
3.3E+4
22
3
88
2.4E+4
6,500
14
3.8E+5
9.3E+7
1,600
8
1
3
1.1E+11
90
29
2.1E+5
2
7,300
7.7E+4
1.4E+10
Associated Decay Chain"
-
-
-
-
Pa-233 (27 d)
-
Bi-210(5d)
Po-210(138y)
-
-
-
-
-
-
U-240-100%, 14)
Np-240
Rn-222 (4 d)
Po-218(3min)
Pb-214(~100%, 27min)
Bi-214(20min)
Po-214(~100%, 1 min)
Ac-228 (6 h)
Rh-106(30s)
Te-125m(23%, 58 d)
-
-
Y-90 (64 h)
-
Ra-224 (4 d)
Rn-220 (56 s)
Po-216(0.2s)
Pb-212(11h)
Bi-212(61 min)
[Po-212(64%, 0.3 |js)
TI-208 (36%, 3 min)]
Ra-225(15d)
Ac-225(10d)
Fr-221 (5 min)
At-217(32ms)
Bi-213(46min)
[Po-213(98%, 4 |js)
TI-209 (2%, 2 min)]
Pd-209 (3 h)
-
-
Terminal Nuclide or
Radionuclide0
Nuclide
Ne-22
Mo-94
Co-59
Cu-53
U-233
Ac-227
Pb-206
Sm-147
U-234
U-235
U-236
Am-241
U-238
Pu-240
Pb-210
Th-228
Pd-106
Te-125
Nd-143
Eu-151
Zr-90
Ru-99
Pb-208
Bi-209
Ra-226
Ra-228
Half-life (yr)
stable
stable
stable
stable
1.6E+5
22
stable
1.1E+11
2.4E+5
7E+8
2.3E+6
432 y
4.5E+9
6,500
22
2
stable
stable
stable
stable
stable
stable
stable
stable
1,600
6
2-5
-------
Table 2.1. Radionuclides Included in Generic Soil Screening Analysis
Principal Radionuclide"
Nuclide
TI-204
U-232
U-233
U-234
U-235+D
U-236
U-238+D
Zn-65
Half-life (yr)
4
72
1.6E+5
2.4E+5
7.0E+8
2.3E+6
4.5E+9
0.7
Associated Decay Chain"
-
-
-
-
Th-231 (26 h)
-
Th-234 (24 d)
[Pa-234m (99.8%, 1 min)
Pa-234 (0.2%, 7 h)]
-
Terminal Nuclide or
Radionuclide0
Nuclide
Pb-204 (97%)
Hg-204 (3%)
Th-228
Th-229
Th-230
Pa-231
Th-232
U-234
Cu-65
Half-life (yr)
stable
stable
2
7,300
8E+4
3.4E+4
1.4E+10
2.4E+5
stable
Radionuclides with half-lives greater than six months. "+D" designates principal radionuclides with associated decay chains.
The chain of decay products of a principal radionuclide extending to (but not including) the next principal radionuclide or a stable
nuclide. Half-lives are given in parentheses. Branches are indicated by square brackets with branching ratios in parentheses.
The principal radionuclide or stable nuclide that terminates an associated decay chain.
A hyphen indicates that there are no associated decay products.
The branching decay for Pu-241 and Cm-243 involves multiple principal radionuclides and associated radionuclides.
2-6
-------
Table 2.2. Radionuclide Cancer Morbidity - Slope Factors (1)
Radionuclide
Ac-227+D
Ag-108m
Ag-110m
Am-241
Am-243+D
Bi-207
C-14
Cd-109
Ce-144+D
CI-36
Cm-243
Cm-244
Co-57
Co-60
Cs-134
Cs-135
Cs-137+D
Eu-152
Eu-154
Eu-155
Fe-55
Gd-153
H-3
1-129
K-40
Mn-54
Na-22
Nb-94
Ni-59
Ni-63
Np-237+D
Pa-231
Pb-210+D
Pm-147
Pu-238
Pu-239
Pu-240
Pu-241
Pu-242
Pu-244+D
Ra-226+D
Ra-228+D
Ru-106+D
Sb-125+D
Sm-147
Sm-151
Sr-90+D
Tc-99
Th-228+D
Th-229+D
Th-230
Th-232
TI-204
U-232
U-233
U-234
U-235+D
U-236
U-238+D
Slope Factor (Morbidity Risk Coefficient)
Lifetime Excess Cancer Risk per Unit Exposure
Water
Ingestion
(risk/pCi)
4.86E-10
8.14E-12
9.88E-12
1.04E-10
1.08E-10
5.66E-12
1.55E-12
5.00E-12
3.53E-11
3.30E-12
9.47E-11
8.36E-11
1.04E-12
1.57E-11
4.22E-11
4.74E-12
3.04E-11
6.07E-12
1.03E-11
1.90E-12
8.62E-13
1.52E-12
5.07E-14
1.48E-10
2.47E-11
2.28E-12
9.62E-12
7.77E-12
2.74E-13
6.70E-13
6.74E-11
1.73E-10
1.27E-09
1.69E-12
1.31E-10
1.35E-10
1.35E-10
1.76E-12
1.28E-10
1.44E-10
3.86E-10
1.04E-09
4.22E-11
5.13E-12
3.74E-11
5.55E-13
7.40E-11
2.75E-12
3.00E-10
5.28E-10
9.10E-11
1.01E-10
5.85E-12
2.92E-10
7.18E-11
7.07E-11
7.18E-11
6.70E-11
8.71 E-11
Food
Ingestion
(risk/pCi)
6.53E-10
1.12E-11
1.37E-11
1.34E-10
1.42E-10
8.14E-12
2.00E-12
6.70E-12
5.19E-11
4.44E-12
1.23E-10
1.08E-10
1.49E-12
2.23E-11
5.14E-11
5.88E-12
3.74E-11
8.70E-12
1.49E-11
2.77E-12
1.16E-12
2.22E-12
6.51E-14
3.22E-10
3.43E-11
3.11E-12
1.26E-11
1.11E-11
3.89E-13
9.51E-13
9.10E-11
2.26E-10
3.44E-09
2.48E-12
1.69E-10
1.74E-10
1.74E-10
2.28E-12
1.65E-10
1.90E-10
5.15E-10
1.43E-09
6.11E-11
7.21E-12
4.77E-11
8.07E-13
9.53E-11
4.00E-12
4.22E-10
7.16E-10
1.19E-10
1.33E-10
8.25E-12
3.85E-10
9.69E-11
9.55E-11
9.76E-11
9.03E-11
1.21E-10
Soil
Ingestion
(risk/pCi)
1.16E-09
1.92E-11
2.37E-11
2.17E-10
2.32E-10
1.49E-11
2.79E-12
1.14E-11
1.02E-10
7.66E-12
2.05E-10
1.81E-10
2.78E-12
4.03E-11
5.81 E-11
7.18E-12
4.33E-11
1.62E-11
2.85E-11
5.40E-12
2.09E-12
4.26E-12
9.25E-14
2.71E-10
6.18E-11
5.14E-12
1.97E-11
2.05E-11
7.33E-13
1.79E-12
1.62E-10
3.74E-10
2.66E-09
4.88E-12
2.72E-10
2.76E-10
2.77E-10
3.29E-12
2.63E-10
3.14E-10
7.30E-10
2.29E-09
1.19E-10
1.32E-11
7.59E-11
1.59E-12
1.44E-10
7.66E-12
8.09E-10
1.29E-09
2.02E-10
2.31E-10
1.54E-11
5.74E-10
1.60E-10
1.58E-10
1.63E-11
1.49E-10
2.10E-10
Inhalation
(risk/pCi)
2.09E-07
2.67E-11
2.83E-11
2.81 E-08
2.70E-08
2.10E-11
7.07E-12
2.19E-11
1.10E-10
2.50E-11
2.69E-08
2.53E-08
2.09E-12
3.58E-11
1.65E-11
1.86E-12
1.19E-11
9.10E-11
1.15E-10
1.48E-11
7.99E-13
6.55E-12
5.62E-14
6.07E-11
1.03E-11
5.88E-12
3.89E-12
3.77E-11
4.66E-13
1.64E-12
1.77E-08
4.55E-08
1.39E-08
1.61 E-11
3.36E-08
3.33E-08
3.33E-08
3.34E-10
3.13E-08
2.93E-08
1.16E-08
5.23E-09
1.02E-10
1.93E-11
6.88E-09
4.88E-12
1.13E-10
1.41 E-11
1.43E-07
2.25E-07
2.85E-08
4.33E-08
2.45E-12
1.95E-08
1.16E-08
1.14E-08
1.01 E-08
1.05E-08
9.35E-09
External Exposure
(risk/yr per
PCi/g soil)
1.47E-06
7.19E-06
1.30E-05
2.76E-08
6.36E-07
7.08E-06
7.83E-12
8.73E-09
2.44E-07
1.74E-09
4.19E-07
4.85E-11
3.55E-07
1.24E-05
7.10E-06
2.36E-11
2.55E-06
5.30E-06
5.83E-06
1.24E-07
0
1.62E-07
0
6.10E-09
7.97E-07
3.89E-06
1.03E-05
7.29E-06
0
0
7.97E-07
1.39E-07
4.21 E-09
3.21 E-11
7.22E-11
2.00E-10
6.98E-11
4.11E-12
6.25E-11
1.51E-06
8.49E-06
4.53E-06
9.66E-07
1.81E-06
0
3.60E-13
1.96E-08
8.14E-11
7.76E-06
1.17E-06
8.19E-10
3.42E-10
2.76E-09
5.98E-10
9.82E-10
2.52E-10
5.43E-07
1.25E-10
1.14E-07
Notes
2
2
2
2
3
2
2
4
5
2
2
2
2
2
2
2
2
2
2
2
2
2-7
-------
Radionuclide
Zn-65
Water
Ingestion
(risk/pCi)
1.17E-11
Food
Ingestion
(risk/pCi)
1.54E-11
Soil
Ingestion
(risk/pCi)
2.45E-11
Inhalation
(risk/pCi)
5.81E-12
External Exposure
(risk/yr per
PCi/g soil)
2.81 E-06
Notes
Notes:
1. A curie (Ci), the customary unit of activity, is equal to 3.7 x 1010 nuclear transformations per second. 1 picocurie (pCi) = 10"
12 Ci. If required, slope factors in Table D.1 can be converted into the International System (SI) units of becquerels (1 Bq = 1
nuclear transformation per second) by multiplying each inhalation, ingestion, or external exposure value by 27.03. Users can
calculate cancer risks using slope factors expressed in either customary units or SI units with equivalent results, provided that
they also use air, water, food and soil concentration values in the same system of units.
2. For each radionuclide listed, slope factors correspond to the risks per unit intake or exposure for that radionuclide only,
except when marked with a "+D". In these cases, the risks from associated short-lived radioactive decay products (i.e., those
decay products with radioactive half-lives less than or equal to 6 months) are also included, based on an assumption of
secular equilibrium. These decay chains are identified in Table C.1 of Attachment C.
3. The inhalation slope factor listed represents inhalation of C-14 as a particulate. Alternative values for inhalation of C-14 as
a gas are 3.36E-15 risk/pCi for carbon monoxide and 1.99E-14 risk/pCi for carbon dioxide.
4. The inhalation slope factor for H-3 represents inhalation of titiated water vapor, which is considered the most likely form in
the environment. Alternative values of inhalation of H-3 include 1.99E-13 risk/pCi for particulates, 5.62E-18 risk/pCi for
elemental hydrogen gas, and 1.28E-13 risk/pCi for organic forms. Similarly, the ingestion slope factor values for H-3
represent ingestion of tritiated water, which is considered the most likely form in the environment. Alternative values for
ingestion of organically bound forms of H-3 in water, food, and soil are 1.12E-13 risk/pCi, 1.44E-13 risk/pCi, and 2.02E-13
risk/pCi, respectively.
5. The food ingestion slope factor for 1-129 represents ingestion of milk. For ingestion of non-dairy foodstuffs, a lower value of
1.93E-10 risk/pCi ingested would apply. The inhalation slope factor for 1-129 represents inhalation of particulates; alternative
values for inhalation of 1-129 vapor are 1.24E-10 for inhalation of methyl iodide and 1.60E-10 for inhalation of other
compounds in vapor form.
-------
Table 2.3 Radionuclide Drinking Water MCLs
Radionuclide
Ac-227
Ag-108m
Ag-110m
Am-241
Am-243
Bi-207
C-14
Cd-109
Ce-144
CI-36
Cm-243
Cm-244
Cm-248
Co-57
Co-60
Cs-134
Cs-135
Cs-137
Eu-152
Eu-154
Eu-155
Fe-55
Gd-153
H-3
1-129
K-40
Mn-54
Na-22
Nb-94
Ni-59
Ni-63
Np-237
Pa-231
Pb-210
Pm-147
Pu-238
CurrentMCLab
(pCi/L)
90
15
15
200
2,000
600
30
700
15
15
15
1,000
100
80
900
200
200
60
600
2,000
600
20,000
1
300
400
300
50
15
15
587
15
Proposed MCL
(pCi/L)
Risk Base Limit
(RBL)e
(pCi/L)
0.24
5.8
1.9
6.1
0.054
Mass Equivto MCL,
Proposed MCL, or
RBL (mg/L)
3.3E-12
2.2E-10
1.9E-11
4.4E-09
7.5E-08
4.4E-09
4.5E-07
2.3E-10
9.1E-12
2.1E-05
2.9E-10
1.9E-10
3.5E-06
1.2E-10
8.9E-11
6.2E-11
7.8E-04
2.3E-09
1.1E-09
2.3E-10
1.3E-09
8.3E-10
1.7E-10
2.1E-09
5.7E-06
2.7E-4
3.9E-11
6.4E-11
3.3E-8
3.7E-06
8.5E-10
2.1E-05
3.2E-07
7.1E-13
6.3E-10
8.8E-10
2-9
-------
Table 2.3 Radionuclide Drinking Water MCLs
Radionuclide
Pu-239
Pu-240
Pu-241
Pu-242
Pu-244
Ra-226
Ra-228
Ru-106
Sb-125
Sm-147
Sm-151
Sr-90
Tc-99
Th-228
Th-229
Th-230
Th-232
TI-204
U-232
U-232
U-233
U-233
U-234
U-234
U-235
U-235
U-236
U-236
U-238
U-238
Zn-65
CurrentMCLab
(pCi/L)
15
15
15
15
5C
5C
30
300
15
1,000
8
900
15
15
15
15
300
300
Proposed MCL
(pCi/L)
20d
(20 /4)/l)d
20d
(20 /.g/l)d
20d
(20 /4)/l)d
20d
(20 /.g/l)d
20d
(20 /4)/l)d
20d
(20 /.g/l)d
Risk Base Limit
(RBL)e
(pCi/L)
27
Mass Equivto MCL,
Proposed MCL, or
RBL (mg/L)
2.4E-07
6.6E-08
2.6E-10
3.8E-06
8.5E-04
5.1E-09
1.8E-11
9.0E-12
2.9E-10
6.5E-01
3.8E-08
5.9E-11
5.3E-05
1.8E-11
7.1E-08
7.4E-07
1.4E-01
6.5E-10
9.4E-10
2.0E-02
2.1E-06
2.0E-02
3.2E-06
2.0E-02
9.3E-03
2.0E-02
3.1E-04
2.0E-02
6.0E-02
2.0E-02
3.6E-11
Notes:
Existing MCL is 4 mrem/yr to the whole body or an organ, combined from all beta and photon emitters.
Existing MCL is 15 pCi/L, with the concentration level combined for all alpha emitters, except radon and uranium.
Existing MCL is 5 pCi/L combined for Ra-226 and Ra-228.
Preferred EPA proposed MCL standard is 20 ^g/l and 20 pCi/l for uranium, with EPA soliciting comments on options of 40 ^g/l and
40 pCi/l, and 80 (jgfi and 80 pCi/l. The preferred proposed MCL standard for uranium of 20 ^g/l and 20 pCi/l is represented in this
table.
Risk Based Limits are calculated for 30-year exposure duration and 10"6 risk.
2-10
-------
2.1.1 Additive Risk. For all pathways except migration to ground water, SSLs correspond to a 10~6 lifetime
cancer risk level. While the pathways included in the analysis are considered to represent a residential setting, SSLs
are not calculated for a specific scenario (i.e., SSLs are not summed over a set of pathways). EPA believes that
setting a 10~6 risk level for individual radionuclides and pathways generally will lead to cumulative risks within the
10~4 to 10~6 range for the combinations of radionuclides typically found at Superfund sites.
2.1.2 ToxJCOkinetiCS Of Radionuclides. The exposure assumptions used to develop SSLs are representative
of a chronic exposure pathway. Exposures necessary to cause acute radiation effects are many orders of magnitude
greater than those associated with SSLs. Therefore, it is extremely unlikely that even massive intakes of
contaminated soil could cause such effects. It should be noted, however, that the slope factors used to calculate S SLs
in this report are applicable to either chronic or acute exposure to a radionuclide. That is, the slope factor may be
interpreted either as the average risk per unit exposure to members of a population exposed throughout life to a
constant concentration of a radionuclide in soil, or as the average risk per unit exposure to members of a population
acutely exposed to the radionuclide in soil.
2.2 Direct Ingestion of Soil
Calculation of SSLs for direct ingestion of soil is based on the methodology presented for residential land use in
RAGS HHEM, Part B (U.S. EPA, 1991b). Briefly, this methodology backcalculates a soil concentration level from
a target risk. A number of studies have shown that inadvertent ingestion of soil is common among children 6 years
old and younger (Calabrese et al., 1989; Davis et al., 1990; Van Wijnen et al., 1990). Therefore, the approach uses
an age-adjusted soil ingestion factor that takes into account the difference in daily soil ingestion rates and exposure
duration for children from 1 to 6 years old and others from 7 to 31 years old. The higher intake rate of soil by
children leads to a lower, or more conservative, risk-based concentration compared to an adult-only assumption.
RAGS HHEM, Part B also uses this age-adjusted approach.
For radionuclides, both the magnitude and duration of exposure are important. Duration is critical because the
toxicity criteria are based on "lifetime average daily dose." Therefore, the total intake, whether it be over 5 years
or 50 years, is averaged over a lifetime of 70 years. To be protective of exposures to radionuclides in the residential
setting, RAGS HHEM, Part B (U.S. EPA, 1991b) and EPA focus on exposures to individuals who may live in the
same residence for a "high-end" period of time (e.g., 30 years). As mentioned above, exposure to soil is higher
during childhood and decreases with age. Thus, Equation 1 uses the RAGS HHEM, Part B time-weighted average
soil ingestion rate for children and adults; the derivation of this factor is shown in Equation 2.
2-11
-------
Screening Level Equation for Ingestion of Radionuclides in Residential Soil
(Source: RAGS HHEM, Part B; U.S. EPA, 1991b)
SSL -
TR
SFs * IRs * IE-3 *EF*ED
(1)
Parameter/Definition (units)
TR/target cancer risk (unitless)
SFs/soil ingestion slope factor
(pCi)-1
IRs/soil ingestion rate (mg/d)
1 E-3/conversion factor (g/mg)
EF/exposure frequency (d/yr)
ED/exposure duration (yr)
Default
ID'6
See Table 2.2
120 (age-averaged)
350
30
Source: RAGS HHEM, Part B; U.S. EPA, 1991b
Equation for Age-Adjusted Soil Ingestion Factor, IRS
soil/agel-6 *
soil/age7-31 * ^ age 7
ED
(2)
Parameter/Definition (units)
IRs/age-adjusted soil ingestion factor (mg/d)
IRsoi,/age16 /ingestion rate of soil age 1-6
(mg/d)
EDage16/exposure duration during ages 1-6
(yr)
IRsoi,/age7 31 /ingestion rate of soil age 7-31
(mg/d)
EDage7 31/exposure duration during ages 7-31
(yr)
ED/exposure duration (yr)
Default
120
200
6
100
24
30
Source: RAGS HHEM, Part B (U.S. EPA, 1991b).
Because of the impracticability of developing site-specific input parameters (e.g., soil ingestion rates) for direct soil
ingestion, SSLs are calculated using the defaults listed in Equations 1 and 2. Appendix A lists these generic SSLs
for direct ingestion of soil.
2-12
-------
2.3 Inhalation of Fugitive Dusts
The models and assumptions used to calculate SSLs for the inhalation pathway are presented in Equations 3 and 4,
along with the default parameter values used to calculate the generic SSLs presented in Appendix A. Particular
attention is given to the emissions portion and the dispersion portion (Q/F) of the particulate emission factor (PEF)
equation, which has been revised since originally presented in RAGS HHEM, Part B. The default PEF presented
here is the same as that given in the SSG for chemicals, which allows for the presence of a family garden associated
with the ingestion of homegrown produce pathway in a residential setting. The available radionuclide-specific
human health benchmarks used in these equations are presented in Section 2.1.
2.3.1 Screening Level Equation for Inhalation of Fugitive Dusts. Equation 3 is used to calculate
SSLs for the inhalation of fugitive dusts containing radionuclides.
Screening Level Equation for Inhalation of Radionuclides in Fugitive Dusts in Residential Soil
SSL -
TR
SF * IR. * (—— ) * IE+3 *EF*ED* [ET + (ET * DF)]
i i ^ ° z
(3)
Parameter/Definition (units)
TR/target cancer risk (unitless)
SF, /inhalation slope factor (pCi~1)
IR, /inhalation rate (m3/d)
PEF/particulate emission factor (m3/kg)
1E+3/conversion factor (g/kg)
EF/exposure frequency (d/yr)
ED/exposure duration (yr)
ETo/exposure time fraction, outdoor (unitless)
ET/exposure time fraction, indoor (unitless)
DF/dilution factor for indoor inhalation, (unitless)
Default
1Q-6
See Table 2. 2
20
1.32E+09 (Equation 3)
-
350
30
0.073
0.683
0.4
Source
U.S. EPA, 1991b
See Part 2.1
U.S. EPA, 1991b
U.S. EPA, 1991b
-
U.S. EPA, 1991b
U.S. EPA, 1991b
U.S. EPA, 1997b
U.S. EPA, 1997b
Alonza, 1979
To calculate inhalation SSLs, the particulate emission factor must be calculated. The derivation of PEF has been
updated since RAGS FiHEM, Part B was published and is discussed fully in Sections 2.3.3. The PEF equation can
be broken into two separate models: models to estimate the emissions of dusts, and a dispersion model (reduced to
the term Q/C) that simulates the dispersion of radionuclides in the atmosphere.
2.3.2 Dispersion Model. The box model in RAGS HHEM, Part B has been replaced with a Q/C term derived
from a modeling exercise using meteorologic data from 29 locations across the United States.
The dispersion model used in the Part B guidance is based on the assumption that emissions into a hypothetical box
will be distributed uniformly throughout the box. To arrive at the volume within the box, it is necessary to assign
values to the length, width, and height of the box. The length (LS) was the length of a side of a contaminated site
with a default value of 45 m; the width was based on the windspeed in the mixing zone (V) with a default value of
2.25 m (based on a windspeed of 2.25 m/s); and the height was the diffusion height (DH) with a default value of 2
m.
However, the assumptions and mathematical treatment of dispersion used in the box model may not be applicable
to a broad range of site types and meteorology and do not utilize state-of-the-art techniques developed for regulatory
2-13
-------
dispersion modeling. EPA was very concerned about the defensibility of the box model and sought a more defensible
dispersion model that could be used as a replacement to the Part B guidance and had the following characteristics:
• Dispersion modeling from a ground-level area source
• Onsite receptor
• A long-term/annual average exposure point concentration
• Algorithms for calculating the exposure point concentration for area sources of different sizes and shapes.
To identify such a model, EPA held discussions with the EPA Office of Air Quality Planning and Standards
(OAQPS) concerning recent efforts to develop a new algorithm for estimating ambient air concentrations from low
or ground-level, nonbuoyant sources of emissions. The new algorithm is incorporated into the Industrial Source
Complex Model (ISC2) platform in both a short-term mode (AREA-ST) and a long-term mode (AREA-LT). Both
models employ a double numerical integration over the source in the upwind and crosswind directions. Wind tunnel
tests have shown that the new algorithm performs well with onsite and near-field receptors. In addition, subdivision
of the source is not required for these receptors.
Because the new algorithm provides better concentration estimates for onsite and for near-field receptors, a revised
dispersion analysis was performed (Appendix D; EQ, 1994). The AREA-ST model was run for 0.5-acre and 30-acre
square sources with a full year of meteorologic data for 29 U.S locations selected to be representative of the national
range of meteorologic conditions (EQ, 1993). Additional modeling runs were conducted to address a range of square
area sources from 0.5 to 30 acres in size (Table 2.4). The Q/C values in Table 2.4 for 0.5- and 30-acre sources differ
slightly from the values in Appendix D due to differences in rounding conventions used in the final model runs.
To calculate site-specific SSLs, select a Q/C value from Table 2.4 that best represents a site's size and meteorologic
condition.
To develop a reasonably conservative default Q/C for calculating generic SSLs, a default site (Minneapolis, MN)
was chosen that best approximated the 90th percentile of the 29 normalized concentrations (kg/m3 per g/m2-s). The
inverse of this concentration results in a default Q/C value of 90.80 g/m2-s per kg/m3 for a 0.5-acre site.
2-14
-------
Table 2.4 . QIC Values by Source Area, City, and Climatic Zone
QIC (g/rrT-s per kg/m")
Zone 1
Seattle
Salem
Zone II
Fresno
Los Angeles
San Francisco
Zone III
Las Vegas
Phoenix
Albuquerque
Zone IV
Boise
Winnemucca
Salt Lake City
Casper
Denver
Zone V
Bismark
Minneapolis
Lincoln
Zone VI
Little Rock
Houston
Atlanta
Charleston
Raleigh-Durham
Zone VII
Chicago
Cleveland
Huntington
Harrisburg
Zone VIM
Portland
Hartford
Philadelphia
Zone IX
Miami
0.5 Acre
82.72
73.44
62.00
68.81
89.51
95.55
64.04
84.18
69.41
69.23
78.09
100.13
75.59
83.39
90.80
81.64
73.63
79.25
77.08
74.89
77.26
97.78
83.22
53.89
81.90
74.23
71.35
90.24
85.61
1 Acre
72.62
64.42
54.37
60.24
78.51
83.87
56.07
73.82
60.88
60.67
68.47
87.87
66.27
73.07
79.68
71.47
64.51
69.47
67.56
65.65
67.75
85.81
73.06
47.24
71.87
65.01
62.55
79.14
74.97
2 Acre
64.38
57.09
48.16
53.30
69.55
74.38
49.59
65.40
53.94
53.72
60.66
77.91
58.68
64.71
70.64
63.22
57.10
61.53
59.83
58.13
60.01
76.08
64.78
41.83
63.72
57.52
55.40
70.14
66.33
5 Acre
55.66
49.33
41.57
45.93
60.03
64.32
42.72
56.47
46.57
46.35
52.37
67.34
50.64
55.82
61.03
54.47
49.23
53.11
51.62
50.17
51.78
65.75
55.99
36.10
55.07
49.57
47.83
60.59
57.17
10 Acre
50.09
44.37
37.36
41.24
53.95
57.90
38.35
50.77
41.87
41.65
47.08
60.59
45.52
50.16
54.90
48.89
44.19
47.74
46.37
45.08
46.51
59.16
50.38
32.43
49.56
44.49
43.00
54.50
51.33
30 Acre
42.86
37.94
31.90
35.15
46.03
49.56
32.68
43.37
35.75
35.55
40.20
51.80
38.87
42.79
46.92
41.65
37.64
40.76
39.54
38.48
39.64
50.60
43.08
27.67
42.40
37.88
36.73
46.59
43.74
2-15
-------
2.3.3 Particulate Emission Factor. The particulate emission factor relates the concentration of
contaminant in soil with the concentration of dust particles in the air. This guidance addresses dust generated from
open sources, which is termed "fugitive" because it is not discharged into the atmosphere in a confined flow stream.
Other sources of fugitive dusts that may lead to higher emissions due to mechanical disturbances include unpaved
roads, tilled agricultural soils, and heavy construction operations.
Both the emissions portion and the dispersion portion (Q/C) of the PEF equation have been updated since RAGS
HHEM, Part B.
As in Part B, the emissions part of the PEF equation is based on the "unlimited reservoir" model from Cowherd et
al. (1985) developed to estimate particulate emissions due to wind erosion. The unlimited reservoir model is most
sensitive to the threshold friction velocity, which is a function of the mode of the size distribution of surface soil
aggregates. This parameter has the greatest effect on the emissions and resulting concentration. For this reason, a
conservative mode soil aggregate size of 500 (im was selected as the default value for calculating generic SSLs.
The mode soil aggregate size determines how much wind is needed before dust is generated at a site. A mode soil
aggregate size of 500 (im yields an uncorrected threshold friction velocity of 0.5 m/s. This means that the windspeed
must be at least 0.5 m/s before any fugitive dusts are generated. However, the threshold friction velocity should be
corrected to account for the presence of nonerodible elements. In Cowherd et al. (1985), nonerodible elements are
described as:
. . . clumps of grass or stones (larger than about 1 cm in diameter) on the surface (that will) consume
part of the shear stress of the wind which otherwise would be transferred to erodible soil.
Cowherd et al. describe a study by Marshall (1971) that used wind tunnel studies to quantify the increase in the
threshold friction velocity for different kinds of nonerodible elements. His results are presented in Cowherd et al.
as a graph showing the rate of corrected to uncorrected threshold friction velocity vs. Lc, where Lc is a measure of
nonerodible elements vs. bare, loose soil. Thus, the ratio of corrected to uncorrected threshold friction velocity is
directly related to the amount of nonerodible elements in surface soils.
Using a ratio of corrected to uncorrected threshold friction velocity of 1, or no correction, is roughly equivalent to
modeling "coal dust on a concrete pad," whereas using a correction factor of 2 corresponds to a windspeed of 19 m/s
at a height of 10 m. This means that about a 43-mph wind would be required to produce any particulate emissions.
Given that the 29 meteorologic data sets used in this modeling effort showed few windspeeds at, or greater than, 19
m/s, EPA felt that it was necessary to choose a default correction ratio between 1 and 2. A value of 1.25 was selected
as a reasonable number that would be at the more conservative end of the range. This equates to a corrected threshold
friction velocity of 0.625 m/s and an equivalent windspeed of 11.3 m/s at a height of 7 meters. The default PEF is
the same as for chemicals. Another, key assumption in the derivation of the PEF is that the 1A acre lot has only 50%
vegetative cover. Although the ingestion of homegrown produce is not quantitatively evaluated in the SSG for
chemicals, the assumption of 50% vegetative cover allows for the presence of a family garden.
Q/C values are needed to calculate the PEF (Equation 4); use the Q/C value in Table 2.4 that best represents a site's
size and meteorologic conditions. Cowherd et al. (1985) describe how to obtain site-specific estimates of V, Um, Ut,
andF(x).
Meteorologic conditions (i.e., the intensity and frequency of wind) affect both the dispersion and emissions of
particulate matter. For this reason, a separate default Q/C value was derived for particulate matter [nominally 10 (im
and less (PM10)] emissions for the generic SSLs. The PEF equation was used to calculate annual average
concentrations for each of 29 sites across the country. To develop a reasonably conservative default Q/C for
calculating generic SSLs, a default site (Minneapolis, MN) was selected that best approximated the 90th percentile
concentration.
The results produced a revised default PEF Q/C value of 90.80 g/m2-s per kg/m3 for a 0.5-acre site (see Appendix
D;EQ, 1994). The generic PEF derived using the default values in Equation 4 is 1.32 x 109m3/kg, which corresponds
to a receptor point concentration of approximately 0.76 (ig/m3. This represents an annual average emission rate based
on wind erosion that should be compared with chronic health criteria; it is not appropriate for evaluating the potential
for more acute exposures.
2-16
-------
Derivation of the Participate Emission Factor
PEF = QIC *
3,600
0.036 * (l-V) * (Um/Ut)3 * F(x)
(4)
Parameter/Definition (units)
Default
Source
PEF/particulate emission factor (m3/kg)
Q/C/inverse of mean cone, at center of square source
(g/m2-s per kg/m3)
V/fraction of vegetative cover (unitless)
U^/mean annual windspeed (m/s)
U/equivalent threshold value of windspeed at 7 m (m/s)
F(x)/function dependent on (JJ(Jt derived using
Cowherd et al. (1985) (unitless)
1.32x109
90.80
0.5 (50%)
4.69
11.32
0.194
Table 3 (for 0.5-acre source
in Minneapolis, MN)
U.S. EPA, 1991 b
EQ, 1994
U.S. EPA, 1991 b
U.S. EPA, 1991b
2.4 External Exposure from Radionuclides in Soil
Individuals residing on a contaminated site will be exposed to photons emitted by those radionuclides present in the
soil. In modeling external exposure to contaminated soil, the RAGS/HHEM Part B model (EPA9 la) does not account
for the following processes:
• radioactive decay and progeny ingrowth (i.e., radioactive daughters);
• correction factors for the geometry of the contaminated soil;
• depletion of the contaminated soil horizon by environmental processes, such as leaching, erosion, or plant
uptake; and
• corrections for shielding by clean cover material.
The RAGS/HHEM Part B model effectively assumes that an individual is continually exposed to a non-depleting
source term with a geometry that is effectively an infinite slab. The concept of an "infinite slab" means that the
thickness of the contaminated zone and its aerial extent are so large that it behaves as if it were infinite in its physical
dimensions. In practice, soil contaminated to a depth greater than about 15 cm and with an aerial extent greater than
about 1,000 m2 will create a radiation field comparable to that of an infinite slab. For calculation of SSLs for a
residential setting, an adjustment for small areas is considered to be an important modification to the RAGS/HHEM
Part B model, since in most residential settings the assumption of an infinite slab source will result in overly
conservative SSLs. Thus, an area correction factor, ACF, has been added to the model for the calculation of SSLs.
For the purposes of this report, adjustments for clean cover are not needed since, in all cases, it is assumed that the
contaminated soil extends to the surface. The model provides adjustments for indoor occupancy and associated
shielding effects by the simple application of a shielding factor and indoor occupancy time adjustment.
2-17
-------
Screening Level Equation for External Exposure to Radionuclides in Soil
SSL =
TR
SF *
EF_
365
= ED * ACF * [ETo * (ETi * GSF)]
(5)
Parameter/Definition (units)
TR/target cancer risk (unitless)
SFe /external exposure slope factor
(g/pCi/yr)
EF/exposure frequency (d/yr)
ED/exposure duration (yr)
ACF/area correction factor
ET0/exposure time fraction, outdoor
(unitless)
ET/exposure time fraction, indoor (unitless)
GSF/gamma shielding factor
Default
10-6
See Table 2. 2
350
30
0.9
0.073
0.683
0.4
Source
U.S. EPA, 1991b
See Part 2.1
U.S. EPA, 1991 b
U.S. EPA, 1991 b
ANL, 1993b
U.S. EPA, 1997b
U.S. EPA, 1997b
U.S. EPA, 1996d
With the exception of the area correction factor, default values are used for all input parameters in Equation 5 to
calculate generic external exposure SSLs. The amount of data required to derive site-specific values for these
parameters makes their collection and use impracticable for calculation of simple site-specific SSLs. Therefore, site-
specific data are generally not available for this exposure pathway. An area correction factor less than 0.9 will only
apply to very small sites (i.e., those with an area less than 1,000 m2). Alternative area correction factors are discussed
in Section 5.1.
2.4.1. Gamma Shielding Factor - GSF (unitless)
The gamma shielding factor is the ratio of the external gamma radiation level indoors on site to the radiation level
outdoors on-site. It is based on the fact that a building provides shielding against penetration of gamma radiation.
Therefore, the calculation of the risk posed by gamma radiation from radionuclides in the soil should take into
account this shielding effect.
EPA's previous gamma shielding factor default value, taken from RAGS/HHEM Part B (U.S. EPA, 1991b), is 0.8,
which assumes that the external gamma radiation level indoors is 20% lower than the outdoor gamma radiation level.
This value is based on information presented in two EPA reports, Natural Radiation in the United States (U.S. EPA,
1972) and Population Exposure to External Natural Radiation Background in the United States (U.S. EPA, 1981).
Based on a review of the literature provided in these reports, EPA concluded that external background exposures in
frame dwellings are 70 to 80% of outdoor values. This conclusion is based on empirical data for natural background
radiation. As such, it includes the contribution from both terrestrial and cosmic radiation and from the radionuclides
present in structure material.
A further review of the literature performed in the EPA report, Reassessment of Radium and Thorium Soil
Concentrations and Annual Dose Rates (U.S. EPA, 1996d) reveals numerous publications that address indoor/outdoor
gamma ray shielding factors as applied to radioactive fallout from nuclear weapons and reactor accidents. In U.S.
EPA, 1981, the authors performed a review of experimentally measured reduction factors from fallout. The authors
concluded that "reduction factors of 0.4 to 0.2 are recommended as representative values for above-ground lightly
constructed (wood frame) and heavily constructed (block and brick) homes, respectively." On the basis of this
review, U.S. EPA 1996d, suggests that a default gamma shielding factor of 0.4 based solely on the contribution of
terrestrial radiation might be a more appropriate value to use at sites with soil contaminated with radionuclides than
the previous EPA default of 0.8 which also included the effects of cosmic radiation and the inherent radioactivity in
structure materials. Based on this rationale, the value of 0.4 is adopted in this guidance as EPA's new default gamma
shielding factor.
2-18
-------
2.5 Ingestion of Homegrown Produce
Persons living on a contaminated site may ingest radioactive material by consumption of plants grown in a family
garden. In this model, the fruits and vegetables primarily become contaminated by root uptake of radionuclides
contained in the pore water of the soil in which the plants are growing.
Residential gardens that provide a significant fraction of the family diet are not uncommon, particularly in more rural
areas. However, it is quite uncommon that such a family garden would be totally self-sufficient. To be self-
sufficient, the family would have to raise all of its vegetables onsite—that is, grow its own potatoes, get all fruit from
an onsite orchard, can or freeze enough of the summer crop to provide all food during the winter, and raise all grain
used for bread and cereal. The model uses available consumption data for home grown produce from USDA surveys.
The model accounts for root uptake with a simple soil-to-plant transfer factor. These soil-to-plant transfer factors
have been developed based upon the assumption that the entire plant root system is wholly exposed to contaminated
soil. If the plant roots extend to a depth of 100 cm but the radionuclide contaminants are confined to the upper 15
cm, an initial assumption may be that only 15% of the root system is active in accumulating contaminants and that
the reported soil-to-plant transfer factors should be reduced by a correction factor of 0.15. However, the equation
for calculation of SSLs for this pathway does not apply any reduction to the soil-to-plant transfer factors. The basis
for this assumption is as follows.
Most plant root systems are in fact very active in the upper soil horizon, especially in the upper 15 cm of soil. This
point is illustrated in a number of ways: 1) by illustrations of root morphology and growth habit, 2) positive
physiological factors including the availability of water, oxygen and nutrients near the soil surface, 3) negative
physiological or agronomic factors—including subsurface soil compaction, subsurface zones of acidity, perched
water tables, hypoxia, etc., 4) interactions with soil microbes—with a special focus on mychorrizal fungi, and 5) split
root studies. Thus, roots commonly proliferate in the upper layers of soil. If one assumes that a plant is actively
growing, then ion uptake characteristics and lateral root growth strongly suggest that simply attributing 15% of root
uptake activity to the upper 15 cm of the soil is not a sound approach. Environmental forces may influence root
growth to one or more meters in depth, but more so for obtaining water than nutrients. In reality, the upper 15 cm
of soil may include 50% or more of the root system—and thus 50% or more of the ion uptake (SC&A, 1994).
The decision to not include air deposition or rain-splash does not affect any radionuclides because the increase in
concentration from this route is not significant or is markedly reduced when peeling, washing, cooking, and other
food preparation processes are taken into consideration (U.S. EPA, 1994k). The decision to not include the irrigation
pathway is only an issue when there is medium to heavy irrigation using contaminated water for a radionuclide with
a long half-life, and an insignificant contribution from external exposure. The model also makes a conservative
assumption to ignore the decay between harvest and ingestion and any removal during food processing.
The model does not include any special calculations for estimating concentrations of 3H and 14C in plants. Such
calculations assume that a state of equilibrium exists among the concentrations of 3H and 14C in all environmental
media—air, water, food products, and body tissues. This assumption may be conservative for a radioactively
contaminated site with a finite area, but may be appropriate for an individual pathway, such as soil-to-plant pathway
(ANL93a). For these calculations, the 3H concentration in the plant is assumed to be the same as that in the
contaminated water to which the plant is exposed. Similarly, the specific activity of 14C in the plant (i.e., pCi/g of
14C per gram of carbon in the plant) is the same as that of the ambient CO2.
Other models use a factor, known as the contaminated plant fraction (CPF), to adjust for the fraction of fruits and
vegetables obtained from the contaminated site relative to the total consumption rate. The ingestion rate used in these
models is a total ingestion rate, which, when multiplied by the contaminated plant fraction, gives the ingestion rate
of contaminated fruits and vegetables. However, the calculation of SSLs for this pathway makes use of published
estimates of individual home produced fruit and vegetable intake, thus eliminating the need for a CPF.
Default values are used for all input parameters in Equation 5 to calculate SSLs for this pathway. The amount of data
required to derive site-specific values for these parameters makes their collection and use impracticable for
calculation of simple site-specific SSLs. Therefore, site-specific data are generally not available for this exposure
pathway. The generic SSLs presented in the TBD are recommended for all sites.
2-19
-------
Screening Level Equation for Ingestion of Radionuclides in Homegrown Produce
SSL =
TR
SF * (IRf + IRv) * 7F *EF*ED
(6)
Parameter/Definition (units)
TR/target cancer risk (unitless)
SFP /produce ingestion slope factor (pCi)"1
IRf/home grown fruits ingestion rate (g/d)
IRiv/home grown vegetables ingestion rate
(g/d)
TFp/soil-to-plant transfer factor (pCi/g plant
per pCi/g soil)
EF/exposure frequency (d/yr)
ED/exposure duration (yr)
Default
10'6
See Table 2.2
189
147
See Part 3. 5. 5
350
30
Source
U.S. EPA,
1991b
See Part 2.1
U.S. EPA,
1997b
U.S. EPA,
1997b
U.S. EPA,
1997b
U.S. EPA,
1991b
U.S. EPA,
1991b
2.6 Migration to Ground Water
The methodology for calculating SSLs for the migration to ground water pathway was developed to identify
radionuclide concentrations in soil that have the potential to contaminate ground water. Migration of radionuclides
from soil to ground water can be envisioned as a two-stage process: (1) release of contaminant in soil leachate and
(2) transport of the contaminant through the underlying soil and aquifer to a receptor well. The SSL methodology
considers both of these fate and transport mechanisms.
The methodology incorporates a standard linear equilibrium soil/water partition equation to estimate radionuclide
release in soil leachate (see Sections 2.6.1 through 2.6.4) and a simple water-balance equation that calculates a
dilution factor to account for dilution of soil leachate in an aquifer (see Section 2.6.5). The dilution factor represents
the reduction in soil leachate radionuclide concentrations by mixing in the aquifer, expressed as the ratio of leachate
concentration to the concentration in ground water at the receptor point (i.e., drinking water well). Because the
infinite source (i.e., steady-state) assumption can result in mass-balance violations for soluble contaminants and small
sources, mass-limit models are provided that limit the amount of contaminant migrating from soil to ground water
to the total amount of contaminant present in the source (see Section 2.7).
SSLs are backcalculated from acceptable ground water concentrations (i.e., MCLs, see Section 2.1). First, the
acceptable ground water concentration is multiplied by a dilution factor to obtain a target leachate concentration.
For example, if the dilution factor is 10 and the acceptable ground water concentration is 10 pCi/L, the target soil
leachate concentration would be 100 pCi/L. The partition equation is then used to calculate the total soil
concentration (i.e., SSL) corresponding to this soil leachate concentration.
The methodology for calculating S SLs for the migration to ground water pathway was developed under the following
constraints:
• Because of the large nationwide variability in ground water vulnerability, the methodology should
be flexible, allowing adjustments for site-specific conditions if adequate information is available.
• To be appropriate for early-stage application, the methodology needs to be simple, requiring a
minimum of site-specific data.
• The methodology should be consistent with current understanding of subsurface processes.
2-20
-------
• The process of developing and applying SSLs should generate information that can be used and built
upon as a site evaluation progresses.
Flexibility is achieved by using readily obtainable site-specific data in standardized equations; conservative default
input parameters are also provided for use when site-specific data are not available. In addition, more complex
unsaturated zone fate-and-transport models have been identified that can be used to calculate SSLs when more
detailed site-specific information is available or can be obtained (see Part 3). These models can extend the
applicability of SSLs to subsurface conditions that are not adequately addressed by the simple equations (e.g., deep
water tables; clay layers or other unsaturated zone characteristics that can attenuate contaminants before they reach
ground water).
The SSL methodology was designed for use during the early stages of a site evaluation when information about
subsurface conditions may be limited. Because of this constraint, the methodology is based on conservative,
simplifying assumptions about the release and transport of contaminants in the subsurface (see Highlight 2).
Highlight 2: Simplifying Assumptions for the Migration to Ground Water Pathway
The source is infinite (i.e., steady-state concentrations will be maintained in ground water over the exposure
period of interest).
Contaminants are uniformly distributed throughout the zone of contamination.
Soil contamination extends from the surface to the water table (i.e., adsorption sites are filled in the
unsaturated zone beneath the area of contamination).
There is no chemical or biological degradation in the unsaturated zone.
Equations in this document do not account for decay, however electronic version of these equations will
account for decay in the unsaturated zone.
Equilibrium soil/water partitioning is instantaneous and linear in the contaminated soil.
The receptor well is at the edge of the source (i.e., there is no dilution from recharge downgradient of the
site) and is screened within the plume.
The aquifer is unconsolidated and unconfined (surficial).
Aquifer properties are homogeneous and isotropic.
Chelating orcomplexing agents not present.
No facilitated transport (e.g., colloidal transport) of inorganic contaminants in aquifer.
Although simplified, the S SL methodology described in this section is theoretically and operationally consistent with
the more sophisticated investigation and modeling efforts that are conducted to develop soil cleanup goals and
cleanup levels for protection of ground water at Superfund sites. SSLs developed using this methodology can be
viewed as evolving risk-based levels that can be refined as more site information becomes available. The early use
of the methodology at a site will help focus further subsurface investigations on areas of true concern with respect
to ground water quality and will provide information on soil characteristics, aquifer characteristics, and radionuclide
properties that can be built upon as a site evaluation progresses.
2.6.1 Development Of Soil/Water Partition Equation The methodology used to estimate radionuclide
release in soil leachate is based on the Freundlich equation, which was developed to model sorption from liquids to
solids. The basic Freundlich equation applied to the soil/water system is:
2-21
-------
= CS/C"W (7)
where
Kd = Freundlich soil/water partition coefficient (L/kg)
Cs = concentration sorbed on soil (mg/kg)
Cw = soil leachate concentration (mg/L)
n = Freundlich exponent (dimensionless).
Assuming that adsorption is linear with respect to concentration (n=l) and rearranging to backcalculate a sorbed
concentration (Cs):
For SSL calculation, Q is the target soil leachate concentration.
Adjusting Sorbed Soil Concentrations to Total Concentrations. To develop a screening level for
comparison with contaminated soil samples, the sorbed concentration derived above (Cs) must be related to the total
concentration measured in a soil sample (Ct). In a soil sample, contaminants can be associated with the solid soil
materials or the soil water, and the soil air as follows (Feenstra et al., 1991):
where
(9)
Mt = total contaminant mass in sample (mg)
Ms = contaminant mass sorbed on soil materials (mg)
Mw = contaminant mass in soil water (mg)
Ma = contaminant mass in soil air (mg).
Furthermore,
and
t = CtPbV
sp,
s = CsPbVsp,
= CW6WVSP,
a = Ca6aVsp,
(10)
(11)
(12)
(13)
where
pb =
Vsp =
6W =
Ca =
6 =
dry soil bulk density (kg/L)
sample volume (L)
water-filled porosity (Lwatei/Lsoll)
concentration on soil pore air (mg/Lsoll)
air-filled soil porosity (Lan/Lsoll).
For contaminated soils (with concentrations below Csat), Ca may be determined from Cw and the dimensionless
Henry's law constant (H') using the following relationship:
2-22
-------
thus
Cw H
(14)
= CwH10BVsp
(15)
Substituting into Equation 9:
or
cs-ct
CJf'Qa
(16)
(17)
Substituting into Equation 8 and rearranging:
c,= cw*K
0 + 0 *H'
(18)
For most radionuclides there is no significant vapor pressure and the dimensionless Henry's Law constant (H1) may
be assumed to be zero, (except for radon). When H' is zero, Equation 18 becomes:
Soil-Water Partition Equation for Migration to Ground Water: Inorganic Contaminants:
SSL =
(19)
Parameter/Definition (units)
Default
Source
SSL /screening level in soil (mg/kg)
CJtarget soil leachate concentration
(mg/L)
Kd/soil-water partition coefficient
(L/kg)
ejwater-filled soil porosity
n/total soil porosity (Lpore/Lsoil)
pb/dry soil bulk density (kg/L)
p,7soil particle density (kg/L)
( MCL ) x 20 DAF
radionuclide -specific
0.3 (30%)
0.43
1.5
2.65
Table 2.3 ( MCL); Section 2.6.2 (DAF for
0.5-acre source)
see Section 3.3.1
U.S. EPA/ORD
1 - Pb/Ps
U.S. EPA, 1991 b
U.S. EPA, 1991b
Since MCLs for radionuclides are given in radiological activity units of picocuries per liter of water (pCi/1), equation
19 can also be written as:
SSL =
(20)
2-23
-------
Parameter/Definition (units)
Default
Source
SSL /screening level in soil
(pCi/g)
CJtarget soil leachate concentration
(pCi/L)
1x10"3/conversion factor (kg/g)
Kd/soil-water partition coefficient
(L/kg)
ejwater-filled soil porosity
n/total soil porosity (Lpore/Lsoil)
pb/dry soil bulk density (kg/L)
p,7soil particle density (kg/L)
( MCL) x 20 DAF
radionuclide -specific
0.3 (30%)
0.43
1.5
2.65
Table 2.3 ( MCL); Section 2.6.2 (DAF for
0.5-acre source)
see Section 3.3.1
U.S. EPA/ORD
1 - Pb/Ps
U.S. EPA, 1991 b
U.S. EPA, 1991b
Either Equation 19 or Equation 20 can be used to calculate SSLs (total soil concentrations, Ct) corresponding to soil
leachate concentrations (Cw) equal to the target contaminant soil leachate concentration. The equation assumes that
soil, water and solids are conserved during sampling.
Defaults. The User's Guide (U.S. EPA, 1996a) describes how to develop site-specific estimates of the soil
parameters needed to calculate SSLs.
Sensitivity analyses have shown that soil bulk density (pb) has too limited a range for surface soils (generally between
1.3 and 1.7 g/cm3) to affect results with nearly the significance of soil moisture conditions. Therefore, a default bulk
density of 1.50 g/cm3, the mode of the range given for U.S. soils in the Superfund Exposure Assessment Manual (U.S.
EPA, 1988), was chosen to calculate generic SSLs. This value is also consistent with the mean porosity (0.43) for
loam soil presented in Carsel and Parrish (1988). The default value of ps (2.65 g/cm3) was taken from U.S. EPA
(1988) as the particle density for most soil mineral material.
Because migration to ground water SSLs are not particularly sensitive to soil water content (see Section 2.5.7), a
value for average water-filled soil porosity that is more typical of subsurface conditions (0.30) was used. This value
is between the mean field capacity (0.20) of Class B soils (Carsel et al., 1988) and the saturated volumetric water
content for loam (0.43).
Kd varies by radioactive element and soil type, as discussed below.
2.6.2 Inorganics (Metals, including radionuclides)—Partition Theory. Equations 19 or 20 can be
used to estimate SSLs for radionuclides for the migration to ground water pathway. The derivation of Kd values is
much more complicated for metals, including radionuclides, than for organic compounds. Unlike organic compounds,
for which Kd values are largely controlled by a single parameter (soil organic carbon), Kd values for metals are
significantly affected by a variety of soil conditions. The most significant parameters are pH, oxidation-reduction
conditions, iron oxide content, soil organic matter content, cation exchange capacity, and major ion chemistry. The
number of significant influencing parameters, their variability in the field, and differences in experimental methods
result in a wide range of Kd values for individual metals reported in the literature (over several orders of magnitude).
Thus, it is much more difficult to derive generic Kd values for metals than for organics. For a more detailed
discussion of the Kd model concept, see Understanding Variation in Partition Coefficient, Kd Values, Volume I &
Volume II (U.S. EPA , 1999b).
2.6.3 Assumptions for Soil/Water Partition Theory. The following assumptions are implicit in the SSL
partitioning methodology. These assumptions and their implications for SSL accuracy should be read and
understood before using this methodology to calculate SSLs.
2-24
-------
1. There is no contaminant loss due to volatilization or degradation. The source is considered to be infinite;
i.e., these processes do not reduce soil leachate concentrations overtime. This is a conservative assumption,
especially for smaller sites.
2. Adsorption is linear with concentration. The methodology assumes that adsorption is independent of
concentration (i.e., the Freundlich exponent =1). This assumption is valid at low concentrations (e.g., at levels
close to the MCL) for most chemicals. As concentrations increase, however, the adsorption isotherm can depart
from the linear.
3. The system is at equilibrium with respect to adsorption. This ignores adsorption/desorption kinetics by
assuming that the soil and pore water concentrations are at equilibrium levels. In other words, the pore-water
residence time is assumed to be longer than the time it takes for the system to reach equilibrium conditions.
This assumption is conservative. If equilibrium conditions are not met, the concentration in the pore water
will be less than that predicted by the methodology. The kinetics of adsorption are not adequately understood
for a sufficient number of chemicals and site conditions to consider equilibrium kinetics in the methodology.
4. Adsorption is reversible. The methodology assumes that desorption processes operate in the same way as
adsorption processes, since most of the partition coefficients are measured by adsorption experiments rather
than by desorption experiments. In actuality, desorption is slower to some degree than adsorption and, in some
cases, radionuclides can be irreversibly bound to the soil matrix. In general, the significance of this effect
increases with increasing Kd. Note, adsorption is only one of several chemical mechanisms that may reduce
the aqueous concentrations in soil or ground water. For a detailed discussion of these chemical mechanisms,
see U.S. EPA, 1999b.
This assumption is conservative. Slower desorption rates and irreversible sorption will result in lower pore-
water concentrations than that predicted by the methodology. Again, the level of knowledge on desorption
processes is not sufficient to consider desorption kinetics and degree of reversibility for all of the subject
radionuclides.
2.6.4 Dilution/Attenuation Factor Development. As radionuclides in soil leachate move through soil
and ground water, they are subjected to physical, chemical, and biological processes that tend to reduce the eventual
contaminant concentration at the receptor point (i.e., drinking water well). These processes include adsorption onto
soil and aquifer media, chemical transformation (e.g., hydrolysis, precipitation), biological degradation, and dilution
due to mixing of the leachate with ambient ground water. The reduction in concentration can be expressed succinctly
by a DAF, which is defined as the ratio of radionuclides concentration in soil leachate to the concentration in ground
water at the receptor point. When calculating SSLs, a DAF is used to backcalculate the target soil leachate
concentration from an acceptable ground water concentration (e.g., MCL). For example, if the acceptable ground
water concentration is 10 pCi/L and the DAF is 10, the target leachate concentration would be 100 pCi/L.
The SSL methodology addresses only one of these dilution-attenuation processes: radionuclide dilution in ground
water. A simple equation derived from a geohydrologic water-balance relationship has been developed for the
methodology, as described in the following subsection. The ratio factor calculated by this equation is referred to as
a dilution factor rather than a DAF because it does not consider processes that attenuate radionuclides in the
subsurface (i.e., adsorption and degradation processes). This simplifying assumption was necessary for several
reasons.
First, the infinite source assumption (constant source concentrations maintained over period of interest) results in
all subsurface adsorption sites being eventually filled and no longer available to attenuate radionuclides. Second,
soil contamination extends to the watertable, eliminating attenuation processes in the unsaturated zone. Additionally,
the receptor well is assumed to be at the edge of the source, minimizing the opportunity for attenuation in the aquifer.
Finally, chemical-specific biological and chemical degradation rates are not known for many of the SSL chemicals;
2-25
-------
where they are available they are usually based on laboratory studies under simplified, controlled conditions. Because
natural subsurface conditions such as pH, redox conditions, soil mineralogy, and available nutrients have been shown
to markedly affect natural chemical and biological degradation rates, and because the national variability in these
properties is significant and has not been characterized, EPA does not believe that it is possible at this time to
incorporate these degradation processes into the simple site-specific methodology for national application.
If adsorption or degradation processes are expected to significantly attenuate radionuclide concentrations at a site
(e.g., for sites with deep water tables or soil conditions that will attenuate radionuclides), the site manager is
encouraged to consider the option of using more sophisticated fate and transport models. Many of these models can
consider adsorption and degradation processes and can model transient conditions necessary to consider a finite
source size. Part 3 of this document presents information on the selection and use of such models for SSL
application.
The dilution factor model assumes that the aquifer is unconfined and unconsolidated and has homogeneous and
isotropic properties. Unconfined (surficial) aquifers are common across the country, are vulnerable to contamination,
and can be used as drinking water sources by local residents. Dilution model results may not be applicable to
fractured rock or karst aquifer types. The site manager should consider use of more appropriate models to calculate
a dilution factor (or DAF) for such settings.
In addition, the simple dilution model does not consider facilitated transport. This ignores processes such as colloidal
transport, transport via solvents other than water (e.g., NAPLs), and transport via dissolved organic matter (DOM).
However, the transport via solvents other than water is operative only if certain site-specific conditions are present.
Transport by DOM and colloids has been shown to be potentially significant under certain conditions in laboratory
and field studies. Although much research is in progress on these processes, the current state of knowledge is not
adequate to allow for their consideration in SSL calculations.
If there is the potential for the presence of NAPLs in soils at the site or site area in question, SSLs should not be used
for this area (i.e., further investigation is required). If NAPLs are suspected in site soils, refer to U.S. EPA (1992c)
for additional guidance on how to estimate the potential for DNAPL occurrence in the subsurface.
Dilution Model Development. EPA evaluated four simple water balance models to adjust SSLs for dilution
in the aquifer. Although written in different terms, all four options reviewed can be expressed as the same simple
water balance equation to calculate a dilution factor, as follows:
Option 1 (ASTM):
dilution factor = (1 + Ugw d/IL) (21)
where
Ugw = Darcy ground water velocity (m/yr)
d = mixing zone depth (m)
I = infiltration rate (m/yr)
L = length of source parallel to flow (m).
For Darcy velocity:
Ugw = Ki (22)
where
K = aquifer hydraulic conductivity (m/yr)
i = hydraulic gradient (m/m).
2-26
-------
Thus
dilution factor = 1 + (Kid/IL) (23)
Option 2 (EPA Ground Water Forum):
dilution factor = (Qp + QA)/QP (24)
where
Qp = percolation flow rate (m3/yr)
QA = aquifer flow rate (m3/yr)
For percolation flow rate:
QP = IA (25)
where
A = facility area (m2) = WL.
For aquifer flow rate:
QA = WdKi (26)
where
W = width of source perpendicular to flow (m)
d = mixing zone depth (m).
Thus
dilution factor = (IA + WdKi)/IWL
dilution factor = 1 + (Kid/IL) (27)
Option 3 (Summers Model):
Cw = (QpCpKQp + QJ (28)
where
Cw = ground water contaminant concentration (mg/L)
Cp = soil leachate concentration (mg/L)
given that
2-27
-------
Cw = Cp/dilution factor
I/dilution factor = Qp/(Qp + QJ
or
dilution factor = (Qp + QA)/QP (see Option 2)
Option 4 (EPA ORD/RSKERL):
dilution factor = (Qp + QA)/QP = RX/RL (29)
where
R = recharge rate (m/yr) = infiltration rate (I, m/yr)
X = distance from receptor well to ground water divide (m)
(Note that the intermediate equation is the same as Option 2.)
This option is a longer-term option that is not considered further in this analysis because valid X values are not
currently available either nationally or for specific sites. EPA is considering developing regional estimates for these
parameters.
Dilution Model Input Parameters. As shown, all three options for calculating contaminant dilution in ground
water can be expressed as the same equation (equation 30):
Ground Water Dilution Factor
dilution factor = 1 + (Kid/IL) (30)
Parameter/Definition (units)
K/aquifer hydraulic conductivity (m/yr)
i/hydraulic gradient (m/m)
d/mixing zone depth (m)
I/infiltration rate (m/yr)
L/source length parallel to ground water flow (m)
Mixing Zone Depth (d). Because of its dependence on the other variables, mixing zone depth is estimated with
the method used for the MULTIMED model (Sharp-Hansen et al., 1990). The MULTIMED estimation method was
selected to be consistent with that used by EPA's Office of Solid Waste for the EPA Composite Model for Landfills
(EPACML). The equation for estimating mixing zone depth (d) is as follows:
d = (2 ccv L)°5 + da {1 - exp[(- L I)/(VS ne dj]} (31)
2-28
-------
where
av = vertical dispersivity (m/m)
Vs = horizontal seepage velocity (m/yr)
ne = effective aquifer porosity (Lpore/Laqmfer)
da = aquifer depth (m).
The first term, (2avL)05, estimates the depth of mixing due to vertical dispersivity (dKV) along the length of ground
water travel. Defining the point of compliance with ground water standards at the downgradient edge of the source,
this travel distance becomes the length of the source parallel to flow L. Vertical dispersivity can be estimated by the
following relationship (Gelhar and Axness, 1981):
ccv = 0.056 CCL (32)
where
aL = longitudinal dispersivity = 0.1 xr
XT = horizontal distance to receptor (m).
Because the potential receptor is assumed to have a well at the edge of the facility, xr = L and
ccv = 0.0056 L (33)
Thus
dav = (0.0112L2)05 (34)
The second term, da {1 - exp[(-LI) / (Vsneda)]}, estimates the depth of mixing due to the downward velocity of
infiltrating water, dlv. In this equation, the following substitution may be made:
Vs = Ki/ne (35)
so
dlv = da{l-exp[(-LI)/(KIda)]} (36)
Thus, mixing zone depth is calculated as follows:
d = dav + dlv (37)
2-29
-------
Estimation of Mixing Zone Depth
d = (0.0112 L2)05 + da {1 - exp[(- L I)/(K i da)]}
(38)
Parameter/Definition (units)
d/mixing zone depth (m)
L/source length parallel to ground water flow (m)
I/infiltration rate (m/yr)
K/aquifer hydraulic conductivity (m/yr)
i/hydraulic gradient (m/m)
da/aquifer thickness (m)
Incorporation of this equation for mixing zone depth into the SSL dilution equation results in five parameters that
must be estimated to calculate dilution: source length (L), infiltration rate (I), aquifer hydraulic conductivity (K),
aquifer hydraulic gradient (i), and aquifer thickness (da). Aquifer thickness also serves as a limit for mixing zone
depth. The User's Guide for Radionuclides (U.S. EPA, 1999a) describes how to develop site-specific estimates for
these parameters. Parameter definitions and defaults used to develop generic SSLs are as follows:
• Source Length (L) is the length of the source (i.e., area of contaminated soil) parallel to ground water flow
and affects the flux of contaminant released in soil leachate (IL) as well as the depth of mixing in the aquifer.
The default option for this parameter assumes a square, 0.5-acre contaminant source. This default was
changed from 30 acres in response to comments to be more representative of actual contaminated soil
sources (see Section 1.3.4). Increasing source area (and thereby area) may result in a lower dilution factor.
Appendix A includes an analysis of the conservatism associated with the 0.5-acre source size.
• Infiltration Rate (I). Infiltration rate times the source area determines the amount of contaminant (in soil
leachate) that enters the aquifer over time. Thus, increasing infiltration decreases the dilution factor. Two
options can be used to generate infiltration rate estimates for SSL calculation. The first assumes that
infiltration rate is equivalent to recharge. This is generally true for uncontrolled contaminated soil sites but
would be conservative for capped sites (infiltration < recharge) and nonconservative for sites with an
additional source of infiltration, such as surface impoundments (infiltration > recharge). Recharge estimates
for this option can be obtained from Aller et al. (1987) by hydrogeologic setting, as described in Section
2.5.6. of the 1996 TBD (EPA 1996b). Recent additional information may also be found in "Estimation of
Infiltration Rate in the Vadose Zone: Compilation of Simple Mathematical Models. Volume I" (U.S. EPA
1998c)
The second option is to use the HELP model to estimate infiltration, as was done for OSW's EPACML and
EPA's Composite Model for Leachate Migration with Transformation Products (EPACMTP) modeling
efforts. The Soil Screening Guidance (U.S. EPA, 1995c) provides information on obtaining and using the
HELP model to estimate site-specific infiltration rates.
• Aquifer Parameters. Aquifer parameters needed for the dilution factor model include hydraulic
conductivity (K, m/yr), hydraulic gradient (i, m/m), and aquifer thickness (da, m). The User's Guide (U.S.
EPA, 1996a) describes how to develop aquifer parameter estimates for calculating a site-specific dilution
factor.
2.6.5 Default Dilution-Attenuation Factor. EPA has selected a default DAF of 20 to account for
contaminant dilution and attenuation during transport through the saturated zone to a compliance point (i.e., receptor
well). This assumption was developed in Section 2.5.6 of the 1996TBD(EPA 1996b). At most sites, this adjustment
will more accurately reflect a radionuclide's threat to ground water resources than assuming a DAF of 1 (i.e., no
2-30
-------
dilution or attenuation). EPA selected a DAF of 20 using a "weight of evidence" approach. This approach considers
results from OSW's EPACMTP model as well as results from applying the SSL dilution model described in Section
2.6.4 of this report to 300 ground water sites across the country.
The default DAF of 20 represents an adjustment from the DAF of 10 presented in the December 1994 draft Soil
Screening Guidance (U.S. EPA, 1994h) to reflect a change in default source size from 30 acres to 0.5 acre. A DAF
of 20 is protective for sources up to 0.5 acre in size. Analyses presented in Appendix A indicate that it can be
protective of larger sources as well. However, this hypothesis should be examined on a case-by-case basis before
applying a DAF of 20 to sources larger than 0.5 acre.
2.7 Mass-Limit Model Development
This section describes the development of models to solve the mass-balance violations inherent in the infinite source
(i.e., steady-state) models used to calculate SSLs for the migration to ground water exposure pathway. The model
developed is not a finite source model per se, but is designed for use with the current infinite source (i.e., steady-
state) model to provide a lower, mass-based limit for SSLs for the migration to ground water.
The mass-limit model calculates a soil concentration that corresponds to the release of all radionuclides present
within the source, at a constant health-based concentration, over the duration of exposure. These concentration limits
are used as a minimum concentration for each SSL; below this concentration, a receptor point concentration time-
averaged over the exposure period cannot exceed the health-based concentration on which it is based.
2.7.1 Mass Balance Issues. Infinite source (i.e., steady-state) models are subject to mass balance violations
under certain conditions. Depending on a compound's solubility and the size of the source, modeled leaching rates
can result in a source being depleted in a shorter time than the exposure duration (or the flux over a 30- or 70-year
duration would release a greater mass of radionuclides than are present). Several commenters to the December 1994
draft Soil Screening Guidance for chemicals expressed concern that it is unrealistic for total emissions over the
duration of exposure to exceed the total mass of contaminants in a source.
In summary, the mass-limit approach offers the following advantages:
• It corrects the possible mass-balance violation in the infinite-source SSLs.
• It does not require development of a finite source model to calculate SSLs.
• It is appropriate for screening, being based on the conservative assumption that all
of the radionuclide present leaches over the period of exposure.
• It is easy to develop and implement, requiring only very simple algebraic equations
and input parameters that are, with the exception of source depth, already used to
calculate SSLs.
The derivation of this model is described below. It should be noted that the American Industrial Health Council
(AIHC) independently developed identical models to solve the mass-balance violation as part of their public
comments on the Soil Screening Guidance.
2.7.2 Migration to Ground Water Mass-Limit Model. For the migration to ground water pathway, the
mass of radionuclide leached from a contaminant source over a fixed exposure duration (ED) period can be calculated
as
2-31
-------
(39)
where
M! = mass of contaminant leached (g)
Cw = leachate contaminant concentration (mg/L or g/m3)
I = infiltration rate (m/yr)
As = source area (m2)
ED = exposure duration (yr).
The total mass of contaminants present in a source can be expressed as
MT = Ct*pb*As*ds (40)
where
MT = total mass of contaminant present (g)
Ct = total soil contaminant concentration (mg/kg or g/Mg, dry basis)
pb = dry soil bulk density (kg/L or Mg/m3)
As = source area (m2)
ds = source depth (m).
To avoid a mass balance violation, the mass of contaminant leached cannot exceed the total mass of contaminants
present (i.e., M! cannot exceed MT). Therefore, the maximum possible contaminant mass that can be leached from
a source (assuming no volatilization or degradation) is MT and the upper limit for M! is
M, = MT
or
Ct*pb*As*ds (41)
Rearranging to solve for the total soil concentration (Q) corresponding to this situation (i.e., maximum possible
leaching), as equation 41 becomes:
Mass-Limit Model for Migration to Ground Water Pathway
C */*££>
' -
c -
2-32
-------
Parameter/Definition (units)
C/screening level in soil (mg/kg)
CJtarget soil leachate concentration (mg/L)
I/infiltration rate (m/yr)
ED/exposure duration (yr)
pb/dry soil bulk density (kg/L)
ds/average source depth (m)
Default
(MCL or HBL) x 20 DAF
site-specific
70
1.5
site-specific
This soil concentration (C^ represents a lower limit for soil screening levels calculated for the migration to ground
water pathway. It represents the soil concentration corresponding to complete release of soil contaminants over the
ED time period at a constant soil leachate concentration (Cw). Below this Ct, the soil leachate concentration averaged
over the ED time period cannot exceed Cw.
SSLs for radionuclides may also be calculated in traditional radiological units of pCi/g, as follows:
C *I*ED * 1x10
SSL = —
(43)
Parameter/Definition (units)
Default
SSL/screening level in soil (pCi/g)
CJtarget soil leachate concentration (pCi/L)
I/infiltration rate (m/yr)
ED/exposure duration (yr)
1x10-3/conversion factor (kg/g)
pb/dry soil bulk density (kg/L)
ds/average source depth (m)
(MCL or HBL) x 20 DAF
site-specific
70
1.5
site-specific
2-33
-------
Part 3: UNSATURATED ZONE MODELS FOR RADIONUCLIDE FATE AND TRANSPORT
In an effort to provide useful information for model application, EPA's NRMRL/SPRD in Ada, Oklahoma, conducted
an evaluation of five unsaturated zone fate and transport models for radionuclides. The results of this effort follow.
The models reviewed are only a subset of the potentially appropriate models available to the public and are not meant
to be construed as having received EPA approval. Other models also may be applicable to SSL development,
depending on site-specific circumstances.
Each of the unsaturated zone models selected for evaluation are capable, to varying degrees, of simulating the
transport and transformation of chemicals in the subsurface. Even the most unique site conditions can be simulated
by either a single model or a combination of models. However, the intended uses and the required input parameters
of these models vary. The models evaluated include:
HYDRUS
MULTIMED_DP 1.0
FECTUZ
CHAIN
CHAIN 2D
The applications, assumptions, and input requirements for the five models evaluated are described in this section.
The model descriptions include model solution method (i.e., analytical, numerical), purpose of model, and methods
used by the model to simulate water flow, decay reaction, and radionuclide transport. Each description is
accompanied by a table of required input parameters. Input parameters discussed include soil properties, chemical
properties, meteorological data, and other site information. In addition, certain input control parameters may be
required, such as time stepping, grid discritization information, and output format. Information on determining the
general applicability of the models to subsurface conditions is provided, followed by an assessment of each model's
potential applicability to the soil screening process.
HYDRUS. Information on the HYDRUS model (Version 6.0) was obtained from Simunek et al. (1998). HYDRUS
is a finite-element model for one-dimensional solute and heat transport simulations in variably saturated media. The
flow equation incorporates a sink term to account for water uptake by plant roots. The boundary conditions for flow,
and heat and solute transport can vary with time. A finite source also can be modeled. Soil parameters are described
by the van Genuchten parameters. The model also considers hysteresis in the water movement. Solute transport and
transformation incorporates molecular diffusion, hydrodynamic dispersion, linear equilibrium reactions between the
liquid and gaseous phases, nonlinear nonequilibrium partitioning (sorption) between the solid and liquid phases, zero-
order production, and first-order decay/degradation reaction. The input parameters required by HYDRUS are
presented in Table 3.1.
HYDRUS considers up to six solutes/radionuclides in the transport equation. For decay reaction, the radionuclides
can be either coupled in a unidirectional chain or may move independently of each other. In the former case, two
or more radionuclides are coupled in a sequential chain reaction. In the later case, each decay reaction is independent
of each other. A specific distribution coefficient (or soil-water partition coefficient, a Kd value) as well as a specific
decay rate for each radionuclide is required for the HYDRUS model simulation.
3-1
-------
Table 3.1. Input Parameters Required for HYDRUS
Soil properties
Number of soil
materials
Site characteristics
Uniform orstepwise
rainfall intensity
Pollutant properties
Molecular diffusion
coefficient
Root uptake
parameters
Potential transpiration
rate
Depth of soil layers
Saturated water
content
Volumetric fraction of
solid phase
Volumetric fraction of
organic matter
Residual water content Thermal dispersivity
Dispersivity
Freundlich isotherm
coefficients*
Freundlich isotherm
exponents*
Saturated hydraulic
conductivity
Soil bulk density
van Genuchten
retention parameter, a
van Genuchten
retention parameter, p
Rescaling factors for
hydraulic
properties
Empirical parameters First order rate
for thermal conductivity constants (dissolved)
calculations
Osmotic coefficient
Pressure head where
transpiration is
reduced by 50%
Root density as a
function of depth
Power function in
stress-response
function (van
Genuchten or Feddes)
Volumetric heat
capacities of solid
phase, organic matter,
and liquid phase
Number of solutes
Contaminant
concentrations in soil
First order rate
constants (adsorbed)
Decay coefficient
(dissolved)
Decay coefficient
(adsorbed)
Heat transport properties
Volumetric solid phase
fraction
Volumetric organic
fraction of each soil
Coefficients of the
thermal conductivity
functions
Volumetric heat
capacity of solid phase
for each soil
Volumetric heat
capacity of liquid
phase for each soil
Volumetric heat
capacity of organic
matter for each soil
Longitudinal and
transverse thermal
dispersivities for each
soil
Thermal initial and
boundary conditions
t Freundlich isotherm becomes a linear isotherm and Freundlich isotherm coefficient becomes a distribution coefficient (Kd value) when
Freundlich isotherm exponent is 1.
M ULTIM ED_DP. Information on the MULTIMED_DP 1.0 (Multimedia exposure assessment model, including fate
and transformation products) model was obtained from Liu et al. (1995), Salhotra et al. (1995) and Sharp-Hansen
et al. (1995). MULTIMED_DP was initially developed as a multimedia fate and transport model to simulate
contaminant migration from a waste disposal unit (MULTIMED) through different pathways in air, surface water,
soil, and ground water. It has been modified to simulate the transport and fate of first and second-generation
transformation products through the unsaturated and saturated zones. In MULTIMED_DP Version 1.0, the model
3-2
-------
now has an option to allow it to be used for unsaturated zone transport alone. The capacity for Monte Carlo
simulation has been extended to the unsaturated zone, as well as to the saturated zone. Contamination of a surface
stream due to the complete interception of a steady-state saturated zone plume is simulated by the surface water
module. The air emissions and the atmosphere dispersion modules simulate the movement of chemicals into the
atmosphere. For this review, only the fate and transport of pollutants from the soil to the ground water pathway is
considered in detail. In MULTIMED_DP, infiltration of waste into the unsaturated or saturated zones can be
simulated using a landfill module or by direct infiltration to the unsaturated or saturated zones. The unsaturated flow
model simulates a one-dimensional steady flow with a semi-analytical solution and includes the option to consider
seasonal variability in precipitation and evapotranspiration while retaining the assumption of steady-state. Transport
in the unsaturated zone considers the effects of advection, dispersion, linear or nonlinear sorption, volatilization,
hydrolysis, biodegradation, and first-order chemical decay. It can address steady or time variable infiltration and a
finite or infinite source. A one-dimensional uniform steady flow in the saturated zone is assumed. The saturated
transport module is also one-dimensional, but considers three-dimensional dispersion, linear adsorption, first-order
decay, and dilution due to recharge. Mixing in the underlying saturated zone is based on a specified vertical
dispersivity, the length of the disposal facility parallel to the flow direction, the thickness of the saturated zone, the
ground water velocity, and the infiltration rate. The parameters required for the unsaturated and saturated zone
transport in MULTIMED_DP are presented in Table 3.2.
The MULTIMED_DP model can handle parent, daughter, and granddaughter species chain decay. The chain decay
reactions, for example, can be either one parent with one or two daughters or one parent with one daughter and two
granddaughters. Effective/overall decay rate, individual decay rates , and distribution coefficient are required for
the parent, daughter, and granddaughter species.
Table 3.2. Input Parameters Required for MULTIMED
Site characteristics/Source characteristics
Recharge rate
Infiltration rate
Depth of unsaturated zone
Area of waste disposal unit
Length scale of facility
Width scale of facility
Duration of pulse
Source decay constant
Initial parent concentration at
landfill
Initial daughter concentration at
landfill
Initial granddaughter
concentration at landfill
Unsaturated zone
parameters
Number of physical flow layers
Thickness of each layer
Number of porous materials
Saturated hydraulic conductivity
van Genuchten retention
parameter, a
van Genuchten retention
parameter, p
Residual water content
pH of layer
Soil bulk density
Longitudinal dispersivity
Reference temperature for
air diffusion
Distribution coefficient
Temperature of layer
Brooks and Corey exponent
Air entry pressure head
3-3
-------
Table 3.2. Input Parameters Required for MULTIMED
Saturated zone parameters
Hydraulic gradient
Hydraulic conductivity
Mixing zone depth
Aquifer thickness
Aquifer porosity
Bulk density
Organic carbon content
Longitudinal dispersivity
Transverse dispersivity
Vertical dispersivity
Temperature of aquifer
PH
Parent retardation coefficients
Daughter retardation coefficients
Granddaughter retardation
coefficient
Well distance from site
Angle off-center of well
Well Vertical distance
3-4
-------
Table 3.2. Input Parameters Required for MULTIMED
Parent/Daughter/Granddaughter Decay Module
Parent to daughter 1 dissolved
Phase decay rate constant
Parent to daughter 2 dissolved
phase decay rate constant
General parent dissolved phase
decay rate constant (decay not
contributing to either daughter
product)
Parent to daughter 1 sorbed
phase decay rate constant
Parent to daughter 2 sorbed
phase decay rate constant
General parent sorbed phase
decay rate constant (decay not
contributing to either daughter
product)
Daughterto granddaughter],
j+1 dissolved phase decay rate
constant
Daughter I general dissolved
phase sorbed phase decay rate
constant
Daughter I to granddaughter],
j+1 sorbed phase decay rate
constant
Daughter I general sorbed
phase decay rate constant
Daughter second-order acid-
catalysis hydrolysis rate
constants
Granddaughter k dissolved
phase decay rate constant
Granddaughter k sorbed
phase decay rate constant
Parent general term second-
order acid-catalysis
hydrolysis rate constant at
reference temperature
Parent general term neutral
hydrolysis rate constant at
reference temperature
Parent general term second-
order base-catalysis
hydrolysis rate constant at
reference temperature
Parent to daughter 1 second-
order acid catalysis
hydrolysis rate constant at
reference temperature
Parent to daughter 1 neutral
hydrolysis rate constant at
reference temperature
Parent to daughter 1 base-
catalysis hydrolysis rate
constant at Reference
Temperature
Parent general term neutral
hydrolysis rate constant at
reference temperature
Parent general term second-
order base-catalysis
hydrolysis rate constant at
reference temperature
Parent general term second-
order base-catalysis
hydrolysis rate constant at
reference temperature
Daughter second-order acid-
catalysis hydrolysis rate constants
Daughter neutral catalysis
hydrolysis rate constants
Daughter second-order base-
catalysis hydrolysis rate constant
Granddaughter second-order acid
catalysis hydrolysis rate constant
Granddaughter neutral hydrolysis
rate constant
Granddaughter second-order
base-catalysis hydrolysis rate
constant
Parent to daughter I dissolved
and sorbed phase stoichiometric
coefficients
Daughterto granddaughter m
dissolved and sorbed phase
stoichiometric coefficients
pH of the unsaturated zone or
saturated zone
Reference temperature
Temperature of the unsaturated
or saturated zone
3-5
-------
FECTUZ. FECTUZ is a one-dimensional fate and transport model for the unsaturated zone which is incorporated
into the EPACMTP program (EPA 1995b, 1995c). It is an extension of the VADOFT code (Huyakorn and Buckley,
1987) which is also incorporated into the RUSTIC and PRZM codes. FECTUZ simulates migration of contaminants
from a landfill (or a surface impoundment), through the unsaturated zone to an unconfmed aquifer with a water table
present at some depth. The model allows for finite or infinite sources which can undergo decay using Bateman's
equation (EPA, 1995b). The model can simulate linear and nonlinear adsorption and first order decay. FECTUZ is
limited to the simulation of simple hydrogeological sites.
FECTUZ is capable of chain decay with up to seven radionuclide species (one parent with up to six daughter
products). The chain decay path can be either straight or branched. A daughter product (a species) can have one or
more immediate parent species. The sum of the chemical and biological transformation coefficients gives the overall
decay coefficient. The fraction of each parent that decays into the same daughter product has to be specified. A
distribution coefficient and a decay rate are specified for each individual radionuclide. A radionuclide species with
a zero decay rate indicates it is the end product in the decay chain.
FECTUZ assumes steady state flow in the unsaturated zone which can be composed of one or more uniform soil
layers. It simulates the transport of contaminants in soil using a advective-dispersive equation. FECTUZ employs
three solution options for the transport equation. An analytical solution is used for steady state single species decay
with linear adsorption. A semi-analytical solution is used for transient and steady state chain decay and linear
sorption. A finite element solution is used for chain decay with nonlinear sorption. The input parameters required
for FECTUZ are presented in Table 3.3.
Table 3.3. Input Parameters Required for FECTUZ Module within EPACMTP
Soil properties
Site characteristics
Pollutant properties
Soil bulk density
Saturated water content
Thickness of unsaturated zone
Uniform thickness for
discretized soil layers
Saturated hydraulic conductivity Uniform infiltration rate except
for surface impoundments
Residual water content
van Genuchten retention
parameter, a
van Genuchten retention
parameter, p
Fraction of organic carbon
Constant source or Decaying
source or finite pulsed source
Organic carbon partition
coefficient
Freundlich isotherm coefficients
Dispersivity
Decay coefficient (dissolved)
Decay coefficient (adsorbed)
Parent and daughter species
decay reaction stoichiometry
t Freundlich isotherm becomes a linear isotherm and Freundlich isotherm coefficient becomes a distribution coefficient (Kd value) when
Freundlich isotherm exponent is 1.
CHAIN. CHAIN (van Genuchten, 1985) is a simple program which uses analytical solutions to solve for the
simultaneous one-dimensional advective-dispersive transport of contaminants for up to four members of a sequential
(first-order) decay chain. The model assumes that the soil system is a homogeneous soil system and that the moisture
content and infiltration are constant in time (steady state flow). The adsorbed concentrations of the contaminant are
represented by linear reversible isotherms. The transport contaminant boundary condition is either a constant
3-6
-------
concentration or a source decay given an arbitrary general release mechanism or the specific release mechanism
defined by the Bateman equations (van Genuchten, 1985). The analytical solutions provide the solution in terms of
concentration vs. distance (depth) at selected times as well as concentrations vs. time at selected depths. The input
parameters required for the CHAIN model are presented in Table 3.4.
The consecutive chain reaction included in the transport model can involve up to four species. For example, a reaction
could have parent and daughter products (two species) or a reaction could involve parent, daughter, and
granddaughter products (three species). There is a decay rate associated with each pair of parent-daughter (or
daughter-granddaughter) products. For the last product in the consecutive chain, a decay rate of zero is specified.
Table 3.4. Input Parameters Required for CHAIN Model
Soil properties Site characteristics Pollutant properties
Soil Bulk density Source initial concentrations Retardation values for decay
chain members
Volumetric water content Bateman coefficients for source First order decay coefficients
decay for decay chain members
Pore velocity (flux/water General release Pulse time
content) parameters for
source decay
Dispersion coefficient
CHAIN 2D. CHAIN 2D is a two-dimensional model for the simulation of variably saturated flow, contaminant
transport, and heat transport developed by Simunek and van Genuchten (1994). The water flow is represented by
Richards equation for saturated and unsaturated flow. It contains a sink term for water uptake by plant roots. The
flow region can be composed of non-uniform soils to incorporate the effects of anisotropy. The flow model boundary
conditions can include prescribed head, gradient, flux boundaries, or free drainage and a simplified representation
of nodal drains. Heat and contaminant transport are modeled by the advective-dispersive equation which includes
conduction and convection processes. The contaminant transport simulation can be modified for nonlinear
nonequilibrium reactions between solid and liquid phases and for linear equilibrium reactions between liquid and
gaseous phases. The contaminant transport simulation also includes zero order production and two first-order decay
reactions: one which is independent of other solutes, and one which solutes are in sequential chain decay reactions.
CHAIN 2D can simulate up to six species independent of one another or in a unidirectional chain decay. The
boundary conditions for the contaminant transport could be constant concentration or constant flux. Flow and
transport can be simulated in a vertical or horizontal plane or in an axisymmetrical cylindrical system. The flow and
transport equations are solved using the Galerkin finite element method. The parameters required for CHAIN 2D
model are presented in Table 3.5.
Similar to the HYDRUS model, CHAIN 2D considers up to six solutes/radionuclides in the transport equation, The
radionuclide can be either coupled in a unidirectional chain or may move independently of each other. In the former
case, two or more radionuclides are coupled in a sequential chain reaction. In the latter case, each decay reaction is
independent of each other. A specific distribution coefficient as well as a specific decay rate for each radionuclide
is required for the model simulation.
3-7
-------
Table 3.5. Input Parameters Required for CHAIN 2D
Soil properties
Site characteristics Pollutant properties
Root uptake
parameters
2D cell discritization Transpiration rate
Saturated water content Evaporation/infiltration
rates
Saturated hydraulic
conductivity
Soil bulk density
Initial contaminant
concentrations in soil
Contaminant species
initial and boundary
conditions
van Genuchten retention Initial head conditions
parameter, a
van Genuchten retention Location and rates of
parameter, p pumping/ injection
wells
Residual water content Seepage faces, tile
drains
Ionic or molecular
diffusion coefficient in
free water for each
species
Ionic or molecular
diffusion coefficient in
gas phase for each
species
Longitudinal and
transverse dispersivities
for each species
First order decay
coefficient for each
species in liquid, solid
or gas phase
Zero order rate constant
for each species in
liquid, solid or gas
phase
Adsorption (Freundlich)*
isotherm coefficients
for each species
Source Decay
Root density as a function
of depth
Power function in stress-
response function
Pressure head where
transpiration is reduced by
50%
Heat transport properties
Volumetric solid phase
fraction
Volumetric organic
fraction of each soil
Coefficients of the
thermal conductivity
functions
Volumetric heat
capacity of solid phase
for each soil
Volumetric heat
capacity of liquid phase
for each soil
Volumetric heat
capacity of organic
matter for each soil
Longitudinal and
transverse thermal
dispersivities for each
soil
Thermal initial and
boundary conditions
Freundlich isotherm becomes a linear isotherm and Freundlich isotherm coefficient becomes a distribution coefficient (Kd value) when
Freundlich isotherm exponent is 1.
3-8
-------
3.1 Considerations for Unsaturated Zone Model Selection. The accuracy of a model in a site-specific
application depends on simplifications and assumptions implicit in the model and their relationship to site-specific
conditions. Errors may be introduced from assumptions made when deriving input parameters. Although each of
the five models evaluated has been tested and validated for simulation of water and radionuclide movement in the
unsaturated zone, they are different in purpose and complexity, with certain models designed to simulate very specific
scenarios. A model should be selected to accommodate a site-specific scenario as closely as possible. For example,
if contaminant chain decay is of concern, the model should consider chain decay reaction. After a model is
determined to be appropriate for a site, the contaminant(s) and the conditions to be modeled, the site-specific
information available (or potentially available) should be compared to the input requirements for the model to ensure
that adequate inputs can be developed. The unsaturated zone models addressed in this study use either analytical,
semianalytical, or numerical solution methods. Analytical models represent the simplest models, requiring the least
number of input parameters. They use a closed-form solution for the pertinent equations. In analytical models,
certain assumptions have to be made with respect to the geometry of the system and external stresses. For this reason,
there are few analytical flow models (van der Heijde, 1994). Analytical solutions are common for fate and transport
problems which arise from the solutions of advective-dispersive equations. Analytical models (e.g. CHAIN) may
require the assumption of uniform flow conditions, both spatially and temporally. Semianalytical models (e.g.,
MULTIMED_DP) approximate complex analytical solutions using numerical techniques (van der Heijde, 1994).
Transient or steady-state conditions can be approximated using a semianalytical model. However, spatial variability
in soil or aquifer conditions are usually not accommodated. Numerical models (e.g., HYDRUS or CHAIN-2D) use
approximations of pertinent partial differential equations, such as finite-difference or finite-element methods. The
resolution of the area and time of simulation is defined by the modeler. Numerical models may be used when
simulating time-dependent scenarios, spatially variable soil conditions, and unsteady flow (van der Heijde, 1994).
3.2 Model Applicability to SSLs for Radionuclides. Evaluation of model applicability to SSLs is based on
the following considerations:
1. Whether a model can be used in simulating fate and transport of the five selected radionuclides ~
uranium, strontium, technetium, plutonium, and tritium?
2. Whether a model can simulate fate and transport of the selected radionuclides for a test case (the Las
Cruces Trench Site, New Mexico)?
3. Is a model capable for use in the soil screening process?
4. What are the limitations for a specific model?
To facilitate the evaluation, the characteristics and capabilities of all five models are summarized in Table 3.6. These
characteristics/capabilities include flow and transport processes, site-specific conditions, solution methods,
assumptions, and model outputs. Each characteristic/capability is evaluated and is check marked when that
characteristic/capacity is implemented in the model. Note that Table 3.6 addresses only unsaturated zone fate and
transport model components in spite of the HYDRUS, MULTIMED_DP, and FECTUZ models having saturated zone
flow and transport capabilities. The following text highlights a general description of the evaluation process,
discusses some of the differences between the models, outlines their advantages and disadvantages, and describes
appropriate scenarios for model application.
Model Applicability to the Five Selected Radionuclides: As provided in Table 3.6, all five models address four
essential processes that predominately control the migration of radionuclides in the unsaturated zone — advection
(derived from infiltration), dispersion, sorption and radionuclide (straight chain and/or branched) decay. Therefore,
any of the five models can be used in simulating fate and transport for the five selected radionuclides ~
uranium, strontium, technetium, plutonium, and tritium under the flow and site-specific conditions (e.g., steady-flow
3-9
-------
in uniform soil, or transient flow in a layered soil) which have been implemented in the model. These five
radionuclides exhibit a broad range of physical properties in soil — distribution coefficients (Kd) of a radionuclide
between water solution and soil, and chain decay rates. The simulation results for all five radionuclides using any
of the five models would be similar to the results in the sensitivity analysis for Kd when the simulation time is much
smaller than the magnitude of the half-life of any radionuclide.
Model Applicability to Site-Specific Conditions at the Las Cruces Trench Site: At the Las Cruces Trench Site, soil
properties and solute transport parameters are available (Wierenga et al. 1991; Hills et al. 1991). They are
reproduced in Tables 3.7 and 3.8. Other site-specific conditions are given in Table 3.9. As indicated in Table 3.6,
the HYDRUS, CHAIN 2D, FECTUZ, and MULTIMED_DP models can simulate layered soils. However, when soil
at the site is treated as a uniform soil and the averaged soil properties are used, all five models can simulate
the transport of radionuclides under steady-state water flow conditions at the site (assuming net recharge rate
is known). Only HYDRUS and the CHAIN 2D model can handle transient flow at the site. Among the five models
evaluated, CHAIN does not have a water flow module, FECTUZ considers water flow with a known infiltration rate,
and the HYDRUS, CHAIN 2D and MULTIMED_DP models can calculate infiltration rates from rainfall and
potential evaporation data. Note that the average annual infiltration rate at a site is difficult to measure in the field,
yet it is required for estimating a dilution factor or DAF in the simplified SSL method. Furthermore, it is worthy to
note that the processes/characteristics listed in Table 3.6 are not inclusive enough to accurately simulate the fate and
transport of radionuclides in unsaturated zone. Other processes such as colloid-facilitated transport, pH-facilitated
transport, cation-exchange, or complexation, are not considered. These processes might significantly influence the
radionuclide movement in clay-rich soil, but they are not likely at the Las Cruces Trench Site, where the soils are
predominantly classified as sands, sandy loams, loamy sands and sandy clay loams (Wierenga et al. 1991). For the
purpose of SSL processes, all five models can simulate the transport of a radionuclide to a reasonable extent because
these models consider the four essential processes that probably control the migration of radionulides in the
unsaturated zone — advection (derived from infiltration), dispersion, sorption, and radionuclide (straight and/or
branched) decay. This will be further confirmed in the sensitivity analysis.
Model Capability for the Use in the Soil Screening Process: All five unsaturated models for radionuclides evaluated
herein are PC-based, public domain models. These five models can calculate the leachate radionuclide
concentrations entering ground water (Table 3.10). The leachate concentrations are needed in the SSL
process/comparison. The leachate concentrations as a function of time are used to estimate ground water
concentrations at the receptor well. These receptor point concentrations are then compared with the acceptable
ground-water concentration (e.g. MCL) to determine if a site's soil exceed SSLs. If they do not exceed the acceptable
ground-water concentrations, it might be believed that there is no concern for human health and ecology at the site.
The conclusion of such a comparison is based on the assumptions that —1) the conceptual site model (CSM) is
reasonably developed; 2) site-specific data required by the model are properly collected and meet data quality
standards; and 3) the unsaturated zone model for radionuclide migration is properly used. Therefore, all five selected
unsaturated zone models for radionuclides - CHAIN, HYDRUS, CHAIN 2D, MULTIMED_DP, and FECTUZ, are
basically capable for use in the soil screening levels process. When there is a concern for the uncertainty of the
SSL estimates due to variability of the input parameters, the built-in Monte Carlo simulation in the MULTIMED_DP
and FECTUZ (EPACMTP version) can be employed. For CHAIN, HYDRUS, and CHAIN 2D models, multiple runs
using a set of variable input parameters (either obtained from the data survey or data generation for a know
distribution of a parameter.) is needed.
Using any of these five models for the SSL calculation requires more site-specific data and modeling effort than
using the simple site-specific SSL calculation (see Section 2 in TBD). However, these five models (and other
potential models) can take into account more complex site conditions and can provide more accurate SSL calculation
if data collection is sufficient and model application is properly done. In other words, using models in the SSL
process require less assumptions than using the simple site-specific SSL calculation. However, for both the simple
approach and the more detailed modeling approach the receptor well is assumed to be located at the edge of the waste
3-10
-------
source. The SSLs are intended to correspond to levels of radionuclides in soil such that MCLs will not be exceeded
in the underlying ground water beyond the edge of the waste area. Among those simplifying assumptions listed in
the Highlight 2 of TBD Section 2, only the assumptions of ignoring facilitated transport and complexation of
nuclides with other constituents in soil solution still apply to the model SSL calculation.
Note that two methods can be used to calculate the receptor point concentration using the leachate radionuclide
concentrations to ground water provided by the unsaturated model. One method uses a dilution/attenuation factor
(DAF) to account for the mixing of leachate with ambient ground water. Another uses a saturated zone flow and
transport model to calculate the resulting concentration considering the processes of advective, dispersive, decay,
and biodegradation of the radionuclide in ground water (the MULTIMED_DP and FECTUZ models have a saturated
zone flow and transport component.) When the former method is used, the receptor well is at the edge of the source
(i.e., there is no dilution from recharge downgradient of the site) and is screened within the plume. A default value
of 20 for DAF as proposed in the technical background document (TBD) for SSLs for radionuclide can be used.
Alternatively, the dilution models in the section 2.6.4 in TBD can also be used.
It is worthy to note that in using the simple soil/water partition equation, the SSLs are directly calculated from an
acceptable radionuclide level at the receptor well. The existing total soil radionuclide concentrations are then
compared with the SSLs (TBD, Section 2.6). However, the unsaturated zone models do not perform this direct
calculation. Instead, the existing site radionuclide-contamination is simulated, and the resulting radionuclide
concentrations are then compared to the acceptable radionuclide level at the receptor well.
As described above, all five models can simulate the migration of radionulides in the unsaturated zone to some degree
of satisfaction, depending on the complexity of the site-specific conditions. The final aspect for model applicability
considered here is the complexity of the model and its ease-of-use. As given in Table 3.6, CHAIN is the simplest
and easiest model while the rest are approximately the same in model complexity and ease-of-use. The pre- and post-
processors for EPACMTP (FECTUZ) and MULTIMED (MULTIMED_DP) would provide some degree of help in
the use of the model. Lastly, usability of the model user's manuals is also very important for model applicability,
but it is not evaluated here. The following sections provide the description of the model applicability in the SSL
processes. A summary table is also provided to highlight the key points in the model application to the SSL processes
(Table 3.10).
HYDRUS. The HYDRUS model can simulate chemical movement in layered soils from a finite source. It may be
useful in settings where low-permeability clay layers may attenuate contaminants through adsorption. The model
also considers root zone water uptake and evapotranspiration so that infiltration (net recharge) into the soil can be
obtained from rainfall data. The net infiltration amount needs to be provided by the user for the models CHAIN,
FECTUZ, and MULTIMED-DP. HYDRUS outputs the radionuclide concentration in the soil water as a function
of time and depth (including at the water table), which can be used in the SSLs comparison, along with the amount
of chemical remaining in the soil. The model also outputs cumulative solute flux across the bottom of the soil profile
(water table). Because it can estimate infiltration from rainfall intensities, HYDRUS may be useful in SSL
applications. Grid discretization for HYDRUS version 6.0 requires extra effort. Availability of root water uptake
parameters and potential evaporation might be limited at a site if the root water uptake process is considered.
MULTIMED-DP. MULTIMED-DP simulates simple vertical water movement in the unsaturated zone. Because
an initial soil concentration cannot be specified, either the soil/water partition equation or a leaching test (SPLP) must
be used to estimate soil leachate contaminant concentrations. MULTIMED_DP is appropriate for simulating
contaminant migration in soil and can be used to model vadose zone attenuation of leachate concentrations derived
from a partition equation. MULTIMED_DP outputs the leachate radionuclide concentration at the water table which
is needed for the SSLs comparison. In addition, because it links the output from the unsaturated zone transport
module with a saturated zone module, it can be used to determine the concentration of a radionuclide in a receptor
well. MULTIMED_DP is appropriate for early-stage site simulations because the input parameters required are
3-11
-------
typically available (a database of chemical properties is provided in MULTIMED_DP). This model is the only model
which considers runoff. Therefore, when runoff is an important hydrological process at a site, selection of
MULTIMED_DP might be considered. Uncertainty analyses can be performed using Monte Carlo simulations for
those parameters for which reliable values are not known. Because of the complexity of the model and requirement
of a great amount of input data, expertise in properly using the model for SSLs is essential.
FECTUZ. FECTUZ at one time was a stand-alone unsaturated zone model, but, in recent years the model is only
available as one of the coupled modules in the EPACMTP program (EPA 1995b, 1995c). FECTUZ simulates
migration of contaminants from a landfill (or a surface impoundment), with finite or infinite sources, through the
unsaturated zone and into ground water. The model allows time varying precipitation data for inputs, but a steady-
state water flow is assumed during each precipitation event. The concentrations of radionuclides at a specific time
and space are part of the model outputs and can be used in the SSL process. The implementation of Monte Carlo
simulation in the EPACMTP program also provides the capacity of uncertainty analysis for the FECTUZ model. It
is worthy to note that mixed units for the model inputs (e.g. cm/hr, m/year, cm, m) are used in FECTUZ. Therefore,
caution should be taken for the use of correct values and units in FECTUZ.
CHAIN. The CF£AIN model is simple and easy to use in simulating fate and transport of radionuclides in a uniform
soil under steady unsaturated flow conditions. The model outputs the leachate radionuclide concentrations at the
specified time and depth (including at the water table) and these concentrations can be used in the SSL process. It
is not adequate to use CHAIN for fate and transport of a radionuclide in layered soil. However, the CHAIN model
might be used in a layered soil by using representative uniform soil properties as an approximation of heterogeneous
soil properties. The CHAIN model could be the first choice as a preliminary assessment tool in SSLs evaluating for
radionuclides.
CHAIN 2D. Basically, CHAIN 2D is very similar to HYDRUS 6.0. The major difference is it is a two-dimensional
model. When the assumption of vertical flow in the unsaturated zone is in question, CHAIN 2D may be the choice
to simulate leaching of a radionuclide from the disposal facility to ground water. This is especially true when the
leaching area is small and horizontal flow becomes significant in a highly stratified soil. The CHAIN 2D model
outputs the radionuclide concentration in the soil water as a function of time and depth, which can be used in the
S SLs comparison, along with the amount of chemical remaining in the soil. The model also outputs cumulative solute
flux across the bottom of the soil profile (water table). Because it can estimate infiltration from rainfall, the model
may be useful in SSL applications. Both the CHAIN 2D and HYDRUS models are capable of simulating water
uptake by plant roots. Both models can be used for estimating net recharge when potential evaporation and water
uptake parameters are available.
Examples of Model Application in the Soil Screening Processes
To demonstrate how to apply the five models for radionuclides evaluated herein in the SSL estimation, a conceptual
site model was developed at the Las Cruces Trench Site and the base case simulations were performed. It is assumed
that the Site had been used as a waste disposal/storage facility where radionuclides from tank leaks or improper waste
disposal were released to the soil surface for 1000 days with a total amount for 3xlO"4 mg /cm2 "Tc ("Tc
concentration from the waste source is 1.25x10"2 mg/L). Rainfall infiltration (with a net annual recharge rate of 87
mm/y) is the driving force for the downward migration of radionuclide to the water table beneath. Base values of
the input parameter are given in Table 3.11. It is assumed that the steady-state uniform water flow occurs at the site.
Time-varied leachate concentrations of "Tc entering ground water predicted by the CHAIN, HYDRUS,
MULTIMED_DP, FECTUZ, and CHAIN 2D models are presented in Figures 3.1 and 3.2. The results indicate that
the five models provide similar breakthrough curves except that numerical dispersion is observed using
MULTIMED_DP. Figures 1 and 2 also show how variability of the distribution coefficient (Kd) influence the
migration of "Tc. Increasing Kd values would reduce migration of radionuclides. The breakthrough curves for the
decay (daughter) product "Ru, using the CHAIN and FECTUZ (EPACMTP), are presented in Figure 3.3. The
3-12
-------
breakthrough curves predicted by the CHAIN and FECTUZ models are not distinguishable. Using
MULTIMED_DP, a large numerical dispersion was observed (not shown). It was also found that the breakthrough
curves predicted by the CHAIN model will exhibit a great oscillation if single precision is used in the CHAIN code.
Figure 3.3 is also used to demonstrate the decay reaction implemented in the five models. No attempt was made in
using the HYDRUS and CHAIN 2D models for obtaining the breakthrough curves of the decay products since
simulation times of up to 800 years are required.
For the use of models in the SSL process, the task is to simulate the conceptual site scenario and examine whether
the leachate concentration entering the ground water will result in an exceedance of the radionuclide concentration
over the MCL at the receptor well. The results shown in Figure 3.4 indicate that the calculated concentrations at
the receptor well exceed the MCL (5.3E-5 mg/L) for "Tc. In other words, the site conditions exhibit a soil
radionuclide contamination of "Tc exceeding the SSL and thus has imposed a potential risk to human health and
the environment. Note a few things about Figure 3.4: 1) three receptor well locations (Om, 10m and 30m from the
waste source) were examined; 2) two methods were employed to calculate the dilution effect of the radionuclides
in the saturated zone once they enter the water table. One was the use of the simple dilution factor. Another method
was the use of a saturated zone model (SZM) such as FECTUZ/EPACMTP to simulate the radionuclide transport.
Using the dilution factor, a default DAF value of 20 as well as the dilution factors, at three ground-water velocities
(5 m/y, 10 m/y, and 25 m/y), obtained from equations 7 and 8 in the section 2.5.2 of the SSG user's guide were used.
The parameters used for the dilution factor calculation and the ground water flow and transport simulation using
FECTUZ/EPACMTP are given in Table 3.12.
The details of the model application in the SSL process and the sensitivity analysis are provided in the report entitled
"Evaluation of Computer Models for Simulating Radionuclide Transport in the Unsaturated Zone"
3-13
-------
Table 3.6. Summary Comparisons of the Vadose Zone Model
for Radionuclides in the SSL Process
Model component
HYDRUS MULTIMED- FECTUZ
DP
CHAIN CHAIN 2D
Contaminants
Organics
Metals
Radionuclides (parent)
Radionuclides (progeny)
Non-aqueous phase liquids
Site characteristics
Vadose zone
Layered soil
Sources types
Contaminated soil
Landfill
Surface impoundment
Waste piles
Source term characteristics
Mass balance
Multimedia partitioning
Source decay
Source ingrowth (radionuclides)
Multiple contaminants per simulations
Source release mechanisms
Leaching
Direct release to:
Vadose zone
Groundwater
Surface water
Air
Medium-specific flow
Air Box model (0-D, complete mixing)
Surface Hydrology
Precipitation
Runoff
Infiltration
ET
Surface Water (Stream discharge)
Vadose Zone
Vadose zone (Steady-state infiltration —>soil
Vadose zone (n-D dynamic)
Groundwater
Medium-specific contaminant transport
Atmosphere (emission through diffusion)
Surface water (stream interception and mixing)
Vadose zone (1-D advection and dispersion)
Vadose zone (2-D advection and dispersion)
Groundwater
Homogeneous aquifer (1-D advection and dispersion)
Homogeneous aquifer (2-D advection and dispersion)
Homogeneous aquifer (3-D advection and dispersion)
Medium-specific heat transport
3-14
-------
Contaminant transformations and fate processes
1st order decay (not decay products)
1st order decay (with chained daughter and
granddaughter decay products) -straight chain
1st order decay - branch chain
Non-1 st order decay
Linear partitioning (water/soil)
Nonlinear partitioning (water/soil)
Hydrolysis
Chemical reactions/speciation
Intermedia contaminant fluxes
Surface soil -> Air (volatilization)
Surface soil -> Vadose zone (leaching)
Surface soil -> Overland (erosion, runoff)
Surface water ->Sediment (sedimentation)
Vadose zone -> groundwater (percolation)
Vadose zone -> Air (volatilization)
Groundwater -> Surface water (deposition)
Air-> Surface soil (deposition)
Air -> Surface water (deposition)
Exposure pathways
Human health endpoints
! denotes component is included in model; - denotes component is not included in model.
3-15
-------
Table 3.7. Soil Hydraulic Properties at the Las Cruces Trench Site for SSG Model Evaluation Study
(Modified from Wierenga et al., 1991)
Layers
Depth (cm)
Saturated water
content (cmVcm3)
Residual water
content (cmVcm3)
van Genuchten
alpha coefficient,
a, (cm'1)
van Genuchten
beta coefficient, |3
(-)
Saturated
hydraulic
conductivity, Ks,
(cm/d)
Uniform Soil Model
all
0-600
0.321
0.083
0.055
1.509
270
Layered Soil Model
1
2
3
4
5
6
7
8
9
0-15
15-140
140-205
205-250
250-305
305-370
370-460
460-540
540-600
0.348
0.343
0.336
0.313
0.302
0.294
0.310
0.325
0.306
0.095
0.091
0.085
0.071
0.072
0.090
0.073
0.083
0.078
0.042
0.062
0.060
0.068
0.040
0.070
0.027
0.041
0.047
1.903
1.528
1.574
1.537
1.550
1.711
1.418
1.383
1.432
539
250
267
300
250
334
221
172
226
3-16
-------
Table 3.8. Solute Transport Properties at the Las Cruces Trench Site for SSG Model Evaluation Study
(Modified from Porro and Wierenga, 1993)
Layers
Depth (cm)
Pore velocity, v, (cm/d)
Dispersion coefficient, D,
(cm2/d)
Dispersivity, 8, (cm)
Uniform Soil Model
all
0-500
14.7
62.1
4.53
Layered Soil Model
1
2
3
4
5
6
82
125
220
310
400
500
18.4
16.5
17.1
13.2
10.9
12.3
40.5
66.1
52.5
67.3
84.6
61.9
2.20
4.00
3.06
5.09
7.80
5.04
3-17
-------
Table 3.9. Characteristics of the Las Cruces Trench Site for SSG Model Evaluation Study
(modified from Gee et al., 1994)
Annual
Precipitation
(cm/y)
33.8
Annual
Potential
(Pan)
Evaporation
(cm/y)
239
Annual
Potential
Recharge
(cm/y)
8.7
Average
Daily Max.
Air
Temperature
(°C)
28
Average
Daily Min.
Air
Temperature
(°C)
13
Elevation
(m)
1357
Depth to
Water Table
(m)
60
Geology
Alluvial
Typical Soil
Type
Berino fine
loamy sand
Typical
Vegetation
Creosote
bush
3-18
-------
Table 3.10. Summary of the Use of the Unsaturated Zone Models
for Radionuclides in the SSL Process
Model
Processes, outputs, components
HYDRUS
- provides the leachate radionuclide concentrations entering the ground water so
whether the resulting concentration of the radionuclide at the receptor well would
exceed the acceptable level or not, can be examined
- calculates infiltration which can be used as inputs in the SSL calculation
- considers soil heterogeneity, time-varying infiltration and evapotranspiration
- outputs radionuclide concentration in soil, cumulative flux across water table
- grid discretization for HYDRUS version 6.0 requires extra effort
MULTIMED DP
provides the leachate radionuclide concentrations entering the ground water so
whether the resulting concentration of the radionuclide at the receptor well would
exceed the acceptable level or not, can be examined
uncertainty of model outputs can be examined
considers runoff, evapotranspiration
linked with a saturated flow and transport model
requires a great amount of input data, expertise because of model complexity
FECTUZ
provides the leachate radionuclide concentrations entering the ground water so
whether the resulting concentration of the radionuclide at the receptor well would
exceed the acceptable level or not, can be examined
uncertainty of model outputs can be examined
linked with a saturated flow and transport model
uses mixed units for the input data
CHAIN
provides the leachate radionuclide concentrations entering the ground water so
whether the resulting concentration of the radionuclide at the receptor well would
exceed the acceptable level or not, can be examined
used for simplified radionuclide-contaminated site scenario
simple-to-use, less input data requirement
as a preliminary assessment tool in SSL estimation
CHAIN 2D
provides the leachate radionuclide concentrations entering the ground water so
whether the resulting concentration of the radionuclide at the receptor well would
exceed the acceptable level or not, can be examined
calculates infiltration which can be used as inputs in the SSL calculation
considers soil heterogeneity, time-varying infiltration and evapotranspiration
outputs radionuclide concentration in soil, cumulative flux across water table
considers two-dimensional soil heterogeneity
3-19
-------
Table 3.11. Base Values of Input Parameters for the Unsaturated Zone Models
Parameters
Values
Source-Specific Parameters
Area of disposal facility (m2)
Width of disposal facility (m)
Length of disposal facility (m)
Mass release of (parent) radionuclide "TC (mg/cm2)
Concentration of "Tc in recharge water from waste source (mg/L)
Duration of waste source being completely released (days)
Net recharge rate (mm/y)
Water content (cm3/cm3)
Source decay constant
Soil Properties in Unsaturated Zone (Uniform Soil in Table 3-7)
Saturated hydraulic conductivity, Ks, (cm/d)
Porosity (--)
Residual water content (cm3/cm3)
Saturated water content (cm3/cm3)
Bulk density
van Genuchten alpha coefficient, a, (cm"1)
van Genuchten beta coefficient, P (--)
Depth to water table (m)
Solute Transport Parameters
Decay coefficient for parent (99TC) (1/y)
Decay coefficient for daughter (99Ru) (1/y)
Distribution coefficient for parent (99TC) (ml/g)
Distribution coefficient for daughter (99Ru) (ml/g)
Dispersion Coefficient (cm2/d )
Dispersivity (cm)
400
20
20
3xlQ-4
1.25xlQ-2
1000
87
0.16
270.1
0.358
0.083
0.321
1.70
0.055
1.509
6
3.3xlQ-6
7.9xlQ-9
0.007
5.0
1.0
4.53
3-20
-------
Table 3.12. Input Parameters Used for the Dilution Factor Calculation and Transport Simulation of
Radonuclide in the Saturated Zone using FECTUZ/EPACMTP Model
Parameters Values
Hydraulic Properties in Saturated Zone
Aquifer hydraulic conductivity (m/y) 500
Hydraulic gradient (m/m) 0.005
Porosity 0.32
Bulk density (g/cm3) 1.70
Longitudinal dispersivity (m) 1.0
Transverse dispersivitiy (m) 0.2
Vertical dispersivity (m) 0.04
Aquifer Characteristics
Aquifer (m) 10
Computed Dilution Factor Using Equations 7 & 8 in
the Section 2.5.2 of the SSG User's Guide
Dilution factor for ground-water velocity = 5 m/y 8.1
Dilution factor for ground-water velocity =10 m/y 14.2
Dilution factor for ground-water velocity = 25 m/y 32.4
3-21
-------
c.
Qj,
o
o
o
at
01
0.010
0.008
0.006
0.004
0.002
0
0,010
0.008
0*006
0.004
0.002
0
0.010
0.008
0.006
0.004
0,002
0
0
CHAIN 2D
Distribution Coefficient (mL/g)
0.007
0.25
0.5
HYDRUS Distribution Coefficient (mL/cj)
0,007
0,25
0,5
CHAIN
Distribution Coefficient (mL/g)
0.007
0.25
0.5
10000 20000 30000 40000
Tim© (days)
50000
Figure 3.1, Base case simulations of "TC breakthrough (through the unsaturated zone)
with distribution coefficients using the CHAIN, ifOTKPg,, and CHMN 2D
models.
3-22
-------
0.010
0.008
0.006
0.004
J: 0.002
c
CD
£ 0
3 o.oio
m
en
MULTIMED DP
Distribution Coefficient (mL/g)
0.007
0.25
0.5
0.008
0.006
0.004
0.002
FECTUZ
Distribution Coefficient (mL/g)
0.007
0.25
0.5
--T
0
10000 2000O 30000
Time (days)
40000
50000
Figure 3,2. Base case simulations of "TC breakthrough (Uirough the unsaturated zcne)
with three different distribution coefficients using the FJgQHJS, and ^
models.
3-23
-------
B9
93-
The aaRu Produci of aalc Decay
V*
•**,— *••
s~
_
Concentrotl
3
35
1 X 1 U
Sx 1 Q~3
6x1 CT9
2x10 9
0
i i
Distribution Coefficient = 0,007 mL/g
T L.U. ( U.l£n
j^ LJ * i y
L, IP J% l|Tj
1 1 1 ^^^A^^^.
Time (Days)
3.3. case simulations of the daughter product — "Ru of "TC breakthrough
(through the unsaturated zone) with a distribution coefficient of 0.007 .mL/g using
.EECTIE, and CHAIN model s.
3-24
-------
c
o
o
o
I—
en
U.UUJ
0.002
0.001
0
. ^ ^
[25 m/y)
I I
y = 10 m Estimation
Method
Dilution Method
SZM (5 m/y)
SZM (10 m/y)
SZM (25 m/y)
/\
/- \
• ' , ^ s.
• / ";-\ ^
• // A %
••/'.---- -N. \
• 'J/V"" "^--.'..-^. x^
-
I I
x = 0 m EstTmalion
Method
Dilution Method
MCL=5.3x1Cr5mg/L SZM
f •. ___ _ _— - ,JI JL. lV|
/ 1 -UT — _
/ » SZM
\ DAF=
/,/^A \
/"' ^ ^
_r^"""^^-^
5 m/y) _
10 m/y)
25 m/y)
20
-
0
2000 4.000 6000
Tim© (days)
8000
10000
Figure 3.4. predication of "TC concentration at the receptor well by
employing dilution fector method and using saturated zone model S^SM) —
The numbers in the parenthes2s are ground water ¥elocity. The solid lines are used to
represent the,MQt< (5.3 x 1O5 ml/L) level.
3-25
-------
Part 4: MEASURING RADIONUCLIDE CONCENTRATIONS IN SOIL
The Soil Screening Guidance for Radionuclides includes a sampling strategy for implementing the soil screening
process. Section 4.1 presents the sampling approach for surface soils. This approach provides a simple decision rule
based on comparing the maximum radionuclide concentrations of composite samples with surface soil screening
levels (the Max test) to determine whether further investigation is needed for a particular exposure area (EA). In
addition, this section presents a more complex strategy (the Sign test) that allows the user to design a site-specific
quantitative sampling strategy by varying decision error limits and soil radionuclide variability to optimize the
number of samples and composites. Section 4.2 provides a subsurface soil sampling strategy for developing SSLs
and applying the screening procedure for the migration to ground water exposure pathways.
Section 4.3 describes the technical details behind the development of the SSL sampling strategy, including analyses
and response to public and peer-review comments received on the December 1994 draft Soil Screening Guidance for
chemicals.
The sampling strategy for the soil screening process is designed to achieve the following objectives:
• Estimate mean concentrations of radionuclides of concern for comparison with
SSLs
• Fill in the data gaps in the conceptual site model necessary to develop SSLs.
The soils of interest for the first objective differ according to the exposure pathway being addressed. For the direct
ingestion of soil, inhalation of fugitive dust, external radiation exposure, and ingestion of homegrown produce
pathways, EPA is concerned about surface soils. The sampling goal is to determine average radionuclide
concentrations of surface soils in exposure areas of concern. For the migration to ground water pathway, subsurface
soils are the primary concern. For this pathway, the average radionuclide concentration through each source is the
parameter of interest.
The second objective (filling in the data gaps) applies primarily to the external radiation exposure, inhalation of
fugitive dust, ingestion of homegrown produce, and migration to ground water pathways. For these pathways, the
source area and depth as well as average soil properties within the source are needed to calculate the pathway-specific
SSLs. Therefore, the sampling strategy needs to address collection of these site-specific data.
Because of the difference in objectives, the sampling strategies for surface soils and subsurface soils are addressed
separately. If both surface and subsurface soils are a concern, then surface soils should be sampled first because the
results of surface soil analyses may help delineate source areas to target for subsurface sampling.
At some sites, athird sampling objective may be appropriate. As discussed in the Soil Screening Guidance, applying
SSLs may be difficult at sites where background radionuclide levels are above the SSLs. Where sampling
information suggests that background radionuclide concentrations may be a concern, background sampling may be
necessary. Methods for Evaluating the Attainment of Cleanup Standards - Volume 3: Reference-Based Standards
for Soil and Solid Media (U.S. EPA, 1994e) provides further information on sampling soils to determine background
conditions at a site. The Multi-Agency Radiation Survey and Site Investigation Manual (MARSSIM) (U.S. EPA,
1997c) also provides specific guidance concerning site surveys when background radiation is a consideration.
In order to accurately represent radionuclide distributions at a site, EPA used the Data Quality Objectives (DQO)
4-1
-------
process (Figure 4) to develop a sampling strategy that will satisfy Superfund program objectives. The DQO process
is a systematic data collection planning process developed by EPA to ensure that the right type, quality, and quantity
of data are collected to support EPA decision making. As shown in Sections 4.1.1 through 4.1.6, most of the key
outputs of the DQO process already have been developed as part of the Soil Screening Guidance. The DQO activities
addressed in this section are described in detail in the Data Quality Objectives for Superfund: Interim Final Guidance
(U.S. EPA, 1993b) and the Guidance for the Data Quality Objectives Process (U.S. EPA, 1994c). Refer to these
documents for more information on how to complete each DQO activity or how to develop other, site-specific
sampling strategies.
4.1 Sampling Surface Soils
State the Problem
j,
Identify the Decision
4**
Identify Inputs to the Decision
*
Define the Study Boundaries
4**
Develop a Decision Rule
4/
Specify Limits on Decision
Errors
A sampling strategy for surface soils is presented in this section, organized
by the steps of the DQO process. The first five steps of this process, from
defining the problem through developing the basic decision rule, are
summarized in Table 4.1, and are described in detail in the first five
subsections. The details of the two remaining steps of the DQO process,
specifying limits on decision errors and optimizing the design, have been
developed separately for two alternative hypothesis testing procedures (the
Max test and the Sign test) and are presented in four (4.1.6,4.1.7,4.1.9, and
4.1.10) subsections. In addition, a data quality assessment (DQA) follows
the DQO process step for optimizing the design. The DQA ensures that
site-specific error limits are achieved. Sections 4.1.8 and 4.1.11 describe
the DQA for the Max and Sign tests, respectively. The technical details
behind the development of the surface soil sampling design strategy are
explained in Section 4.3.
4.1.1 State the Problem. In screening, the problem is to identify
the radionuclides and exposure areas (EAs) that do not pose significant risk
to human health so that future investigations can be focused on the areas
and radionuclides of concern at a site.
The main site-specific activities involved in this first step of the DQO
process include identifying the data collection planning team (including
technical experts and key stakeholders) and specifying the available
resources. The list of technical experts and stakeholders should contain all
key personnel who are involved with applying the Soil Screening Guidance
at the site. Other activities in this step include developing the conceptual site model (CSM), identifying exposure
scenarios, and preparing a summary description of the surface soil contamination problem. The User's Guide for
Radionuclides (U.S. EPA, 1999) describes these activities in with more detail.
4.1.2 Identify the Decision. The decision is to determine whether the mean surface soil concentrations
exceed surface soil screening levels for specific radionuclides within EAs. If so, the EA must be investigated further.
If not, no further action is necessary under CERCLA for the specific radionuclides in the surface soils of those EAs.
Optimize the Design for Obtaining
. Data
Figure 4. The Data Quality
Objectives process.
4-2
-------
Table 4.1. Sampling Soil Screening DQOs for Surface Soils
DQO Process Steps
Soil Screening Inputs/Outputs
State the Problem
Identify scoping team
Develop conceptual site model (CSM)
Define exposure scenarios
Specify available resources
Write brief summary of contamination
problem
Site manager and technical experts (e.g., health physicists, risk assessors,
statisticians, soil scientists)
CSM development (described in Step 1 of the User's Guide, U.S. EPA, 1996)
Direct ingestion of soil, inhalation of fugitive dusts, external radiation exposure,
and ingestion of homegrown produce in a residential setting;
Sampling and analysis budget, scheduling constraints, and available
personnel
Summary of the surface soil contamination problem to be investigated at the
site
Identify the Decision
Identify decision
Identify alternative actions
Do mean soil concentrations for particular radionuclides (e.g., radionuclides of
potential concern) exceed appropriate screening levels?
Eliminate area from further study under CERCLA
or
Plan and conduct further investigation
Identify Inputs to the Decision
Identify inputs
Define basis for screening
Identify analytical methods
SSLs for each pathway for specified radionuclides
Measurements of surface soil radionuclide concentration
Soil Screening Guidance for Radionuclides
Feasible analytical methods (both field and laboratory) consistent with
program-level requirements
Define the Study Boundaries
Define geographic areas of field
investigation
Define population of interest
Divide site into strata
Define temporal boundaries of study
Define scale of decision making
Identify practical constraints
The entire NPL site (which may include areas beyond facility boundaries),
except for any areas with clear evidence that no contamination has occurred
Surface soils (usually the top 15 centimeters, but may be deeper where
activities could redistribute subsurface soils to the surface)
Strata may be defined so that radionuclide concentrations are likely to be
relatively homogeneous within each stratum based on the CSM and field
measurements
Temporal constraints on scheduling field visits
Exposure areas (EAs) no larger than 0.5 acre each (based on residential land
use)
Potential impediments to sample collection, such as access, health, and safety
issues
4-3
-------
Table 4.1. Sampling Soil Screening DQOs for Surface Soils
DQO Process Steps Soil Screening Inputs/Outputs
Develop a Decision Rule
Specify parameter of interest ?True mean" (u) individual radionuclide concentration in each EA. (since the
determination of the Ttrue mean" would require the collection and analysis of
many samples, the ?Max Test" uses another sample statistic, the maximum
composite concentration).
Specify screening level Screening levels calculated using available parameters and site data (or
generic SSLs if site data are unavailable).
Specify "if..., then..." decision rule If the ?true mean" EA concentration exceeds the screening level, then
investigate the EA further. If the Ttrue mean" is less than the screening
level, then no further investigation of the EA is required under CERCLA.
4.1.3 Identify Inputs to the Decision. This step of the DQO process requires identifying the inputs to
the decision process, including the basis for further investigation and the applicable analytical methods. The inputs
for deciding whether to investigate further are the SSLs calculated for the site radionuclides for each pathway as
described in Part 2 of this document, and the surface soil concentration measurements for those same radionuclides.
Therefore, the remaining task is to identify radioanalytical laboratory methods and/or field methods for which the
minimum detectable concentrations (MDCs) are less than the SSLs. EPA recommends the use of field methods
where applicable and appropriate as long as quantitation limits are below the SSLs. At least 10 percent of field
samples should be split and sent to a radioanalytical laboratory for confirmatory analysis (U.S. EPA, 1993d).
However, for soil screening purposes, most SSLs for radionuclides are too low to be detected using typical
radiological field methods. Therefore, for most radionuclides, sampling and laboratory analysis must be the primary
means of soil screening for the majority of radionuclides.
4.1.4 Define the Study Boundaries. This step of the DQO process defines the sample population of
interest, subdivides the site into appropriate exposure areas, and specifies temporal or practical constraints on the data
collection. The description of the population of interest must include the surface soil depth.
Sampling Depth. When measuring soil chemical contamination levels at the surface for the ingestion and
inhalation pathways, the Soil Screening Guidance for chemicals suggests that the top 2 centimeters is usually
considered surface soil, as defined by Urban Soil LeadAbatementProject(U.S. EPA 1993f). It further recommends
that additional sampling beyond this depth may be appropriate for surface soils under a future residential use scenario
in areas where major soil disturbances can reasonably be expected as a result of landscaping, gardening, or
construction activities. In this situation, contaminants that were at depth can be moved to the surface. Thus, it is
important to be cognizant of local residential construction practices when determining the depth of surface soil
sampling and to weigh the likelihood of that area being developed.
For the Soil Screening Guidance for Radionuclides, a surface soil depth of 15 cm has been assumed in order to ensure
that SSLs calculated for radionuclides are adequately protective in a residential setting. The residential setting used
to develop the SSLs for each radionuclide and pathway assumes that: 1) there is no clean cover of soil; 2) the top few
centimeters of soil are available for resuspension in air; 3) the top 15 cm of contaminated soil are homogenized by
agricultural activities (e.g., plowing); 4) there is a sufficiently large area and depth of contamination to approximate
an infinite slab source for external exposure purposes; 5) there is enough land for a garden to produce homegrown
fruits and vegetables for the residents; and 6) while the plant root system grows to a depth of 1 meter, most plant
nutrients are obtained from within the upper 15 cm of soil.
4-4
-------
Subdividing the Site. This step involves dividing the site into areas or strata depending on the likelihood of
contamination and identifying areas with similar radionuclide patterns. These divisions can be based on process
knowledge, operational units, historical records, and/or prior sampling. Partitioning the site into such areas and strata
can lead to a more efficient sampling design for the entire site.
For example, the site manager may have documentation that large areas of the site are unlikely to have been used for
waste disposal activities. These areas would be expected to exhibit relatively low variability and the sampling design
could involve a relatively small number of samples. The greatest intensity of sampling effort would be expected to
focus on areas of the site where there is greater uncertainty or greater variability associated with contamination
patterns. When relatively large variability in radionuclide concentrations is expected, more samples are required to
determine with confidence whether the EA should be screened out or investigated further.
Initially, the site may be partitioned into three types of areas:
1. Areas that are not likely to be contaminated
2. Areas that are known to be highly contaminated
3. Areas that are suspected to be contaminated and cannot be ruled out.
Areas that are not likely to be contaminated generally will not require further investigation if this assumption is based
on historical site use information or other site data that are reasonably complete and accurate. (However, the site
manager may also want take a few samples to confirm this assumption). These may be parts of the site that are within
the legal boundaries of the property but were completely undisturbed by hazardous-waste-generating activities. All
other areas need investigation.
Areas that are known to be highly contaminated (i.e., sources) are targeted for subsurface sampling. The information
collected on source area and depth is used to calculate site-specific SSLs for the inhalation and migration to ground
water pathways (see Section 4.2 for more information).
Areas that are suspected to be contaminated (and cannot be ruled out for screening) are the primary subjects of the
surface soil investigation. If a geostatistician is available, a geostatistical model may be used to characterize these
areas (e.g., kriging model). However, guidance for this type of design is beyond the scope of the current guidance
(see Chapter 10 of U.S. EPA, 1989a).
Defining Exposure Areas. After the site has been partitioned into relatively homogeneous areas, each region
that is targeted for surface soil sampling is then subdivided into EAs. An exposure area (EA) is a physical area of
a specified size and shape for which a separate decision will be made as to whether or not the area exceeds the
screening criteria. To facilitate sampling design and ensure that the number of sample data points for a specific site
are relatively uniformly distributed among areas of similar contamination potential, the site is divided into EAs that
share a common history or other characteristics, or are naturally distinguishable from other portions of the site.
Because the SSLs were developed for a residential scenario, EPA assumes the EA is a suburban residential lot
corresponding to 0.5 acre (2,000 m2). For soil screening purposes, each EA should be 0.5 acre or less. To the extent
possible, EAs should be constructed as square or rectangular areas that can be subdivided into squares to facilitate
compositing and grid sampling. If the site is currently residential, then the EA should be the actual residential lot
size. The exposure areas should not be laid out in such a way that they unnecessarily combine areas of high and low
levels of contamination. The orientation and exact location of the EA, relative to the distribution of the radionuclide
in the soil, can lead to instances where sampling of the EA may lead to results above the mean, and other instances,
to results below the mean. Try to avoid straddling radionuclide ?distribution units" within the 0.5 acre EA.
The sampling strategy for surface soils allows investigators to determine mean soil radionuclide concentration across
an EA of interest. An arithmetic mean concentration for an EA best represents the exposure to site radionuclides over
4-5
-------
a long period of time. For risk assessment purposes, an individual is assumed to move randomly across an EA over
time, spending equivalent amounts of time in each location. Since reliable information about specific patterns of
nonrandom activity for future use scenarios is not available, random exposure appears to be the most reasonable
assumption for a residential exposure scenario. Therefore, spatially averaged surface soil concentrations are used
to estimate mean exposure concentrations.
Because all the EAs within a given stratum should exhibit similar radionuclide concentrations, one site-specific
sampling design can be developed for all EAs within that stratum. As discussed above, some strata may have
relatively low variability and other strata may have relatively high variability. Consequently, a different sampling
design may be necessary for each stratum, based upon the stratum-specific estimate of the radionuclide variability.
In addition to determining the mean concentration of a radionuclide in an EA, it is important to identify the presence
of small areas of elevated activity. This is done by the performance of scanning surveys. The sensitivity of scanning
surveys will be insufficient to detect small areas of elevated activity for most radionuclides with levels of
contamination as low as those of the SSLs calculated for large areas of uniform contamination. However, standard
scanning survey techniques may be able to detect SSLs calculated for smaller areas of contamination. Scan surveys
are intended to provide a degree of confidence that any significant areas of elevated activity are identified. Therefore,
scanning surveys should be performed for all EAs prior to sampling. The extent of the survey coverage should be
dictated by the potential for small areas of elevated activity in the EA. EAs with a high potential for small areas of
elevated activity should receive 100% coverage. In EAs with a very low potential for small areas of elevated activity,
scanning surveys should be performed in at least 10% of the area. In such cases, the areas selected for scan should
be those with highest potential based on professional judgement. Due to the limited sensitivity of scan surveys, any
small areas of elevated activity found during the survey should be identified for further investigation (i.e., not
screened out).
4.1.5 Develop a Decision Rule. Ideally, the decision rule for surface soils is:
If the mean radionuclide concentration within an EA exceeds the screening level, then investigate
that EA further.
This "screening level" is the actual numerical value used to compare against the site contamination data. It may be
identical to the SSL, or it may be a multiple of the SSL (e.g., 2 SSL) for a hypothesis test designed to achieve
specified decision error rates in a specified region above and below the SSL. In addition, another sample statistic
(e.g., the maximum concentration) may be used as an estimate of the mean for comparison with the "screening level."
4.1.6 Specify Limits on Decision Errors for the Max Test. Sampling data will be used to support
a decision about whether an EA requires further investigation. Because of variability in radionuclide concentrations
within an EA, practical constraints on sample sizes, and sampling or measurement error, the data collected may be
inaccurate or nonrepresentative and may mislead the decision maker into making an incorrect decision. A decision
error occurs when sampling data mislead the decision maker into choosing a course of action that is different from
or less desirable than the course of action that would have been chosen with perfect information (i.e., with no
constraints on sample size and no measurement error).
EPA recognizes that data obtained from sampling and analysis are never perfectly representative and accurate, and
that the costs of trying to achieve near-perfect results can outweigh the benefits. Consequently, EPA acknowledges
that uncertainty in data must be tolerated to some degree. The DQO process controls the degree to which uncertainty
in data affects the outcomes of decisions that are based on those data. This step of the DQO process allows the
decision maker to set limits on the probabilities of making an incorrect decision.
The DQO process utilizes hypothesis tests to control decision errors. When performing a hypothesis test, a presumed
4-6
-------
or baseline condition, referred to as the "null hypothesis" (H0), is established. This baseline condition is presumed
to be true unless the data conclusively demonstrate otherwise, which is called "rejecting the null hypothesis" in favor
of an alternative hypothesis. For the Soil Screening Guidance for Radionuclides, the baseline condition, or H0, is that
the site needs further investigation.
When the hypothesis test is performed, two possible decision errors may occur:
1. Decide not to investigate an EA further (i.e., "walk away") when the correct decision (with complete
and perfect information) would be to "investigate further"
2. Decide to investigate further when the correct decision would be to "walk away."
Since the site is on the NPL, site areas are presumed to need further investigation. Therefore, the data must provide
clear evidence that it would be acceptable to "walk away." This presumption provides the basis for classifying the
two types of decision errors. The "incorrectly walk away" decision error is designated as the Type I decision error
because one has incorrectly rejected the baseline condition (null hypothesis). Correspondingly, the "unnecessarily
investigate further" decision error is designated as the Type II decision error.
To complete the specification of limits on decision errors, Type I and Type II decision error probability limits must
be defined in relation to the SSL. First a "gray region" is specified with respect to the mean radionuclide
concentration within an EA. The gray region represents the range of radionuclide levels near the SSL, where
uncertainty in the data (i.e., the variability) can make the decision "too close to call." In other words, when the
average of the data values is very close to the SSL, it would be too expensive to generate a data set of sufficient size
and precision to resolve what the correct determination should be. (i.e., Does the average concentration fall "above"
or "below" the SSL?)
The Soil Screening Guidance for Radionuclides establishes a default range for the width and location of the "gray
region": from one-half the SSL (0.5 SSL) to two times the SSL (2 SSL). By specifying the upper edge of the gray
region as twice the SSL, it is possible that exposure areas with mean values slightly higher than the SSL may be
screened from further study. However, EPA believes that the exposure scenario and assumptions used to derive SSLs
are sufficiently conservative to be protective in such cases.
On the lower side of the gray region, the consequences of decision errors at one-half the SSL are primarily financial.
If the lower edge of the gray region were to be moved closer to the SSL, then more exposure areas that were truly
below the SSL would be screened out, but more money would be spent on sampling to make this determination. If
the lower edge of the gray region were to be moved closer to zero, then less money could be spent on sampling, but
fewer EAs that were truly below the SSLs would be screened out, leading to unnecessary investigation of EAs. The
Superfund program chose the gray region to be one-half to two times the SSL after investigating several different
ranges. This range forthe gray region represents abalance between the costs of collecting and analyzing soil samples
and making incorrect decisions. While it is desirable to estimate exactly the exposure area mean, the number of
samples required are much more than project managers are generally willing to collect in a "screening" effort.
Although some exposure areas will have radionuclide concentrations that are between the SSL and twice the SSL
and will be screened out, human health will still be protected given the conservative assumptions used to derive the
SSLs.
The Soil Screening Guidance for Radionuclides establishes the following goals for Type I and Type II decision error
rates:
• Prob ("walk away" when the true EA mean is 2 SSL) = 0.05
• Prob ("investigate further" when the true EA mean is 0.5 SSL) = 0.20.
4-7
-------
This means that there should be no more than a 5 percent chance that the site manager will "walk away" from an EA
where the true mean concentration is 2 SSL or more. In addition, there should be no more than a 20 percent chance
that the site manager will unnecessarily investigate an EA when the mean is 0.5 SSL or less.
These decision error limits are general goals for the soil screening process. Consistent with the DQO process, these
goals may be adjusted on a site-specific basis by considering the available resources (i.e., time and budget), the
importance of screening surface soil relative to other potential exposure pathways, consequences of potential decision
errors, and consistency with other relevant EPA guidance and programs.
Table 4.2 summarizes this step of the DQO process for the Max test, specifying limits on the decision error rates, and
the final step of the DQO process for the Max test, optimizing the design. Figure 5 illustrates the gray region for the
decision error goals: a Type I decision error rate of 0.05 (5 percent) at 2 SSL and a Type II decision error rate of 0.20
(20 percent) at 0.5 SSL.
4.1.7 Optimize the Design for the Max Test. This section provides instructions for developing an
optimum sampling strategy for screening surface soils. It discusses compositing, the selection of sampling points
for composited and uncomposited surface soil sampling, and the recommended procedures for determining the sample
sizes necessary to achieve specified limits on decision errors using the Max test.
Note that the size, shape, and orientation of sampling volume (i.e., ?support") for heterogenous media have a
significant effect on reported measurement values. For instance, particle size has a varying affect on the transport
and fate of radionuclides in the environment and on the potential receptors. Because comparison of data from
methods that are based on different supports can be difficult, defining the sampling support is important in the early
stages of site characterization. This may be accomplished through the DQO process with existing knowledge of the
site, contamination, and identification of the exposure pathways that need to be characterized. Refer to Preparation
of Soil Sampling Protocols: Sampling Techniques and Strategies (U.S. EPA, 1992f) for more information about soil
sampling support.
The SAP developed for surface soils should specify sampling and analytical procedures as well as the development
of QA/QC procedures. To identify the appropriate analytical procedures, the screening levels must be known. If data
are not available to calculate site-specific SSLs, then the generic SSLs in Appendix A should be used.
Compositing. Because the objective of surface soil screening is to ensure that the mean radionuclide
concentration does not exceed the screening level, the physical "averaging" that occurs during compositing is
consistent with the intended use of the data. Compositing allows a larger number of locations to be sampled while
controlling analytical costs because several discrete samples are physically mixed (homogenized) and one or more
subsamples are drawn from the mixture and submitted for analysis. If the individual samples in each composite are
taken across the EA, each composite represents an estimate of the EA mean.
-------
Table 4.2. Sampling Soil Screening DQOs for Surface Soils under the Max Test
DQO Process Steps
Soil Screening Inputs/Outputs
Specify Limits on Decision Errors*
Define baseline condition (null hypothesis)
Define the gray region**
Define Type I and Type II decision errors
Identify consequences
Assign acceptable probabilities of Type I and Type
decision errors
Define QA/QC goals
The EA needs further investigation
From 0.5 SSL to 2 SSL
Type I error: Do not investigate further ("walk away from") an EA
whose true mean exceeds the screening level of 2 SSL
Type II error: Investigate further when an EA's true mean falls below
the screening level of 0.5 SSL
Type I error: potential public health consequences
Type II error: unnecessary expenditure of resources to investigate
further
Goals:
Type I: 0.05 (5%) probability of not investigating further when ?true
mean" of the EA is 2 SSL
Type II: 0.20 (20%) probability of investigating further when Ttrue
mean" of the EA is 0.5 SSL
Radioanalytical laboratory precision and bias requirements
10% laboratory analyses for field methods
Optimize the Design
Determine how to best estimate Ttrue mean"
Determine expected variability of EA surface soil
radionuclide concentrations
Design sampling strategy by evaluating costs and
performance of alternatives
Develop planning documents for the field
investigation
Samples composited across the EA estimate the EA mean (X). Use
maximum composite concentration as a conservative estimate of
the true EA mean.
A conservatively large expected coefficient of variation (CV) from
prior data for the site, field measurements, or data from other
comparable sites and expert judgment. A minimum default CV of
2.5 should be used when information is insufficient to estimate the
CV.
Lowest cost sampling design option (i.e., compositing scheme and
number of composites) that will achieve acceptable decision error
rates
Sampling and Analysis Plan (SAP)
Quality Assurance Project Plan (QAPJP)
Since the DQO process controls the degree to which uncertainty in data affects the outcome of decisions that are
based on that data, specifying limits on decision errors will allow the decision maker to control the probability of
making an incorrect decision when using the DQOs.
The gray region represents the area where the consequences of decision errors are minor (and uncertainty in
sampling data makes decisions too close to call).
A practical constraint to compositing in some situations is the heterogeneity of the soil matrix. The efficiency and
effectiveness of the mixing process may be hindered when soil particle sizes vary widely or when the soil matrix
contains foreign objects, organic matter, viscous fluids, or sticky material. Soil samples should not be composited
if matrix interference among radionuclides is likely (e.g., when the presence of one radionuclide biases analytical
results for another).
Before individual specimens are composited for chemical analysis, the site manager should consider homogenizing
and splitting each specimen. By compositing one portion of each specimen with the other specimens and storing one
4-9
-------
portion for potential future analysis, the spatial integrity of each specimen is maintained. If the concentration of a
radionuclide in a composite sample is high, the splits of the individual specimens from which it was composed can
be analyzed discretely to determine which individual specimen(s) have high concentrations of the radionuclide. This
will permit the site manager to determine which portion within an EA is contaminated without making a repeat visit
to the site.
1.0
0.9 _
0,8 _
0.7 _
0,6
Probability of
Deciding that 0 5
the Mean '
Exceeds the
Screening Level
0.4
0.3
02
TotwtfM*
Decision
Enw
m.
" - -
X
Tolerable Type i
Decision Error'
Rat«s
.
'
Gray Region
(Relatively Large
Decision Error Rates
are Considered
Tolerable.)
_ 0.9
_ 0.6
_ Q-7
_ 0.6
_ 0,5
_ 0.4
_ Q-3
_ 0.2
_ 0.1
0
OSS
1
T
0.5* SSL SSL 2xSSL
True Mean Contaminant Concentration
Figure 5. Design performance goal diagram.
Sample Pattern. The Max test should only be applied using composite samples that are representative of the
entire EA. However, the Sign test (see Section 4.1.9) can be applied with individual, uncomposited samples. There
are several options for developing a sampling pattern for compositing that produce samples that should be
representative. If individual, uncomposited samples will be analyzed for radionuclide concentrations, the N sample
points can be selected using either (1) simple random sampling (SRS), (2) stratified SRS, or (3) systematic grid
sampling (square or rectangular grid) with a random starting point (SyGS/rs). Step-by-step procedures for selecting
SRS and SyGS/rs samples are provided in Chapter 5 of the U.S. EPA (1989a) and Chapter 5 of U.S. EPA (1994e).
If stratified random sampling is used, the sampling rate must be the same in every sector, or stratum of the EA.
Hence, the number of sampling points assigned to a stratum must be directly proportional to the surface area of the
stratum.
Systematic grid sampling with a random starting point is generally preferred because it ensures that the sample points
will be dispersed across the entire EA. However, if the boundaries of the EA are irregular (e.g., around the perimeter
of the site or the boundaries of a stratum within which the EAs were defined), the number of grid sample points that
fall within the EA depends on the random starting point selected. Therefore, for these irregularly shaped EAs, SRS
or stratified SRS is recommended. Moreover, if a systematic trend of contamination is suspected across the EA (e.g.,
a strip of higher contamination), then SRS or stratified SRS is recommended again. In this case, grid sampling would
be likely to result in either over- or under representation of the strip of higher radionuclide levels, depending on the
random starting point.
4-10
-------
For composite sampling, the sampling pattern used to locate the discrete sample specimens that form each composite
sample (N) is important. The composite samples should be formed in a manner that is consistent with the
assumptions underlying the sample size calculations. In particular, each composite sample should provide an
unbiased estimate of the mean radionuclide concentration over the entire EA. One way to construct a valid composite
of C specimens is to divide the EA into C sectors, or strata, of equal area and select one point at random from each
sector. If sectors (strata) are of unequal sizes, the simple average is no longer representative of the EA as a whole.
Five valid sampling patterns and compositing schemes for selecting N composite samples that each consist of C
specimens are listed below:
1. Select an SRS consisting of C points and composite all specimens associated with these points into
a sample. Repeat this process N times, discarding any points that were used in a previous sample.
2. Select an SyGS/rs of C points and composite all specimens associated with the points in this sample.
Repeat this process N times, using a new randomly selected starting point each time.
3. Select a single SyGS/rs of C x N points and use the systematic compositing scheme that is described
in Highlight 3 to form N composites, as illustrated in Figure 6.
4. Select a single SyGS/rs of C x N points and use the random compositing scheme that is described
in Highlight 4 to form N composites, as illustrated in Figure 7.
5. Select a stratified random sample of C x N points and use a random compositing scheme, as
described in Highlight 5, to form N composites, as illustrated in Figure 8.
Methods 1, 2, and 5 are the most statistically defensible, with method 5 used as the default method in the Soil
Screening Guidance for Radionuclides. However, given the practical limits of implementing these methods, either
method 3 or 4 is generally recommended for EAs with regular boundaries (e.g., square or rectangular). As noted
above, if the boundaries of the EA are irregular, SyGS/rs sampling may not result in exactly C xN sample points.
Therefore, for EAs with irregular boundaries, method 5 is recommended. Alternatively, a combination of methods
4 and 5 can be used for EAs that can be partitioned into C sectors of equal area of which K have regular boundaries
and the remaining C - K have irregular boundaries.
Additionally, compositing within sectors to indicate whether one sector of the EA exceeds SSLs is an option that may
also be considered. See Section 4.3.6 for a full discussion.
Sample Size. This section presents procedures to determine sample size requirements forthe Max test that achieve
the site-specific decision error limits discussed in Section 4.1.6. The Max test is based on the maximum
concentration observed in N composite samples that each consist of C individual specimens. The individual
specimens are selected so that each of the N composite samples is representative of the site as a whole, as discussed
above. Hence, this section addresses determining the sample size pair, C and N, that achieves the site-specific
decision error limits. Directions for performing the Max test in a manner that is consistent with DQOs established
for a site are presented later in this section.
Table 4.3 presents the probabilities of Type I errors at 2 SSL and Type II errors at 0.5 SSL (the boundary points of
the gray region discussed in Section 4.1.6) for several sample size options when the variability for concentrations
of individual measurements across the EA ranges from 100 percent to 400 percent (CV = 1.0 to 4.0). Two choices
for the number, C, of specimens per composite are shown in this table: 4 and 6. Fewer than four specimens per
composite is not considered sufficient for the Max test. Fewer than four specimens per composite does not achieve
the decision error limit goals for the level of variability generally encountered at CERCLA sites. More than six
4-11
-------
specimens may be more than can be effectively homogenized into a composite sample.
The number, N, of composite samples shown in Table 4.3 ranges from 4 to 9. Fewer than four samples is not
considered sufficient because, considering decision error rates from simulation results (Section 4.3), the Max text
should be based on at least four independent estimates of the EA mean. More than nine composite samples per EA
is generally unlikely for screening surface soils at Superfund sites. However, additional sample size options can be
determined from the simulation results reported in Appendix I.
Highlights: Procedure for Compositing of Specimens from a Grid Sample
Using a Systematic Scheme (Figure 6)
1. Lay out a square or triangular grid sample over the EA, using a random start. Step-by-step
procedures can be found in Chapter 4 of U.S. EPA (1989a). The number of points in the grid
should be equal to C x N, where C is the desired number of specimens per composite and N is
the desired number of composites.
2. Divide the EA into C sectors (strata) of equal area and shape such that each sector contains
the same number of sample points. The number of sectors (C) should be equal to the number
of specimens in each composite (since one specimen per area will be used in each composite)
and the number of points within each sector, N, should equal the desired number of composite
samples.
3. Label the points within one sector in any arbitrary fashion from 1 to N. Use the same scheme
for each of the other sectors.
4. Form composite number 1 by compositing specimens with the '1' label, form composite number
2 by compositing specimens with the '2' label, etc. This leads to N composite samples that are
subjected to chemical analysis.
• 1
• 3
• 5
• 1
• 3
• 5
• 2
• 4
• 6
• 2
• 4
• 6
• 1
• 3
• 5
• 1
• 3
• 5
• 2
• 4
• 6
• 2
• 4
• 6
Figure 6. Systematic (square grid points) sample with systematic compositing scheme
(6 composite samples consisting of 4 specimens).
4-12
-------
Table 4.3. Probability of Decision Error at 0.5 SSL and 2 SSL Using Max Test
Sample
Size"
4
5
6
7
8
9
4
5
6
7
8
9
CV=1.0a
EO/
F d
C2.0
CV=1.5
E0.5
E2.0
CV = 2.0
E0.5
E2.0
CV = 2.5
Eo.5
E2.0
CV = 3.0
Eo.5
E2.0
CV = 3.5
E0.5
E2.0
CV = 4.0
E0.5
E2.o
C = 4 specimens per composite6
<.01
<.01
<.01
<.01
<.01
<.01
0.08
0.05
0.03
0.01
0.01
0.01
0.02
0.02
0.02
0.03
0.03
0.05
0.11
0.06
0.04
0.02
0.01
0.01
0.09
0.11
0.11
0.12
0.16
0.16
0.13
0.10
0.06
0.04
0.02
0.01
0.14
0.15
0.21
0.25
0.25
0.28
0.19
0.10
0.08
0.05
0.04
0.03
0.19
0.26
0.28
0.31
0.36
0.36
0.20
0.17
0.11
0.08
0.05
0.04
0.24
0.26
0.31
0.36
0.42
0.44
0.26
0.18
0.11
0.09
0.07
0.07
0.25
0.31
0.35
0.41
0.41
0.48
0.30
0.25
0.16
0.15
0.09
0.08
C = 6 specimens per composite
<.01
<.01
<.01
<.01
<.01
<.01
0.08
0.05
0.03
0.01
0.01
0.01
<.01
<.01
0.01
0.01
0.01
0.01
0.11
0.08
0.04
0.02
0.01
0.01
0.03
0.04
0.06
0.06
0.06
0.06
0.12
0.09
0.04
0.02
0.02
0.01
0.08
0.11
0.14
0.14
0.15
0.18
0.16
0.08
0.06
0.04
0.02
0.02
0.15
0.17
0.19
0.23
0.25
0.28
0.17
0.13
0.09
0.06
0.03
0.03
0.26
0.22
0.25
0.29
0.30
0.34
0.20
0.15
0.09
0.08
0.04
0.03
0.23
0.25
0.29
0.37
0.40
0.39
0.27
0.20
0.12
0.08
0.08
0.04
a The CV is the coefficient of variation for individual, uncomposited measurements across the entire EA, including measurement error
b Sample size (N) = number of composite samples
c_ 05 = Probability of requiring further investigation when the EA mean is 0.5 SSL
d E 2.0 = Probability of requiring further investigation when the EA mean is 2.0 SSL
e C = number of specimens per composite sample, where each composite consists of points from a stratified random or systematic grid sample from across the entire
EA
NOTE: All decision error rates are based on 1,000 simulations that assume each composite is representative of the entire EA, that half the EA has concentrations
below the quantitation limit (i.e., SSL/100), and half the EA has concentrations that follow a gamma distribution (a conservative distributional assumption)
4-13
-------
Highlight 4: Procedure for Compositing of Specimens from a Grid Sample Using a
Random Scheme (Figure 7)
1. Lay out a square or triangular grid sample over the EA, using a random start. Step-by-step
procedures can be found in Chapters of U.S. EPA (1989a). The number of points in the
grid should be equal to C x N, where C is the desired number of specimens per composite
and N is the desired number of composites.
2. Divide the EA into C sectors (strata) of equal area and shape such that each sector contains
the same number of sample points. The number of sectors (C) should be equal to the
number of specimens in each composite (since one specimen per area will be used in each
composite) and the number of points within each sector, N, should equal the desired
number of composite samples.
3. Use a random number table or random number generator to establish a set of labels for the
N points within each sector. This is done by first labeling the points in a sector in an arbitrary
fashion (say, points A, B, C,...) and associating the first random number with point A, the
second with point B, etc. Then rank the points in the sector according to the set of random
numbers and relabel each point with its rank. Repeat this process for each sector.
4. Form composite number 1 by compositing specimens with the <1' label, form composite
number 2 by compositing specimens with the <2' label, etc. This leads to N composite
samples that are subjected to chemical analysis.
• 3 92
91 94
• 5 96
• 6 95
91 94
93 92
91 95
96 93
92 94
94 96
93 95
92 91
Figure 7. Systematic (square grid points) sample with random compositing scheme
(6 composite samples consisting of 4 specimens).
4-14
-------
Highlight 5: Procedure for Compositing of Specimens from a Stratified Random Sample
Using a Random Scheme (Figure 8)
1. Divide the EA into C sectors (strata) of equal area, where C is equal to the number of
specimens to be in each composite (since one specimen per stratum will be used in each
composite).
2. Within each stratum, choose N random locations, where N is the desired number of
composites. Step-by-step procedures for choosing random locations can be found in Chapter
5 of U.S. EPA(1989a).
3. Use a random number table or random number generator to establish a set of labels for the N
points within each sector. This is done by first labeling the points in a sector in an arbitrary
fashion (say, points A, B, C,...) and associating the first random number with point A, the
second with point B, etc. Then rank the points in the sector according to the set of random
numbers and relabel each point with its rank. Repeat this process for each sector.
4. Form composite number 1 by compositing specimens with the "I' label, form composite number
2 by compositing specimens with the '2' label, etc. This leads to N composite samples that are
subjected to chemical analysis.
Figure 8. Stratified random sample with random compositing scheme
(6 composite samples consisting of 4 specimens).
4-15
-------
The error rates shown in Table 4.3 are based on the simulations presented in Appendix I. These simulations are based
on the following assumptions:
1. Each of the N composite samples is based on C specimens selected to be representative of the EA
as a whole, as specified above (C = number of sectors or strata).
2. One-half the EA has concentrations below the quantitation limit (which is assumed to be SSL/100).
3. One-half the EA has concentrations that follow a gamma distribution (see Section 4.3 for additional
discussion).
4. Each chemical analysis is subject to a 20 percent measurement error.
The error rates presented in Table 4.3 are based on the above assumptions which make them robust for most potential
distributions of soil radionuclide concentrations. Distribution assumptions 2 and 3 were used because they were
found in the simulations to produce high error rates relative to other potential radionuclide distributions (see Section
3.3). If the proportion of the site below the quantitation limit (QL) is less than half or if the distribution of the
concentration measurements is some other distribution skewed to the right (e.g., lognormal), rather than gamma, then
the error rates achieved are likely to be no worse than those cited in Table 4.3. Although the actual radionuclide
distribution may be different from those cited above as the basis for Table 4.3, only extensive investigations will
usually generate sufficient data to determine the actual distribution for each EA.
Using Table 4.3 to determine the sample size pair (C and N) needed to achieve satisfactory error rates with the Max
test requires an a priori estimate of the coefficient of variation for measurements of the radionuclide of interest across
the EA. The coefficient of variation (CV) is the ratio of the standard deviation of radionuclide concentrations for
individual, uncomposited specimens divided by the EA mean concentration. As discussed in Section 4.1.4, the EAs
should be constructed within strata expected to have relatively homogeneous concentrations so that an estimate of
the CV for a stratum may be applicable for all EAs in that stratum. The site manager should use a conservatively
large estimate of the CV for determining sample size requirements because additional sampling will be needed if the
data suggest that the true CV is greater than that used to determine the sample sizes.
Potential sources of information for estimating the EA or stratum means, variances, and CVs include the following
(in descending order of desirability):
• Data from a pilot study conducted at the site
• Prior sampling data from the site
• Data from similar sites
• Professional judgment.
For more information on estimating variability, see Section 4.3.1 of U.S. EPA (1989a).
4.1.8 Using the DQA Process: Analyzing Max Test Data. This section provides guidance for
analyzing the data for the Max test.
The hypothesis test for the Max test is very simple to implement, which is one reason that the Max test is attractive
as a surface soil screening test. If x1; x2,..., XN represent concentration measurements for N composite samples that
each consist of C specimens selected so that each composite is representative of the EA as a whole (as described in
Section 4.1.7), the Max test is implemented as follows:
If Max (xls x2, ..., XN) >_ 2 SSL, then investigate the EA further;
4-16
-------
If Max (xls x2, ..., XN) < 2 SSL, and the data quality assessment (DQA) indicates that the sample size was
adequate, then no further investigation is necessary.
In addition, the step-by-step procedures presented in Highlight 6 must be implemented to ensure that the site-specific
error limits, as discussed in Section 4 .1.6, are achieved.
If the EA mean is below 2 SSL, the DQA process may be used to determine if the sample size was sufficiently large
to justify the decision to not investigate further. To use Table 4.3 to check whether the sample size is adequate, an
estimate of the CV is needed for each EA. The first four steps of Highlight 6, the DQA process for the Max test,
present a process for the computation of a sample CV for an EA based on the N composite samples that each consist
of C specimens.
However, the sample CV can be quite large when all the measurements are very small (e.g., well below the SSL)
because CV approaches infinity as the EA sample mean (x) approaches zero. Thus, when the composite
concentration values for an EA are all near zero, the sample CV may be questionable and therefore unreliable for
determining if the original sample size was sufficient (i.e., it could lead to further sampling when the EA mean is well
below 2 SSL). To protect against unnecessary additional sampling in such cases, compare all composites against the
equation given in Step 5 of Highlight 6. If the maximum composite sample concentration is below the value given
by the equation, then the sample size may be assumed to be adequate and no further DQA is necessary.
To develop Step 5, EPA decided that if there were no compositing (C=l) and all the observations (based on a sample
size appropriate for a CV of 2.5) were less than the SSL, then one can reasonably assume that the EA mean was not
greater than 2 SSL. Likewise, because the standard error for the mean of C specimens, as represented by the
composite sample, is proportional to 1/^/C, the comparable condition for composite observations is that one can
reasonably assume that the EA mean was not greater than 2 SSL when all composite observations were less than
S SL/ Jc • If this is the case for an EA sample set, the sample size can be assumed to be adequate and no further DQA
is needed. Otherwise (when at lease one composite observation is not this small), use Table 4.3 with the sample CV
for the EA to determine whether a sufficient number of samples were taken to achieve DQOs.
In addition to being simple to implement, the Max test is recommended because it provides good control over the
Type I error rates at 2 SSL with small sample sizes. It also does not need any assumptions regarding observations
below the QL. Moreover, the Max test error rates at 2 SSL are fairly robust against alternative assumptions regarding
the distribution of surface soil concentrations in the EA. The simulations in Appendix I show that these error rates
are rather stable for lognormal or Weibull radionuclide concentration distributions and for different assumptions
about portions of the site with radionuclide concentrations below the MDC.
4-17
-------
Highlight 6: Directions for Data Quality Assessment for the Max Test
Let X.,, x2, ..., XN represent contaminant concentration measurements for N composite samples that each
consist of C specimens selected so that each composite is representative of the EA as a whole. The
following describes the steps required to ensure that the Max test achieves the DQOs established for the
site.
STEP 1: The site manager determines the Type I error rate to be achieved at 2 SSL and the Type II error
rate to be achieved at 0.5 SSL, as described in Section 4.1.6.
1 N
STEP 2: Calculate the sample mean x = — J^ *,
N 2 = 1
STEP 3: Calculate the sample standard deviation
5 = N
STEP 4: Calculate the sample estimate of the coefficient of variation,CV, for individual concentration
measurements from across the EA.
NOTE: This is a conservative approximation of the CV for individual assessments.
SSL
STEP 5: If Max (x1, x2, ..., XN) < , then no further data quality assessment is needed and the EA needs
JC
no further investigation.
Otherwise proceed to Step 6.
STEP 6: Use the value of the sample CV calculated in Step 4 as the true CV of concentrations to
determine which column of Table 4.3 is applicable for determining sample size requirements.
Using the error limits established in Step 1, determine the sample size requirements from this
table. If the required sample size is greaterthan that implemented, further investigation of the EA
is necessary. The further investigation may consist of selecting a supplemental sample and
repeating the Max test with the larger, combined sample.
A limitation of the Max test is that it does not provide as good control over the Type II error rates at 0.5 SSL as it
does for Type I error rates at 2 SSL. In fact, for a fixed number, C, of specimens per composite, the Type II error rate
increases as the number of composite samples, N, increases. As the sample size increases, the likelihood of observing
an unusual sample with the maximum exceeding 2 SSL increases. However, the Type II error rate can be decreased
by increasing the number of specimens per composite. This unusual performance of the Max test as a hypothesis
testing procedure occurs because the rejection region is fixed below 2 SSL and thus does not depend on the sample
size (as it does for typical hypothesis testing procedures).
4.1.9 Specify Limits on Decision Errors for Sign Test. Although the Max test is adequate and
appropriate for selecting a sample size for site screening, there are other alternate methods of screening surface soils.
One such alternate method is the Sign test. While the Sign test is a more complex statistical method than the Max
test, it is based on the same statistical null hypothesis (i.e., the EA requires further investigation). Some EAs that
cannot be screened out with the Max test could be screened out with the Sign test since it uses a less conservative
estimate of the mean concentration than does the Max test.
4-18
-------
A disadvantage of the Max test is its performance when the true EA mean is between 0.5 SSL and the SSL. In this
case, as the sample size increases, the test indicates the decision to investigate further, even though the mean is less
than the SSL. In fact, no test procedure with feasible sample sizes performs well when the true EA mean is in the
"gray region" between 0.5 SSL and 2 SSL (see Section 4.3).
However, improved performance may be achieved in this region through the use of nonparametric statistical tests.
The basic difference between parametric and nonparametric statistical techniques is that parametric techniques, such
as the Max test, use specific assumptions about the probability distributions of the measurement data. Non parametric
techniques (sometimes referred to as distribution-free statistical methods) can be used without assuming a particular
underlying distribution. Thus, nonparametric techniques are often more appropriate in situations when the probability
distribution of the data is either unknown or is some continuous distribution other than the normal distribution.
For radiological measurements at levels associated with SSLs, the data are likely to approach the MDC for typical
radioanalytical methods. There may, therefore, be some measurement data which are at or below MDC, and the
variability of the measurements will be relatively large. Such data are not easily treated using parametric methods,
thus offering an additional advantage to the use of nonparametric techniques.
The Sign test is a nonparametric statistical test recommended for use as an alternative to the Max test. As is the case
for many nonparametric tests, it is based on ranking the measurement data. The data are ordered from smallest to
largest, and assigned numbers (ranks) 1,2, 3,... accordingly. The analysis is then performed on the ranks rather than
on the original measurement values. The advantage of this approach is that the probability that one measurement is
larger (i.e., ranked higher) than another can be computed exactly by combinatorial methods without reference to a
specific probability distribution.
The Soil Screening Guidance for chemicals proposes the use of the Chen test as an alternative to the Max test. This
requires the use of a different null hypothesis and different decision error rates than those for the Max test. An
additional advantage of using the Sign test as the alternative to the Max test is that the same null hypothesis and
decision error rates are used for both tests. Therefore, the null hypothesis for the Sign test is the same as given in
Section 4.1.6, namely that the site needs further investigation. The Type I and Type II decision error rates are the
same as those given in Section 4.1.6:
• Prob ("walk away" when the true EA mean is 2 SSL) = 0.05
• Prob ("investigate further" when the true EA mean is 0.05 SSL) = 0.20.
The DQO Process for the Sign test is therefore the same as that given for the Max test and is shown in Table 4.2.
4.1.10 Optimize the Design Using the Sign Test. This section includes guidance on developing an
optimum sampling strategy for screening surface soils. It discusses compositing, the selection of sampling points
for composited and uncomposited surface soil sampling, and the recommended procedures for determining the sample
sizes necessary to achieve specified limits on decision errors using the Sign test.
The general guidance provided in Section 4.1.7 for optimizing the design for the Max test is also applicable for the
Sign test.
Compositing. The guidance provided in Section 4.1.7 for compositing samples for the Max test is also applicable
for the Sign test.
Sample Pattern. The Sign test can be applied using composite samples that are representative of the entire EA
or with individual uncomposited samples.
4-19
-------
The guidance provided in Section 4.1.7 concerning sample patterns and compositing schemes for the Max test are
also applicable to the Sign test.
Sample Size. This section provides procedures to determine sample size requirements for the Sign test that
achieve the site-specific decision error limits discussed in Section 4.1.6. Calculate the Relative Shift. The initial
step in determining the number of data points for the Sign test is to calculate a term known as the relative shift, A/os.
This is defined as the ratio of the width of the gray region, A, and the standard deviation of the radionuclide in the
EA, os. As specified in Section 4.1.6, the gray region ranges from 0.5 SSL to 2.0 SSL, therefore the width of the gray
region is 1.5 SSL. Also as described in Section 4.1.7, the value of os may be obtained from data from a pilot study
conducted at the site, prior sampling data from the site, data from similar sites, or professional judgement. Values
of the relative shift that are less than one will result in a large number of measurements needed to achieve the site-
specific error limits discussed in Section 4.1.6.
The importance of choosing an appropriate value for os must be emphasized. If the value is grossly underestimated,
the number of data points will be too few to obtain the desired power level for the test and a resurvey may be
recommended. If, on the other hand, the value is overestimated, the number of data points determined will be
unnecessarily large.
Determine Sign p. Sign p is the estimated probability that a random measurement from the EA will be less than
the upper bound of the gray region (i.e., 2.0 SSL) when the EA median is actually at the lower bound of the gray
region (i.e., 0.5 SSL). The Sign p is used to calculate the minimum number of data points necessary for the survey
to meet the DQOs. The value of the relative shift calculated in the previous section is used to obtain the
corresponding value of Sign p from Table 4.4.
Table 4.4. Values of Sign p for Given Values of the Relative Shift, A/a
A/a
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
1.1
Sign p
0.539828
0.579260
0.617911
0.655422
0.691462
0.725747
0.758036
0.788145
0.815940
0.841345
0.864334
A/a
1.2
1.3
1.4
1.5
1.6
1.7
1.8
1.9
2.0
2.5
3.0
Sign p
0.884930
0.903199
0.919243
0.933193
0.945201
0.955435
0.964070
0.971284
0.977250
0.993790
0.998650
If A/o > 3.0, use Signp = 1.000000
4-20
-------
Determine Decision Error Percentiles. The next step in this process is to determine the percentiles, Z\_a and
Z!_B, represented by the selected decision error levels, a and B, respectively (see Table 4.5).
Table 4.5. Percentiles Represented by Selected Values of a and li
a (or B)
0.005
0.01
0.015
0.025
0.05
Zi-aforZ^)
2.576
2.326
2.241
1.960
1.645
a (or p)
0.10
0.15
0.20
0.25
0.30
Zi-aforZ^)
1.282
1.036
0.842
0.674
0.524
Calculate Number of Data Points for Sign Test. The number of data points, N, to be obtained for the Sign
test is next calculated using the following formula:
N =
4 (Sign P - 0.5)2
(36)
The value of N calculated using this equation is an approximation based on estimates of o, so there is some
uncertainty associated with this calculation. In addition, there will be some missing or unusable data from any
sampling and analysis program. The rate of missing or unusable measurements expected to occur in an EA and the
uncertainty associated with the calculation of N should be accounted for during the sampling and analysis program
planning stage. The number of data points should be increased by 20%, and rounded up, over the values calculated
using the above equation to obtain sufficient data points to attain the desired power level with the statistical tests and
allow for possible lost or unusable data. The value of 20% is selected to account for a reasonable amount of
uncertainty in the parameters used to calculate N and still allow flexibility to account for some lost or unusable data.
The recommended 20% correction factor should be applied as a minimum value. Experience and site-specific
considerations should be used to increase the correction factor if required. If the user determines that the 20%
increase in the number of measurements is excessive for a specific site, a retrospective power curve should be used
to demonstrate that the survey design provides adequate power to support the decision.
Obtain Number of Data Points for Sign Test from Table 4.6. Table 4.6 provides a list of the number of data
points necessary to achieve the desired Type I error rate of 0.05 and Type II error rate of 0.20 as a function of the
relative shift, A/a. The values listed in Table 4.6 represent the number of measurements to be performed in each EA.
These values were calculated using above equation for N and increased by 20% to account for missing or unusable
data and uncertainty in the calculated value of N.
4-21
-------
Table 4.6. Values of N for a Given Relative Shift (A/a)
for a = 0.05 and p = 0.20
A/a
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
1.2
1.4
1.6
1.8
2.0
2.5
3.0
N
1170
296
134
77
51
37
28
23
19
16
13
11
10
9
9
8
8
Given an a priori estimate of the standard deviation, os of concentration measurements in the EA, the site manager
can use Table 4.6 to determine a sample size option that achieves the decision error goals for surface soil screening
presented in Section 4.1.6 (i.e., not more than 20 percent chance of error at 0.5 SSL and not more than 5 percent at
2 SSL). For example, suppose that the site manager expects that the ratio of the width of the gray region and the
maximum true os for concentration measurements in an EA is 2. Then Table 4.6 shows that a total of nine samples
will be sufficient to achieve the decision error limit goals.
4.1.11 Using the DQA Process: Analyzing Sign Test Data. Data Quality Assessment (DQA) is
a scientific and statistical evaluation that determines if the data are of the right type, quality, and quantity to support
their intended use. Prior to conducting the Sign test on the measurement data, a preliminary data review should be
conducted.
Preliminary Data Review. The purpose of this phase of the DQA process is to conduct a preliminary evaluation
of the data set to determine that the data are consistent with the underlying assumptions upon which the statistical
test is based. In order to learn about the structure of the data—identifying patterns, relationships, or potential
anomalies—basic statistical quantities (i.e., mean, standard deviation, median) should be calculated and the data
should be plotted (i.e., a frequency plot and a posting plot). In addition, a retrospective power curve should be
constructed to examine if the DQOs have been achieved.
Basic Statistical Quantities
• The mean of the data. The mean of the data for the EA can be compared to 2 times the SSL to get a preliminary
indication of the EA status. If the site is indeed contaminated, this comparison may readily reveal that a EA
contains excess residual radioactivity—even before applying statistical tests. For example, if the EA average
4-22
-------
concentration exceeds 2 times the SSL, then the EA clearly does not meet the screening criteria. On the other
hand, if the maximum EA measurement is less than 2 times the SSL, the EA meets the screening criteria.
Likewise, if every measurement in the EA is less than 2 times the SSL, the EA meets the screening criteria and
no statistical tests are required.
• The sample standard deviation. The value of the sample standard deviation is especially important. If too large
compared to that assumed during the survey design, this may indicate an insufficient number of samples were
collected to achieve the desired power of the statistical test. Again, inadequate power can lead to unnecessary
additional surveys.
• The median of the data. The median is the middle value of the data set when the number of data points is odd,
and is the average of the two middle values when the number of data points is even. Thus, 50% of the data points
are above the median, and 50% are below the median. Large differences between the mean and the median
would be an early indication of skewness in the data. This would also be evident in a histogram of the data. For
example, if the difference between the median and the mean is a small fraction of the sample standard deviation,
the mean and median would not be considered significantly different.
• Minimum, maximum, and range of data. Examining the minimum, maximum, and range of the data may provide
additional useful information. If the difference between the minimum and the maximum is a small number of
standard deviations, the overall range is not unusually large. When there are 30 or fewer data points, values of
the range much larger than about 4 to 5 standard deviations would be unusual. For larger data sets the range
might be wider.
Graphical Data Review
• Posing plot. Spatial dependencies that potentially affect the assumptions can be assessed using posting plots.
A posting plot is simply a map of the EA with the data values entered at the measurement locations. This
potentially reveals heterogeneities in the data—especially possible small areas of elevated activity.
For example, a plot may reveal an obvious trend towards larger values as one moves from left to right across the
EA. This trend may not be apparent from a simple listing of the data. The trend may become even more apparent
if isopleths are added to the posting plot. If the posting plot reveals systematic spatial trends in the EA, the cause
of the trends would need to be investigated. More sophisticated tools for determining the extent of spatial
dependencies are also available, such as the Guidance for Data Quality Assessment: Practical Methods for Data
Analysis, EPA QA/G-9 (e.g., U.S. EPA, 1996c). These methods tend to be complex and are best used with
guidance from a professional statistician.
• Frequency plot. A frequency plot (or a histogram) is a useful tool for examining the general shape of a data
distribution. This plot is a bar chart of the number of data points within a certain range of values. A simple
method for generating a rough frequency plot is the stem and leaf display. The frequency plot will reveal any
obvious departures from symmetry, such as skewness or bimodality (two peaks), in the data distributions for the
EA. Skewness or other asymmetry can impact the accuracy of the statistical tests. A data transformation (e.g.,
taking the logarithms of the data) can sometimes be used to make the distribution more symmetric. The statistical
tests would then be performed on the transformed data. When the underlying data distribution is highly skewed,
it is often because there are a few high areas.
The presence of two peaks in the EA frequency plot may indicate the existence of isolated areas of residual
radioactivity. It may also indicate a mixture of soil types, construction materials, etc. The greater variability in
the data due to the presence of such a mixture will reduce the power of the statistical tests to detect an adequately
remediated EA. These situations should be avoided whenever possible by carefully choosing EAs.
4-23
-------
• Quantile plot. A Quantile plot is constructed by first ranking the data from smallest to largest. Sorting the data
is easy once the stem and leaf display has been constructed. Then, each data value is simply plotted against the
percentage of the samples with that value or less. The slope of the curve in the Quantile plot is an indication of
the amount of data in a given range of values. A small amount of data in a range will result in a large slope. A
large amount of data in a range of values will result in a more horizontal slope. A sharp rise near the bottom or
the top is an indication of asymmetry. Sudden changes in slope, or notably steep areas may indicate peculiarities
in the data needing further investigation.
MARSSIM contains a detailed discussion of graphical data review methods. In addition, Guidance for Data Quality
Assessment: Practical Methods for Data Analysis, EPA QA/G-9 (e.g., U.S. EPA, 1996c) provides additional
information concerning interpretation of these methods.
Retrospective Power Curve
One of the more important assumptions made in the survey design described earlier is that the sample sizes
determined for the tests must be sufficient to achieve the data quality objectives set for the Type I (a) and Type II
(P) error rates. A retrospective power curve should be constructed using the actual number of measurements
obtained, N, and the measured value of o. The resulting power curve shows the probability that the EA could be
screened out using the Sign test versus the screening level. This curve can show if the DQOs have been achieved.
The consequence of inadequate power is that a EA that should be screened out has a higher probability of being
incorrectly deemed to require further investigation.
An effective sampling and analysis plan will slightly overestimate both the number of measurements and the standard
deviation to ensure adequate power. This insures that an EA is not subjected to additional remediation simply
because the measurements is not sensitive enough to detect that residual radioactivity is below the guideline level.
When the null hypothesis is rejected, the power of the test becomes a somewhat moot question. Nonetheless, even
in this case, a retrospective power curve can be a useful diagnostic tool and an aid to designing future surveys.
Analyze the Data Using the Sign Test. Step-by-step instructions forusingthe Sign test to analyze data from
both discrete random samples and pseudo-random samples (e.g., composite samples constructed as described
previously) are provided in Highlight 7.
4-24
-------
Highlight 7: Directions for the Sign Test
STEP 1: List the EA measurements, x.,, x2, ..., XN .
STEP 2: Subtract each measurement, Xj, from the upper bound of the gray area, 2 SSL to obtain the
differences:
D, =2 SSL - x, /=1,2, ...,/V.
STEP 3: Discard each difference that is exactly zero and reduce the sample size, N, by the number of
such zero measurements.
STEP 4: Count the number of positive differences. The result is the Sign test statistic S+. Note that a
positive difference corresponds to a measurement below 2 SSL and contributes evidence that
the EAdoes not require further investigation. The larger the number of positive differences, the
more likely that the null hypothesis may be rejected.
STEP 5: The value of S+ is compared to the critical value for the Sign test, ks!gn, given in Table 4.7,
corresponding to the desired Type I error rate, a, and the number of measurements N. If S+ is
greater than the critical value, ks!gn, in Table 5.7, the null hypothesis is rejected and the EA does
not require further investigation.
Table 4.7. Critical Values ksign for the Sign Test Statistic S+
N
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Type I error rate, a
0.005
4
5
6
7
7
8
9
10
10
11
12
12
13
14
14
15
16
16
0.01
4
5
6
6
7
8
9
9
10
11
11
12
13
13
14
14
15
16
0.025
4
5
5
6
7
7
8
9
9
10
11
11
12
12
13
14
14
15
0.05
4
4
5
6
6
7
8
8
9
9
10
11
11
12
12
13
14
14
0.1
3
4
5
5
6
6
7
8
8
9
9
10
11
11
12
12
13
13
0.2
3
3
4
5
5
6
6
7
7
8
9
9
10
10
11
11
12
12
0.3
3
3
4
4
5
5
6
6
7
7
8
9
9
10
10
11
11
12
0.4
2
3
3
4
4
5
5
6
6
7
7
8
9
9
10
10
11
11
0.5
2
2
3
3
4
4
5
5
6
6
7
7
8
8
9
9
10
10
4-25
-------
Table 4.7. Critical Values ksi for the Sign Test Statistic S+
N
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
Type I error rate, a
0.005
17
18
18
19
19
20
21
21
22
23
23
24
24
25
26
26
27
27
28
29
29
30
30
31
32
32
33
33
34
0.01
16
17
18
18
19
19
20
21
21
22
23
23
24
24
25
26
26
27
27
28
28
29
30
30
31
31
32
33
33
0.025
16
16
17
17
18
19
19
20
20
21
22
22
23
23
24
24
25
26
26
27
27
28
28
29
30
30
31
31
32
0.05
15
15
16
17
17
18
18
19
19
20
21
21
22
22
23
23
24
25
25
26
26
27
27
28
29
29
30
30
31
0.1
14
15
15
16
16
17
17
18
19
19
20
20
21
21
22
22
23
23
24
25
25
26
26
27
27
28
28
29
30
0.2
13
14
14
15
15
16
16
17
17
18
18
19
19
20
21
21
22
22
23
23
24
24
25
25
26
26
27
27
28
0.3
12
13
13
14
14
15
15
16
16
17
17
18
19
19
20
20
21
21
22
22
23
23
24
24
25
25
26
26
27
0.4
12
12
13
13
14
14
15
15
16
16
17
17
18
18
19
19
20
20
21
21
22
22
23
23
24
24
25
25
26
0.5
11
11
12
12
13
13
14
14
15
15
16
16
17
17
18
18
19
19
20
20
21
21
22
22
23
23
24
24
25
4-26
-------
A retrospective power analysis for the test will often be useful, especially when the null hypothesis is not rejected
(see Appendix 1.9). When the null hypothesis is not rejected, it may be because it is in fact true, or it may be because
the test did not have sufficient power to detect that it is not true. The power of the test will be primarily affected by
changes in the actual number of measurements obtained and their standard deviation. A retrospective power curve
should be constructed using the actual number of measurements obtained, N, and the measured value of o. The
resulting power curve shows the probability that the EA could be screened out using the Sign test. This curve can
show if the DQOs have been achieved. The consequence of inadequate power is that an EA that should be screened
out has a higher probability of being incorrectly deemed to require further investigation.
An effective sampling and analysis plan will slightly overestimate both the number of measurements and the standard
deviation to ensure adequate power. This insures that an EA is not subjected to additional remediation simply
because the measurements is not sensitive enough to detect that residual radioactivity is below the guideline level.
When the null hypothesis is rejected, the power of the test becomes a somewhat moot question. Nonetheless, even
in this case, a retrospective power curve can be a useful diagnostic tool and an aid to designing future surveys.
4.1.12 Special Considerations for Multiple Radionuclides. If the surface soil samples collected for
an EA will be tested for multiple radionuclides , be aware that the expected CVs or o for the different radionuclides
may not all be identical. A conservative approach is to base the sample sizes for all radionuclides on the largest
expected CV or o.
4.1.13 Quality Assurance/Quality Control Requirements. Regardless of the sampling approach used,
the Superfund quality assurance program guidance should be followed to ensure that measurement error rates are
documented and within acceptable limits (U.S. EPA, 1993d).
4.1.14 Final Analysis. After either the Max test or the Sign test has been performed for each EA of interest (0.5
acre or less) at an NPL site, the pattern of decisions for individual EAs (to "walk away" or to "investigate further")
should be examined. If some EAs for which the decision was to "walk away" are surrounded by EAs for which the
decision was to "investigate further," it may be more efficient to identify an area including all these EAs for further
study and develop a global investigation strategy.
4.1.15 Reporting. The decision process for surface soil screening should be thoroughly documented as part of
the RI/FS process. This documentation should include a map of the site (showing the boundaries of the EAs and the
sectors, or strata, within EAs that were used to select sampling points within the EAs); documentation of how
composite samples were formed and the number of composite samples that were analyzed for each EA; the raw
analytical data; the results of all hypothesis tests; and the results of all QA/QC analyses.
4.2 Sampling Subsurface Soils
Subsurface soil sampling is conducted to estimate the mean concentrations of radionuclides in each source at a site
for comparison to inhalation and migration to ground water SSLs. Measurements of soil properties and estimates
of the area and depth of contamination in each source are also needed to calculate SSLs for these pathways. Table
4.8 shows the steps in the DQO process necessary to develop a sampling strategy to meet these objectives. Each of
these steps is described below.
4-27
-------
Table 4.8. Soil Screening DQOs for Subsurface Soils
DQO Process Steps
Soil Screening Inputs/Outputs
State the Problem
Identify scoping team
Develop conceptual site model (CSM)
Define exposure scenarios
Specify available resources
Write brief summary of contamination
problem
Site manager and technical experts (e.g., health physicists, risk assessors,
hydrogeologists, statisticians).
CSM development (described in Step 1 of the User's Guide, U.S. EPA, 1999).
migration of radionuclides from soil to potable ground water.
Sampling and analysis budget, scheduling constraints, and available
personnel.
Summary of the subsurface soil contamination problem to be investigated at
the site.
Identify the Decision
Identify decision
Identify alternative actions
Do mean soil concentrations for particular radionuclides (e.g., radionuclides
of potential concern) exceed appropriate SSLs?
Eliminate area from further action or study under CERCLA
or
Plan and conduct further investigation.
Identify Inputs to the Decision
Identify decision
Define basis for screening
Identify analytical methods
migration to ground water SSLs for specified radionuclides
Measurements of subsurface soil radionuclide concentration
Soil Screening Guidance for Radionuclides
Feasible analytical methods (both field and laboratory) consistent with
program-level requirements.
Specify the Study Boundaries
Define geographic areas of field
investigation
Define population of interest
Define scale of decision making
Subdivide site into decision units
Define temporal boundaries of study
Identify (list) practical constraints
The entire NPL site (which may include areas beyond facility boundaries),
except for any areas with clear evidence that no contamination has
occurred.
Subsurface soils
Sources (areas of contiguous soil contamination, defined by the area and
depth of contamination or to the water table, whichever is more shallow).
Individual sources delineated (area and depth) using existing information or
field measurements (several nearby sources may be combined into a single
source).
Temporal constraints on scheduling field visits.
Potential impediments to sample collection, such as access, health, and
safety issues.
Develop a Decision Rule
Specify parameter of interest
Specify screening level
Specify ?if..., then..." decision rule
Mean soil radionuclide concentration in a source (as represented by discrete
radionuclide concentrations averaged within soil borings).
SSLs calculated using available parameters and site data (or generic SSLs if
site data are unavailable).
If the mean soil concentration exceeds the SSL, then investigate the source
further. If the mean soil boring concentration is less than the SSL, then no
further investigation is required under CERCLA.
4-28
-------
Table 4.8. Soil Screening DQOs for Subsurface Soils
DQO Process Steps
^^•^^H
Specify Limits on Decision Errors
Define QA/QC goals
Soil Screening Inputs/Outputs
Radioanalytical laboratory precision and bias requirements
10% laboratory analyses for field methods
Optimize the Design
Determine how to estimate mean
concentration in a source
Define subsurface sampling strategy
by evaluating costs and site-specific
conditions
Develop planning documents for the
field investigation
For each source, the highest mean soil core concentration (i.e., depth-
weighted average of discrete radionuclide concentrations within a boring).
Number of soil borings per source area; number of sampling intervals with
depth.
Sampling and Analysis Plan (SAP)
Quality Assurance Project Plan (QAPJP)
4.2.1 State the Problem. Radionuclides present in subsurface soils at the site may pose significant risk to
human health and the environment by the migration of radionuclides through soils to an underlying potable aquifer.
The problem is to identify the radionuclides and source areas that do not pose significant risk to human health so that
future investigations may be focused on areas and radionuclides of true concern.
Site-specific activities in this step include identifying the data collection planning team (including technical experts
and key stakeholders) and specifying the available resources (i.e., the cost and time available for sampling). The list
of technical experts and stakeholders should contain all key personnel who are involved with applying SSLs to the
site. Other activities include developing the conceptual site model and identifying exposure scenarios, which are fully
addressed in the Soil Screening Guidance for Radionuclides: User's Guide (U.S. EPA, 2000).
4.2.2 Identify the Decision. The decision is to determine whether mean soil concentrations in each source
area exceed migration to ground water SSLs for specific radionuclides. If so, the source area will be investigated
further. If not, no further action will be taken under CERCLA.
4.2.3 Identify Inputs to the Decision. Site-specific inputs to the decision include the average radionuclide
concentrations within each source area and the inhalation and migration ground water SSLs. Calculation of the SSLs
for the two pathways of concern also requires site-specific measurements of soil properties (i.e., bulk density, pH,
and soil texture class) and estimates of the areal extent and depth of contamination.
A list of feasible sampling and analytical methods should be assembled during this step. EPA recommends the use
of field methods where applicable and appropriate. Verify that radioanalytical laboratory methods and field methods
for analyzing the samples exist and that the analytical method detection limits or field method detection limits are
appropriate for the site-specific or generic SSL. MARSSIM (U.S. EPA, 1997c) contains further information
concerning the collection and preparation of samples for analysis as well as a general discussion of radiological
laboratory sample preparation and analysis methods, and in evaluating the results of these analyses.
Field methods will be useful in defining the study boundaries (i.e., area and depth of contamination) during site
reconnaissance and during the sampling effort. For radioactive contaminants, core samples may also be and
monitored intact in the field to determine if layers of radioactivity are present. In addition, the use of a subsurface
sampling technique, which results in a borehole or soil face, may be logged using a gamma scintillation detector.
This enables scanning of the exposed soil surface to identify radioactive contamination within small fractions of hole
depth, thus facilitating the identification of the presence and depth distribution of subsurface radioactivity. This
4-29
-------
information may be used to direct further core sampling and laboratory analysis as warranted. EPA expects field
methods will become more prevalent and useful because the design and capabilities of field portable instrumentation
are rapidly evolving. Documents on standard operating procedures (SOPs) for field methods are available through
NTIS and should be referenced in soil screening documentation if these methods are used.
Soil parameters necessary for SSL calculation are soil texture and bulk density. Some of these parameters can be
measured in the field, others require laboratory measurement. Although laboratory measurements of these parameters
cannot be obtained under the Superfund Contract Laboratory Program, they are readily available from soil testing
laboratories across the country.
Note that the size, shape, and orientation of sampling volume (i.e., ?support") for heterogenous media have a
significant effect on reported measurement values. For instance, particle size has a varying affect on the transport
and fate of radionuclides in the environment and on the potential receptors. Comparison of data from methods that
are based on different supports can be difficult. Defining the sampling support is important in the early stages of site
characterization. This may be accomplished through the DQO process with existing knowledge of the site,
contamination, and identification of the exposure pathways that need to be characterized. Refer to Preparation of
Soil Sampling Protocols: Sampling Techniques and Strategies (U.S. EPA, 1992f) for more information about soil
sampling support.
Soil Texture. The soil texture class (e.g., loam, sand, silt loam) is necessary to estimate average soil moisture
conditions and to estimate infiltration rates. A soil's texture classification is determined from a particle size analysis
and the U.S. Department of Agriculture (USDA) soil textural triangle shown at the top of Figure 9. This
classification system is based on the USDA soil particle size classification at the bottom of Figure 9. The particle
size analysis method in Gee and Bauder (1986) can provide this particle size distribution also. Other particle size
analysis methods may be used as long as they provide the same particle size breakpoints for sand/silt (0.05 mm) and
silt/clay (0.002 mm). Field methods are an alternative for determining soil textural class; an example from Brady
(1990) is also presented in Figure 9.
Dry Bulk Density. Dry soil bulk density (pb) is used to calculate total soil porosity and can be determined for any
soil horizon by weighing athin-walled tube soil sample (e.g., Shelby tube) of known volume and subtracting the tube
weight to estimate field bulk density (ASTM D 293 7). A moisture content determination (ASTM 2216) is then made
on a subsample of the tube sample to adjust field bulk density to dry bulk density. The other methods (e.g., ASTM
D 1556, D 2167, D 2922) are not generally applicable to subsurface soils. ASTM soil testing methods are readily
available in the Annual Book of ASTM Standards, Volume 4.08, Soil and Rock; Building Stones, which is available
from ASTM, 100 Barr Harbor Drive, West Conshohocken, PA, 19428.
pH. Soil pH is used to select site-specific partition coefficients for metals and ionizing organic compounds (see Part
4 ). This simple measurement is made with a pH meter in a soil/water slurry (McLean, 1982) and may be measured
in the field using a portable pH meter.
4-30
-------
Exhibit 9: U.S. Department of Agriculture soil texture classification
100
Percent Clay
Percent Silt
SKlyCIa/
L\a/ \ 70
100
100 90
80
70
60 50 40
Percent Sand
30
20
10
Criteria Used with the Field Method for Determining Soil Texture Classes
(Source: Brady, 1990)
Criterion
Sand
Sandy loam
Loam
Silt loam
Clay loam
Clay
1. Individual grains Yes Yes Some Few
visible to eye
2. Stability of dry Do not form Do not form Easily Moderately
clods
No
Hard and
broken easily broken stable
3. Stability of wet Unstable Slightly stable Moderately Stable
clods
stable
4. Stability of Does not Does not form Does
"ribbon" when form
wet soil rubbed
between thumb
and fingers
Particle Size,
Very stable
No
Very hard
and stable
Very stable
not formBroken appearance Thin, will break Very long,
flexible
mm
0.002 0.05 0.10 0.25 0.5 1.0
U.S.
Department C
of Agriculture
ay Silt
Very Fine Fine
Med. Coarse Very Co
Sand
2.0
arse
u rave I
Source: USDA.
4-31
-------
4.2.4 Define the Study Boundaries. As discussed in Section 4.1.4, areas that are known to be highly
contaminated (i.e., sources) are targeted for subsurface sampling. The information collected on source area and depth
is used to calculate site-specific SSLs for the inhalation and migration to ground water pathways. For the purposes
of this guidance, source areas are defined by area and depth as contiguous zones of contamination. However, discrete
sources that are near each other may be combined and investigated as a single source if site conditions warrant.
4.2.5 Develop a Decision Rule. The decision rule for subsurface soils is:
If the mean concentration of a radionuclide within a source area exceeds the screening level,
then investigate that area further.
In this case "screening level" means the SSL. As explained in Section 4.1.5, statistics other than the mean (e.g., the
maximum concentration) may be used as estimates of the mean in this comparison as long as they represent valid or
conservative estimates of the mean.
4.2.6 Specify Limits on Decision Errors. EPA recognizes that data obtained from sampling and analysis
can never be perfectly representative or accurate and that the costs of trying to achieve near-perfect results can
outweigh the benefits. Consequently, EPA acknowledges that uncertainty in data must be tolerated to some degree.
The DQO process attempts to control the degree to which uncertainty in data affects the outcomes of decisions that
are based on data.
The sampling intensity necessary to accurately determine the mean concentration of subsurface soil contamination
within a source with a specified level of confidence (e.g., 95 percent) is impracticable for screening due to excessive
costs and difficulties with implementation. Therefore, EPA has developed an alternative decision rule based on
average concentrations within individual soil cores taken in a source:
If the mean concentration within any soil core taken in a source exceeds the screening level,
then investigate that source further.
For each core, the mean core concentration is defined as the depth-weighted average concentration within the zone
of contamination (see Section 4.2.7). Since the soil cores are taken in the area(s) of highest contamination within
each source, the highest average core concentration among a set of core samples serves as a conservative estimate
of the mean source concentration. Because this rule is not a statistical decision, it is not possible to statistically define
limits on decision errors.
Standard limits on the precision and bias of sampling and analytical operations conducted during the sampling
program do apply. These are specified by the Superfund quality assurance program requirements (U.S. EPA, 1993d),
which must be followed during the subsurface sampling effort.
If field methods are used, at least 10 percent of field samples should be split and sent to a radioanalytical laboratory
for confirmatory analysis (U.S. EPA, 1993d).
Although the EPA does not require full laboratory sample tracking and quality assurance/quality control (QA/QC)
procedures for measurement of soil properties, routine EPA QA/QC procedures are recommended, including a
Quality Assurance Project Plan (QAPjP), chain-of-custody forms, and duplicate analyses.
4-32
-------
4.2.7 Optimize the Design. Within each source, the Soil Screening Guidance for Radionuclides suggests
taking two to three soil cores using split spoon or Shelby tube samplers. For each soil core, samples should begin
at the ground surface and continue at approximately 2-foot intervals until no contamination is encountered or to the
water table, whichever is shallower. Subsurface sampling depths and intervals can be adjusted at a site to
accommodate site-specific information on surface and subsurface radionuclide distributions and geological
conditions (e.g., very deep water table, very thick uncontaminated unsaturated zone, user well far beyond edge of
site, soils underlain by karst or fractured rock aquifers ).
The number and location of subsurface soil sampling (i.e., soil core) locations should be based on knowledge of likely
surface soil contamination patterns and subsurface conditions. This usually means that core samples should be taken
directly beneath areas of high surface soil contamination. Surface soils sampling efforts and field measurements (e.g.,
soil gas surveys) taken during site reconnaissance will provide information on source areas and high radionuclide
concentrations to help target subsurface sampling efforts. Information in the CSM also will provide information on
areas likely to have the highest levels of contamination. Note that there may be sources buried in subsurface soils
that are not discernible at the surface. Information on past practices at the site included in the CSM can help identify
such areas. Surface geophysical methods also can aid in identifying such areas (e.g., magnetometry to detect buried
drums).
The intensity of the subsurface soil sampling needed to implement the soil screening process typically will not be
sufficient to fully characterize the extent of subsurface contamination. In these cases, conservative assumptions
should be usedto develop hypotheses on likely radionuclide distributions (e.g., the assumption that soil contamination
extends to the water table). Along with knowledge of subsurface hydrogeology and stratigraphy, geostatistics can
be a useful tool in developing subsurface radionuclide distributions from limited data and can provide information
to help guide additional sampling efforts. However, instructions on the use of geostatistics is beyond the scope of
this guidance.
Samples for measuring soil parameters should be collected when taking samples for measuring radionuclide
concentrations. If possible, consider splitting single samples for radionuclide and soil parameter measurements. A
number of soil testing laboratories have provisions in place for handling and testing radioactive samples. However,
if testing radioactive samples is a problem, samples may be taken from clean areas of the site as long as they represent
the same soil texture and series and are taken from the same depth as the radionuclide concentration samples.
The SAP developed for subsurface soils should specify sampling and analytical procedures as well as the
development of QA/QC procedures. To identify the appropriate analytical procedures, the screening levels must be
known. If data are not available to calculate site-specific SSLs, then the generic SSLs in Appendix A should be used.
Finally, soil investigation for the migration to ground water pathway should not be conducted independently of
ground water investigations. Contaminated ground water may indicate the presence of a nearby source area, with
radionuclides leaching from soil into the aquifer.
4.2.8 Analyzing the Data. The mean soil radionuclide concentration for each soil core should be compared
to the SSL for the radionuclide. The soil core average should be obtained by averaging analyses results for the
discrete samples taken along the entire soil core within the zone of contamination (compositing will prevent the
evaluation of radionuclide concentration trends with depth).
If each subsurface soil core segment represents the same subsurface soil interval (e.g., 2 feet), then the average
concentration from the surface to the depth of contamination is the simple arithmetic average of the concentrations
measured for core samples representative of each of the 2-foot segments from the surface to the depth of
contamination or to the water table. However, if the intervals are not all of the same length (e.g., some are 2 feet while
4-33
-------
others are 1 foot or 6 inches), then the calculation of the average concentration in the total core must account for the
different lengths of the intervals.
If c; is the concentration measured in a core sample representative of a core interval of length li; and the n-th interval
is considered to be the last interval in the source area (i.e., the n-th sample represents the depth of contamination),
then the average concentration in the core from the surface to the depth of contamination should be calculated as the
following depth-weighted average ( r ),
n
Ylc
i—i i i
i = \
c= — (37)
If the leach test option is used, a sample representing the average radionuclide concentration within the zone of
contamination should be formed for each soil core by combining discrete samples into a composite sample for the
test. The composites should include only samples taken within the zone of contamination (i.e., clean soil below the
minimum detectable concentration should not be mixed with contaminated soil).
As with any Superfund sampling effort, all analytical data should be reviewed to ensure that Superfund quality
assurance program requirements are met (U.S. EPA, 1993d).
4.2.9 Reporting. The decision process for subsurface soil screening should be thoroughly documented. This
documentation should contain as a minimum: a map of the site showing the contaminated soil sources and any areas
assumed not to be contaminated, the soil core sampling points within each source, and the soil core sampling points
that were compared with the SSLs; the depth and area assumed for each source and their basis; the average soil
properties used to calculate SSLs for each source; a description of how samples were taken and (if applicable) how
composite samples were formed; the raw analytical data; the average soil core radionuclide concentrations compared
with the SSLs for each source; and the results of all QA/QC analyses.
4.3 Basis for the Surface Soil Sampling Strategies: Technical Analyses Performed
This section describes a series of technical analyses conducted to support the sampling strategy for surface soils
outlined in the Soil Screening Guidance for chemicals. Section 4.3.1 describes the sample design procedure presented
in the December 1994 draft guidance (U.S. EPA, 1994h). The remaining sections describe the technical analyses
conductedto develop the final SSL sampling strategy. Section 4.3.2 describes an alternative, nonparametric procedure
that EPA considered but rejected for the soil screening strategy.
Section 4.3.3 describes the simulations conducted to support the selection of the Max test and the Chen test in the
final Soil Screening Guidance for chemicals. These simulation results also can be used to determine sample sizes for
site conditions not adequately addressed by the tables in Section 4.1. Quantitation limit and multiple comparison
issues are discussed in Sections 4.3.4 and 4.3.5, respectively. Section 4.3.6 describes a limited investigation of
compositing samples within individual EA sectors or strata.
4-34
-------
4.3.1 1994 Draft Guidance Sampling Strategy. The DQO-based sampling strategy in the 1994 draft Soil
Screening Guidance assumed a lognormal distribution for contaminant levels over an EA and derived sample size
determinations from lognormal confidence interval procedures by C. E. Land (1971). This section summarizes the
rationale for this approach and technical issues raised by peer review.
For the 1994 draft Soil Screening Guidance, EPA based the surface soil SSL methodology on the comparison of the
arithmetic mean concentration over an EA with the SSL. As explained in Section 4.1, this approach reflects the type
of exposure to soil under a future residential land use scenario. A person moving randomly across a residential lot
would be expected to experience an average concentration of contaminants in soil.
Generally speaking, there are few nonparametric approaches to statistical inference about a mean unless a symmetric
distribution (e.g., normal) is assumed, in which case the mean and median are identical and inference about the
median is the same as inference about the mean. However, environmental contaminant concentration distributions
over a surface area tend to be skewed with a long right tail, so symmetry is not plausible. In this case the main options
for inference about means are inherently parametric, i.e., they are based on an assumed family of probability
distributions.
In addition to being skewed with a long right tail, environmental contaminant concentration data must be positive
because concentration measurements cannot be negative. Several standard two-parameter probability models are
nonnegative and skewed to the right, including the gamma, lognormal, and Weibull distributions. The properties of
these distributions are summarized in Chapter 12 of Gilbert (1987).
The lognormal distribution is the distribution most commonly used for environmental contaminant data (see, e.g.,
Gilbert, 1987, page 164). The lognormal family can be easy to work with in some respects, due to the work of Land
(1971,1975) on estimating confidence intervals for lognormal parameters, which are also described in Gilbert (1987).
The equation for estimating the Land upper confidence limit (UL) for a lognormal mean has the form
(38)
where y and sy are the average and standard deviation of the sample log concentrations. The lower confidence limit
(LL) has a similar form. The factor H depends on sy and n and is tabulated in Gilbert (1987) and Land (1975). If the
data truly follow a lognormal distribution, then the Land confidence limits are exact (i.e., the coverage probability
of a 95 percent confidence interval is 0.95).
The problem formulation used to develop SSL DQOs in the 1994 draft Soil Screening Guidance tested the null
hypothesis H0: ^ > 2 SSL versus the alternative hypothesis H^ (i < 2 SSL, with a Type I error rate of 0.05 (at 2 SSL),
and a Type II error rate of 0.20 at 0.5 SSL (^ represents the true EA mean). That is, the probability of incorrectly
deciding not to investigate further when the true mean is 2 SSL was set not to exceed 0.05, and the probability of
incorrectly deciding to investigate further when the true mean is 0.5 SSL was not to exceed 0.20.
This null hypothesis can be tested at the 5 percent level of significance by calculating Land's upper 95 percent
confidence limit for a lognormal mean, if one assumes that the true EA concentrations are lognormally distributed.
The null hypothesis is rejected if the upper confidence limit falls below 2 SSL.
Simulation studies of the Land procedure were used to obtain sample size estimates that achieve these DQOs for
different possible values of the standard deviation of log concentrations. Additional simulation studies were
4-35
-------
conducted to calculate sample sizes and to investigate the properties of the Land procedure in situations where
specimens are composited.
All of these simulation studies assumed a lognormal distribution of site concentrations. If the underlying site
distribution is lognormal, then the composites, viewed as physical averages, are not lognormal (although they may
be approximately lognormal). Hence, correction factors are necessary to apply the Land procedure with compositing,
if the individual specimen concentrations are assumed lognormal. The correction factors were also developed through
simulations. The correction factors are multiplied by the sample standard deviation, sy, before calculating the
confidence limit and conducting the test.
Procedures for estimating sample sizes and testing hypotheses about the site mean using the Land procedure, with
and without compositing, are described in the 1994 draft Technical Background Document (U.S. EPA, 1994i).
A peer review of the draft Technical Background Document identified several issues of concern:
• The use of a procedure relying strongly on the assumption of a lognormal distribution
• Quantitation limit issues
• Issues associated with multiple hypothesis tests where multiple contaminants are present in site soils.
The first issue is of concern because the small sample sizes appropriate for surface soil screening will not provide
sufficient data to validate this assumption. To address this issue, EPA considered several alternative approaches and
performed extensive analyses. These analyses are described in Sections 4.3.2 and 4.3.3. Section 4.3.3 describes
extensive simulation studies involving a variety of distributions that were done to compare the Land, Chen, and Max
tests and to develop the latter two as options for soil screening.
4.3.2 Test Of Proportion Exceeding a Threshold. One of the difficulties noted for the Land test,
described in Section 4.3.1, is its strong reliance on an assumption of lognormality (see Section 4.3.3). Even in cases
where the assumption may hold, there will rarely be sufficient information to test it.
A second criticism of applying the Land test (or another test based on estimating the mean) is that values must be
substituted for values reported as less than a quantitation limit (
-------
choosing
P0 = l- P0 (EA needs further investigation)
versus
Hj! P < P0 (EA does not need further investigation).
The test is based on concentration data from a grid sample of N points in the EA (without compositing). Let p
represent the proportion of these n points with observed concentrations greater than or equal to 2 SSL. The test is
carried out by choosing a critical value, pc, to meet the desired Type I error rate, that is,
a = Prob(p
-------
when only a small portion of the EA has very high levels (i.e., a hot spot). In that case, the observed
p will converge for increasing n to that proportion of the EA that is contaminated; it would do the
same if the concentration levels in that same portion were just slightly above 2 SSL. A test based on
a mean for large samples, however, is able to distinguish between these two situations; by its very
nature, a test based on a proportion of measurements exceeding a single threshold level cannot.
For these reasons, the test described here based on the proportion of observations exceeding 2 SSL was not selected
for inclusion in the current guidance.
4.3.3 Relative Performance of Land, Max, and Chen Tests. A simulation study was conducted to
compare the Land, Chen, and Max tests and to determine sample sizes necessary to achieve DQOs. This section
describes the design of the simulation study and summarizes its results. Detailed output from the simulations is
presented in Appendix I.
Treatment Of Data Below the Quantitation Limit. Review of quantitation limits for 110 chemicals showed
that for more than 90 percent of the chemicals, the quantitation limit was less than 1 percent of the ingestion SSL.
In such cases, the treatment of values below the QL is not expected to have much effect, as long as all data are used
in the analysis, with concentrations assigned to results below the QL in some reasonable way. In the simulations, the
QL was assumed to be SSL/100 and any simulated value below the QL was set equal to 0.5 QL. This is a conservative
assumption based on the comparison of ingestion SSLs with QLs.
Decision Rules. For the Land procedure, as discussed in Section 4.3.1, the null hypothesis H0: H > 2 SSL (where
H represents the true mean concentration for the EA) can be tested at the 5 percent level by calculating Land's upper
95 percent confidence limit for a lognormal mean. The null hypothesis is rejected (i.e., surface soil contaminant
concentrations are less than 2 SSL), if this upper confidence limit falls below 2 SSL. This application of the Land
(1971) procedure, as described in the draft 1994 Guidance, will be referred to as the "SSL DQOs" and the "original
Land procedure."
For the Max test, one decides to walk away if the maximum concentration observed in composite samples taken from
the EA does not exceed 2 SSL. As indicated in Section 4.1.6, it is viewed as providing a test of the original null
hypothesis, H0: \a>2 SSL. The Max test does not inherently control either type of error rate (i.e., its critical region
is always the region below 2 SSL, not where concentrations below a threshold that achieve a specified Type I error
rate). However, control of error rates for the Max test can be achieved through the DQO process by choice of design
(i.e., by choice of the number N of composite samples and choice of the number C of specimens per composite).
The Chen test requires that the null hypothesis have the form H0: H < \IQ, with the alternative hypothesis as H]: H >
Ho (Chen, 1995). Hypotheses or DQOs of this form are referred to as "flipped hypotheses" or "flipped DQOs" because
they represent the inverse of the actual hypothesis for SSL decisions. In the simulations, the Chen method was applied
with HO = 0.5 SSL at significance levels (Type I error rates) of 0.4, 0.3, 0.2, 0.1, 0.05, 0.025, and 0.01. In this
formulation, a Type I error occurs if one decides incorrectly to investigate further when the true site mean, H, is at
or below 0.5 SSL.
The two formulations of the hypotheses are equivalent in the sense that both allow achievement of soil screening
DQOs. That is, working with either formulation, it is possible to control the probability of incorrectly deciding to
walk away when the true site mean is 2 SSL and to also control the probability of incorrectly deciding to investigate
further when the true site mean is 0.5 SSL.
In addition to the original Land procedure, the Chen test, and the Max test, the simulations also include the Land test
of the flipped null hypothesis H0: H < 0.5 SSL at the 10 percent significance level. This Land test of the flipped
hypothesis was included to investigate how interchanging the null and alternative hypotheses affected sample sizes
4-38
-------
for the Land and Chen procedures.
Simulation Distributions. In the following description of the simulations, parameter acronyms used as labels
in the tables of results are indicated by capital letters enclosed in parentheses.
Each distribution used for simulation is a mixture of a lower concentration distribution and a higher concentration
distribution. The lower distribution represents the EA in its natural (unpolluted) state, and the higher distribution
represents contaminated areas. Typically, all measurements of pollutants in uncontaminated areas are below the QL.
Accordingly, the lower distribution is assumed to be completely below the QL. For the purposes of this analysis, it
is unnecessary to specify any other aspect of the lower distribution, because any measurement below the QL is set
equal to 0.5 QL.
A parameter between 0 and 1, called the mixing proportion (MIX), specifies the probability allocated to the lower
distribution. The remaining probability (1-MIX) is spread over higher values according to either a lognormal,
gamma, or Weibull distribution. The parameters of the higher distribution are chosen so that the overall mixture has
a given true EA mean (MU) and a given coefficient of variation (CV). Where s is the sample standard deviation, ^
is the sample mean, and C is the number of specimens per composite sample, CV is defined as:
CV = — or CV = " ' (43)
The following parameter values were used in the simulations:
EA mean (MU) = 0.5 SSL or 2 SSL
EA coefficient of variation (CV) = 1, 1.5, 2, 2.5, 3, 3.5, 4, 5, or 6 (i.e., 100 to 600 percent)
Number of specimens per composite (C) = 1, 2, 3, 4, 5, 6, 8, 9, 12, or 16
Number of composites chemically analyzed (N) = 4, 5, 6, 7, 8, 9, 12, or 16.
The true EA mean was set equal to 0.5 SSL or 2 SSL in order to estimate the two error rates of primary concern. Most
CVs encountered in practice probably will lie between 1 and 2.5 (i.e., variability between 100 and 250 percent). This
expectation is based on data from the Hanford site (see Hardin and Gilbert, 1993) and the Piazza Road site (discussed
in Section 4.3.6). EPA believes that the most practical choices for the number of specimens per composite will be
four and six. In some cases, compositing may not be appropriate (the case C = 1 corresponds to no compositing). EPA
also believes that for soil screening, a practical number of samples chemically analyzed per EA lies below nine, and
that screening decisions about soils in each EA should not be based on fewer than four chemical analyses.
For a given CV, there is a theoretical limit to how large the mixing proportion can be. The values of the mixing
proportion used in the simulations are shown below as a function of CV. The case MIX = 0 corresponds to an EA
characterized by a gamma, lognormal, or Weibull distribution. A value of MIX near 1 indicates an EA where all
concentrations are below the QL except those in a small portion of the EA. Neither of these extremes implies an
extreme overall mean. If MIX = 0, the contaminating (higher) distribution can have a low mean, resulting in a low
overall mean. If MIX is near 1 (i.e., a relatively small contamination area), a high overall mean can be obtained if
the mean of the distribution of contaminant concentrations is high enough.
4-39
-------
cv
1.0
1.5
2.0
2.5
3.0
3.5
4.0
5.0
6.0
Values of MIX
used in the simulations
0, 0.49
0,0.50
0,0.50,0.75
0,0.50,0.85
0,0.50,0.85
0,0.50,0.90
0,0.50,0.90
0,0.50,0.95
0.0.50.0.95
Treatment Of Measurement Error. Measurement errors were assumed to be normally distributed with mean
0 (i.e., unbiased measurements) and standard deviation equal to 20 percent of the true value for each chemically
analyzed sample. (Earlier simulations included measurement error standard deviations of 10 percent and 25 percent.
The difference in results between these two cases was negligible.)
Number Of Simulated Samples. Unique combinations ofthe simulation parameters considered (i.e., 2 values
of the EA mean, 10 values forthe number of specimens per composite, 8 values forthe number of composite samples,
25 combinations of CV and MIX, and 3 contamination models—lognormal, gamma, Weibull), result in a total of
12,000 simulation conditions. One thousand simulated random samples were generated for each ofthe 12,000 cases
obtained by varying the simulation parameters as described above. The average number of physical samples simulated
from an EA for a hypothesis test (i.e., the product CN) was 56.
The following 10 hypothesis tests were applied to each ofthe 12 million random samples:
Chen test at significance levels of 0.4, 0.3, 0.2, 0.1, 0.05, 0.025, and 0.01
Original Land test ofthe null hypothesis H0: (i > 2 SSL at the 5 percent significance level
• Land test of the flipped null hypothesis H0: (i < 0.5 SSL at the 10 percent
significance level
• Maximum test.
These simulations involved generation of approximately 650 million random numbers.
Simulation Results. A complete listing ofthe simulation results, with 150 columns and 59 lines per page,
requires 180 pages and is available from EPA on a 3.5-inch diskette.
Representative results for gamma contamination data, with eight composite samples that each consist of six
specimens, are shown in Table 4.9. The gamma contamination model is recommended for determining sample size
requirements because it was consistently seen to be least favorable, in the sense that it required higher sample sizes
to achieve DQOs than either ofthe lognormal or Weibull models. Hence, sample sizes sufficient to protect against
a gamma distribution of contaminant concentrations are also protective against a lognormal or Weibull distribution.
4-40
-------
Table 4.9 Comparison of Error Rates for Max Test, Chen Test
(at .20 and .10 Significance Levels), and Original Land Test,
Using 8 Composites of 6 Samples Each, for Gamma Contamination Data
MU/SSL
C=6 N=8 CV=4
0.5
0.5
0.5
2.0
2.0
2.0
C=6 N=8 CV=3
0.5
0.5
0.5
2.0
2.0
2.0
C=6 N=8 CV=2
0.5
0.5
0.5
2.0
2.0
2.0
C=6 N=8 CV=1
0.5
0.5
2.0
2.0
MU = True EA Mean
MIX
.00
.50
.90
.00
.50
.90
.00
.50
.85
.00
.50
.85
.00
.50
.75
.00
.50
.75
.00
.49
.00
.49
Max test
.35
.40
.40
.06
.06
.04
.24
.25
.23
.04
.03
.03
.07
.06
.04
.02
.02
.01
.00
.00
.01
.01
0.20 Chen test 0.
.18
.22
.19
.10
.11
.16
.18
.19
.22
.03
.03
.06
.22
.19
.19
.00
.00
.00
.20
.20
.00
.00
10 Chen test
.09
.11
.09
.18
.18
.29
.10
.10
.11
.06
.05
.12
.11
.09
.10
.00
.01
.01
.10
.12
.00
.00
Land test
.99
.99
.98
.00
.00
.01
.93
.94
.99
.00
.00
.00
.57
.68
.85
.01
.00
.00
.01
.12
.02
.00
- see subsection entitled ?Simulation Distributions" in Section 5.3 .3.
MIX = Mixing Proportion - see subsection entitled
C = Number of specimens in a composite.
N = Number of composites analyzed.
CV = EA coefficient of variation L/c\s
where s = sample
X
standard deviation and % =
?Simulation Distributions" in
mean sample concentration
Section 5.3 .3
Table 4.9 shows that the original Land method is unable to control the error rates at 0.5 SSL for gamma distributions.
This limitation of the Land method was seen consistently throughout the results for all nonlognormal distributions
tested. This limitation led to removal of the Land procedure from the Soil Screening Guidance.
Earlier simulation results for gamma and Weibull distributions did not censor results below the QL and used pure
unmixed distributions. In these cases, as the sample size N increased, with all other factors fixed, the Land error rates
at 0.5 SSL increased toward 1. Normally, the expectation is that as the sample size increases, information increases,
and error rates decrease.
When using data from a Weibull or gamma distribution, the Land confidence interval endpoints converge to a value
that does not equal the true site mean, (ix , and results in an increase in error rates. This phenomenon is easily
demonstrated, as follows. Let X denote the concentration random variable, let Y = ln(X) denote its logarithm. Let
\iy and oy denote the mean and standard deviation of logarithms of the soil concentrations. Then, as the sample size
increases, the Land confidence interval endpoints (UL and LL) converge to
4-41
-------
UL = LL = exp(^ + -) (44)
If X is lognormally distributed, this expression is the mean of X. If X has a Weibull or gamma distribution, this
expression is not the mean of X. This inconsistency accounts for the increase in error rates with sample size.
Table 4.9 also shows the fundamental difference between the Max test and the Chen test. For the Max test, the
probability of error in deciding to walk away when the EA mean is 2.0 SSL is fairly stable, ranging from 0.01 to 0.06
across the different values of the CV. On the other hand, these error rates vary more across the CV values for the
Chen test (e.g., from 0.00 to 0.29 for Chen test at the 0.10 significance level). This occurs because the Chen test is
designed to control the other type of error rate (at 0.5 SSL). The Max test is presented in the 1995 Soil Screening
Guidance (U.S. EPA, 1995c) because of its simplicity and the stability of its control over the error rate at 2 SSL.
Table 4.10 shows error rate estimates for four to nine composite samples that each consist of four, six, or eight
specimens for EAs with CVs of 2,2.5, 3, or 3.5, and assuming a gamma distribution. Table 4.10 should be adequate
for most SSL planning purposes. However, more complete simulation results are reported in Appendix I.
Planning for CVs at least as large as 2 is recommended because it is known that CVs greater than 2 occur in practice
(e.g., for two of seven EAs in the Piazza Road simulations reported in Section 4.3.6). One conclusion that can be
drawn from Table 4.11 is that composite sample sizes of four are often inadequate. Further support for this conclusion
is reported in the Piazza Road simulations discussed in Section 4.3.6.
Conclusions. The primary conclusions from the simulations are:
• For distributions other than lognormal, the Land procedure is prone to decide to investigate further
at 0.5 SSL, when the correct decision is to walk away. It is therefore unsuitable for surface soil
screening.
• Both the Max test and the Chen test perform acceptably under a variety of distributional assumptions
and are potentially suitable for surface soil screening.
4-42
-------
Table 4.10. Error Rates of Max Test and Chen Test at .2 (C20) and .1 (C10)
Significance Level for CV = 2, 2.5, 3, 3.5
CV = 2.0
N
C = 4
4
4
5
5
6
6
7
7
8
8
9
9
C = 6
4
4
5
5
6
6
7
7
8
8
9
9
C = 8
4
4
5
5
6
6
7
7
8
8
9
9
MU/SSL
0.5
2.0
0.5
2.0
0.5
2.0
0.5
2.0
0.5
2.0
0.5
2.0
0.5
2.0
0.5
2.0
0.5
2.0
0.5
2.0
0.5
2.0
0.5
2.0
0.5
2.0
0.5
2.0
0.5
2.0
0.5
2.0
0.5
2.0
0.5
2.0
Max
.09
.13
.11
.10
.11
.06
.12
.04
.16
.02
.16
.01
.03
.14
.04
.09
.06
.04
.06
.02
.06
.02
.06
.01
.02
.12
.03
.07
.02
.04
.03
.03
.04
.02
.04
.01
C20
.20
.08
.21
.05
.21
.03
.20
.03
.19
.02
.21
.01
.20
.03
.20
.02
.20
.01
.20
.00
.19
.00
.20
.00
.21
.02
.22
.01
.18
.00
.20
.00
.20
.00
.21
.00
C10
.11
.16
.10
.11
.12
.08
.10
.05
.09
.03
.11
.02
.12
.08
.10
.05
.11
.02
.09
.01
.09
.01
.10
.01
.13
.05
.11
.02
.09
.01
.11
.00
.10
.00
.11
.00
CV = 2.5
Max
.14
.19
.15
.10
.21
.08
.25
.05
.25
.04
.28
.03
.08
.16
.11
.09
.14
.06
.12
.05
.15
.02
.18
.02
.06
.15
.05
.09
.08
.06
.09
.04
.11
.02
.11
.02
C20
.18
.17
.18
.09
.20
.08
.22
.04
.20
.03
.20
.03
.21
.08
.17
.04
.21
.03
.19
.02
.20
.01
.22
.01
.19
.04
.20
.02
.21
.01
.20
.01
.21
.01
.21
.00
C10
.09
.28
.09
.18
.10
.14
.11
.09
.09
.07
.09
.06
.12
.17
.09
.10
.10
.07
.10
.04
.10
.03
.11
.02
.10
.09
.11
.06
.11
.02
.11
.01
.11
.01
.10
.01
CV = 3.0
Max
.19
.20
.26
.17
.28
.11
.31
.08
.36
.05
.36
.04
.15
.17
.17
.13
.19
.09
.23
.06
.25
.03
.28
.03
.10
.17
.11
.09
.13
.07
.18
.04
.17
.04
.20
.01
C20
.18
.21
.20
.19
.21
.13
.20
.11
.20
.08
.18
.07
.20
.14
.20
.10
.20
.07
.22
.06
.19
.03
.20
.02
.21
.09
.20
.04
.19
.04
.21
.02
.21
.01
.19
.00
C10
.08
.33
.08
.30
.11
.23
.09
.18
.10
.14
.09
.13
.10
.24
.10
.18
.10
.14
.10
.10
.10
.05
.11
.04
.10
.17
.10
.10
.10
.07
.11
.04
.10
.03
.10
.01
CV = 3.5
Max
.24
.26
.26
.18
.31
.11
.36
.08
.42
.07
.44
.07
.16
.20
.22
.15
.25
.09
.29
.08
.30
.04
.34
.03
.14
.19
.17
.12
.20
.08
.22
.05
.26
.03
.30
.02
C20
.20
.29
.20
.23
.19
.18
.18
.14
.20
.13
.22
.12
.17
.19
.20
.13
.20
.10
.21
.09
.19
.06
.19
.05
.18
.14
.19
.08
.20
.07
.20
.05
.19
.03
.23
.02
C10
.10
.42
.09
.36
.09
.28
.10
.23
.09
.21
.12
.20
.08
.33
.10
.24
.10
.19
.10
.14
.10
.11
.09
.09
.08
.25
.09
.17
.10
.13
.11
.09
.10
.06
.12
.04
MU = True EA Mean - see subsection entitled ?Simulation Distributions" in Section 5.3 .3.
MIX= Mixing Proportion - see subsection entitled ?Simulation Distributions" in Section 5.3 .3
C = Number of specimens in a composite.
N = Number of composites analyzed.
CV = EA coefficient of variation
where s = sample standard deviation and x = mean sample concentration
4-43
-------
4.3.4 Treatment of Observations Below the Limit of Quantitation. Test procedures that are based
on estimating a mean contaminant level for an EA, such as the Land and Chen procedures, make use of each
measured concentration value. For this reason, the use of all reported concentration measurements in such
calculations should be considered regardless of their magnitude—that is, even if the measured levels fall below a
quantitation level. One argument for this approach is that the QL is itself an estimate. Another is that some value
will have to be substituted for any censored data point (i.e., a point reported as 2 SSL
versus
Hji mean concentration of a given chemical < 2 SSL.
The default value for the probability of a Type I error is a = 0.05, while the default value for the power of the test
at 0.5 SSL is l-(3 = 0.80. The test is applied separately for each chemical, so that these probabilities apply for each
individual chemical. Thus, there is an 80 percent probability of walking away from an EA (i.e., rejecting H0) when
only one chemical is being tested and its true mean level is 0.5 SSL and a 5 percent probability of walking away if
its true mean level is 2 SSL.
However, the Soil Screening Guidance does not explicitly address the following issues:
What is the composite probability of walking away from an EA if there are multiple
contaminants?
and
If such probabilities are unacceptable, how should one compensate when testing for multiple
contaminants within a single EA?
4-44
-------
The answer to the first question cannot be determined, in general, since the concentrations of the various
contaminants will often be dependent on one another (e.g., this would be expected if they originated from the same
source of contamination). The joint probability of walking away can be determined, however, if one makes the
simplifying assumption that the contaminant concentrations for the different chemicals are independent
(uncorrelated). In that case, the probability of walking away is simply the product of the individual rejection
probabilities.
For two chemicals (Chemical A and Chemical B, say), this is:
Prjwalking away from EA} = Prjreject H0 for Chemical A} x Pr{reject H0 for Chemical B}.
While these joint probabilities must be regarded as approximate, they nevertheless serve to illustrate the effect on
the error rates when dealing with multiple contaminants.
Assume (for illustrative purposes only) that the probabilities for rejecting the null hypothesis (walking away from
the EA) for each single chemical appear as follows:
True concentration
0.2 SSL
0.5 SSL
0.7 SSL
1.0 SSL
1.5 SSL
2.0 SSL
Probability of rejecting H0
0.95
0.80(default1-p)
0.60
0.50
0.20
0.05 (default a)
Let C(A) denote the concentration of Chemical A divided by the SSL, and let P(A) denote the corresponding
probability of rejecting H0. Define C(B) and P(B) similarly for Chemical B. Assuming independence, the joint
probabilities of rejecting the null hypothesis (walking away) are as shown in Table 45.11.
Table 4.11. Probability of "Walking Away" from an EA
When Comparing Two Chemicals to SSLs
Chemical A
C(A)
0.2
0.5
0.7
1.0
1.5
2.0
P(A)
0.95
0.80
0.60
0.50
0.20
0.05
Chemical B
C(B) = 0.2
P(B) =.95
0.90
0.76
0.57
0.48
0.19
0.05
C(B) = 0.5
P(B) = .80
0.76
0.64
0.48
0.40
0.16
0.04
C(B) =
P(B) =
0.57
0.48
0.36
0.30
0.12
0.03
0.7 C(B) = 1.0
.60 P(B) = .50
0.48
0.40
0.30
0.25
0.10
0.03
C(B) = 1.5
P(B) = .20
0.19
0.16
0.12
0.10
0.04
0.01
C(B) =
P(B) =
0.05
0.04
0.03
0.03
0.01
<0.01
2.0
.05
These probabilities demonstrate that the test procedure will tend to be very conservative if multiple chemicals are
involved—that is, all of the chemical concentrations must be quite low relative to their SSL in order to have a high
probability of walking away from the EA. On the other hand, there will be a high probability that further
investigation will be called for if the mean concentration for even a single chemical is twice the SSL.
4-45
-------
A potential problem occurs when there are several chemicals under consideration and when all or most of them have
levels slightly below the SSL (e.g., near 0.5 SSL). For instance, if each of six independent chemicals had levels at
0.5 SSL, the probability of rejecting the null hypothesis would be 80 percent for each such chemical, but the
probability of walking away from the EA would be only (0.80)6 = 0.26.
If the same samples are being analyzed for multiple chemicals, then the original choice for the number of such
samples ideally should have been based on the worst case (i.e., the chemical expected to have the largest variability).
In this case, the probability of correctly rejecting the null hypothesis at 0.5 SSL for the chemicals with less variability
will be higher. The overall probability of walking away will be greater than shown above if all or some of the
chemicals have less variability than assumed as the basis for determining sample sizes. Here, the sample size will
be large enough for the probability of rejecting the null hypothesis at 0.5 SSL to be greater than 0.80 for these
chemicals.
The probability values assumed above for deciding that no further investigation is necessary for individual chemicals,
which are the basis for these conclusions, are equally applicable for the Land, Chen, and Max tests. They simply
represent six hypothetical points of the power curves for these tests (from 0.2 SSL to 2.0 SSL). Therefore, the
conclusions are equally applicable for each of the hypothesis testing procedures that have been considered in the
current guidance for screening surface soils.
If the surface soil concentrations are positively correlated, as expected when dealing with multiple chemicals, then
it is likely that either all the chemicals of concern have relatively high concentrations or they all have relatively low
concentrations. In this case, the probability of making the correct decision for an EA would be greater than that
suggested by the above calculations that assume independence of the various chemicals.
However, the potential problem of several chemicals having concentrations near 0.5 SSL is not precluded by
assuming positive correlations. In fact, it suggests that if the EA average for one chemical is near 0.5 SSL, then the
average for others is also likely to be near 0.5 S SL, which is exactly the situation where the probability of not walking
away from the EA can become large because there is a high probability that H0 will be rejected for at least one of
these chemicals.
An alternative would be to use multiple hypothesis testing procedures to control the overall error rate for the set of
chemicals (i.e., the set of hypothesis tests) rather than the separate error rates for the individual chemicals. Guidance
for performing multiple hypothesis tests is beyond the scope of the current document. Obtain the advice of a
statistician familiar with multiple hypothesis testing procedures if the overall error rates for multiple chemicals is of
concern for a particular site. The classical statistical guidance regarding this subject is Simultaneous Statistical
Inference (Miller, 1991).
4.3.6 Investigation Of Compositing Within EA Sectors. If one decides that an EA needs further
investigation, then it is natural to inquire which portion(s) of the EA exceed the screening level. This is a different
question than simply asking whether or not the EA average soil concentration exceeds the SSL. Conceivably, this
question may require additional sampling, chemical analysis, and statistical analysis. A natural question is whether
this additional effort can be avoided by forming composites within sectors (subareas) of the EA. The sector with the
highest estimated concentration would then be a natural place to begin a detailed investigation.
The simulations to investigate the performance of rules to decide whether further investigation is required, reported
in Section 4.3.3, make specific assumptions about the sampling design. It is assumed that N composite samples are
chemically analyzed, each consisting of C specimens selected to be statistically representative of the entire EA. The
key point, in addition to random sampling, is that composites must be formed across sectors rather than within
sectors. This assumption is necessary to achieve composite samples that are representative of the EA mean (i.e., have
the EA mean as their expected value).
4-46
-------
If compositing is limited to sectors, such as quadrants, then each composite represents its sector, rather than the entire
EA. The simulations reported in Section 4.3.3, and sample sizes based on them, do not apply to this type of
compositing. This does not necessarily preclude compositing within sectors for both purposes, i.e., to test the
hypothesis about the EA mean and also to indicate the most contaminated sector. However, little is known about the
statistical properties of this approach when applying the Max test, which would depend on specifics of the actual
spatial distribution of contaminants for a given EA. Because of the lack of extensive spatial data sets for
contaminated soil, there is limited basis for determining what sample sizes would be adequate for achieving desired
DQOs for various sites. However, one spatial data set was available and used to investigate the performance of
compositing within sectors at one site.
Piazza Road Simulations. Data from the Piazza Road NPL site were used to investigate the properties of tests
of the EA mean based on compositing within sectors, as compared to compositing between sectors. The investigation
of a single site cannot be used to validate a given procedure, but it may indicate whether further investigation of the
procedure is worthwhile.
Seven nonoverlapping 0.4-acre EAs were defined within the Piazza Road site. Each EA is an 8-by-12 grid composed
of 14'xl4' squares. The data consist of a single dioxin measurement of a composite sample from each small square.
These measurements are regarded as true values for the simulations reported in this section. Measurement error was
incorporated in the same fashion as for the simulations reported in Section 4.3.3.
Each of the seven EAs was subdivided into four 4-by-6 sectors, six 4-by-4 sectors, eight 4-by-3 sectors, twelve 2-by-4
sectors, and sixteen 2-by-3 sectors. Results are presented here for the cases of four, six, and eight sectors because
composites of more than eight specimens are expected to be used rarely, if at all.
Table 4.12 presents the "true" mean and CV for each EA, computed from all 96 measurements within the 0.4-acre
EA. The CVs range from 1.0 to 2.2. Note that two of the seven CVs equal or exceed 2 at this site. This supports
EPA's belief that at many sites it is prudent, when planning sample size requirements for screening, to assume a CV
of at least 2.5 and to consider the possibility of CVs as large as 3 or 3.5.
As data on variability within EAs for different sites and contaminant conditions accrue overtime, it will be possible
to base the choice of procedures on a larger, more comprehensive database, rather than just a single site.
Appendix J contains results of simulations from the seven Piazza Road EAs. Sampling with replacement from each
sector was used, because this was felt to be more consistent with the planned compositing. To estimate the error rates
at 0.5 SSL and 2 SSL for each EA, the SSL was defined so that the site mean first was regarded as 0.5 SSL and then
was regarded as 2 SSL.
Table 4.12. Means and CVs for Dioxin Concentrations
for 7 Piazza Road Exposure Areas
EA
1
2
3
4
5
6
7
Mean of EA
2.1
2.4
5.1
4.0
9.3
15.8
2.8
CV of EA
1.0
1.6
1.1
1.2
2.0
2.2
1.4
N
96
96
96
96
96
96
96
4-47
-------
Notation for Results from Piazza Road Simulations. The following notation is used in Appendix J. The
design variable (DBS) indicates whether compositing was within sector (DES=W) or across sectors (DES=X). As
in Section 4.3.3, C denotes the number of specimens per composite, and N denotes the number of composite samples
chemically analyzed. Results in Appendix J are for the Chen test at the 10 percent significance level and for the Max
test. The true mean and CV are shown in the header for each EA.
Results and Conclusions from Piazza Road Simulations. Although the results from a single site
cannot be assumed to apply to all sites, the following observations can be made based on the Piazza. Road simulations
reported in Appendix J.
• The error rate at 0.5 SSL for the Chen test, using compositing across sectors (DES=X), is generally
close to the nominal rate of 0.10. For compositing within sectors (DES=W), the error rate for Chen
at 0.5 SSL is generally much lower than the nominal rate.
• Except for plans involving only four analyses (N = 4), the error rate at 2 SSL is always below 0.05
for the Chen test. For the Max test, the error rate at 2 S SL fluctuated between 0 and 16 percent. The
error rate at 2 SSL is smaller for the Chen test at the 10 percent significance level than for the Max
test in virtually all cases. The only two exceptions to this are for compositing within sector
(DES=W) in EA No. 6.
• This observation provides further support for the conclusion drawn from the simulations reported
in Section 4.3.3: plans involving only four analyses can result in high error rates in determining the
mean contaminant concentration of an EA with the Max test. In most cases the error rates of concern
to EPA (at 2 SSL) are 0.10 or larger.
• In general, error rates estimated from Piazza Road simulations for compositing across sectors are at
least as small as would be predicted on the basis of the simulation results reported in Section 4.3.3.
• The simulation results show that compositing within sectors using the Max test may be an option for
site managers who want to know whether one sector of an EA is more contaminated than the other.
However, use of the Max test when compositing within sectors may lead the site manager to draw
conclusions about the mean contaminant concentration in that sector only, not across the entire EA.
4-48
-------
Part 5: RADIONUCLIDE-SPECIFIC PARAMETERS
Radionuclide-specific parameters required for calculating soil screening levels include an area correction
factor (ACF) for the external radiation exposure pathway, and the soil-water partition coefficient for
radionuclides (Kd). This part of the background document describes the collection and compilation of these
parameters for the SSL radionuclides.
5.1 Area Correction Factor - ACF (unitless)
The risk model used for the external radiation exposure pathway in this guidance effectively assumes that
an individual is exposed to a source geometry that is effectively an infinite slab. The concept of an "infinite
slab" means that the thickness of the contaminated zone and its aerial extent are so large that it behaves as
if it were infinite in its physical dimensions. In practice, soil contaminated to a depth greater than about 15
cm and with an aerial extent greater than about 1,000 m2 (i.e., one-quarter acre) will create a radiation field
approaching that of an infinite slab.
This infinite slab assumption has been used in the calculation of radionuclide slope factors presented in
Section 2.1. For very small areas of contamination, this will result in overly conservative estimates of risk.
For calculation of SSLs for a residential setting, an adjustment for source area is considered to be an
important modification for Superfund sites. Thus, an area correction factor, ACF, has been added to the SSL
calculation.
Table 5.1 provides recommended ACFs for radionuclides as a function of source area calculated using
MicroShield V5.01.1 Since the default source size is 0.5-acre (i.e., 2,000 m2), the default ACF for SSL
equations is set at 0.9. The calculations assume for a uniform layer of contamination 15 cm deep with a soil
density of 1.6 g/cm3. A single recommended value is considered suitable for all radionuclides over the range
of source areas since EPA's analysis shows that ACFs vary little from one radionuclide to another. For other
source areas, recommended ACFs are presented in Table 5.1.
1 Grove Engineering, Rockville, MD.
5-1
-------
Table 5.1 Recommended Area Correction
Factors as Function of Source Area
Source Area
(m2)
10,000
5,000
2,000
1,000
500
100
50
10
ACF
1.00
0.94
0.90
0.88
0.86
0.75
0.66
0.40
EPA's analysis of ACFs is found in Table 5.2, which provides examples of ACFs for several radionuclides
as a function of source area calculated using MicroShield V5.01. The calculations assume for a uniform
layer of contamination 15 cm deep with a soil density of 1.6 g/cm3. Strong gamma-ray emitters like 60Co
have relatively large slope factors for this pathway relative to the slope factors for weak photon emitters like
239Pu. However, as noted in the table, ACFs vary little from one radionuclide to another over the range of
source areas shown. Users that have one of the radionuclides in Table 5.2 as a contaminant at their site may
use the radionuclide specific ACF that is appropriate for their source area rather than the value found in
Table 5.1.
Table 5.2 Area Correction Factors as Function of Source Area
for Selected Radionuclides Calculated Using MicroShield
Source
Area
(m2)
10,000
5,000
2,000
1,000
500
100
50
10
231 Am
1.00
0.93
0.89
0.87
0.85
0.76
0.69
0.44
60Co
1.00
0.95
0.92
0.90
0.87
0.75
0.66
0.38
137Cs
1.00
0.95
0.92
0.90
0.87
0.76
0.67
0.39
239pu
1.00
0.94
0.90
0.89
0.86
0.78
0.71
0.45
226Ra+D
1.00
0.95
0.92
0.90
0.87
0.75
0.66
0.38
-n,
1.00
0.94
0.89
0.88
0.86
0.77
0.70
0.44
238U+[)
1.00
0.94
0.91
0.89
0.86
0.76
0.68
0.41
5.2 Soil-Water Distribution Coefficients (Kd) for Radionuclides
As with organic chemicals, development of SSLs for inorganics (including radionuclides) requires a soil-
water partition coefficient or distribution cofficient (Kd) for each constituent. The soil- water distribution
coefficient is typically defined in fate and contaminant transport calculation as the ratio of the contaminant
5-2
-------
concentrations associated with the solid to the contaminant concentration in the surrounding aqueous solution
when the system is at equilibrium (EPA 1999b). However, the simple relationship between soil organic
carbon content and sorption observed for organic chemicals does not apply to inorganics (including
radionuclides). The soil-water distribution coefficient (Kj) for inorganics (including radionuclides) is
affected by numerous geochemical parameters and processes, including pH; sorption to clays, organic matter,
iron oxides, and other soil constituents; oxidation/reduction conditions; major ion chemistry; and the
chemical form of the radionuclide. The number of significant influencing parameters, their variability in the
field, and differences in experimental methods result in as much as seven orders of magnitude variability in
measured metal Kd values reported in the literature (see Table 43 in the Soil Screening Guidance: Technical
Background Document (EPA 1996b)). This variability makes it much more difficult to derive generic Kd
values for metals (including radionuclides) than for organics. Therefore, it is recommended that Kd values
be measured for site-specific conditions. If the Kd is not measured site-specifically, then a conservative Kd
should be used in calculating SSLs.
Tables 5.3 in the TBD (also C.2a in the User Guide) and C.2b in the User Guide list the default Kd values
for each element. Table 5.2 is derived from the EPA Office of Radiation and Indoor Air's 1999 final
document Understanding Variation In Partition Coefficient, Kd Values, Volume 1: The Kd Model of
Measurement, and Application of Chemical Reaction Codes, & Volume 2: Review of Geochemistry and
Available Kd Values for Cadmium, Cesium, Chromium, Lead , Plutonium, Radon, Strontium, Thorium,
Tritium, and Uranium. This document is intended to provide technical information to EPA and/or DOE
transport modelers on the key geochemical processes affecting contaminant transport through soil and
sediments. Particular attention is directed at providing an understanding of: 1) the use of distribution
coefficients, 2) the difference between the original thermodynamic Kd parameter derived from the ion-
exchange literature and its "empiricized" use in contaminant transport codes, 3) the explicit and implicit
assumptions underlying the use of the Kd paramter in contaminant transport codes, and 4) methods for site-
specific measurements.
When estimating migration of contaminants from soil to groundwater for a contaminant which is not
represented with a default Kd value in either Table 5.3 or Tables C.2a and C.2b in the User Guide, site
decision-makers should measure a site-specific Kd. Site decision-makers also may develop site-specific Kds
to more accurately estimate contaminant migration rather than using the default values in Tables 5.4 to 5.9
or either Tables C.2a or C.2b or in the User Guide.
The elements chosen for study in EPA (1999b) include the following: chromium, cadmium, cesium, lead,
plutonium, radon, strontium, thorium, tritium, and uranium. The selection of these contaminants by EPA
staff was based on two criteria. First, the elements had to be one of high priority to the site remediation or
risk assessment activities of EPA and/or DOE (EPA 1993). Second, six nonexclusive categories were
developed based on chemical behavior:
1) Cations - cadmium, cesium, plutonium, strontium, thorium, and uranium;
2) Anions - chromium (as CrO42");
3) Radionuclides - cesium, plutonium, radon, strontium, thorium, tritium, and uranium;
4) Conservatively Transported Contaminants - tritium and radon;
5) Nonconservatively Transported Contaminants - other than tritium and radon; and
6) Redox Sensitive Elements - chromium, plutonium, and uranium.
By categorizing the contaminants in this manner, general geochemical behaviors of one contaminant may
5-3
-------
be extrapolated by analogy to other contaminants in the same category. For example, contaminants present
primarily in anionic form, such as Cr(VI), tend to adsorb to a limited extent to geological materials (soil,
rocks, and geological deposits, hence forth simply referred to as soils). Thus, one might generalize that other
anions, such as nitrate, chloride, and U(VI)-anionic complexes, would also adsorb to a limited extent.
Literature on the absorption of these three solutes shows no or very little adsorption. Volume 2, EPA (1999)
identifies, when possible, minimum and maximum conservative Kd values for each contaminants as a
function of key geochemical processes affecting sorption. The tables that follows reflect this conservatism
in using the lower bounding Kd values for these contaminants. Site specific measurements of Kd values
should be used in determining SSLs. If Kd is not measured site-specifically, then conservative Kd values
should be used from the following tables when calculating SSLs.
Table 5.3 Default Kd Values for Selected Elements, based
on EPA (1999b).
Element
Cs
H
Pu
Rn
Kd value*
(ml/g)
10
0
5
0
Element
Sr
Th
U
Kd value
(ml/g)
1
20
0.4
*Note: Kd values were given of units in ml/g inEPA (1999b). The
units in other parts of this TBD are given in L/kg. However, these
are equivalent units, that is: 1 ml/g = 1 L/kg
The Kd values in Table 5.3 reflect the most conservative values provided for each element in EPA (1999).
Each of these values are based on a chemical mechanism that was considered to provide the most
conservative Kd value for that element. Users that have measured pH values at their site that differ from the
range given in this report, may want to consult Tables 5.4 to 5.9 for alternative Kds that are still conservative.
Table 5.4 shows the estimated conservative Kd values for cesium based on cation exchange capacity (CEC)
(in meg/100 g) or clay content (in wt.%). This table is for systems containing low concentrations of mica-
like minerals (i.e., <5% of the clay-size fraction). This table pertains to systems consisting of natural soils
(as opposed to pure mineral phases), low ionic strength (<0.1 M), low humic material concentrations (<5
mg/1), no organic chelates (such as EDTA), and oxidizing conditions.
Table 5.4 Kd values for cesium as a function of cation exchange capacity or clay
content low mica content soils
CEC (meg/100 g) / Clay Content (wt.%)
Kd value (ml/g)
<3/<4
10
3-10/4-20
30
10-50/20-60
80
Source: EPA (1999b)
5-4
-------
Table 5.5 also shows the estimated conservative Kd values for cesium based on CEC or clay content, but for
systems containing high concentrations of mica-like minerals (i.e., >5% of the clay-size fraction).
Table 5.5 Kd values for cesium as a function of cation exchange capacity
or clay content for high mica content soils
CEC (meg/100 g) / Clay Content (wt.%)
Kd value (ml/g)
<3/<4
30
3-10/4-20
70
10-50/20-60
210
Source: EPA(1999b)
Table 5.6 shows the EPA (1999b) look-up-table for plutonium, which is a function of clay content (in wt.%)
and dissolved carbonate (in meq/1).
Table 5.6 Kd values for plutonium as a function of soluble carbonate
and soil clay values
Clay Content (wt.%)
Soluble Carbonate
(meg/100 g)
K,, value (ml/a)
0-30
0.1 -2
5
3-4
80
5-6
130
31 -50
0.1 -2
380
3-4
1.440
5-6
2.010
51 -70
0.1 -2
620
3-4
1.860
5-6
2.440
Source: EPA (1999b)
The look-up-table for strontium, Table 5.7, requires knowledge of the CEC (or clay content) and pH of the
system in order to select the appropriate conservative strontium Kd value. This table pertains to systems
consisting of natural soils (as opposed to pure mineral phases), low ionic strength (<0.1 M), low humic
material concentrations (<5 mg/1), no organic chelates (such as EDTA), and oxidizing conditions.
Table 5.7 Kd values for strontium as a function of CEC, clay content, and pH
CEC (meq/1 00 g)
Clay Content (wt.%)
PH
Kd value (ml/g)
<3
<4
<5
1
5-8
2
8-10
3
3-10
4-20
<5
10
5-8
15
8-10
20
10-50
20-60
<5
100
5-8
200
8-10
300
Source:) EPA (1999b)
The look-up-table for thorium, Table 5.8, is based on plume thorium concentrations and pH. This table
pertains to systems consisting of natural soils (as opposed to pure mineral phases), low ionic strength (<0.1
M), low humic material concentrations (<5 mg/1), no organic chelates (such as EDTA), and oxidizing
conditions.
5-5
-------
Table 5.8 Kd values for thorium as a function of pH
and dissolved thorium concentrations
PH
Dissolved Th, M
Kd value (ml/g)
3-5
<1026
62
>io-26
300,000
5-8
<10-26
1,700
>10-26
300,000
8-10
io-26
300,000
Source: EPA(1999b)
Table 5.9 shows the EPA (1999b) look-up-table for uranium, which is a function of clay content (in wt.%)
and dissolved carbonate (in meq/1).
Table 5.9 Kd values for uranium as a function of pH
PH
Kd value (ml/g)
3
<1
4
0.4
5
25
6
100
7
63
8
0.4
9
<1
10
<1
Source: EPA(1999b)
5-6
-------
Part 6: REFERENCES
Aller, L., T. Bennett, J.H. Lehr, R.J. Petty, and G. Hackett. 1987. DRASTIC: A Standardized System for Evaluating
Ground Water Pollution Potential Using Hydrogeologic Settings. Prepared for U. S. EPA, Office of Research
and Development, Ada, OK. National Water Well Association, Dublin, OH. EPA-600/2-87-035.
American Society for Testing Materials. 1995. Annual Book of ASTM Standards. Volume 4.08. Soil and Rock
Building Stones. Philadelphia, PA.
Argonne National Laboratory. 1993. Manual for Implementing Residual Radioactive Materials Guidelines Using
RESRAD, Version 5.0, Argonne National Laboratory, Argonne, IL. ANL/EAD/LD-2.
Brady, N.C. 1990. The Nature and Properties of Soils. Macmillan Publishing Company, New York, NY.
Calabrese, E.J., H. Pastides, R. Barnes, et al. 1989. How much soil do young children ingest: an epidemiologic study.
Petroleum Contaminated Soils, Vol. 2. E.J. Calabrese and P.T. Kostecki, eds. pp. 363-417. Lewis Publishers,
Chelsea, MI.
Carsel, ILF., and R.S. Parrish. 1988. Develop ing joint probability, distributions of soil water retention characteristics.
Water Resources Research, 24(5):755-769.
Carsel, ILF., R.S. Parrish, R.L. Jones, J.L. Hansen, and R.L. Lamb. 1988. Characterizing the uncertainty of pesticide
leaching in agricultural soils, J. of Contam. Hyd 2:111-124
Coughtry, P.J., D. Jackson, and M.C. Thorne. 1985. Radionuclide Distribution and Transporting Terrestrial and
Aquatic Ecosystems: A Compedium of Data. A.A. Balkem, Netherlands, Vol. 6.
Cowherd, C., G. Muleski, P. Engelhart, and D. Gillette. 1985. Rapid Assessment of Exposure to Paniculate
Emissions from Surface Contamination. Prepared for U.S. EPA, Office of Health and Environmental
Assessment, Washington, DC. EPA/600/8-85/002. NTIS PB85-192219 7AS.
Davis, S., P. Waller, R. Buschom, J. Ballou, and P. White. 1990. Quantitative estimates of soil ingestion in normal
children between the ages of 2 and 7 years: population-based estimates using Al, Si, and Ti as soil tracer
elements. Arch. Environ. Health 45:112-122.
EQ (Environmental Quality Management). 1993. Evaluation of the Dispersion Equations In the Risk Assessment
Guidance for Superfund (RAGS): Volume I—Human Health Evaluation Manual (Part B, Development of
Preliminary Remediation Goals). Contract No. 68-02-D 120. Prepared for Office of Emergency and Remedial
Response, Washington, DC.
EQ (Environmental Quality Management). 1994. Revisions to VF and PEF Equations. Memorandum #5099-3. .
Prepared for Office of Emergency and Remedial Response, Washington, DC.
EQ (Environmental Quality Management) and Pechan (E.H. Pechan and Associates). A Comparison of Soil
Volatilization Models in Support of Superfund Soil Screening Development. Contract No. 68-03-0035.
Prepared for Office of Emergency and Remedial Response, Washington, DC.
R-l
-------
Feenstra. S., D.M. Mackay, and J-A. Cherry. 1991. A method for assessing residual NAPL based on organic
chemical concentrations in soil samples. Ground Water Monitoring Rev. 11(2): 128-136.
Gee, G.W., P.J. Wierenga, B.J. Andraski, M.H. Young, M.J. Payer, and M.L. Rockhold. 1994, Variations in water
balance and recharge potential at three western desert sites. Soil Science Society of America Journal, 58(1):
63-72.
Gelhat, L.W., and C.J. Axness. 1981. Stochastic Analysis of Macro-Dipersion in Three-Dimensionally
Heterogeneous Aquifers. Report No. H-8. Hydraulic Research Program. New Mexico Institute of Mining
and Technology, Socorro, NM.
Gilbert, Richard O. 1987. Statistical Methods for Environmental Pollution Monitoring. New York: VanNostrand
Reinhold.
Hardin J.W., and ILO. Gilbert. 1993. Comparing Statistical Tests for Detecting Soil Contamination Greater than
Background. Pacific Northwest Laboratory (PNL-8989), Richard, WA.
Hills, R.G., P.J. Wierenga, D.B. Hudson and M.R. Kirkland. 1991. The second Las Cruces trench experiment:
Experimental results and two-dimensional flow predictions. Water Re sour. Res. 27:2707-2718.
Huyakorn, P.S., and J.E. Buckley. 1987. Finite Element Code for Simulating One-Dimensional Flow and Solute
Transport in the Vadose Zone. Technical Report prepared for U.S. EPA, Athens, GA.
Kolig, H.P. 1993. Environmental Fate Constants for Organic Chemicals under Consideration for EPA 's Hazardous
Waste Identification Projects. Office of Research and Development,. U.S. Environmental Protection Agency,
Washington, D.C.
Land, C.E. 1971. Confidence intervals for linear functions of the normal mean and variance. Annals of Mathematical
Statistics 42:1187-1205.
Land, C.E. 1975 Tables of confidence limits for linear functions of the normal mean and variance. Selected Tables
in Mathematical Statistics, 3:385-419. American Mathematical Society, Providence, RI.
Liu, S., W.B. Mills, and B. Baena. 1995. Multimedia Exposure Assessment Modeling Including Fate and
Transformation Products (MULTIMED_ DP 1.0): Implementation, Tests and User's Manual. Technical
Report prepared for U.S. EPA, Athens, GA.
Marshall, J. 1971. Drag measurements in roughness arrays of varying density and distribution. Agricultural
Meteorology, 8:269-292.
Miller, R.G. 1991. Simultaneous Statistical Inference. 2nd edition. Springer-Vefiag, New York.
Porro, I., and P.J. Wierenga. 1993. Transient and steady state solute transport through a large unsaturated soil
column. Ground Water. 31(2): 193-200.
RTI (Research Triangle Institute). 1994. Chemical Properties for Soil Screening Levels. Draft Report. Prepared for
Office of Emergency and Remedial Response, U.S. Environmental Protection Agency, Washington, DC.
Salhotra, A.M., P. Mineart, S. Sharp-Hansen, T. Allison, R. Johns, and W.B. Mills. 1995. Multimedia Exposure
Assessment Model (MULTIMED 2.0) for Evaluating the Land Disposal of Wastes— Model Theory.
Technical Report prepared for U.S. EPA, Athens, GA.
R-2
-------
SC&A (S. Cohen & Associates). 1994. Root Uptake, Models, Plant Physiology, and Estimates for Plant
Accumulation of Radionuclide Contaminants fromSoils. Contract No. 68D20155. Prepared for Office of
Radiation and Indoor Air, Washington, DC.
Sharp-Hansen, S., C. Travers, P. Hummel, and T. Allison. 1990. A Subtitle D Landfill Application Manual for the
Multimedia Exposure Assessment Model (MULTIMED). EPA Contract No. 68-03-3513. Environmental
Research Laboratory, Office of Research and Development, U.S. Environmental Protection Agency, Athens,
GA.
Sharp-Hansen, S., C. Travers, P. Hummel, T. Allison, R. Johns, and W. Mills. 1995. A Subtitle D'Landfill Application
Manual for the Multimedia Exposure Assessment Model (MULTIMED 2.0). Technical Report prepared for
U.S. EPA, Athens, GA.
Simunek, J. and M. Th. van Genuchten. 1994. The CHAIN 2D Code for Simulating the Two-Dimensional Movement
of Water, Heat and Multiple Solutes in Variably Saturated Porous Media, Version 1.1. Research Report No.
136, U.S. Salinity Laboratory, USDA, ARS. U.S. Department of Agriculture, Riverside, CA.
Simunek, J., K. Huang, and M. Th. van Genuchten. 1998. The HYDRUS Code for Simulating the One-Dimensional
Movement of Water, Heat, and Multiple Solutes in Variably-Saturated Media. Version 6.0, Research Report
No. 144, U.S. Salinity Laboratory, Agricultural Research Service, U.S. Department of Agriculture, Riverside,
CA.
USDA. 1988. Dataset: National Food Consumption Survey 1987/88 Household Food Use. U.S. Department of
Agriculture. Washington, D.C. 1987/88 NFCS Database
U.S. DOC. 1977. Local Climatological Data-Annual Summaries for 1977, U.S. Department of Commerce, National
Oceanic and Atmospheric Administration/Environmental Data Service/National Climatic Data Center.
U.S. EPA (Environmental Protection Agency). 1972. Natural Radiation in the United States, ORP/SID 72-1.
U.S. EPA (Environmental Protection Agency). 1976. National Interim Primary Drinking Water Regulations, Office
of Water Supply, EPA-5 70/9-76-003.
U.S. EPA (Environmental Protection Agency). 1981. Population Exposure to External Natural Radiation
Background in the United States, ORP/SED-80-12.
U.S. EPA (Environmental Protection Agency). 1986. Method 9100: Saturated Hydraulic Conductivity, Saturated
Leachate Conductivity, and Intrinsic Permeability. Test Methods for Evaluating Solid Waste, Volume 1C:
Laboratory Manual Physical/Chemical Methods, EPA SW-846, Part 3 of 4, 3rd ed. Washington, DC.
U.S. EPA (Environmental Protection Agency). 1988. SuperfundExposure AssessmentManual. OSWER Directive
9285.5-1. EPA/540/G-88/001. Office of Emergency and Remedial Response, Washington, DC. NTIS PB89-
135859.
U.S. EPA (Environmental Protection Agency). 1989b. Risk Assessment Guidance for Superfund (RAGS): Volume
1: Human Health Evaluation Manual (HHEM), Part A, Interim Final. Office of Emergency and Remedial
Response, Washington, DC. OSWER Directive 9285.7-02B. EPA/540/1-89/002. NTIS PB90-155581.
R-3
-------
U.S. EPA (Environmental Protection Agency). 1989a. Methods for Evaluating the Attainment of Soil Cleanup
Standards. Volume I: Soils and Solid Media. EPA 230/02-89-042. Statistical Policy Branch, Office of
Policy, Planning and Evaluation, Washington, DC.
U.S. EPA (Environmental Protection Agency). 1990a. Guidance on Remedial Actions for Superfund Sites with PCB
Contamination. Office of Emergency and Remedial Response, Washington, DC.. OSWER Directive 9355.4-
01. EPA/540/G-90-007. NTIS PB91-1921206.
U.S. EPA (Environmental Protection Agency). 1990b. Sampler's Guide to the Contract Laboratory Program.
Office of Emergency and Remedial Response, Washington, DC. OSWER Directive 9240.0-06.
EPA/540/P/90-006. NTIS PB91-921330CDH.
U.S. EPA (Environmental Protection Agency). 1991a. Human Health Evaluation Manual (HHEM), Supplemental
Guidance: Standard Default Exposure Factors. Publication 9285.6-03. Office of Emergency and Remedial
Response, Washington, DC. NTIS PB91-921314.
U.S. EPA (Environmental Protection Agency). 1991b. Risk Assessment Guidance for Superfund (RAGS), Volume
1: Human Health EvaluationManual (HHEM), PartB, Development of Risk-BasedPreliminary Remediation
Goals. OSWER Directive No. 9287.01B. EPA/540/R-92/003. Publication 9285.7-01B. Office of
Emergency and Remedial Response, Washington, DC. NTIS PB92-963333.
U.S. EPA (Environmental Protection Agency). 1991c. Role of the Baseline Risk Assessment in Superfund Remedy
Selection Decisions. Publication 9355.0-30. Office of Emergency and Remedial Response, Washington, DC.
NTISPB91-921359/.
U.S. EPA (Environmental Protection Agency). 1992a. Estimating Potential for Occurrence ofDNAPL at Superfund
Sites. Office of Emergency and Remedial Response, Washington, DC. Publication 9355.4-07FS. NTIS
PB92-963338.
U.S. EPA (Environmental Protection Agency). 1992b. Technical Support Document for Land Application of Sewage
Sludge. Volumes I and 2. EPA 822/R-93-001 A, B, Office of Water, Washington, DC.
U.S. EPA (Environmental Protection Agency). 1992c. Handbook ofRCRA Ground-Water Monitoring Constituents:
Chemical and Physical Properties. EPA/530-R-92-022. Office of Solid Waste, Washington, DC. NTIS
PB92-233287
U.S. EPA (Environmental Protection Agency). 1992&.Preparation of Soil SamplingProtocols: Sampling Techniques
and Strategies. Office of Research and Development, Washington, DC. EPA/600/SR-92/128.
U.S. EPA (Environmental Protection Agency). 1993a. Background Document for EPA 's Composite Model for
Leachate Migration with Transformation Products, EPACMTP. Office of Solid Waste, Washington, D.C.
July.
U.S. EPA (Environmental Protection Agency). 1993b. Data Quality Objectives for Superfund: Interim Final
Guidance. OSWER Directive No. 9355.9-01. Publication 9255.9-01. Office of Emergency and Remedial
Response, Washington, DC. EPA 540-R-93-071. NTIS PB94-963203.
U.S. EPA (Environmental Protection Agency). 1993c. Quality Assurance for Superfund Environmental Data
Collection Activities. Quick Reference Fact Sheet. OSWER Directive No. 9200.2-16FS. Office of
Emergency and Remedial Response, Washington, DC. NTIS PB93-963273.
U.S. EPA (Environmental Protection Agency). 1993d. The Urban Soil Lead Abatement Demonstration Project. Vol
I: Integrated Report Review Draft. EPA/600/P-93/001a7. NTIS PB96-168-356.
R-4
-------
U.S. EPA (Environmental Protection Agency). 1994a. Revised Interim Soil Lead Guidance for CERCLA andRCRA
Corrective Action Facilities. Office of Solid Waste and Emergency Response, Washington, B.C. Directive
9355.4-12. EPA 540/7-94-043. NTIS PB94-963-282.
U.S. EPA (Environmental Protection Agency). 1994b. Framework for Assessing Ground Water Modeling
Applications. EPA-500-B-94-003. Resource Management and Information Staff, Office of Solid Waste and
Emergency Response, Washington, DC. NTIS PB95-125-142.
U.S. EPA (Environmental Protection Agency). 1994c. Ground Water Modeling Compendium, Second Edition, EPA-
5 00-B-94-004. Resource Management and Information Staff, Office of Solid Waste and Emergency
Response, Washington, DC. NTIS PB95-104-154.
U.S. EPA (Environmental Protection Agency). 1994d. Guidance for the Data Quality Objectives Process. EPA
QA/G-4. Quality Assurance Management Staff, Office of Research and Development, Washington, DC.
EPA-600-R-96-0/55.
U.S. EPA (Environmental Protection Agency). 1994e. Methods for Evaluating the Attainment of Cleanup
Standards—Volume 3: Reference-Based Standards for Soils and Solid Media. EPA 230-R-94-004.
Environmental Statistics and Information Division, Office of Policy, Planning, and Evaluation, Washington,
DC.
U.S. EPA (Environmental Protection Agency). 1994f Soil Screening Level Guidance. Quick Reference Fact Sheet.
Draft. Office of Emergency and Remedial Response, Washington, DC. OSWER Directive No. 9344.4-14FS.
NTIS PB93-963508.
U.S. EPA (Environmental Protection Agency). 1994g. Technical Background Document for Soil Screening
Guidance. Draft. EPA/540-R-94-106. Office of Emergency and Remedial Response, Washington, DC. NTIS
PB95-963532.
U.S. EPA (Environmental Protection Agency). 1994h. Radon Mitigation Standards. Office of Air and Radiation,
Washington, DC. EPA 402-R-93-078.
U.S. EPA (Environmental Protection Agency). 1994i. Radiation Site Cleanup Regulations: Technical Support
Document for the Development of Radiation Cleanup Levels for Soil - Review Draft. Office of Air and
Radiation, Washington, DC. EPA 402-R-96-011A.
U.S. EPA (Environmental Protection Agency). 1995a. Drinking Water Regulations and Health Advisories. Office
of Water, Washington, DC. EPA/822/R-95/001.
U.S. EPA, 1995b, EPA's Composite Model for Leachate Migration with Transformation Products, EPACMTP.
Background Document, Office of Solid Waste, Washington, DC.
U.S. EPA, 1995c. EPA's Composite Model for Leachate Migration with Transformation Products, EPACMTP,
User's Guide, Office of Solid Waste, Washington, DC.
U.S. EPA (Environmental Protection Agency). 1996a. Soil Screening Guidance: User's Guide. EPA/540/R-96/018.
Office of Emergency and Remedial Response, Washington, DC. OSWER Directive No. 9355.4-23. NTIS
PB96-963505.
U.S. EPA (Environmental Protection Agency). 1996b. Soil Screening Guidance: Technical Background Document.
OSWER Directive No. 9355.4-17A. EPA/540/R95/128. Office of Emergency and Remedial Response,
Washington, DC. NTIS PB96-963502.
R-5
-------
U.S. EPA (Environmental Protection Agency). 1996c. Guidance for Data Quality Assessment: Practical Methods
for Data Analysis, EPA/600/R-96/084, EPA QA/G-9, QA96 Version.
U.S. EPA (Environmental Protection Agency). 1996d. Reassessment of Radium and Thorium Soil Concentrations
and Annual Dose Rates, Office of Radiation and Indoor Air, Washington, DC.
U.S. EPA (Environmental Protection Agency). 1997a. Establishment of Cleanup Levels for CERCLA Sites with
Radioactive Contamination. OSWER Directive 9200.4-18. Office of Emergency and Remedial Response,
Washington, DC. EPA/540/7/97/003. NTIS PB97-963210.
U.S. EPA (Environmental Protection Agency). 1997b. Exposure Factors Handbook. Office of Research and
Development, Washington, DC. EPA/600/P-95/002Fa.
U.S. EPA (Environmental Protection Agency). \991c.Multi-AgencyRadiationSurveyandSiteInvestigationManual
(MARSSIM). Office of Radiation and Indoor Air, Washington, DC. EPA 402-R-97-016.
U.S. EPA (Environmental Protection Agency). 1998a. Use of Soil Cleanup Criteria in 40 CFR 192 as Remediation
Goals for CERCLA Sites. OSWER Directive 9200.4-25. EPA/540/7-97/032. Office of Emergency and
Remedial Response, Washington, DC. NTIS PB97 -963308.
U.S. EPA (Environmental Protection Agency). 1998b. Uranium Issues Workshop. Washington, DC. June 23-24,
1998.
U.S. EPA (Environmental Protection Agency). 1998c. Estimation of Infiltration Rate in the Vadose Zone:
Compilation of Simple Mathematical Models. Volume I. EPA/600-R-97-128a.
U.S. EPA (Environmental Protection Agency). 1999a. Health Risks from Low-Level Environmental Exposure to
Radionuclides, Federal Guidance Report No. 13,, Office of Radiation and Indoor Air, Washington, DC. EPA
402-R-90-001.
U.S. EPA (Environmental Protection Agency). 1999b, Understanding Variation in Partition Coefficient, Kd Values,
Volume I & Volume II, Office of Radiation and Indoor Air, EPA-402-R-99-004A&B. NTIS PB2000108438
andPB2000 108439
U.S. EPA (Environmental Protection Agency). 2000a. Remediation Goals for Radioactively Contaminated CERCLA
Sites Using the Benchmark Dose Cleanup Criteria in 10 CFR Part 40 Appendix A, I, Criterion 6(6). Office
of Emergency and Remedial Response, Washington, DC. OSWER Directive No. 9200.4-35P. NTIS
PB2000-104362.
U.S. EPA (Environmental Protection Agency). 2000b. Soil Screening Guidance: Users Guide. Office of Emergency
and Remedial Response, Washington, DC. OSWER Directive No. 9355.4-16A. NTIS PB2000 963307
VwderHeiide,P.M.l994.IdentificationandCompilationofUnsaturated/VadoseZoneModels.EPA/600/R-94/028.
R.S. Kerr Environmental Research Laboratory, Office of Research and Development, U.S. Environmental
Protection Agency, Ada, OK.
Van Genuchten, M. Th. 1985. Convective-Dispersive Transport of Solutes Involved in Sequential First- Order Decay
Reactions. Computers and Geosciences. 11(2): 129-147.
Van Wijnen, J.H., P. Clausing, and B. Brunekreef 1990. Estimated soil ingestion by children. Environ Research
51:147-162.
R-6
-------
Wierenga, P.J., R.G. Hills and D.B. Hudson. 1991. The Las Cruces Trench Site: Experimental results and one
dimensional flow predictions. Water Resour. Res. 27:2695-2705.
R-7
-------
APPENDIX A
Generic SSLs
Tables A. 1 and A.2 provide generic SSLs for 60 radionuclides in units of pCi/g and mg/kg respectively. These set
of generic SSLs are derived using default values in the standardized equations presented in Part 2 of this document.
Tables A.3 and A.4 provide a second set of generic SSLs. These were derived using the electronic version of the
standardized equations in Part 2 to account for radionuclide decay and ingrowth. The default values (listed in Table
A.5) are conservative and are likely to be protective for the majority of site conditions across the nation.
However, the generic SSLs are not necessarily protective of all known human exposure pathways, reasonable land
uses, or ecological threats. Thus, before applying generic SSLs at a site, it is extremely important to compare the
conceptual site model (see the User's Guide) with the assumptions behind the SSLs to ensure that the site conditions
and exposure pathways match those used to develop generic SSLs (see Parts 1 and 2 and Table A.5). If this
comparison indicates that the site is more complex than the SSL scenario, or that there are significant exposure
pathways not accounted for by the SSLs, then generic SSLs are not sufficient for a full evaluation of the site. A more
detailed site-specific approach will be necessary to evaluate the additional pathways or site conditions.
Generic SSLs are presented separately for major pathways of concern in both surface and subsurface soils. The
pathways include external radiation exposure, inhalation of fugitive dusts, ingestion of homegrown produce, direct
ingestion of soil, and migration to ground water.
The last two columns present SSLs for the migration to ground water pathway calculated using two different DAFs
(dilution-attenuation factors). The second to last column presents generic SSL values for the migration to ground
water pathway developed using a default DAF of 20 to account for natural processes that reduce contaminant
concentrations in the subsurface (see Section 2.6.4).
The last column contains the generic S SLs for the migration to ground water pathway developed assuming no dilution
or attenuation between the source and the receptor well (i.e, a DAF of 1). These values can be used at sites where
little or no dilution or attenuation of soil leachate concentrations is expected at a site (e.g., sites with shallow water
tables, fractured media, karst topography, or source size greater than 30 acres).
Generally, if an SSL is not exceeded for a pathway of concern, the user may eliminate the pathway or areas of the
site from further investigation. If more than one exposure pathway is of concern, the lowest SSL should be used.
Analysis of Effects of Source Size on Generic SSLs
A large number of commenters on the December 1994 Soil Screening Guidance for chemicals suggested that most
contaminated soil sources were 0.5 acre or less. Before changing this default assumption from 30 acres to 0.5 acre,
the Office of Emergency and Remedial Response (OERR) conducted an analysis of the effects of changing the area
of a chemically-contaminated soil source on generic S SLs calculated for the inhalation and migration to ground water
exposure pathways. This analysis includes:
An analysis of the sensitivity of SSLs to a change in source area from 30 acres to 0.5 acre
Mass-limit modeling results showing the depth of contamination for a 30-acre source that
corresponds to a 0.5-acre SSL
A-l
-------
All equations, assumptions, and model input parameters used in this analysis are consistent with those described in
Part 2 of the Technical Background Document for chemicals unless otherwise indicated. Chemical properties used
in the analysis are described in Part 5 of the Technical Background Document for chemicals.
In summary, the results of this analysis indicate that:
• The SSLs are not particularly sensitive to varying the source area from 30 acres to 0.5 acre. This
reduction in source area lowers SSLs for the inhalation pathway by about a factor of 2 and lowers
SSLs for the migration to ground water pathway by a factor of 2.9 under typical hydrogeologic
conditions.
• Half-acre SSLs calculated for 43 volatile and semivolatile contaminants using the infinite source
(i.e., steady-state) models correspond to mass-limit SSLs for a 30-acre source uniformly
contaminated to a depth of about 1 to 21 meters (depending on contaminant and pathway); the
average depth is 8 meters for the inhalation pathway (21 contaminants) and 11 meters for the
migration to ground water pathway (43 contaminants).
Sensitivity Analysis. For the inhalation pathway, source area affects the Q/C value (a measure of dispersion),
which directly affects the final SSL and is not chemical-specific. Higher Q/C values result in higher SSLs. As shown
in Table 2.4 (Section 2.3.2), the effect of area on the Q/C value is not sensitive to meteorological conditions, with
the ratio of a 0.5-acre Q/C to a 30-acre Q/C ranging from 1.93 to 1.96 over the 29 conditions analyzed. Decreasing
the source area from 30 acres to 0.5 acre will therefore increase inhalation SSLs by about a factor of 2.
For the migration to ground water pathway, source area affects the DAF, which also directly affects the final SSLs
and is not chemical-specific. The sensitivity analysis for the dilution factor is more complicated than for Q/C because
increasing source area (expressed as the length of source parallel to ground water flow) not only increases infiltration
to the aquifer, which decreases the dilution factor, but also increases the mixing zone depth, which tends to increase
the dilution factor. The first effect generally overrides the second (i.e., longer sources have lower dilution factors)
except for very thick aquifers (see Section 2.6.5).
The sensitivity analysis described in Section 2.6.5 shows that the dilution model is most sensitive to the aquifer's
Darcy velocity (i.e., hydraulic conductivity x hydraulic gradient). For a less conservative Darcy velocity (90th
percentile), decreasing the source area from 30 acres to 0.5 acre increased the dilution factor by a factor of 3.1 (see
Table 2.8, Section 2.6.5). For the conditions analyzed, decreasing the source area from 30 acres to 0.5 acre affected
dilution factor from no increase to a factor of 4.3 increase. No increase in dilution factor for a 0.5-acre source was
observed for the less conservative (higher) aquifer thickness (46 m). In this case the decrease in mixing zone depth
balances the decrease in infiltration rate for the smaller source.
Mass-Limit Analysis. The infinite source (i.e., steady-state) assumption is one of the more conservative
assumptions inherent in the SSL models, especially for small sources. This assumption should provide adequate
protection for sources with larger areas than those used to calculate SSLs. To test this hypothesis the SSL mass-limit
models (Section 2.7) were used to calculate, for 43 volatile and semivolatile chemicals, the depth at which a mass-
limit SSL for a 30-acre source is equal to a 0.5 acre infinite-source SSL.
The mass limit model is a simple mass-balance model that calculate SSLs based on the conservative assumption that
the entire mass of contamination in a source leaches (migration to ground water model) over the exposure period of
interest. This model was developed to correct the mass-balance violation in the infinite source (i.e., steady-state)
models for highly soluble contaminants.
A-2
-------
References
U.S. EPA (Environmental Protection Agency). 1990. Guidance on Remedial Actions for Sites with PCB
Contamination. Office of Solid Waste and Emergency Response, Washington, DC. NTIS PB91-921206CDH.
U.S. EPA (Environmental Protection Agency). 1994. Revised Interim Soil Lead Guidance for CERCLA Sites and
RCRA Corrective Action Facilities. Office of Solid Waste and Response, Washington, DC. Directive 9355.4-12.
A-3
-------
Table A.1. Generic (no accounting for decay) SSLs for Radionuclides (pCi/g)
Radionuelide
Ac-227+D
Ag-108m
Ag-llOm
Am-241
Am-243+D
Bi-207
C-14
Cd-109
Ce-144+D
Cl-36
Cm-243
Cm-244
Co-57
Co-60
Cs-134
Cs-135
Cs-137+D
Eu-152
Eu-154
Eu-155
Fe-55
Gd-153
H-3
1-129
K-40
Mn-54
Na-22
Nb-94
Ni-59
Ni-63
Np-237+D
Pa-231
Pb-210+D
Pm-147
Pu-238
Pu-239
Pu-240
Pu-241
Pu-242
Pu-244+D
Ra-226+D
Ra-228+D
Ru-106+D
Sb-125+D
Sm-147
Sm-151
Sr-90+D
IngestiOn of
Homegrown
Produce
8.62E-01
8.38E-01
6.85E-01
1.05E+01
9.92E+00
1.73E+00
1.28E-01
7.00E-01
1.36E+01
1.59E-02
1.14E+01
1.30E+01
1.18E+01
7.89E-01
6.85E-01
5.99E+00
9.41E-01
6.47E+01
3.78E+01
2.03E+02
1.21E+03
2.54E+02
4.51E+00
2.19E-01
1.37E-01
1.51E+00
2.23E+00
1.27E+01
7.24E+01
2.96E+01
7.74E-01
6.23E-01
4.09E-02
2.27E+02
8.33E+00
8.09E+00
8.09E+00
6.18E+02
8.53E+00
7.41E+00
6.83E-02
2.46E-02
7.68E-01
1.95E+01
1.18E+01
6.98E+02
4.92E-02
., Direct
Ingestion of
Soil . '•
6.84E-01
4.13E+01
3.35E+01
3.66E+00
3.42E+00
5.33E+01
2.84E+02
6.96E+01
7.78E+00
1.04E+02
3.87E+00
4.39E+00
2.86E+02
1.97E+01
1.37E+01
1.11E+02
1.83E+01
4.90E+01
2.78E+01
1.47E+02
3.80E+02
1.86E+02
8.58E+03
2.93E+00
1.28E+01
1.54E+02
4.03E+01
3.87E+01
1.08E+03
4.43E+02
4.90E+00
2.12E+00
2.98E-01
1.63E+02
2.92E+00
2.88E+00
2.87E+00
2.41E+02
3.02E+00
2.53E+00
1.09E+00
3.47E-01
6.67E+00
6.01E+01
1.05E+01
4.99E+02
5.51E+00
Inhalation
ofFugitiv« .
••Dusts
8.69E+01
6.80E+05
6.42E+05
6.46E+02
6.73E+02
8.65E+05
2.57E+06
8.29E+05
1.65E+05
7.26E+05
6.75E+02
7.18E+02
8.69E+06
5.07E+05
1.10E+06
9.76E+06
1.53E+06
2.00E+05
1.58E+05
1.23E+06
2.27E+07
2.77E+06
3.23E+08
2.99E+05
1.76E+06
3.09E+06
4.67E+06
4.82E+05
3.90E+07
1.11E+07
1.03E+03
3.99E+02
1.31E+03
1.13E+06
5.40E+02
5.45E+02
5.45E+02
5.44E+04
5.80E+02
6.20E+02
1.57E+03
3.47E+03
1.78E+05
9.41E+05
2.64E+03
3.72E+06
1.61E+05
External _ "'
Radiation
: Exposure. .
7.59E-02
1.55E-02
8.58E-03
4.04E+00
1.75E-01
1.58E-02
1.43E+04
1.28E+01
4.57E-01
6.41E+01
2.66E-01
2.30E+03
3.14E-01
9.00E-03
1.57E-02
4.73E+03
4.38E-02
2.11E-02
1.91E-02
9.00E-01
a
6.89E-01
a
1.83E+01
1.40E-01
2.87E-02
1.08E-02
1.53E-02
a
a
1.40E-01
8.03E-01
2.65E+01
3.48E+03
1.55E+03
5.58E+02
1.60E+03
2.72E+04
1.79E+03
7.39E-02
1.31E-02
2.46E-02
1.16E-01
6.17E-02
a
3.10E+05
5.69E+00
•Migration to 'Ground _Watef
: M DAF ' ': 1 DAF
b b
3.36E-OF 1.68E-02C
5.22E+00 2.61E-01
2.52E+00 1.26E-01
2.52E+00 1.26E-01
b b
4.00E+01 2.00E+00
3.48E+01 1.74E+00
2.11E+01 1.06E+00
b b
2.59E+01 1.29E+00
2.59E+01 1.29E+00
6.00E+00 3.00E-01
6.00E-01 3.00E-02
1.63E+01 8.16E-01
1.84E+02 9.18E+00
4.08E+01 2.04E+00
b b
b b
b b
1.32E+02 6.60E+00
b b
8.00E+01 4.00E+00
4.60E-03 2.30E-04
b b
3.06E+01 1.53E+00
b b
b b
2.05E+02 1.03E+01
3.42E+01 1.71E+00
9.00E-02 4.50E-03
b b
6.70E-03 3.35E-04
b b
1.56E+00 7.80E-02
1.56E+00 7.80E-02
1.56E+00 7.80E-02
2.81E+00C 1.40E-01C
1.56E+00 7.80E-02
1.56E+00 7.80E-02
3.20E-01 1.60E-02
3.20E-01 1.60E-02
3.12E+00 1.56E-01
b b
b b
b b
1.92E-01 9.60E-03
A-4
-------
Table A.1. Generic (no accounting for decay) SSLs for Radionuclides (pCi/g)
Radionuclide
Tc-99
Th-228+D
Th-229+D
Th-230
Th-232
Tl-204
U-232
U-233
U-234
U-235+D
U-236
U-238+D
Zn-65
IngettiOn of
Homegrown
Produce
7.04E-02
3.34E+00
1.97E+00
1.18E+01
1.06E+01
8.53E-01
1.46E+00
5.81E+00
5.90E+00
5.77E+00
6.24E+00
4.65E+00
2.29E-01
., Direct
Ingestion of
Soil . '•
1.04E+02
9.81E-01
6.15E-01
3.93E+00
3.44E+00
5.15E+01
1.38E+00
4.96E+00
5.02E+00
4.87E+01
5.33E+00
3.78E+00
3.24E+01
Inhalation
offugitiv* .
•••Dusts
1.29E+06
1.27E+02
8.07E+01
6.37E+02
4.19E+02
7.41E+06
9.31E+02
.57E+03
.59E+03
.80E+03
.73E+03
.94E+03
3.13E+06
External _ "'
Radiation
: Exposure. .
1.37E+03
1.44E-02
9.54E-02
1.36E+02
3.26E+02
4.04E+01
1.87E+02
1.14E+02
4.43E+02
2.06E-01
8.93E+02
9.79E-01
3.97E-02
•Migration to 'Ground _Watef
: M DAF ' ': 1 DAP
3.73E+00 1.86E-01
6.06E+00 3.03E-01
6.06E+00 3.03E-01
6.06E+00 3.03E-01
6.06E+00 3.03E-01
b b
2.40E-01d 1.20E-02d
2.40E-01d 1.20E-02d
2.40E-01d 1.20E-02d
2.40E-01d 1.20E-02d
2.40E-01d 1.20E-02d
2.40E-01d 1.20E-02d
1.80E+00 9.00E-02
— * Properties for this radionuclide are such that this pathway is not a concern at any soil concentration.
— b SSL cannot be calculated since a default Kj has not been specified for this radionuclide.
— ° SSL calculated based on risk based limit for this radionuclide.
—d SSL calculated based on proposed MCL of 20 pCi/1 (activity) for uranium.
A-5
-------
Table A.2. Decay Corrected SSLs for Radionuclides (pCi/g)
Hadionuelide
Ac-227+D
Ag-108m
Ag-llOm
Am-241
Am-243+D
Bi-207
C-14
Cd-109
Ce-144+D
Cl-36
Cm-243
Cm-244
Co-57
Co-60
Cs-134
Cs-135
Cs-137+D
Eu-152
Eu-154
Eu-155
Fe-55
Gd-153
H-3
1-129
K-40
Mn-54
Na-22
Nb-94
Ni-59
Ni-63
Np-237+D
Pa-231
Pb-210+D
Pm-147
Pu-238
Pu-239
Pu-240
Pu-241
Pu-242
Pu-244+D
Ra-226+D
Ra-228+D
Ru-106+D
Sb-125+D
Sm-147
Sm-151
Sr-90+D
Tc-99
Th-228+D
Ingeition of
Homegrown
Produce
1.34E+00
9.09E-01
2.08E+01
1.08E+01
9.93E+00
2.25E+00
1.28E-01
1.15E+01
3.62E+02
1.59E-02
1.61E+01
2.19E+01
3.31E+02
3.17E+00
6.91E+00
5.99E+00
1.30E+00
1.28E+02
9.86E+01
8.65E+02
9.35E+03
7.96E+03
9.29E+00
2.19E-01
1.37E-01
3.67E+01
1.79E+01
1.27E+01
7.24E+01
3.29E+01
7.74E-01
6.23E-01
6.29E-02
1.80E+03
9.36E+00
8.10E+00
8.10E+00
1.17E+03
8.53E+00
7.41E+00
6.88E-02
9.15E-02
1.60E+01
1.47E+02
1.18E+01
7.82E+02
6.89E-02
7.04E-02
3.63E+01
Direct
Ingestion of
Soil
1.06E+00
4.48E+01
1.02E+03
3.75E+00
3.43E+00
6.92E+01
2.85E+02
1.14E+03
2.08E+02
1.04E+02
5.45E+00
7.38E+00
7.99E+03
7.92E+01
1.38E+02
1.11E+02
2.54E+01
9.69E+01
7.26E+01
6.26E+02
2.93E+03
5.84E+03
1.77E+04
2.93E+00
1.28E+01
3.76E+03
3.22E+02
3.87E+01
1.08E+03
4.93E+02
4.90E+00
2.12E+00
4.59E-01
1.29E+03
3.28E+00
2.88E+00
2.87E+00
4.56E+02
3.02E+00
2.53E+00
1.09E+00
1.29E+00
1.39E+02
4.52E+02
1.05E+01
5.59E+02
7.71E+00
1.04E+02
1.07E+01
Inhalation
• of Fugitive ' ;
: Dusts
1.35E+02
7.37E+05
1.95E+07
6.62E+02
6.74E+02
1.12E+06
2.57E+06
1.36E+07
4.41E+06
7.26E+05
9.51E+02
1.21E+03
2.43E+08
2.04E+06
1.11E+07
9.76E+06
2.12E+06
3.95E+05
4.12E+05
5.22E+06
1.75E+08
8.69E+07
6.66E+08
2.99E+05
1.76E+06
7.51E+07
3.73E+07
4.82E+05
3.90E+07
1.23E+07
1.03E+03
3.99E+02
2.01E+03
8.95E+06
6.07E+02
5.46E+02
5.46E+02
1.03E+05
5.80E+02
6.20E+02
1.58E+03
1.29E+04
3.70E+06
7.07E+06
2.64E+03
4.17E+06
2.25E+05
1.29E+06
1.38E+03
', External ' ; "
Radiation
: Exposure
1.18E-01
1.68E-02
2.61E-01
4.14E+00
1.76E-01
2.05E-02
1.43E+04
2.09E+02
1.22E+01
6.41E+01
3.75E-01
3.87E+03
8.80E+00
3.62E-02
1.59E-01
4.73E+03
6.07E-02
4.16E-02
4.99E-02
3.83E+00
a
2.16E+01
a
1.83E+01
1.40E-01
6.98E-01
8.67E-02
1.53E-02
a
a
1.40E-01
8.03E-01
4.08E+01
2.76E+04
1.74E+03
5.58E+02
1.60E+03
5.13E+04
1.79E+03
7.39E-02
1.32E-02
9.15E-02
2.40E+00
4.63E-01
a
3.47E+05
7.97E+00
1.37E+03
1.57E-01
Migration to Ground Water
20DAF, - •• 1DAF. '"
b b
4.0E-OT 2.0E-02C
1.6E+02 7.9E+00
2.6E+00 1.3E-01
2.5E+00 1.3E-01
b b
4.0E+01 2.0E+00
5.7E+02 2.8E+01
5.6E+02 2.8E+01
b b
3.6E+01 1.8E+00
4.3E+01 2.2E+00
1.7E+01 8.4E+00
2.4E+00 1.2E-01
1.6E+02 8.2E+00
1.8E+02 9.2E+00
5.7E+01 2.8E+00
b b
b b
b b
l.OE+03 5.1E+01
b b
1.6E+02 8.2E+00
4.6E-03 2.3E-04
b b
7.4E+02 3.7E+01
b b
b b
2.1E+02 l.OE+01
3.8E+01 1.9E+00
9.0E-02 4.5E-03
b b
1.1E-02 5.5E-04
b b
.8E+00 8.8E-02
.6E+00 7.8E-02
.6E+00 7.8E-02
1.0E+01C 5.0E-01C
.6E+00 7.8E-02
.6E+00 7.8E-02
3.2E-01 1.6E-02
1.2E+00 5.9E-02
6.4E+01 3.2E+00
b b
b b
b b
2.7E-01 1.3E-02
3.7E+00 1.9E-01
6.6E+01 3.3E+00
A-6
-------
Table A.2. Decay Corrected SSLs for Radionuclides (pCi/g)
Kadionuelide
Th-229+D
Th-230
Th-232
Tl-204
U-232
U-233
U-234
U-235+D
U-236
U-238+D
Zn-65
Ingeition of
Homegrown
Produce
1.97E+00
1.18E+01
1.06E+01
4.71E+00
1.68E+00
5.81E+00
5.90E+00
5.77E+00
6.24E+00
4.65E+00
7.11E+00
Direct
Ingestion of
Soil
6.16E-01
3.93E+00
3.44E+00
2.85E+02
1.59E+00
4.96E+00
5.02E+00
4.87E+01
5.33E+00
3.78E+00
1.01E+03
Inhalation
• of Fugitive ' ;
: Dusts
8.08E+01
6.37E+02
4.19E+02
4.09E+07
1.07E+03
1.57E+03
1.59E+03
1.80E+03
1.73E+03
1.94E+03
9.72E+07
', External ' ; "
Radiation
: Exposure
9.55E-02
1.36E+02
3.26E+02
2.23E+02
2.15E+02
1.14E+02
4.43E+02
2.06E-01
8.93E+02
9.79E-01
1.24E+00
Migration to Ground Water
20DAF - •• IDA!- '"
6.1E+00 3.0E-01
6.1E+00 3.0E-01
6.1E+00 3.0E-01
b b
2.8E-01d .4E-02d
2.4E-01d .2E-02d
2.4E-01d .2E-02d
2.4E-01d .2E-02d
2.4E-01d .2E-02d
2.4E-01d .2E-02d
5.6E+01 2.8E+00
— a Properties for this radionuclide are such that this pathway is not a concern at any soil concentration.
— b SSL cannot be calculated since a default Kj has not been specified for this radionuclide.
— ° SSL calculated based on risk based limit for this radionuclide.
—d SSL calculated based on proposed MCL of 20 pCi/1 (activity) for uranium.
A-7
-------
Table A.3. Generic (no accounting for decay) SSLs for Radionuclides (mg/kg)
Radiottuelide
Ac-227+D *
Ag-108m *
Ag-llOm*
Am-241
Am-243+D *
Bi-207
C-14
Cd-109
Ce-144+D *
Cl-36
Cm-243
Cm-244
Co-57
Co-60
Cs-134
Cs-135
Cs-137+D *
Eu-152
Eu-154
Eu-155
Fe-55
Gd-153
H-3
1-129
K-40
Mn-54
Na-22
Nb-94
Ni-59
Ni-63
Np-237+D *
Pa-231
Pb-210+D *
Pm-147
Pu-238
Pu-239
Pu-240
Pu-241
Pu-242
Pu-244+D*
Ra-226+D*
Ingestion of
Homegrown
Produce
1.20E-08
3.22E-08
1.45E-10
3.06E-06
4.98E-05
3.81E-08
2.86E-08
2.72E-10
4.26E-09
4.81E-07
2.22E-07
1.61E-07
1.40E-09
6.99E-10
5.29E-10
5.20E-03
1.08E-08
3.66E-07
1.43E-07
4.38E-07
5.05E-07
7.21E-08
4.69E-10
1.24E-03
1.96E-02
1.95E-10
3.58E-10
6.78E-05
8.97E-04
5.01E-07
1.10E-03
1.32E-05
5.37E-10
2.45E-07
4.87E-07
1.31E-04
3.56E-05
6.00E-06
2.17E-03
4.18E-01
6.92E-08
•. Direct . .-
Ingestion
of Soil
9.48E-09
1.59E-06
7.07E-09
1.07E-06
1.72E-05
1.17E-06
6.36E-05
2.70E-08
2.44E-09
3.14E-03
7.51E-08
5.42E-08
3.38E-08
1.74E-08
1.06E-08
9.61E-02
2.11E-07
2.77E-07
1.06E-07
3.16E-07
1.58E-07
5.29E-08
8.94E-07
1.66E-02
1.84E+00
2.00E-08
6.45E-09
2.07E-04
1.34E-02
7.51E-06
6.96E-03
4.50E-05
3.91E-09
1.75E-07
1.71E-07
4.64E-05
1.26E-05
2.34E-06
7.69E-04
1.43E-01
1.10E-06
Inhalation
of Fugitive
" .Dusts
1.20E-06
2.61E-02
1.35E-04
1.88E-04
3.38E-03
1.90E-02
5.74E-01
3.22E-04
5.18E-05
2.20E+01
1.31E-05
8.88E-06
1.03E-03
4.49E-04
8.51E-04
8.49E+03
1.76E-02
1.13E-03
5.99E-04
2.64E-03
9.45E-03
7.87E-04
3.37E-02
1.70E+03
2.53E+05
3.99E-04
7.48E-04
2.57E+00
4.83E+02
1.88E-01
1.46E+00
8.47E-03
1.71E-05
1.22E-03
3.16E-05
8.80E-03
2.40E-03
5.28E-04
1.48E-01
3.50E+01
1.59E-03
; , External
Radiation
: Exposure, •
1.05E-09
5.96E-10
1.81E-12
1.18E-06
8.81E-07
3.47E-10
3.18E-03
4.96E-09
1.43E-10
1.95E-03
5.16E-09
2.85E-05
3.73E-11
7.97E-12
1.21E-11
4.11E+00
5.04E-10
1.19E-10
7.26E-11
1.94E-09
a
1.96E-10
a
1.04E-01
2.01E-02
3.71E-12
1.74E-12
8.18E-08
a
a
1.99E-04
1.70E-05
3.48E-07
3.75E-06
9.03E-05
9.00E-03
7.03E-03
2.64E-04
4.55E-01
4.17E-03
1.33E-08
. Migration to Ground Water
: 20DAF i 1DAF'
b
1.28E-08C
1.10E-09
7.39E-07
1.26E-05
b
9.00E-06
1.33E-08
6.41E-09
b
5.00E-07
3.28E-07
7.20E-10
5.34E-10
1.26E+14
1.59E-01
4.69E-07
b
b
b
5.48E+12
b
8.40E-09
2.28E-05
b
3.98E-09
b
b
2.53E-03
5.81E-07
1.26E-04
b
8.80E-11
b
9.15E-08
2.50E-05
6.86E-06
2.70E-08C
3.95E-04
8.84E-02
3.26E-07
b
6.38E-10C
5.51E-11
3.70E-08
6.30E-07
b
4.50E-07
6.67E-10
3.20E-10
b
2.50E-08
1.64E-08
3.60E-11
2.67E-11
6.32E+12
7.96E-03
2.35E-08
b
b
b
2.74E+11
b
4.20E-10
1.31E-06
b
1.99E-10
b
b
1.27E-04
2.91E-08
6.30E-06
b
4.40E-12
b
4.58E-09
1.25E-06
3.43E-07
1.35E-09C
1.98E-05
4.42E-03
1.63E-08
-------
Table A.3. Generic (no accounting for decay) SSLs for Radionuclides (mg/kg)
Radionuclide
Ra-228+D*
Ru-106+D*
Sb-125+D*
Sm-147
Sm-151
Sr-90+D*
Tc-99
Th-228+D*
Th-229+D*
Th-230
Th-232
Tl-204
U-232
U-233
U-234
U-235+D*
U-236
U-238+D*
uranium
Zn-65
Ingestion of
Homegrown
Produce
9.04E-11
2.28E-10
1.89E-08
5.15E+02
2.66E-05
3.61E-10
4.16E-06
4.07E-09
9.26E-06
5.87E-04
9.70E+01
1.84E-09
6.84E-08
6.03E-04
9.47E-04
2.67E+00
9.64E-02
1.39E+01
2.78E-11
•. Direct . .-
Ingestion
Off Soil
1.27E-09
1.98E-09
5.83E-08
4.56E+02
1.90E-05
4.04E-08
6.12E-03
1.20E-09
2.90E-06
1.95E-04
3.15E+01
1.11E-07
6.47E-08
5.15E-04
8.06E-04
2.26E+01
8.24E-02
1.13E+01
3.94E-09
Inhalation
of Fugitive
" .Dtlsts
1.27E-05
5.28E-05
9.12E-04
1.15E+05
1.42E-01
1.18E-03
7.60E+01
1.55E-07
3.80E-04
3.16E-02
3.84E+03
1.60E-02
4.36E-05
1.62E-01
2.56E-01
8.33E+02
2.67E+01
5.79E+03
3.80E-04
; , External
Radiation
: Exposure, .
9.04E-11
3.43E-11
5.98E-11
a
1.18E-02
4.18E-08
8.09E-02
1.75E-11
4.49E-07
6.76E-03
2.99E+03
8.73E-08
8.73E-06
1.18E-02
7.11E-02
9.52E-02
1.38E+01
2.92E+00
4.83E-12
. Migration to Ground. Water
: 20DAF i IDA!'
1.15E-09
9.36E-10
b
b
b
1.42E-09
2.19E-04
7.27E-09
2.87E-05
2.99E-04
5.66E+01
b
1.13E-08d
2.52E-05d
3.84E-05d
1.12E-01d
3.72E-03d
7.20E-01d
2.40E-01e
2.16E-10
5.76E-11
4.68E-11
b
b
b
7.08E-11
1.10E-05
3.64E-10
1.43E-06
1.49E-05
2.83E+00
b
5.64E-10d
1.26E-06d
1.92E-06d
5.58E-03d
1.86E-04d
3.60E-02d
1.20E-02e
1.08E-11
*Note: The "+D" weight based concentrations only account for the concentration of the first isotope in a series and does not represent the total radioactivity.
— * Properties for this radionuclide are such that this pathway is not a concern at any soil concentration.
— b SSL cannot be calculated since a default Kj has not been specified for this radionuclide.
— ° SSL calculated based on risk based limit for this radionuclide.
—d SSL calculated based on proposed MCL of 20 pCi/1 (activity) for uranium.
—' SSL calculated based on proposed MCL of 20 ug/1 (mass) for uranium.
A-9
-------
Table A.4. Decay Corrected SSLs for Radionuclides (mg/kg)
Radionueiide
Ac-227+D *
Ag-108m *
Ag-llOm*
Am-241
Am-243+D *
Bi-207
C-14
Cd-109
Ce-144+D *
Cl-36
Cm-243
Cm-244
Co-57
Co-60
Cs-134
Cs-135
Cs-137+D *
Eu-152
Eu-154
Eu-155
Fe-55
Gd-153
H-3
1-129
K-40
Mn-54
Na-22
Nb-94
Ni-59
Ni-63
Np-237+D *
Pa-231
Pb-210+D *
Pm-147
Pu-238
Pu-239
Pu-240
Pu-241
Pu-242
Pu-244+D*
Ra-226+D*
Ingestion of
Homegrown
Produce
1.85E-08
3.49E-08
4.39E-09
3.14E-06
4.99E-05
4.95E-08
2.87E-08
4.44E-09
1.14E-07
4.81E-07
3.13E-07
2.71E-07
3.92E-08
2.81E-09
5.34E-09
5.20E-03
1.50E-08
7.25E-07
3.74E-07
1.86E-06
3.89E-06
2.26E-06
9.68E-10
1.24E-03
1.96E-02
4.74E-09
2.86E-09
6.78E-05
8.97E-04
5.58E-07
1.10E-03
1.32E-05
8.25E-10
1.94E-06
5.47E-07
1.31E-04
3.56E-05
1.13E-05
2.17E-03
4.18E-01
6.97E-08
; , 'Direct ,' ' ,
: Ingestion ,
of Soil
1.47E-08
1.72E-06
2.14E-07
1.09E-06
1.72E-05
1.52E-06
6.37E-05
4.42E-07
6.52E-08
3.14E-03
1.06E-07
9.12E-08
9.47E-07
7.01E-08
1.07E-07
9.61E-02
2.92E-07
5.48E-07
2.76E-07
1.35E-06
1.22E-06
1.66E-06
1.84E-06
1.66E-02
1.84E+00
4.85E-07
5.16E-08
2.07E-04
1.34E-02
8.35E-06
6.96E-03
4.50E-05
6.02E-09
1.39E-06
1.92E-07
4.64E-05
1.26E-05
4.43E-06
7.69E-04
1.43E-01
1.11E-06
Inhalation
Of Fugitive
• ' Dusts '
1.87E-06
2.83E-02
4.11E-03
1.93E-04
3.38E-03
2.47E-02
5.75E-01
5.26E-03
1.38E-03
2.20E+01
1.84E-05
1.49E-05
2.88E-02
1.81E-03
8.59E-03
8.49E+03
2.43E-02
2.23E-03
1.56E-03
1.12E-02
7.28E-02
2.47E-02
6.94E-02
1.70E+03
2.53E+05
9.71E-03
5.98E-03
2.58E+00
4.83E+02
2.09E-01
1.46E+00
8.47E-03
2.63E-05
9.66E-03
3.55E-05
8.80E-03
2.40E-03
9.98E-04
1.48E-01
3.50E+01
1.60E-03
; , External . • •
Radiation
• Exposure
1.63E-09
6.46E-10
5.50E-11
1.21E-06
8.82E-07
4.51E-10
3.19E-03
8.11E-08
3.83E-09
1.95E-03
7.28E-09
4.79E-05
1.04E-09
3.21E-11
1.23E-10
4.11E+00
6.98E-10
2.36E-10
1.89E-10
8.25E-09
a
6.14E-09
a
1.04E-01
2.01E-02
9.02E-11
1.39E-11
8.18E-08
a
a
1.99E-04
1.70E-05
5.34E-07
2.98E-05
1.01E-04
9.00E-03
7.04E-03
4.99E-04
4.55E-01
4.17E-03
1.34E-08
Migration to Groundwater
20- DAI .- '. " IDAf
b b
1.4E-08C 7.1E-10C
1.1E-09 5.5E-11
7.4E-07 3.7E-08
1.3E-05 6.3E-07
b b
9.0E-06 4.5E-07
1.3E-08 6.7E-10
6.4E-09 3.2E-10
b b
5.0E-07 2.5E-08
3.3E-07 1.6E-08
7.2E-10 3.6E-11
5.3E-10 2.7E-11
1.3E-08 6.3E-10
1.6E-01 8.0E-03
4.7E-07 2.3E-08
b b
b b
b b
5.5E-08 2.7E-09
b b
8.4E-09 4.2E-10
2.6E-05 1.3E-06
b b
4.0E-09 2.0E-10
b b
b b
2.5E-03 1.3E-04
5.8E-07 2.9E-08
1.3E-04 6.3E-06
b b
9.4E-11 4.7E-12
b b
9.2E-08 4.6E-09
2.5E-05 1.2E-06
6.9E-06 3.4E-07
5.2E-08C 2.6E-09C
4.0E-04 2.0E-05
8.8E-02 4.4E-03
3.3E-07 1.6E-08
A-10
-------
Table A.4. Decay Corrected SSLs for Radionuclides (mg/kg)
Radionueiide
Ra-228+D*
Ru-106+D*
Sb-125+D*
Sm-147
Sm-151
Sr-90+D*
Tc-99
Th-228+D*
Th-229+D*
Th-230
Th-232
Tl-204
U-232
U-233
U-234
U-235+D*
U-236
U-238+D*
Zn-65
Ingestion of
Homegrown
Produce
3.36E-10
4.74E-09
1.42E-07
5.15E+02
2.97E-05
5.05E-10
4.16E-06
4.43E-08
9.27E-06
5.87E-04
9.70E+01
1.02E-08
7.88E-08
6.03E-04
9.47E-04
2.67E+00
9.64E-02
1.39E+01
8.65E-10
; , 'Direct ,' ' ,
: Ingestion ,
of Soil
4.73E-09
4.12E-08
4.38E-07
4.56E+02
2.13E-05
5.66E-08
6.12E-03
1.30E-08
2.90E-06
1.95E-04
3.15E+01
6.15E-07
7.45E-08
5.15E-04
8.06E-04
2.26E+01
8.24E-02
1.13E+01
1.23E-07
Inhalation
Of Fugitive
• ' Dusts '
4.74E-05
1.10E-03
6.85E-03
1.15E+05
1.59E-01
1.65E-03
7.60E+01
1.69E-06
3.80E-04
3.16E-02
3.84E+03
8.84E-02
5.01E-05
1.62E-01
2.56E-01
8.33E+02
2.67E+01
5.79E+03
1.18E-02
; , External . • •
Radiation
• Exposure
3.36E-10
7.13E-10
4.49E-10
a
1.32E-02
5.84E-08
8.10E-02
1.91E-10
4.50E-07
6.76E-03
2.99E+03
4.82E-07
l.OOE-05
1.18E-02
7.11E-02
9.52E-02
1.38E+01
2.92E+00
1.50E-10
Migration to Groundwater
20- DAI .- '. " IDAf
1.2E-09 5.8E-11
9.4E-10 4.7E-11
b b
b b
b b
1.4E-09 7.1E-11
2.2E-04 1.1E-05
7.3E-09 3.6E-10
2.9E-05 1.4E-06
3.0E-04 1.5E-05
5.7E+01 2.8E+00
b b
LlE-08d 5.6E-06d
2.5E-05d 1.2E-06d
3.9E-05d 1.9E-03d
l.lE-08d 5.6E-03d
3.7E-03d 1.9E-04d
7.1E-01d 3.6E-02d
2.2E-10 1.1E-11
*Note: The "+D" weight based concentrations only account for the concentration of the first isotope in a series and does not represent the total radioactivity.
— a Properties for this radionuclide are such that this pathway is not a concern at any soil concentration.
— b SSL cannot be calculated since a default Kj has not been specified for this radionuclide.
— ° SSL calculated based on risk based limit for this radionuclide.
—d SSL calculated based on proposed MCL of 20 pCi/1 (activity) for uranium.
—' SSL calculated based on proposed MCL of 20 ug/1 (mass) for uranium
A-ll
-------
Table A.5. Generic SSLs: Default Parameters and Assumptions
SSL Pathway
Parameter
External Inhalation Ingestion of Migration to
Radiation of Fugitive Homegrown Ground
Exposure Dust Produce Water
Default
Source Characteristics
Continuous vegetative cover
Roughness height
Source area (A) O
Source length (L)
Source depth O
O
O
O
O
50 percent
0.5 cm for open terrain; used to
derive Ut7
0.5 acres (2,024 m2); used to
derive ACF for EXT, CPF for
IMP, and L for MTG,
45 m (assumes square source)
Extends to water table (i.e., no
attenuation in unsaturated zone)
for MTG
Soil Characteristics
Soil texture
Dry soil bulk density (pb)
Soil porosity (n)
Vol. soil water content (Ow)
Soil pH
Mode soil aggregate size
Threshold windspeed @ 7m
(Ut.7)
O
O
O
O Loam; defines soil
characteristics/parameters
• 1.5 kg/L, used to calculate ACF
for EXT
O 0.43
• 0.15 (INH); 0.30 (MTG)
O 6.8; used to determine pH-
specific Kd
0.5 mm; used to derive Ut7
11.32m/s
Meteorological Data
Mean annual windspeed (
Air dispersion factor (Q/C)
Fugitive particulate Q/C
4.69 m/s (Minneapolis, MN)
90th percentile conterminous
U.S.
90.801; Minneapolis, MN; 0.5-
acre source
Hydrogeologic Characteristics (DAF)
Hydrogeologic setting
Dilution/attenuation factor
(DAF)
O Generic (national); surficial
aquifer
• 20
• Indicates parameters used in the SSL equations.
O Indicates parameters/assumptions needed to estimate SSL equation parameters.
EXT = External radiation exposure
INH = Inhalation pathway
IMP = Ingestion of homegrown produce
MTG = Migration to ground water pathway
A-12
-------
APPENDIX B
Sample Measurement Units, Activity and Mass
Introduction
Typically units of decay rate instead of mass are used to quantify the concentration of radioactive material
in soil because the carcinogenic risks of exposure to soils contaminated with radioactive materials are related
more to the decay rate of the material than to its mass. For example, one gram of 226Ra has a decay rate
(activity) of 3.7x1010 transformations (also referred to as disintegrations) per second, while one gram of 137Cs
has a decay rate of 3.2xl012 transformations per second. Since it is the energy emitted by the radioactive
material during radioactive decay and the frequency of the decay that is usually of public health concern, and
generally not the chemical properties of the radioactive material, it is more meaningful for health assessment
purposes to quantify radioactive material according to decay rate. In addition, radioactive materials are
detected and quantified by the type of radiation emitted and number of disintegrations (per unit time), not
by their unique chemistry, as is the case for non-radioactive material. For these reasons, the concentration
of radioactive material in soil is typically expressed in units of decay rate, pCi/g.
When and How to Calculate Mass
The equations presented in this TBD and the User Guide calculate SSLs for radionuclides in the traditional
units of pCi/g. Mass units, however, provide insight and information into treatment selection, treatment
compatibility, and treatment effiency, particularly for remedial actions involving mixed waste. For example,
remediation goals expressed in mass are important for designing and evaluating treatment technologies such
as soil separation, pump and treat, as well as subsurface barriers. Typically units for expressing mass in
environmental media for soil and water are mg/kg for soil and mg/1 for water. These mass units also can be
expressed as parts per million (ppm) for soil and water, which is equivalent to mg/kg and mg/1. SSLs in
pCi/g may be converted to mg/kg, and MCLs in pCi/1 may be converted to their mass equivalent in mg/1, by
the following equations:
SSL(mglkg) = 2.8x1 O^2 *A * Tm * SSL(pCi/g)
MCL(mgll) = 2.8x10 l5 *A * T , *MCL(pCi/l)
where 2.8xlO~12 for soil or 2.8xlO~15 for water is a conversion factor, A is the radionuclide atomic weight in
g/mole, and T1/2 is the radionuclide half-life in years. To put the relationship between SSLs expressed in
these units into perspective, examine the SSL in mg/kg corresponding to a SSL of 1 pCi/g for a long-lived
radionuclide such as 238U and a relatively short-lived one such as 60Co. 238U has a half-life of 4.5 lxlO+9 y,
so a 1 pCi/g SSL would be equivalent to 3 mg/kg. On the other hand, a 1 pCi/g SSL for 60Co, which has a
half-life of 5.26 y, is equivalent to about IxlO"9 mg/kg. As noted in Attachment C, Table C.I, most
B-l
-------
radionuclides, for which generic SSLs have been calculated, have half-lives ranging from a few years to
10,000 years. Appendix A in the TBD shows that most SSLs for radionuclides are in fact less than 1 pCi/g
so the equivalent SSL mg/kg values are even smaller. Therefore, at SSL levels, the masses of most
radionuclides are extremely small values. Attachment D in the User Guide also provides the mass equivalent
values for the MCLs.
Background Information on Using Mass
One important issue associated with using mass to characterize the quantities of radioactive material in the
environment is that many elements, such as uranium, have several isotopes of the same element. For
example, if one were to perform atomic absorption analysis of a water sample, and it revealed the presence
of 1 mg/kg of uranium, there would be no way of knowing how much of the uranium in the sample is 238U,
234U, or 235U, all of which are present in the environment naturally and due to anthropogenic activities. The
potential public health and environmental impact of a given concentration of uranium in the environment will
depend on the specific isotopes of uranium that are present, which could vary considerably depending on
whether we are dealing with naturally-occurring uranium or uranium that may have been enriched in 235U
as part of the uranium fuel cycle or part of weapons production. It is also important to note that the same
mass of each uranium isotope has significantly different levels of radioactivity. A mass of 1 mg/kg of 238U
has an activity of 0.33 pCi/g, while the same mass of 235U has 2.1 pCi/g and 6,200 pCi/g of 234U.
Also, many radioactive elements are present in the environment along with their stable counterpart. One
example is potassium, which is naturally- occurring in the environment, ranging from 0.1 to 1% in limestones
to 3.5% in granite. In addition, a typical 70 kg adult contains 130 g of potassium. A very small fraction
(0.01%) of this potassium is the naturally-occurring radioactive isotope 40K. If one were to measure the
amount of 40K in soil and assume that 40K made up all of the elemental potassium, the mass of the elemental
potassium would be underestimated by 10,000 fold.
Since the potential adverse effects of radioactive material are due to its disintegration rate, measurement of
the mass of a given element present may not accurately represent the amount of radioactivity present and,
therefore, its potential radio-toxicity.
Use of Mass in Remediation and Technology Selection
The measurement of the radioactivity present often will be a misrepresentation of the total mass of the given
element and should not be used alone to calculate the treatment required for remediation technologies, since
technologies are essentially chemical/physical. Doing so may underestimate the total mass of the given
element and lead to errors in the amount of treatment or reactants required for remediation since technologies
are chemically/physically based. For example, to design and implement a subsurface Permeable Reaction
Wall for the Uranium isotopes described above, it would be necessary to know the total mass of the Uranium
isotopes as well as the other aqueous reactive elements to calculate the equivalent amounts of sorption or
precipitation reactants that would be required to remove or reduce the aqueous Uranium species from
contaminated groundwater (EPA 2000a & EPA 1999a). The same considerations would be necessary for
other ground water or water treatment technologies for dissolved concentrations of elements and their
isotopic forms. For example in a pump and treat ground water extraction system that utilizes ion exchange
(chemical separation) or reverse osmosis (physical separation), chemical mass measurements would be used
to determine the amount and type of reactants materials, exchange capacity and effectiveness (EPA 1996).
Much the same can be said for mobility limiting or mobility reduction technologies such as chemical
B-2
-------
solidification/stabilization treatability studies or treatments (EPA, 2000b). Also, mass measurements are
important in the determination of Partition coefficients, Kd values that are essential in fate and transport, risk
assessment modeling, and remediation calculations. Kd values are expressed in mass units for the inorganic
elements and isotopes (EPA 1999b). Partition coefficients, Kd values, are the same value for all forms of the
element and isotopes.
In summary, given that risk or exposure is the basis for remedial actions, mass measurements are often
required for determining, designing and selecting a remediation technology. This contrasts with the need
for radiation specific isotopic measurements required in risk and exposure analysis. Users should note the
different applications and perspectives with their corresponding measurements units of mass and activity.
References
U.S.EPA. 1996. Technology Screening Guide for Radioactively'-Contaminated Sites. EPA/402/R-96/017.
Office of Radiation and Indoor, Washington, DC.
U.S.EPA. 1999a. DRAFT, Field Demonstration of Permeable Reactive Barriers To Remove Dissolved
Uranium from Ground Fry Canyon, Utah. Office of Radiation and Indoor Air, Washington, DC.
U.S.EPA. 1999b. Understanding Variation In Partition Coefficient, Kd Values. Volume 1 &2. EPA/402/R-
99-004A PB2000 108438 and EPA/402/R-99-004A PB2000 108439. Office of Radiation and Indoor Air.
U.S.EPA. 2000a. In Situ Permeable Reactive barriers: Application and Deployment. Training Manual.
EPA542/B-00/001. Office of Solid Waste and Emergency Response, Washington, DC.
U.S.EPA.2000b. Testing Stabilization/Solidification processes For Mixed Waste. EPA/402/R/00/008. Office
of Radiation and Indoor Air, Washington, DC.
B-3
------- |