Guidance for Sample Collection for In Vitro Bioaccessibility Assay for Lead (Pb) in Soil
United States OSWER 9200.3-1 00
Environmental
Protection Agency
j Guidance for Sample Collection for In Vitro
Bioaccessibility Assay for Lead (Pb) in Soil
March 2015
-------
Table of Contents
1. Terminology and Application 1
2. Procedure 1
3. Sampling Materials and Handling 4
4. Quality Assurance/Quality Control 5
5. Health and Safety 5
6. References 6
Appendix A 8
-------
Guidance for Sample Collection for In Vitro Bioaccessibility Assay for Lead (Pb) in Soil
1.0 Terminology and Application
Definitions:
Absolute Unavailability (ABA): Fraction of an ingested dose of lead that is absorbed from
the gastrointestinal tract and enters the blood and tissues.
Relative Unavailability (RBA): Ratio of the absolute bioavailability of lead in soil to that of
a water soluble reference lead compound (lead acetate).
In Vitro Bioaccessibility (IVBA): Fraction of total amount of lead in a soil sample that is
soluble in a gastric-like (i.e., low pH) extraction medium.
The purpose of this document is to provide guidance on the collection of soil samples for measurement of
lead IVBA (SW-846, Method 1340) (U.S. EPA, 2013c). The IVBA assay is used as a rapid and
inexpensive method for predicting soil lead RBA (U.S. EPA, 2007c). Estimates of lead RBA are used to
adjust bioavailability parameters in lead risk assessment models used in site risk assessment (e.g.,
Integrated Exposure Uptake Biokinetics [IEUBK] model for Lead in Children) (U.S. EPA, 2013a). Soil
lead RBA is dependent on physical and chemical properties of the lead in soil and co-occurring elements
at any particular site or location within a given site. As a result, site-specific estimates of soil lead RBA
that provide representative coverage of the site are recommended for increasing confidence in estimates
of risk related to site-specific lead exposures. Sampling plans for estimating soil lead RBA with the IVBA
assay should provide a statistically robust estimate of RBA for decision units at the site. Typically, this
can be achieved by measuring IVBA in a statistical subsample of soils collected as part of the sampling
plan for estimating exposure point concentrations (EPCs) for soil lead. This guidance provides
recommendations for data collection requirements, sampling material handling, QA/QC requirements,
and health and safety requirements for assessments of site-specific soil lead RBA with the IVBA assay.
2.0 Procedure
Data Collection Requirements: A sampling plan for a site should be developed that considers potential
soil exposure pathways for the site and any existing site data; for example, if the site is a residential area,
then evaluation of exposure pathways in children's play areas, gardens, and the drip lines of homes should
be given special attention (U.S. EPA, 2003a). If existing sampling data are available for a site, the
information could assist in targeting the sampling locations where there is likely exposure to these
contaminated areas.
Typically soil samples are collected, submitted for metals analysis, and the samples are archived while
data are collected and reviewed. Based on the analytical results, a subset of the samples is selected for
IVBA assay. At other sites, sample locations could be identified in the sampling plan and IVBA samples
collected and analyzed without previous knowledge of lead concentrations at the site, although total metal
analysis should be collected and conducted concurrently.
X-ray fluorescence (XRF) could be used to screen samples in the field because there is significant cost
saving related to time and financial resources by eliminating the collection of samples that do not meet a
prior criteria for IVBA analysis. There are many advantages of field screening for lead and other metals
-------
including a reduction of both laboratory and field work. Soils with little to no metals are not collected,
shipped, or processed by laboratory staff. Large fluctuations in soil lead concentrations within a site when
determined by XRF in the field could be used as justification for collection of additional samples in order
to form composites samples in the laboratory. The use of the XRF would allow samplers to immediately
collect additional samples which may not be possible following laboratory analyses. Field screening with
XRF therefore reduces the turnaround time required to generate IVBA results and reduces the need for
additional field deployments as well as generating much less waste (fewer sample reduces shipping cost,
processing time, number of analyses, and analytical waste). Field operators of portable XRF instruments
should ensure they are following appropriate protocols to obtain reliable results (SW-846, Method 6200,
U.S. EPA, 2007b).
When collecting samples for in vitro bioaccessibility assay, it is important to note site and sample
medium characteristics that may indicate differences in the bioavailability of the lead or indicate that
interferences might be present. For example, the lead IVBA assay (SW-846, Method 1340) may not
reliably predict RBA of lead in soils that have been amended with phosphate (Scheckel et al., 2013). If
phosphate at a site is a concern, it would be worthwhile to analyze the samples for the phosphate
concentration. When collecting soils from residential properties it may not be advisable to make
composite soil samples from a garden (potentially fertilized with phosphorus) with the surrounding
property. Likewise it may not be advisable to composite soil samples from the drip line of a home
(possible source of lead contaminated paint) with the remainder of the property (potentially different lead
source).
In addition to total metals analysis and IVBA assay, the samples might also be submitted for lead
speciation analysis and animal bioavailability studies. Lead speciation analysis is meant to determine the
exact chemical form(s), or species, of lead in a sample, as opposed to the total lead concentration.
Speciation analysis may be informative in explaining variability in IVBA across the site, identifying
sources of contamination of the soil, and assessing the potential mobility of lead in the soil. While IVBA
assay is meant to be a faster and less expensive alternative to in vivo animal bioavailability studies, there
may be cases (such as potential interference from soil amendment applications [e.g., phosphate], untested
lead phases, etc.) when the animal study would be necessary. It is important to ensure that sufficient
material is collected for each sample so that additional analyses could be performed. If additional analyses
are determined to be necessary, such as lead speciation analysis or in vivo animal bioavailability studies,
consultation with the Technical Review Workgroup (TRW) Lead Committee is recommended.
Prior to sampling, a determination must be made as to whether the soil is regulated or quarantined by the
U.S. Department of Agriculture (USDA) Animal and Plant Health Inspection Service (APHIS)/Plant
Protection and Quarantine (PPQ) (USDA, 2014). Take special care to segregate regulated or quarantined
soil samples from the non-regulated or non-quarantined samples. To determine if the soils collected are
regulated or quarantined contact the State Plant Health Director
(http://www.aphis.usda.gov/wps/portal/aphis/ourfocus/planthealth? ldmy&urile=wcm%3apath%3a%2Fap
his content library%2Fsa our focus%2Fsa plant health%2Fsa program overview%2Fct sphd).
-------
Number of Samples: The number of samples to collect and analyze for FVBA will depend on the Data
Quality Objectives (DQO) for the study. Factors that should be considered in estimating the number of
samples include:
• goals of the RBA assessment;
• size and characteristics of the decision units at the site;
• expected variability in RBA within decision units, based on available data or bounding
assumptions (U.S. EPA, 2007d); and
• acceptable limits on decision errors.
Project managers should consult with U.S. EPA "Guidance on Systematic Planning Using the Data
Quality Objectives Process" or other appropriate guidance when developing DQOs (U.S. EPA, 2006).
In general, sample size estimates for RBA assessments can be based on the same types of power analyses
used to evaluate statistical hypotheses in estimating EPCs at decision units (DUs) (see Appendix A). To
reduce the cost of analyzing numerous discrete samples, an incremented sampling plan may be a cost
effective approach (ITRC, 2012).
Sampling Depth: The appropriate sampling depth for a site will depend on the expected exposure
pathway for a site. For most scenarios involving exposure to contaminated surface soil, EPA recommends
a sampling depth of the top 0-1 inches of soil below organic litter and sod for lead exposure analysis
(http://www.epa.gov/superfund/lead/ieubkfaq.htm). With this rather shallow sample depth it could be
challenging to obtain sufficient sample mass for discrete samples especially if the material is particularly
course. Incremental composite sampling can provide larger masses for shallow samples. If there are other
exposure scenarios for a site, other sampling depth intervals that would represent these scenarios should
be collected.
Sample Preparation: To help ensure that sufficient sample material is available for analysis, the field
samplers should consider sieving the material in the field to remove larger debris. Sieve screens No. 4
(4.72 mm) or No. 10 (2.0 mm) would be sufficient for removing larger debris in the field.
Sample Mass: For metals analysis, SW-846 recommends that a minimum of 200 g of soil be collected
and 2 g of sample be used for the digestions (SW-846, Chapter 3 Inorganic Analytes, Table 3-2, U.S.
EPA, 2007a). Method 1340 specifies that 1 g of dried and sieved soil sample be used for IVBA assay for
lead for a single replicate (U.S. EPA, 2013c). Additional replicates may be required if the assay does not
meet performance specifications for FVBA. The amount of sample required will depend on the particle
size distribution of the soil and the moisture content of the soil following course sieving in the field. If the
samples will be submitted for animal bioavailability studies or speciation analysis, the laboratories that
will be conducting these analyses should be consulted on the amount of sample materials they require to
determine the sample mass needed. For further assistance in determining the sample mass for in vivo
bioavailability and in vitro bioaccessibility assays, please contact the TRW Lead Committee.
Selection of Samples for IVBA: As stated previously, samples for FVBA assay can be designated as part
of the sampling plan for estimating EPCs, or they can be selected based on the results from XRF field
screening or total metals analysis. The strategy used to select samples for IVBA assay from XRF results
or total metals data will depend on the intended use of the IVBA data. If the intended use is for screening,
it may be appropriate to select only those samples that have lead concentrations exceeding the risk-based
-------
concentrations used in screening. If the FVBA data are to be used to estimate risk for the site or a DU at
the site, a representative statistical subsample should be selected. Samples selected for IVBA assay should
have a total lead concentration less than 50,000 mg/kg (SW-846, Method 1340). If the in vitro
bioaccessibility assay needs to be performed on a sample with a concentration greater than 50,000 mg/kg,
the lab performing the assays should be informed of the samples concentrations so that the amount of soil
used in the IVBA assay can be adjusted to be within the appropriate lead concentration range.
3.0 Sampling Materials and Handling
Sample Containers: The analytical laboratory/program that will be conducting the metals analysis
should be consulted about the appropriate sample container and size required. For the in vitro
bioaccessibility assay there are no specific sample container requirements. If no sample container is
specified by the metals analysis laboratory, then appropriate containers include glass jars, wide-mouth
HDPE jars, plastic zippered bags, or any other container that is clean and free of contaminants can be
used. A single one-gallon plastic zippered bag should provide sufficient sample material for at least the
metals analysis and in vivo bioaccessibility assay for most soils. Two-gallon plastic zippered bags may be
required for sandy soils and soils with rocks passing through the sieve in the field. If using wide-mouth
HDPE jars, a 1000-mL jar should provide sufficient sample, but collect multiple jars per sample if the soil
is particularly course. There will be considerable cost reduction using plastic zippered bags compared to
HDPE bottle (both cost of sample containers and shipping).
Sampling Equipment: Collection of surface soil samples may be accomplished with a stainless steel
cylindrical punch which will capture a constant diameter core for the sampling depth of interest. Sampling
using a kick-style cylindrical punch may reduce sample time in the field due to the ease of use. Kick-style
punches are not recommended for sandy soils because the soil readily falls out of the probe. Likewise
soils with heavy clay content or rocks are not recommended due to the difficulty in removing clay soils
from punch and rocky soil will be rejected at the soil surface. For these reasons using plastic or stainless
steel spades, trowels, or spoons may be preferable but the sampler should ensure that a sample is collected
evenly across the sampling depth. Once the samples are collected, they should be placed in suitable
containers for shipment. Any equipment that is not disposable should be thoroughly decontaminated and
appropriately stored after sampling. If the exposure pathway being investigated requires deeper sampling
depths than 0-1 inches, equipment such as augers, split spoon samplers, and backhoes may be necessary
(U.S. EPA, 2000). If sampling at depth, care should be taken during sampling to account for any soil
compaction as a result of sampling.
Labeling, Shipping and Storage Temperature, and Hold Times: Sample ID numbering, labeling,
documentation, and chain of custody should follow the requirements of the analytical laboratory/program
that will be conducting the metals analysis. The samples may be shipped at ambient temperature unless
specified otherwise by the analytical laboratory/program.
EPA recommends a hold time of 6 months for metals samples. EPA 9200.2-86 recommends that all
samples be archived after metal analysis and retained for further analysis, including in vivo bioavailability
assay, for 6 months (U.S. EPA, 2012). The samples may be stored at ambient temperature unless
specified otherwise by the analytical laboratory/program.
Laboratory Sample Preparation: Once in the lab, the samples should be blended and completely dried
at <40°C in an air-drying oven for approximately 5 days to a constant mass. After drying, any clumps in
the sample should be gently broken and then fine sieved. However, samples should not be ground by ball
-------
mill or any other grinding method which could result in reduction in the particle sizes of the collected
soils.
To ensure composite samples are representative of all of the component locations, the entire composite
sample should be processed (i.e. dried and fine sieved). Following sieving, each sample should be
thoroughly mixed using ASTM standard D6051-96 (2006) or ITRC Incremental Sampling Methodology
(2012) and then transferred to a suitable storage container (U.S. EPA, 2013b).
Total metals analysis and other analyses should be conducted on the same dried, sieved, and homogenized
sample material that will also be used for the in vitro bioaccessibility assay. To split a sample into
equivalent aliquots for the different analyses, the processed soil should be passed through a riffle splitter
and the aliquots collected in clean, 250 ml high-density polyethylene bottles (U.S. EPA, 2003b). Samples
that have been dried and sieved can be submitted for total metals analysis, metals speciation, IVBA assay,
and in vivo animal bioavailability studies but should not be used for analysis of other contaminants of
concern.
4.0 Quality Assurance/Quality Control
The field samplers should consult with the metals analysis laboratory/EPA program to determine in
advance the requirements for blanks, duplicates, and matrix spikes for the metals analysis samples. For
the IVBA assay, Method 1340 does not require field blanks, field duplicates, or matrix spikes to be
prepared or collected by field samplers. Material for the matrix spike and duplicates for Method 1340 can
be taken from the samples at the laboratory's discretion and will not require that samplers collect and
designate separate matrix spike and duplicates in the field.
Samplers should take thorough field notes and should retain any photographs taken, logbooks, and notes
following the sampling event. The field group should make note of any differences in the media between
the sample locations and indicate if there is any potential interferents (i.e., phosphate amended soils)
present.
5.0 Health and Safety
When working with potentially hazardous materials, follow U.S. EPA, Occupational Safety and Health
Administration (OSHA), and any contractor's corporate health and safety procedures, in addition to the
procedures specified in the site-specific Health and Safety Plan.
-------
6.0 References
ASTM (American Society for Testing and Materials). 2006. ASTM D6051-96. Standard Guide for
Composite Sampling and Field Subsampling for Environmental Waste Management Activities. Book of
Standards Volume: 11.04. ASTM International: West Conshohocken, PA. Available online:
http://www.astm.org/Standards/D6051 .htm.
ITRC (Interstate Technology and Regulatory Council). 2012. Incremental Sampling Methodology.
ISM-1. Interstate Technology and Regulatory Council, Incremental Sampling Methodology Team:
Washington, DC. February. Available online: www.itrcweb.org/ism-l/pdfs/ISM-l 021512_Final.pdf.
Scheckel, K. G., Diamond, G., Maddaloni, M., Partridge, C., Serda, S., Miller, B. W., Klotzbach, J., and
Burgess, M. 2013. Amending Soils with Phosphate as Means to Mitigate Soil Lead Hazard: A Critical
Review of the State of the Science. J. Toxicol. Environ. Health B 16(6): 337-380.
USDA (U.S. Department of Agriculture). 2014. State Plant Health Directors. U.S. Department of
Agriculture, Animal and Plant Health Inspection Service: Washington, DC. Available online:
http://www.aphis.usda.gov.
U.S. EPA (U.S. Environmental Protection Agency). 1996. Soil Screening Guidance: User's Guide. U.S.
Environmental Protection Agency, Office of Superfund Remediation and Technology Innovation:
Washington, DC. OSWER 9355.4-23. July. Available online:
http://www.epa.gov/superfund/health/conmedia/soil/pdfs/ssg496.pdf.
U.S. EPA (U.S. Environmental Protection Agency). 2000. Standard Operating Procedure: Soil Sampling.
U.S. Environmental Protection Agency, Environmental Response Team: Washington, DC. SOP #2012.
February. Available online: http://www.epa.gov/region9/toxic/noa/eldorado/pdf/EPA-ERT-SOIL-SOP-
2012.pdf.
U.S. EPA (U.S. Environmental Protection Agency). 2003a. Superfund Lead-Contaminated Residential
Sites Handbook. U.S. Environmental Protection Agency, Office of Superfund Remediation and
Technology Innovation: Washington, DC. OSWER 9285.7-50. August. Available online:
http://epa.gov/superfund/lead/products/handbook.pdf.
U.S. EPA (U.S. Environmental Protection Agency). 2003b. Guidance for Obtaining Representative
Laboratory Analytical Subsamples from Particulate Laboratory Samples. U.S. Environmental Protection
Agency, Office of Research and Development: Washington, DC. EPA/600/R-03/027. November.
Available online: http://www.epa.gov/esd/cmb/pdf/guidance.pdf.
U.S. EPA (U.S. Environmental Protection Agency). 2006. Guidance on Systematic Planning Using the
Data Quality Objectives Process. EPA QA/G-4. U.S. Environmental Protection Agency, Office of
Environmental Information: Washington, DC. EPA/240/B-06/001. February. Available online:
http://www.epa.gov/QUALITY/qs-docs/g4-final.pdf
U.S. EPA (U.S. Environmental Protection Agency). 2007a. SW-846. Chapter Three - Inorganic Analytes.
Revision 4. U.S. Environmental Protection Agency, Office of Resource Conservation and Recovery:
Washington, DC. February. Available online:
http://www.epa.gov/wastes/hazard/testmethods/sw846/pdfs/chap3.pdf
-------
U.S. EPA (U.S. Environmental Protection Agency). 2007b. Method 6200 - Field Portable X-ray
Fluorescence Spectrometry for the Determination of Elemental Concentrations in Soil and Sediment. SW-
846, Update VI. Revision 0. U.S. Environmental Protection Agency, Office of Resource Conservation
and Recovery: Washington, DC. February. Available online:
http://www.epa.gov/osw/hazard/testmethods/sw846/pdfs/6200.pdf.
U.S. EPA (U.S. Environmental Protection Agency). 2007c. Estimation of Relative Bioavailability of Lead
in Soil and Soil-Like Materials Using In Vivo and/« Vitro Methods. U.S. Environmental Protection
Agency, Office of Superfund Remediation and Technology Innovation: Washington, DC. OSWER
9285.7-77. May. Available online: http://www.epa.gov/superfund/bioavailability/lead_tsd.pdf.
U.S. EPA (U.S. Environmental Protection Agency). 2007d. Guidance for Evaluating the Oral
Bioavailability of Metals in Soils for Use in Human Health Risk Assessment. U.S. Environmental
Protection Agency, Office of Superfund Remediation and Technology Innovation: Washington, DC.
OSWER 9285.7-80. May. Available online:
http://www.epa.gov/superfund/bioavailability/bio_guidance.pdf.
U.S. EPA (U.S. Environmental Protection Agency). 2012. Standard Operating Procedure for an In Vitro
Bioaccessibility Assay for Lead in Soil. U.S. Environmental Protection Agency, Office of Superfund
Remediation and Technology Innovation: Washington, DC. EPA 9200.2-86. April. Available online:
http://www.epa.gov/superfund/bioavailabilitv/pdfs/EPA Pb IVBA SOP 040412 FINAL SRC.pdf
U.S. EPA (U.S. Environmental Protection Agency). 2013a. Frequent Questions from Risk Assessors on
the Integrated Exposure Uptake Biokinetic (IEUBK) Model. U.S. Environmental Protection Agency,
Office of Superfund Remediation and Technology Innovation: Washington, DC. Available online:
http://www.epa.gov/superfund/lead/ieubkfaq.htmtfrecommendations.
U.S. EPA (U.S. Environmental Protection Agency). 2013b. The Roles of Project Managers and
Laboratories in Maintaining the Representativeness of Incremental and Composite Soil Samples. U.S.
Environmental Protection Agency, Office of Solid Waste and Emergency Response: Washington, DC.
OSWER 9200.1-117FS. June. Available online: http://www.clu-
in.org/download/char/RolesofPMsandLabsinSubsampling.pdf.
U.S. EPA (U.S. Environmental Protection Agency). 2013c. Method 1340 - In Vitro Bioaccessibility
Assay for Lead in Soil. SW-846, Update VI. Revision 0. U.S. Environmental Protection Agency, Office
of Resource Conservation and Recovery: Washington, DC. November. Available online:
http://www.epa.gov/osw/hazard/testmethods/sw846/pdfs/1340.pdf
-------
Appendix A
Guidance for Sample Collection for In Vitro Bioaccessibility Assay for Lead (Pb) in Soil
Objectives:
Predicting the minimum sample number needed to estimate the RBA-adjusted mean soil Pb concentration
involves evaluating and setting limits on the probability of two types of errors. We define the null
hypothesis as:
H0: RBA-adjusted mean soil Pb concentration > Risk based concentration (RBC)
And the alternate as:
Ht: RBA-adjusted mean soil Pb concentration < Risk based concentration (RBC)
A Type 1 error occurs if we reject H0 when it is true; we conclude that the RBA-adjusted mean soil Pb
concentration is less than the RBC, when it is actually greater than the RBC. A Type I error could result
in underestimating risk at the site.
A Type 2 error occurs if we accept H0 when it is false; we conclude that the RBA-adjusted mean soil Pb
concentration exceeds the RBC, when it is actually less than the RBC. A Type 2 error could result in
overestimating risk at the site.
The objective of a sample number assessment is to identify sample numbers that are expected to satisfy
specified requirements for Type 1 and Type 2 error rates. These error rates depend on several factors:
• the difference between the mean soil Pb concentration and the RBC;
• the variability in the soil Pb concentration; and
• the mean and variability of the soil RBA at the site.
Larger sample numbers will be required to achieve a given error rate when the actual RBA-adjusted mean
soil Pb is closer to the RBA, or when variability (i.e., standard deviation) of the soil Pb concentrations or
RBA at the site are higher.
Assumptions for calculating sample number:
An example of sample number calculation is presented here. Assumptions in the analysis are as follows:
1. The underlying distribution of measured Pb concentrations in discrete soil samples at the decision
unit (DU) is lognormal (the incremental-composite sampling [ICS] design should collect
adequate samples to ensure a normal distribution of the concentrations of multiple composites).
2. Distribution of measured RBA within a DU is normal (e.g., single source of Pb contamination
and uniform soil characteristics). An analysis of FVBA measurements of soil RBA at 10 different
sites, at which multiple IVBA measurements were made (range: 12-86 discrete samples per site),
showed that the average coefficient of variation (standard deviation/mean) was 13% (range: 5-
22) (update to EPA TRW, 2003 that included data from Bunker Hill).
3. The RBA-adjusted mean soil Pb concentration for the DU is:
Mean RBA
Adjusted Mean PbSoil = Mean PbSoil • •
0.8
-------
where PbS is the soil Pb concentration and 0.8 is the default values for RBA in the IEUBK
Model.
4. For evaluating Type 1 error, we assume that the RBA-adjusted mean soil Pb concentration at the
DU exceeds the RBC. For evaluating Type 2 error, we assume that the RBA-adjusted mean soil
Pb concentration at the DU is below the RBC.
5. An acceptable Type 1 error rate is 5% (i.e., the probability of concluding that the RBA-adjusted
mean soil Pb concentration is less than the RBC, when it is actually greater than the RBC, is
equal to or less than 5%).
6. An acceptable Type 2 error rate is 20% (i.e., the probability of concluding that the RBA-adjusted
mean soil Pb concentration is greater than the RBC, when it is actually less than the RBC, is
equal to or less than 20%). We are typically less concerned about a Type 2 error than a Type 1
error (overestimating risk) than a Type 1 (underestimating risk).
7. The ICS design consists of «=C composites are collected at the DU, each composite consisting of
«=I increments, and «=R composites are randomly selected for FVBA analysis.
8. The estimated mean soil Pb concentration for the DU is the mean of measured Pb concentrations
of «=C composites.
9. The estimated mean RBA for the DU is based on the mean of measured IVBA of «=R
composites.
10. Values assumed for soil Pb concentration, RBC, and RBA for evaluating Type 1 and Type 2 error
rates are presented in Table A-l.
A Monte Carlo Simulation (MCS) was used to estimate Type 1 and Type 2 error rates. The MCS
consisted of 10,000 random draws from soil Pb concentration and RBA distributions (see Table A-l) and
calculation of 10,000 corresponding values for the mean RBA-adjusted soil Pb concentration. The Type 1
error rate is the number of means that are less than the RBC (divided by 10,000) when the assumed (true)
concentration equals or exceeds the RBC (see Figure A-l). The Type 2 error rate is the number of means
that are greater than or equal to the RBC (divided by 10,000), when the true mean is less than the RBC.
Predictions:
Type 1 and Type 2 error rates for various ICS designs are presented in Table A-2. The single composite
design is equivalent to a discrete sampling design with 7=n discrete samples per DU. The estimated
probability distribution of the RBA-adjusted mean soil Pb concentration for the sampling design C=3,
1=20, and R=l is shown in Figure A-l. A plot of the Type 1 error rates corresponding to various
combinations of C, I, and R is shown in Figure A-2.
As noted previously, error rates depend on the values selected for the various parameters listed in Table
A-l. This is illustrated in Figure A-3 which shows the probability of rejecting H0 as a function of
increasing mean RBA-adjusted soil Pb concentration for a design in which 3 composites of 30 increments
each are collected. When the mean soil Pb concentration is well below 400 ppm (<200 ppm), the
probability of rejecting H0 is 100% (Type 1 error = 0). Similarly, when it is well above 400 ppm
(>600 ppm) the probability of rejecting H0 is 0% (Type 2 error = 0). However, at a soil Pb concentration
of 500 ppm, the probability of rejecting H0 is 5%, even though the mean exceeds the 400 ppm RBA
(Type 1 error = 5%).
-------
Figure A-3 also shows the effect of variability in RBA on the Type 1 error rate. Three coefficients of
variation are shown (0.15, 0.30, 0.50). If the coefficient of variation is 0.50 (RBA=0.6±0.30), rather than
0.15 (RBA=0.6±0.09), the Type 1 error rate at a 500 ppm mean soil concentration increases from 5% to
18%. In order to decrease the Type 1 error rate to an acceptable 5%, the number of increments in each of
the 3 composites would have to increase from 20 to 60. If the coefficient of variation is 0.30
(RBA=0.6±0.18), a 5% Type 1 error rate can be achieved with 25 increments in each of 3 composites.
Conclusions:
1. If the mean RBA-adjusted soil Pb concentration is 500±500 ppm and the mean soil Pb RBA is
0.60±0.09, an acceptable Type 1 error (5%) is predicted with:
a. 1 composite made up of 60 increments;
b. 2 composites made up of 30 increments; or
c. 3 composites made up of 20 increments.
2. If 3 composites of 20 increments are collected, RBA assessment of a single randomly selected
composite would yield an acceptable Type 1 error rate. A minimum of 30 increments has been
recommended (ITRC, 2012).
3. Higher variability in RBA will require a larger number of increments per composite to achieve an
acceptable Type 1 error rate.
a. If the RBA coefficient of variation is 0.30 (RBA=0.60±0.18), 25 increments would be
needed per composite.
b. If the RBA coefficient of variation is 0.50 (RBA=0.60±0.0.30), 60 increments would be
needed per composite.
4. A larger number of increments will be needed if the actual mean soil Pb concentration is closer to
the RBC, and fewer will be needed if the actual mean Pb concentration is further from the RBC.
5. In general, for most risk assessment applications, acceptable Type I error rate can be expected if
ITRC (2012) recommendations are followed (30 increments per composite).
10
-------
TABLE A-l. Parameter Values for Sample Number Calculation
Parameter
Soil Pb RBC (ppm)
Mean RB A-adjusted soil Pb
concentration (ppm)a
Mean RBC-adjusted soil Pb
standard deviation (ppm)
Mean soil RB A
Soil RB A standard deviation
False Negative
Assessment
400
500
500b
0.60
0.09C
False Positive
Assessment
400
300
300b
0.60
0.09C
Basis
OSWER screening level corresponding
to P10=5% (approximately)
Assumption Type 1 error = RBC x 1.25
Assumption Type 2 error = RBC x 0.75
CV=1 for Bunker Hill soil (or dust)
Site-wide median (U.S. EPA OSRTI
TRW)
CV=0. 15, based on median CV for 1 1
IEUBK model default RBA
0.80
0.80
sites (TRW: Estimation of Lead
Bioavailability in Soil and Dust: Update
to the Default Values for the Integrated
Exposure Uptake Biokinetic Model for
Lead in Children (11/02/11) plus Bunker
Hill(CV=0.11)
IEUBK Model
aRBA-adjusted mean soil Pb=Mean soil Pb x mean soil RBA/0.8, where 0.8 is the IEUBK Model default soil RBA
bSoil Pb distribution: lognormal (mean, SD).
°RBA distribution Normal (mean, SD, min, max), with min=0, max=l.
TABLE A-2. Error Rates for ICS Designs
Number of
Composites for
Pb Analysis
(C)a
1
1
1
1
1
1
2
2
2
2
3
3
3
3
3
3
3
3
Number of
Composites for
IVBA Analysis
(R)
i
i
i
i
i
i
2
2
2
2
3
3
3
3
1
1
1
1
Number of
Increments per
Composite
(I)
10
20
40
50
60
80
10
20
30
40
5
10
20
30
5
10
20
30
Typel
Error Rate
(%)
29
18
9.0
6.6
4.1
3.2
18
8.5
4.5
2.6
22
13
4.6
1.8
23
13
5.3
2.3
Type 2
Error Rate
(%)
14
8.2
3.1
2.0
1.3
0.5
8.1
3.3
1.3
0.5
10
4.8
1.3
0.5
10
5.4
1.5
0.6
alf we are interested only in estimating the mean soil Pb concentration (i.e., not the upper confidence limit of the mean), a single
composite of 7=n increments is equivalent to 7=n discrete samples.
11
-------
0.007 i
RBA-adjusted Soil Mean (pp...
399 616
AdjSoil AM C
Minimum 298.73
Maximum 876.90
Mean 499.96
Std Dev 66.68
Values 10000
FIGURE A-l. Probability distribution (vertical axis) of estimated mean RBA-adjusted soil Pb
concentration (horizontal axis) based on a 3 composite samples consisting of 20 increments with RBA
measured on 1 randomly selected composite (C3xI20xRl). Cumulative distribution (percentile) is shown
at the top of the graph. Actual soil Pb RBA is 0.60, actual mean soil Pb concentration is 500 ppm; RBC is
400 ppm. The probability of obtaining estimates that are less than 400 ppm (which would lead to Type 1
errors) is approximately 5%. In this case, a Type 2 error is not possible because the true mean exceeds the
RBC.
12
-------
20 40 60
Number of Increments
80
100
FIGURE A-2. Type 1 error rate (%) predicted for increasing number of increments for 1, 2, or 3
composite samples (20 increments per composite). Mean soil Pb RBA is 0.60±0.09, mean soil Pb
concentration is 500±500 ppm; RBC is 400 ppm. The single composite design (ClxRl) is equivalent to a
discrete sampling design with 7=n discrete samples per DU.
13
-------
100
0
0 200 400 600 800
Mean Soil Pb (ppm)
1000
FIGURE A-3. Probability of rejecting H0 as the mean RBA-adjusted soil Pb concentration increases
when the coefficient of variation of RBA is 0.15 (RBA=0.60±0.09), 0.30 (RBA=0.60±0.18), or 0.50
(RBA=0.60±0.30). Soil Pb coefficient of variation is 1.0; RBC is 400 ppm. The area under the probability
curve, to the right of the vertical line representing the RBC is the Type 1 error. Sample design is 3
composites of 20 increments per composite, with a single composite for RBA (C3xI20xRl).
14
------- |