Using Qualified Data To Document An Observed Release and Observed Contamination

PB94-963311
&EPA
United States
Environmental Protection
Agency
Solid Waste and
Emergency Response
EPA 540-F-94-028
OSWER 9285.7-14FS
PB94-963311
November 1996
Using Qualified Data to Document
an Observed Release and Observed
Contamination
Office of Emergency and Remedfat Response C5204G)
Quick Reference Fact Sheet
This fact sheet discusses the use of the U.S. Environmental Protection Agency's (EPA) Contract Laboratory Program
(CLP) data and other sources of data qualified with a T, "IT, or "UF qualifier t: flag. This guidance provides a
management decision tool for tbe optional use of qualified data to document ar observed release and observed
contamination by chemical analysis under EPA's Hazard Ranking System (HRS). The analyte and sample matrix (i.e.,
soil or water) specific adjustment factors gwera intais feet siwetalfow biased CLP and o»
-------
CLP analytical data be reviewed, or validated by EPA or
third party reviewers, to ensure the data are of known
and documented quality and that the determination be
discussed in a dat- didation report that accompanies the
analytical results. Jased on this data validation, CLP
data are classified into three categories: (1) data for
which all quality control (QC) requirements have passed
contract-required acceptance criteria; (2) data for which
at least one QC requirement has not met acceptance
criteria; and (3) data for which most or all QC
requirements have not met acceptance criteria. Data in
the first category typically are not qualified. Data in me
second category are often qualified with a T qualifier
and, as discussed in mis fact sheet, are usually usable for
HRS purposes. Data in the third category are usually
qualified by an "R" qualifier and are not usable for HRS
purposes.

Whether data are placed into the second or third
category is determined by the amount of bias associated
with the analytical results. Data validation evaluates
biases resulting from laboratory analytical deficiencies or
sample matrices to determine whether the data are
usable. Bias indicates that the reported concentration is
either higher or lower than the true concentration, and
the data validation report identifies the direction of the
bias or if the bias is unknown.

The EPA CLP also sets imniirnim quantitation limits for
all analytes; the Contract Required Quantitation Limit
(CRQL) for organic analytes and the Contract Required!
Detection Limit (CRDL) for inorganic analytes. For
HRS purposes and for this fact sheet, the term CRQL
refers to both the contract required quantitation. limit and;
the contract required detection limit. (40 CFR Part 300,
App. A). The CRQLs are substance specific levels that
a CLP laboratory must be able to routinely and reliably
detect in specific sample matrices (i.e., soil, water,
sediment). The CRQLs are usually set above most
instrument detection limits (DDLs) and method detection
limits (MDLs).

CONSIDERATIONS FOR NON-CLP DATA

Because various laboratories and analytical methods may
be used to develop non-CLP data, the following list
provides the general information sufficient for
determining whether non-CLP data are usable for HRS
purposes.
(1) Identification of the method used for analysis.
Methods include RCRA methods, SW-846, EPA
methods, etc.
(2) Quality control (QQ data. Check each method of
analysis to determine if specific QC requirements
are defined. If not, seek out another method.
(3) Instrument-generated data sheets for sample results.
These data sheets would be the equivalent of Form
I's in CLP data.
(4) MDLs and sample quantization Umits(SQLs). The
analytical method should provide the MDL. The
SQL is an adjusted MDL using sample specific
measurements such as percent moisture and
weight.
(5) Data validation report.

USE OF BIASED QUALIFIED DATA

In the past, all qualified data have been inappropriately
perceived by some people as data of low confidence or
poor quality and have not been used for HRS evaluation.
With careful assessment of the nature of the analytical
biases or QC deficiencies in the data on a case-by-case
basis, qualified data can represent an additional resource
of data for establishing an observed release. Further, the
D.C. District Court of Appeals in 1996 upheld EPA's
case-by-case approach to assess data quality. In
reviewing UK: use of qualified data to identify an
observed release, die Court stated that if mere are
deficiencies in die data, "...die appropriate response is
lo review (he deficiencies on a 'case-by-case basis' to
determine their impact on 'usability of the data.*" The
Court also stated with regards to data quality that,
"...EPA does not face a standard of absolute
perfection....Rather, it is statutorily required to 'assure,
to the marimnm extent feasible,' that it 'accurately
assesses Use relative degree of risk' posed by sites"
IBoani of Regents of the University of Washington, et
aL, v. EPA, No. 95-1324, slip op. at 8-10 (D.C. Cir.
June 25, !996).J

As discussed in this fact sheet, the application of
adjustment factors to "I" qualified data can serve as a
management decision tool to "adjust," or take into
account, the analytical uncertainty in the data indicated
by the qualifier, thereby making qualified data usable for
HRS evaluation. The use of adjustment factors to
account for me larger uncertainty in T* qualified data is
a conservative approach enabling a quantitative
comparison of tne data for use ia documenting an
observed release, ft saould be sated tim the use of
-------
*
adjustment factors only addresses analytical variability
and does not take into account variabilities which may be
introduced during field sampling. Some guidelines for
using the adjustment factor approach are discussed in
Exhibit 1.

CLP QA/QC PROCEDURES

CLP qualifiers are applied to analytical data based on the
results of various Quality Assurance/Quality Control
(QA/QC) procedures used at the laboratory. EPA
analytical methods use a number of QA/QC mechanisms
during sample analysis in order to assess qualitative and
quantitative accuracy (Contract Laboratory Program
Statement of Work for Inorganic Analyses, Document
No. ILM02.0; Contract Laboratory Program Statement
of Work for Organic Analyses, Document No. OLM1.8;
Quality Assurance/Quality Control Samples,
Environmental Response Team Quality Assurance
Technical Information Bulletin; Test Methods for
Evaluating Solid Waste (SW-846): Physical and
Chemical Methods^ Document No. SW-846). To assess
data quality, the laboratory uses matrix spikes, matrix
spike duplicates, laboratory control samples, surrogates,
blanks, laboratory duplicates, and quarterly blind
performance evaluation (PE) samples. The Agency
assumes that if biases are found in the QA/QC samples,
the field sample concentrations may also be biased.

Surrogates are chemically similar to the analytes of
interest. They are added or "spiked" at a known
concentration into the field samples before analysis.
Also, selected target analytes are "spiked" into samples
at a specified frequency to assess potential interferences
from the sample matrix. These samples are called
matrix spikes. Comparison of the known concentration
of the surrogates and matrix spikes with their actual
analytical results reflects the analytical accuracy.
Because the surrogates are expected to behave similarly
to the target analytes, they may indicate bias caused by
interferences from the sample matrices. These types of
interferences from the sample matrix are known as
matrix effects (CLP National Functional Guidelines for
Inorganic Data Review, Publication 9240.1-05-01; CLP
National Functional Guidelines for Organic Data
Review, Publication 9240.1-05; Test Methods for
Evaluating Solid Waste (SW-846): Physical and
Chemical Methods, Document No. SW-846).

Laboratory controfsamples are zero blind samples which
contain known concentrations of specific analytes and are
analyzed in the same batch as field samples. Their
results are used to measure laboratory accuracy. Blanks
are analyzed to detect any extraneous contammarion
introduced either in the field or in the laboratory.

Laboratory duplicates are created when one sample
undergoes two separate analyses. The duplicate results
are compared to determine laboratory precision.
Quarterly blind PE samples are single blind samples that
evaluate the laboratory's capability of performing the
specified analytical protocol.

CLP and other EPA analytical methods include
specifications for acceptable aaalyte identification, target
analytes, and minimum and tnarimnm percent recovery
of the QA/QC compounds. Data are validated according
to guidelines which set performance criteria for
instrument calibration, analyte identification, and
identification and recovery of QA/QC compounds (CLP
Statement of Work and SW-846). The National
Functional Guidelines far Data Review, used in EPA
validation, was designed for the assessment of data
generated under the CLP organic and inorganic
analytical protocols (CLP Statement of Work; National
Functional Guidelines for Data Review). The guidelines
do not preclude die validation of field and other non-
CLPdata. Thus, many EPA Regions have also adapted
the National Functional Guidelines for Data Review to
validate aon-CLP data. Data which do not meet the
guidelines' performance criteria are qualified to indicate
bias or QA/QC deficiencies. The data validation report
usually explains why the data were qualified and
inrffcafeg the bias direction when it can be determined.
Validated data that are not qualified are considered
unbiased and can be used at their reported numerical
value for HRS evaluation.

QUALIFIER DEFINITIONS

Most EPA validation guidelines use the data qualifiers
presented in Exhibit 2 (CLP National Functional
Guidelines for Data Review). Olher qualifiers besides
these may be used; the validation report should always
be checked for me exact list of qualifiers and their
meanings.

It should be emphasized that not meeting one or some of
the contract required QA/QC acceptance criteria is iwten
an indication that the sample was difficult to analyze, not
that there is low confidence in the analysis (i.e., the
-------
EXHIBIT 1
GUIDELINES FOR THE USE OF ADJUSTMENT FACTORS
The use of adjustment factors identified in this fact sheet is a management tool for the optional use of "J"
qualified data generated under CLP or other sources of data to document an observed release.
Adjusted qualified data should be used with non-qualified data whenever possible.
EPA maintains a "worst sites first" policy for placing sites on the NPL (Additional Guidance on "Worst
Sites" and "NPL Caliber Sites' to assist in SACM Implementation, OSWER Directive 9320.2-07).
EPA Regions should use adjustment factors with discretion on a case-by-case basis and should always
carefully consider the use of qualified data in borderline cases.
Resampling and/or reanalysis may be warranted if qualified data do not appear adequate to document an
observed release.
EPA Regions may substitute higher adjustment factors based on documented, justifiable reasons but may
never use a lower adjustment factor vafue.
The adjustment factors should only be applied to anafytes listed in the tables. These adjustment factors
should not be interpolated or extrapolated to develop factors for analytes not listed in the tables.
• The adjustment factors apply only to T qualified data above the CRQL.
Detection below the CRQL is treated as non-quantifiable for HRS purposes.
* "UJ" data may be used under strict circamstances as explained in this fact sheet.
The adjustment factors only apply to biased T qualified data, not to other "J" qualified data.
The adjustment factors do not apply to "N", "HP, or "R" qualified data. These data can not be used to
document an observed release for HRS purposes.
analysis is "under control" and can be adopiafe for HRS
decision making). Often "J", "IT, and "UJ" qualified
data fall into this category.

There are instances when qualified data cannot be used
since the uncertainty of the results is unknown. For
example, violations of laboratory instrument calibration
and tuning requirements, and gross violations of holding
times reflect the possibility that the results are of
unknown quality (i.e., the analysis is "out of control").
Most often these data would be qualified with an "R" or
an "N" (not usable for HRS purposes).

USING "U" QUALIFIED DATA

The "U" qualifier simply means that the reported
concentration of the analyte was at or below the CRQL-
there can be confidence that the true concentration is at
or below the quantitation limit. Therefore, "U"
qualified data can be used for establishing background
levels. If the release sample concentration is above this
level, as specified in the HRS, an observed release can
be established. The quantitation limit for that analyte
could be used as a maximum background concentration
if a more conservative background level seems
appropriate.

USING «S" QUALIFIED DATA

As discussed previously, some "J" qualified data can be
used in establishing an observed release if the uncertainty
in the reported values is documented. Qualified data
should always be carefully examined by the Regions to
determine the reasons for qualification before use in
HRS evaluation. Resampling and/or reanalysis may be
warranted if qualified data only marginally document an
observed release. Whenever possible, qualified data
should be used in conjunction with non-qualified data.
-------
As described in Exhibit 2, T qualified data indicates
that bias has been detected in the sample analysis and
although the analyte is definitively present, the reported
concentration is art estimate. Depending on the reasons
and the direction of bias, with the use of adjustment
factors, "J" qualified data can represent data of known
and documented quality sufficient for use in establishing
an observed release and observed contamination under
the HRS.

USING "UJ" QUALIFIED DATA

A combination of the "U" and "J" qualifiers indicates
that the reported value may not accurately represent the
concentration necessary to positively detect the analyte in
the sample. Under limited conditions, "UJ" qualified
data can be used to represent background concentrations
for establishing an observed release. These conditions
are: instances when there is confidence that the
background concentration is not detectable above the
CRQL, the background concentration is biased high, and
the sample measurement establishing the observed
release equals or exceeds the CRQL.

DIRECTION OF BIAS IN "J" QUALIFIED DATA

It is important to understand the direction of bias
associated with "J" qualified data before using the data
to document an observed release. Qualified data may
have high, low, or unknown bias. A low bias means
that the reported concentration is likely an underestimate
of die true concentration. For example, data may be
biased low when sample holding times for volatile
organic compounds (VOCs) are moderately exceeded or
when recovery of QA/QC compounds is significairdy
less than the amount introduced into the sample. Low
surrogate recovery would also indicate a low bias. A
high bias means the reported concentration is likely an
overestimate of the true concentration. For example,
data may be biased high when recovery of QA/QC
compounds is significantly higher than the amount in the
sample. A bias is unknown when it is impossible to
ascertain whether the concentration is an overestimate or
an underestimate. For example, an unknown bias could
result when surrogate recoveries exceed method recovery
criteria and matrix spike/matrix spike duplicate
compounds below method recovery criteria fail the
relative percent difference (RPD) criteria in the same
sample.

Despite the bias, certain qualified data may be used
without application of adjustment factors for determining
an observed release under certain circumstances. The
following are examples of using "J" qualified data
without adjustment factors:

• Low bias release samples are likely to' be
underestimates of true concentrations. If the
reported concentration of a low bias release sample
is three times above unbiased background levels,
these release samples would still meet the HRS
criteria. The true concentrations would still be
three times above the background level.

• High bias background samples are likely to be
overestimates of true concentrations. If the
reported concentration of unbiased release samples
are three times above the reported background
concentration, they would still meet the HRS
observed release criteria because they would still
be three times above the true background
concentration.

The above examples show that both low bias "J"
qualified release samples at their reported concentrations
and high bias "J" qualified background samples may be
used at their reported concentrations in these situations.

High bias release samples may not be used at then-
reported concentrations because they are an overestimate
of true concentrations in this situation; resampling and/or
re-analysis of the release samples should be considered.
The true difference in the background and release
concentrations may be less than the HRS criteria for
establishing an observed release. The reported
concentration for low bias background concentrations
may not be compared to release samples because it is
most likely an underestimate of background level; the
release sample concentration may not significantly
exceed the true background concentration. However, in
lieu of re-sampling and/or re-analysis, high bias release
data, anrf low bias background data may be used with
adjustment factors which compensate for the probable
uncertainty in the analyses.

ADJUSTMENT FACTORS FOR BIASED "J"
QUALIFIED DATA

Applying adjustment factors to "J" qualified data will
enable EPA to be more confident that the increase in
contaminant concentrations between the background and
-------
EXHIBIT 2
EPA CLP DATA QUALIFIERS AND THEIR USABILITY FOR DOCUMENTING AN OBSERVED RELEASE
Usable*
Not Usable
"U" The substance or analyte was analyzed for, but
no quantifiable concentration was found at or
above the CRQL (CLP National Functional
Guidelines for Data Review).
"N" The analysis indicates the presence of an analyte
for which there is presumptive evidence to make
a "tentative identification'' (CLP National
Functional Guidelines for Data Review).
"J" The analyte was positively identified-the
associated numerical value is the approximate
concentration of the analyte in the sample. The
"J" Qualifier indicates that one or more QA/QC
requirements have not met contract required
acceptance criteria, but the instrumentation was
functioning properly during the analysis. For
example, a "J" qualifier may indicate that the
sample was difficult to analyze or mat the value
may lay near the low end of the linear range of
the instrument. " J" data are considered biased,
but provide definitive analyte identification (CLP
National Functional Guidelines for Data
Review),
"R" The sample results are rejected due to serious
deficiencies in the ability to analyze the sample
and meet QC criteria. The presence or absence
of the analyte can not be verified and the result
has been rejected. A sample result may be
qualified with an "R" qualifier when the
instrument did not remain "in control" or the
stability or sensitivity of the instrument were not
maintained during the analysis (CLP National
Functional Guidelines for Data Review).
"UJ" The analyte was not quantifiable at or above the
CRQL. In addition to not being quantifiable,
one or more QA/QC requirements have not met
contract acceptance criteria (CLP National
Functional Guidelines for Data Review).
"NJ" The analysis indicates the presence of the
analyte that has been "tentatively identified" and
the associated numerical value represents its
approximate concentration (CLP National
Functional Guidelines far Data Review).
Usable under certain circumstances as explained in this fact sheet.
release samples is due to a release. The adjustment
factors are applied as "safety factors" to compensate for
analytical uncertainty, allowing biased data to be used
for determining an observed release. Dividing the high
bias result by an adjustment factor ueflates it from the
high end of the acceptable range towards a low bias
value. Multiplying a low bias concentration by an
adjustment factor inflates it to the high end of the
acceptable range.

Tables 1 through 4 (pages 11-18) present analyte and
matrix-specific adjustment factors to address the
analytical uncertainty when determining an observed
release using high bias release samples and low bias
background data. The factors are derived from percent
recoveries of matrix spikes, surrogates, and laboratory
control samples in the CLP Analytical Results Database
(CARD) from January 1991 to March 1996. A total of
32,447 samples were reviewed for volatile organic
analytes; 32,913 samples for semivolatile organic
analytes; 59,508 samples for pesticides/PCB analytes;
and 5,954 samples for inorganic analytes.

The range of CARD data for each analyte includes 97
percent of all percent recoveries in the database,
discarding outliers. The adjustment factors are ratios of
percent recovery values at the 98.5 and 1.5 percentiles.
The ratios generally show a consistent pattern.

Adjustment factors have been determined for all analytes
in the CLP Target Compound List (organic analytes) and
Target Analyte List (inorganic analytes). A tiered
approach was used to derive the organic adjustment
factors. Percent recoveries for surrogates were
-------
examined first, followed by matrix spike recoveries.
When both matrix spike and surrogate data were
available for the same analyte, the larger adjustment
factor (representing more extreme high and low percent
recoveries) was used. Laboratory control samples were
used to calculate the inorganic adjustment factors.
Quarterly blind sample data were not used to determine
adjustment factors because of the small data set
available. A default adjustment factor of 10 was used
for analytes when percent recovery data were
unavailable.

Adjustment factors do not correct the biased sample
concentration to its true value, as such "correction" is
not possible. CARD data do not differentiate and
quantify individual sources of variation. Instead, the
ratio of percentile used to develop adjustment factors
represents a "worst-case" scenario. Adjustment factors
either inflate background values to the nigh end of the
range or deflate release data to the low end. Therefore,
adjustment factors compensate or adjust for the apparent
analytical variability when comparing a high bias value
to a low bias value (see Exhibit 3).

USING THE ADJUSTMENT FACTORS

This section of the fact sheet demonstrates how
adjustment factors can be used with UJ" qualified data
for HRS scoring purposes, including documentation and
detection limit issues.

Documentation Reauirements for Using Qualified Data
In using "J" qualified data to determine an observed
release, include a discussion of "J" qualifiers from the
data validation report and cite it as a reference in the site
assessment report or HRS documentation record. If
adjustment factors are applied to "J" qualified data.
reference and cite this fact sheet. These steps will
ensure that the direction of bias is documented and will
demonstrate how biases have been adjusted.

Detection Limit Restrictions
Adjustment factors may only be applied to T qualified
data with concentrations above the CLP CRQL for
organics or CRDL for inorganics. "J" qualified data
with concentrations below the CRQL can not be used to
document an observed release except as specified in the
previous section entitled "Using "UJ" Qualified Data."

Application of Factors
Exhibit 3 shows how to apply the factors to "J" qualified
data. Multiply low bias background sample results by
the analyte-specific adjustment factor or the default factor
of 10 when an analyte-specific adjustment factor is not
available. The resulting new background value
effectively becomes a high bias value that may be used
to determine an observed release. Divide high bias
release sample data by the analyte-specific adjustment
factor or the default factor of 10 when an analyte-
specific adjustment factor is not available. The resulting
new release sample value effectively becomes a low bias
value that may be used to determine an observed release.

Note: High bias background data, low bias release data,
and unbiased data may be used at their reported
concentrations.

Note: Adjusted release and background values must still
meet HRS criteria (e.g., release concentration must be
at least three times above background level) to determine
an observed release.

Examples Using Trichloroethene in Soil and Water
1. Release water sample is unbiased, background
water sample is unbiased but all data are qualified
with a "J" due to an contractual laboratory error
not analytical error.

Background sample value: 12 /tg/L (J) no bias
Release sample value: 40 pg/L (J) no bias

The CRQL for trichloroethene is 10 pg/Kg for soil and
10 /tg/L for water.

In this example, the qualification of the data is not
related to bias in the reported concentrations. Thus,
using adjustment factors is not needed and an observed
release is established if all other criteria are met.

2. Release soil sample data is biased low, background
soil .'fflnple data is biased high.

Background sample value: 12 ngfKg (J) high bias
Release sample value: 40 /*g/Kg (J) low bias

In this example, the direction of bias indicates that the
true release value may be higher and the true
background value may be lower than reported values.
The release sample concentration still exceeds
background by more than three times, so an observed
release is established, provided all other HRS criteria are
met. Using adjustment factors is not needed.
-------
EXHIBIT 3
USE OF ADJUSTMENT FACTORS FOR M" QUALIFIED DATA
Type of Sample
Background
Sample
Release
Sample
Type of Bias
No Bias
Low Bias
High Bias
Unknown Bias
No Bias
Low Bias
High Bias
Unknown Bias
Action Required
None: Use concentration without faaor
Multiply concentration by factor
None: Use concentration without factor
Multiply concentration by factor
None: Use concentration without factor
None: Use concentration without factor
Divide concentration by factor
Divide concentration by factor
3. Release soil sample data is unbiased, background
soil sample is biased law.

Background sample value: 12 jtg/Kg (J) low bias
Release sample value: 30 fig/Kg no bias

In this example, the true background value is assumed to
be less than the reported value; however, an observed
release may still be possible. To use the data to establish
an observed release, multiply the background sample
data value by the adjustment factor given for
trichloroethene in soil (2.11). No adjustment factor is
needed for the release sample.

New background sample value:
(12 jug/Kg) x (2.11) = 25.32 /tg/Kg (J) high bias

The release sample concentration does not meet or
exceed the new background level by three times, so an
observed release is not established.

4. Release -water sample data is biased high,
background water sample data is unbiased.

Background sample value: 15 ^g/L no bias
Release sample value: 70 pgfL (J) high bias

In this example, the true release value may be lower
than the reported value; however, an observed release
may still be possible. To use tne data to establish an
observed release, divide the release sample by the
adjustment factor for trichloroethene in water (1.66).
No adjustment factor is needed for the background
sample.

New release sample value:
(70 /tg/L) -*- (1.66) = 42.17 /tg/L (J) low bias

The new release sample concentration does not meet or
exceed the background level by three times, so an
observed release is not established.

5. Release soil sample data has unknown bias;
background soil sample data has unknown bias.

The following example is the most conservative
approach to using adjustment factors with qualified data.
Background sample value:
Release sample value:
325 pg/Kg (J) unknown bias
m mis example, it is not possible to determine from the
reported values if an observed release is possible. To
use the data to establish an observed release, divide the
release sample value and multiply the background
sample value by the adjustment factor given for
tricbloroethene in soil (2.11).

New release sample value:
(325 /%/Kg) * (2.11) = 154.03 /ig/Kg (J) low bios

New background sample value:
(20 jig/Kg) x (2.11) = 42.2 pg/Kg (J) high bios
-------
The new release sample is at least three times the new
background concentration, so an observed release is
established, provided all other HRS criteria are met.

ISSUES wrra USING ADJUSTMENT FACTOR
APPROACH

Some issues were raised regarding the application of
adjustment factors to qualified data during the Agency's
internal review process.

One issue is that "J" qualifiers are added to analytical
results for many reasons that may or may not affect the
accuracy and precision of the analytical result. The
application of an adjustment factor to "J" qualified data
in which bias is not affected could be considered overly
conservative.

All qualified data should be carefully evaluated to
determine if the data are biased. Based on the reasons
for bias, the use of an adjustment factor should 01% be
considered as a management tool that provides a quick
screening of the data for site assessment, not a means for
correcting the biased value to a true value. Application
of adjustment factors are intended for use with qualified
data reported at or above the CRQL and may not be
applicable to data which are qualified but technically
sound. As stated previously, qualified data should
always be carefully reviewed on a case-by-case basis
prior to use in HRS evaluation.

Another issue is the validity of "10" as a default
adjustment factor. A default adjustment factor of 10 was
a policy decision based on the range of adjustment
factors and an industry approach. The default was
chosen in order to account for the maximum' variability
regardless of the direction of the bias. Therefore, the
default value of 10 is generally considered to be a
conservative adjustment factor. EPA reviewed the use
of the default value of 10 and determined mat this "alue
was conservative.

Even if using adjustment factors is sometimes overly
conservative, this approach is preferable to not using the
data at all. EPA maintains a "worst sites first" policy
that only the sites considered most harmful' to human
health and/or the environment should be listed. EPA
considers the use of adjustment factors appropriate as a
inanageme^: decision tool. However, discretion is
needed when applying adjustment factors. The use of
adjustment factors may not be appropriate in all cases.
USE OF OTHER ADJUSTMENT FACTORS

EPA Regions may substitute higher, but never lower,
adjustment factor values for the ones listed in this fact
sheet on a case-by-case basis when technically justified.
For example, other adjustment factors may be applied to
conform with site-specific Data Quality Objectives
(DQOs) or with Regional Standard Operating Procedures
(SOPs) {Data Quality Objectives Process for Superfiatd,
Publication 9355.9-01).

SUMMARY

For site assessment purposes, EPA Regions should not
automatically discard "J" qualified data. However, site-
specific data usability Determinations may result hi the
data's not being used.

Data qualified under me EPA's CLP or from other
sources of validated data may be used to demonstrate an
observed release if certain measures are taken to ensure
mat the bias of die data qualifier is adjusted using the
factor approach specified in this fact sheet. (This fact
sheet provides a management decision tool for making
qualified data usable for documenting an observed
release.) The analyte and matrix-specific adjustment
factors provided in Tables 1 through 4 of this fact sheet
present these adjustment factors.

The scope of mis fact sheet is limited to the situations
described in Exhibit 1. The use of qualified analytical
data without me adjustment factors presented in ibis fact
sheet is limited. Higher adjustment factors may be
substituted by EPA Regions on a case-by-case basis
when tecnmcaHy justified by she-specific DQOs or
SOPs.
-------
REFERENCES

1. U.S. Government Printing Office, FederalRegister,
Part H, Environmental Protection Agency, 40 OFR
Pan 300, Hazard Ranking System, Final Rule,
December 14, 1990.

2. U.S. Environmental Protection Agency, Hazard
Ranking System Guidance Manual, Office of Solid
Waste and Emergency Response, PB92-963377,
November 1992.

3. U.S. Environmental Protection Agency, 1995.
Establishing an Observed Release. Office of
Emergency and Remedial Response. PB94-9633I4.

4. U.S. Environmental Protection Agency, 1995,
Establishing Areas of Observed, Contamination.
Office of Emergency and Remedial Response.
PB94-963312.

5. U.S. Environmental Protection Agency, 1995.
Establishing Background Levels. Office of
Emergency and Remedial Response. PB94-9633J3.

6. U.S. Environmental Protection Agency, 1994. CLP
National Functional Guidelines for Inorganic Data
Review. Office of Solid Waste and Emergency
Response. Publication 9240.1-05-01.

7. U.S. Environmental Protection Agency, 1993. CLP
National Functional Guidelines for Organic Data
Review. Office of Solid Waste and Emergency
Response. Publication 9240.1-05.

8. U.S. Environmental Protection Agency, 1991.
Contract Laboratory Program Statement of Work
for Inorganic Analysis. Document No. ILM02.0.

9. U.S. Environmental Protection Agency, 1991.
Contract Laboratory Program Statement of Work
for Organic Analysis. Document No. OLM1.8.

10. U.S. Environmental Protection Agency, 1993.
Additional Guidance on "Worst Sites" and "NPL
Caliber Sites" to Assist in SACM Implementation.
Office of Emergency and Remedial Response.
PB94-963206.
11. Board of Regents of the University of Washington,
etal., v. EPA, No. 95-1324, slip op. at 10 (D.C.
Cir. June 25, 1996). 10.

12. U.S. Environmental Protection Agency, 1991.
Guidance for Performing Preliminary Assessments
Under CERCLA. Office of Solid Waste and
Emergency Response. Publication 934S.O-01-A.

13. U.S. Environmental Protection Agency, 1992.
Guidance for Performing Site Inspections Under
CERCLA. Office of Solid Waste and Emergency
PB92-963375.
14. U.S. Environmental Protection Agency, 1992.
Quality Assurance/Quality Control Samples,
Environmental Response Team Quality Assurance
Technical Information Bulletin.

15. U.S. EmaroDmental Protection Agency, 1986. Test
Methods for Evaluating Solid Waste (SW-846):
Physical and Chemical Methods. Office of Solid
Waste and Emergency Response. Document No.
SW-846.

16. U.S. Environmental Protection Agency, 1993.
Data Quality Objectives Process for Superfund.
Office of Emergency and Remedial Response.
Publication 9355.9-01.
10
-------
TABLE 1
FACTORS FOR VOLATILE ORGANIC ANALYTES
VOLATILE
ORGANIC
ANALYTES
1,1, i-TRICHLOROETHANE
1, 1,2,2-TETRACHLOROETHANE
1 , 1 ,2-TRICHI-OROETHANE
1 , 1-DICHLOROETHANE
1 , 1-DICHLOROETHENE
1 .2-DICHLOROETHANE-D4
1,2-DICHLOROETHENE (TOTAL)
1 ,2-DICHLOROPROPANE
2-BUTANONE
2-HEXANONE
4-METHYL-2-PENTANONE
ACETONE
BENZENE
BROMODICHLOROMETHANE
BROMOFORM
BROMOFLUOROBENZENE
BROMOMETHANE
CARBON DISULFEDE
SOIL MATRIX
Number of
CARD
Samples
Reviewed
—
—
—
—
7,031
32,446
—
—
—
—
—
—
7,024
—
—
32,444
—
—
Factor
10.0
10.0
10.0
10.0
2.71
152
10.0
10.0
10.0
10.0
10.0
10.0
1.97
10.0
10.0
1.7
10.0
10.0
WATER MATRIX
Number of
CARD Samples
Reviewed
—
—
_.
—
5,015
25,516
—
—
—
—
.«
—
5,001
—
—
25,518
—
—
Factor
10.0
10.0
10.0
10.0
2.35
1.38
10.0
10.0
10.0
10.0
10.0
"10.0
1.64
10.0
10.0
1.26
10.0
10.0
11
-------
TABLE 1
FACTORS FOR VOLATILE ORGANIC ANALYTES
VOLATILE
ORGANIC
ANALYTES
CARBON TETRACHLORIDE
CHLOROBENZENE
CHLOROETHANE
CHLOROFORM
CH1X)ROMETHANE
CIS-1 ,3-DICHLOROPROPENE
DEBROMOCHLOROMETHANE
ETHYLBENZENE
METHYLENE CHLORIDE
STYRENE
TE rRACHLOROETHENE
TOLUENE-D8
TRANS-1 ,3-DICHLOROPROPENE
TRICHLOROETHENE
VINYL CHLORIDE
XYLENE (TOTAL)
SOIL MATRIX
Number of
CARD
Samples
Reviewed
—
7,018
—
—
—
—
—
—
—
—
—
32,447
—
6,988
—
—
Factor
10.0
2.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
1.63
10.0
2.11
10.0
10.0
WATER MATRIX
Number of
CARD Samples
Reviewed
—
5,015
—
—
—
—
—
—
—
—
—
25,526
—
4,938
—
—
Factor
10.0
1.54
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
1.21
10.0
1.66
10.0
10.0
12
-------
TABLE 2
FACTORS FOR SEMIVOLATILE ORGANIC ANALYTES
SEJWIVOLAT1LE
ORGANIC
ANALYTES
1 ,2,4-TRICHLOROBENZENE
1 .2-DICHLOROBENZENE-D4
1 ,3-DICHLOROBENZENE
1 ,4-DICHLOROBENZENE
2,2'-OXYBIS(l-CHLOROPROPANE)
2,4,6-TRIBROMOPHENOL
2,4,5-TRICHIjOROPHENOL
2,4,6-TRICHLOROPHENOL
2,4-DICHLOROPHENOL
2,4-DIMETHYLPHENOL
2,4-DINITROPHENOL
2,4-DINITROTOLUENE
2,6-DINITROTOLUENE
2-CHLORONAPHTHALENE
2-CHLOROPHENOL-D4
2-FLUOROBIPHENYL
2-FLUORPHENOL
2-METHYLNAPHTHALENE
2-METHYLPHENOL
2-NITROANILINE
2-NITROPHENOL
3,3'-DICHLOROBENZIDINE
3-NITROANILINE
^,6-DINITRO-2-METHYLPHENOL
4-BROMOPHENYL-PHENYLETHER
SOIL MATRIX
Number of CARD
Samples Reviewed
6,792
32,848
—
6,796
—
32,605
—
—
_.
—
—
6,798
—
__
32,798
32,913
32,781
—
~
—
—
—
—
_
__
Factor
4.83
4-22
10.0
6.0
10.0
9.38
10.0
10.0
10.0
10.0
, 10.0
4.88
10.0
10.0
4.08
3.38
5.05
10.0
10,0
10.0
10.0
10.0
10.0
10.0
10.0
WATER MATRIX
Number of CARD
Samples
Reviewed
4,605
21,506
—
4,599
—
21,509
—
—
—
—
—
4,623
—
—
21,506
21,532
21,511
—
—
—
—
—
—
—
—
Factor
3.71
3.0
10.0
3.85
10.0
3.57
10.0
10.0
10.0
10.0
10.0
3.52
10.0
10.0
2.92
2.84
3.34
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
13
-------
TABLE 2
FACTORS FOR SBVKVOLATILE ORGANIC ANALYTES
SEMJVOLATILE
ORGANIC
ANALYTES
4-CHLORO-3-METHYLPHENOL
4-CHLOROANOJNE
4^HLOROPHENTiTv-PHENYLETHER
4-METHYLPHENOL
4-NITROANEJNE
4-NITROPHENOL
ACENAPHTHENE
ACENAPHTHYLENE
ANTHRACENE
BENZO(A)ANTHRACENE
BENZO(A)PYRENE
BENZO(B)FLUORANTHENE
BENZO(G,H,I)PERYLENE
BENZO(K)FLUORANTHENE
BIS(2-CHLOROETHOXY)METHANE
BIS(2-CHLX>ROETHYL)ETHER
BIS(2-ETHYLHEXYL)PHTHALATE
BUTYLBENZYLPHTHALATE
CARBAZOLE
CHRYSENE
DI-N-BUTYLPHTHALATE
DI-N-OCTYLPHTHALATE
DffiENZ(A,H)ANTHRACENE
DEBENZOFURAN
DIETmTPHTHALATE
SOIL MATRIX
Number of CARD
Samples Reviewed
6J15
—
—
—
—
6,627
6,773
—
__
—
—
— .
—
—
—
_
—
—
—
—
—
—
—
—
—
Factor
6.26
10.0
10.0
10.0
10.0
9.33
4.68
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10,0
10.0
10.0
10.0
10.0
10.0
10.0
WATER MATRIX
Number of CARD
Samples
Reviewed
4,609
—
—
—
—
4,586
4,600
—
_
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
__
Factor
4.46
10.0
10.0
10.0
10.0
5.96
3.63
10.0
10.0
10.0
10.0
10.0
10.0
10.0
iO.O
10.0
10.0
10.0
10.0
3.0.0
10.0
10.0
10.0
10,0
10.0
14
-------
TABLE 2
FACTORS FOR SEMIVOLATILE ORGANIC ANALYTES
SEMIVOLATILE
ORGANIC
ANALYTES
DIMETHYLPHTHALATE
FLUORANTHENE
FLUORENE
HEXACHLOROBENZENE
HEXACHLOROBUTADffiNE
HEXACHLOROCYCLOPENTADffiNE
HEXACHLOROETHANE
INDENOC1 ,2,3-CD)PYRENE
ISOPHORONE
M-NITROSO-DI-N-PROPYLAMINE
N-NITROSODIPHENYLAMINE(1)
NAPHTHAJJENE
NimOBENZENE-DS
PENTACHLOROPHENOL
PHENANTHRENE
PHENOL-D5
PYRENE
TERPHENYL-D14
SOIL MATRIX
Number of CARD
Samples Reviewed
—
—
—
—
—
—
— .
—
—
6,725
—
—
32,867
6,59?
—
32,855
6,543
32,899
Factor
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
4.92
10.0
10.0
3.96
12,5
10.0
3.85
11.86
435
WATER MATRIX
Number of CARD
Samples
Reviewed
—
—
—
—
—
—
—
—
—
4,513
—
—
21,533
4,550
—
21,489
4,612
21,541
Factor
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
1C.O
4.0
10.0
10.0
.2.73
10.12
10.0
3.53
5.67
6.32
15
-------
TABLE 3
FACTORS FOR PESTICIDES/PCB AMALYTES
VOLATILE
ORGANIC
ANALYTES
4,4'-DDD
4,4'-DDE
4,4'-DDT
ALDRIN
ALPHA-BHC
ALPHA-CHLORDANE
AROCLOR-1016
AROCLOR-1221
AROCLOR-1232
AROCLOR-1242
AROCLOR-1248
AROCLOR-1254
AROCLOR-1260
BETA-BHC
DECACHLOROBIPHENYL
DELTA-BHC
DIELDREN
SOIL MATRIX
Number of
CARD
Samples
Reviewed
—
—
5,343
5,526
—
—
—
—
__
—
—
—
—
—
57,315
—
5,539
Factor
10.0
10.0
12.82
14.26
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
17.79
10.0
11.93
WATER MATRIX
Number of
CARD Samples
Reviewed
—
—
3,850
3,829
—
—
—
—
—
—
—
—
—
—
33,592
—
3,861
Factor
10.0
10.0
7.14
6.63
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
4.87
16
-------
TABLE 3
FACTORS FOR PESTICIDES/PCB ANALYTES
VOLATILE
ORGANIC
ANALYTES
ENDOSULFAN !
ENDOSULFAN H
ENDOSULFAN SULFATE
ENDRIN
ENDRIN ALDEHYDE
ENDRIN KETONE
GAMMA-BHC (LINDANE)
GAMMA-CHLORDANE
HEPTACHLOR
HEPTACHLOR EPOXIDE
METHOXYCHLOR
TETRACHLORO-M-XYLENE
TOXAPHENE
SOIL MATRIX
Number of
CARD
Samples
Reviewed
—
—
_
5,521
—
—
5,545
—
5,548
_
—
59,508
—
Factor
10.0
10.0
ib.o
14.13
10.0
10.0
11.79
10.0
7.88
10.0
10.0
8.5
10.0
WATER MATRIX
Number of
CARD Samples
Reviewed
—
—
—
3,850
—
—
3,832
—
3,836
—
—
33,787
—
Factor
10.0
10.0
10.0
5.33
10.0
10.0
10.0
10.0
5.26
10.0
10.0
5.29
10.0
17
-------
TABLE 4
FACTORS FOR INORGANIC ANALYTES
INORGANIC
ANALYTES
ALUMINUM
ANTIMONY
ARSENIC
BARIUM
BERYLLIUM
CADMIUM
CALCIUM
CHROMIUM
COBALT
COPPER
CYANIDE
IRON
LEAD
MAGNESIUM
MANGANESE
MERCURY
NICKEL
POTASSIUM
SELENIUM
SILVER
SODIUM
THALLIUM
VANADIUM
ZINC
SOEL MATRIX
Number of
CARD
Samples
Reviewed
5387
5392
5675
5360
5399
5385
5383
5389
5392
5394
3281
5391
5982
5397
5395
5954
5400
3874
:S2Q
5392
5024
5621
5393
5404
Factor
1.66
1.98
1.74
3.99
1.28
1.41
1.28
1.29
1.25
1.22
1.55
1.34
1.44
1.23
1.24
1.83
1.35
17.49
228
L74
25,43
1M
1.34
1.50
WATER MATRIX
Number of
CARD
Samples
Reviewed
6208
6170
6303
6201
6208
6166
6201
6210
6212
6205
225
6216
6384
6210
6214
256
6210
6175
6278
6215
6195
6253
6212
6224
Factor
1.30
1.27
1.35
1.25
1.25
1.29
1.24
1.30
1.27
1.25
1.36
1.27
1.31
1.24
1.28
1.50
1.29
1.24
1.41
1.42
1.26
1.37
1.25
1.29
18
-------