J United States
^£ Environmental Protection Agency
LOVE CANAL
EMERGENCY
DECLARATION AREA
HABITABILITY STUDY
FINAL REPORT
VOLUME III
Soil Assessment--
Indicator Chemicals
TECHNICAL REVIEW COMMITTEE
U.S. Environmental Protection Agency Region II
U.S. Department of Health and Human Services/
Centers for Disease Control
New York State Department of Health
New York State Department of Environmental Conservation
-------
VOLUME III
Soil Assessment-
Indicator Chemicals
Prepared for
U.S. EPA REGION II
26 Federal Plaza
New York, New York 10278
Prepared by
CH2MHILL SOUTHEAST, Inc.
P.O. Box 4400
Reston, Virginia 22090
Under Contract No. 68-01-7251
May 1988
-------
VOLUME III
SOIL ASSESSMENT—INDICATOR CHEMICALS
ERRATA
Below is a list of errata for this report. Some of these
are straightforward typographical errors, but others involve
a change in content (Note—negative line numbers indicate
number of lines from the bottom of the page.)
SECTIONS 1-6
Page Line Erratum
1-2 2 "The EDA comparison areas" should read
"The EDA sampling areas"
1-3 1 "A total of 5320" should read "A total of
5329."
4-5 3 The random selection of sampling sites was
done at the neighborhood level in the EDA,
not at the sampling area level, as implied.
6-2 See attached Table 6-5, which shows the
corrections.
6-45 3 The superscripts "a,b,c" should read
"a,b,c,d"
6-48 3 The superscripts "a,b" should read "a,b,c"
APPENDIX A
There is a statement on page A-6, line 9, that ties the
habitability decision to the "minimum unacceptable
difference." It should be emphasized that the "delta" of
one order of magnitude that was used in the design of the
study (performed by CH2M HILL) is in no way related to any
suggested rule for a habitability decision (which will be
made by the Commissioner of Health for the State of New
York.
Other Appendix A errata are as follows:
III-l
-------
Page Line Erratum
A-3 "mean" should read "median"
A-5 "mean" should read "median"
A-8 Delete this page; it is a repeat of page A-6.
A-9 -19 "sampling area-" should read "EDA sampling
area-"
-16 "neighborhood" should read "sampling area"
-13 "neighborhood" should read "sampling area"
-12 "neighborhood" should read "sampling area"
A-10 10 "neighborhood" should read "sampling area"
APPENDIX B
Page Line Erratum
B-2 -14 The text after "analytical methods?" should
start a new paragraph that is not part of
question number 3.
B-ll 24 "(Anderson and McLean)" should read
"(Anderson and McLean, 1974)"
B-13 8 "extension to Eq. B.3" should read "extension
to Eq. B.2"
B-14 -1 In the lower right-hand box, on the last
line, the subscript "k" (lower case) should
be "K" (upper case)
B-15 13 "sampling areas" should read "(sampling
areas) "
B-16 5 "standard normal variable" should read
"standard normal random variable"
B-20 Table B-4—LCICs are mislabeled. Labels
should be:
1,2 Dichlorobenzene (1)
1,2,4 Trichlorobenzene (2)
1,2,3,4 Tetrachlorobenzene (3)
Chloronaphthalene (4)
alpha-BHC (5)
III-2
-------
delta-BHC (6)
beta-BHC (7)
gamma-BHC (8)
APPENDIX F
In Figure F-l, "Blank QC" should be "Blind QC," and in
Figure F-3, abbreviations "EMDC" should be "EMPC," and
"Performance Handling Blank" should be "Preparation Handling
Blank."
APPENDIX K
Tables K-l, K-2, and K-4 to K-14 are missing some or all of
the superscripts in the title that indicate the relevant
footnotes.
Table K-l 2 is a duplicate of Table K-ll in Volume III.
WDR355/061
III-3
-------
ERRATA
Table 6-5
NUMBER OF FIELD QC SAMPLES RECEIVED AND ANALYZED BY LABORATORY
Analytical,
Laboratory0
1
2
3
4
6
7
8
Totals
Field QC Samples
Received
FHB
14
12
11
12
11
10
6
h
Split
22
17
14^
K
11°
9
4
6
PHB
8
7
10
9
6
7
9
SSB
19
18
19
21
18
17
18
Field QC Samples
Analyzed
FHB
14+3d
12
11
0
11
10
6
Split PHB SSB
22+19
17
14
0
9
4
6
76
83
56
130
67
73J
FHB = Field handling blank
PHB = Preparation handling blank
SSB = Shipping and storage blank
Laboratory 5 was not selected for study participation.
One split was not sent to Laboratory 4.
CA total of 57 PHBs were used; however, one was unextrudable.
Three of Laboratory 4's FHBs were analyzed by Laboratory 1.
Q
Nine FHBs were not analyzed by Laboratory 4.
PHBs and SSBs were stored but not analyzed.
gA field spLit extracted by Laboratory 4 was analysed by Laboratory 1.
A total of 83 pairs of splits were sent for analysis. One-half of each
split was treated as a field sample and used in the statistical analysis.
The other half is represented in this table.
XValid results were obtained for both halves of 64 pairs of splits.
WDR355/061
III-4
-------
CONTENTS
Section Page
ACKNOWLEDGEMENTS x
LIST OF ACRONYMS AND ABBREVIATIONS xii
1.0 SUMMARY 1-1
2.0 INTRODUCTION 2-1
2.1 Background 2-1
2.2 Habitability Criteria 2-4
2.3 Pilot Study 2-6
2.4 NYSDOH Soil Study 2-9
2.5 Peer Review 2-10
3.0 GOALS 3-1
4.0 DESIGN 4-1
4,1 Project Organization 4-1
4.2 Study Design 4-1
4.3 Data QA/QC Design 4-6
5.0 METHODS 5-1
5.1 Sample Collection and Preparation 5-1
5.2 Sample Analysis 5-3
5.3 Data and Information Systems Development 5-4
5.4 QA/QC 5-5
5.5 Statistical Analysis 5-8
6.0 RESULTS 6-1
6.1 Sampling and Analytical Activities 6-1
6.2 QA/QC Data Assessment 6-14
6.3 Detection Limit and Analytical Results 6-17
6.4 Statistical Comparison of Sampled Areas 6-36
REFERENCES R-l
-------
Figures
CONTENTS
(continued)
Page
APPENDIX A: Statistical Concepts A-l
APPENDIX B: Development of a Statistical
Approach for the Sampling Area Comparisons B-1
APPENDIX C: Analytical Method Development C-l
APPENDIX D: Holding Time Study D-l
APPENDIX E: Data and Information Systems
Development E-l
APPENDIX F: QA/QC Program F-l
APPENDIX G: NEIC/TechLaw Participation,
Soil Assessment-Indicator Chemicals G-1
APPENDIX H: EMSL-LV/LEMSCo Involvement,
Soil Assessment-Indicator Chemicals H-1
APPENDDC I: Supporting Information on Results
of S ampling and Analytical Activitie s I-1
APPENDIX!: Summary of QA/QC Results J-l
APPENDIX K: Supplementary Results for the
Statistical Comparisons K-l
2-1 Soil Assessment—Indicator Chemicals, Love
Canal Emergency Declaration Area 2-3
2-2 Soil Assessment-Indicator Chemicals, Love
Canal EDA Neighborhoods 2-7
2-3 Soil Assessment-Indicator Chemicals, EDA
S ampling Areas 2-12
VI
-------
CONTENTS
(continued)
Figures (continued) Page
2-4 Soil Assessment—Indicator Chemicals, Location
of EDA and Comparison Areas in New
York State 2-13
4-1 Soil Assessment-Indicator Chemicals, Project
Organization 4-2
6-1 Soil Assessment—Indicator Chemicals,
Locations of Analyzed Samples, EDA 6-6
6-2 Soil Assessment-Indicator Chemicals,
Locations of Analyzed Samples,
Cheektowaga Comparison Area 6-7
6-3 Soil Assessment-Indicator Chemicals,
Locations of Analyzed Samples,
Tonawanda Comparison Area 6-8
6-4 Soil Assessment—Indicator Chemicals,
Locations of Analyzed Samples,
Census Tract 221 Comparison Area 6-9
6-5 Soil Assessment-Indicator Chemicals,
Locations of Analyzed Samples,
Census Tract 225 Comparison Area 6-10
6-6a-h Soil Assessment-Indicator Chemicals,
Histograms Showing LCIC Concentrations
by Sampled Area 6-27
6-7 Soil Assessment-Indicator Chemicals,
Box Plot Example 6-35
6-8a-h Soil Assessment-Indicator Chemicals,
Box Plots Showing LCIC Concentrations
by Sampled Area 6-37
6-9 Soil Assessment-Indicator Chemicals,
Summary of Univariate Comparisons:
EDA Sampling Area to Comparison Area 6-50
VII
-------
CONTENTS
(continued)
Figures (continued) Page
6-10 Soil Assessment-Indicator Chemicals,
Summary of Multivariate Comparisons for
LCICs as a Group: EDA Sampling Area to
Comparison Area 6-51
6-11 Soil Assessment-Indicator Chemicals
Summary of Statistical Comparisons:
Comparison Area to Comparison Area 6-52
Tables
6-1 Number and Type of Field S amples Collected
and Sent for Analysis 6-2
6-2 Number of Field Samples Collected by Sampling
Area, Neighborhood, and Sampling Team 6-3
6-3 Number of Field Samples Received and
Analyzed by Laboratory 6-5
6-4 Summary of LCIC Results Used for Statistical
Analyses 6-11
6-5 Number of Field QC Samples Received and
Analyzed by Laboratory 6-12
6-6 QC Data Summary 6-15
6-7a to h Summary of Analytical Results by LCIC
and Sampling Area 6-19
6-8 Results of Nonparamerric Univariate
Statistical Comparisons with Observations
Classified as Good 6-45
6-9 Results of Nonparametric Multivariate
Statistical Comparisons with Observations
Classified as Good 6-48
vm
-------
CONTENTS
(continued)
Tables (continued) Page
6-10 Summary of Nonparametric Univariate
Statistical Comparisons with Observations
Classified as Good 6-49
IX
-------
ACKNOWLEDGEMENTS
The authors gratefully acknowledge the efforts of all organizations and in-
dividuals who participated in the conceptualization, design, and implemen-
tation of this study. Special acknowledgement for their continuing guidance
goes to the U.S. Environmental Protection Agency (EPA) Region II and the
other agencies composing the Technical Review Committee (TRC): the U.S.
Department of Health and Human Services/Centers for Disease Control
(DHHS/CDC), the New York State Department of Health (NYSDOH), and
the New York State Department of Environmental Conservation (NYSDEC).
Various agency personnel also made the following contributions:
• EPA Region Il-direction of the overall study
• EPA Environmental Monitoring Systems Laboratory in Las Vegas
and its contractor, Lockheed Engineering and Management Services
Company—assistance with revision of analytical method; technical
audits; review of analytical data; quality assurance oversight
• EPA National Enforcement Investigation Center, and its subcontrac-
tor, TechLaw, Inc.--evidentiary audits of the sample collection,
preparation, and analysis programs
• DHHS/CDC--assistance with revision of the analytical method and
review of statistical design and analysis
• NYSDEC-liaison with the public throughout the study
• NYSDOH—acquisition of homeowner permission to collect samples
and distribution of analytical results to homeowners
CH2M HILL served as prime contractor and, under the direction of EPA
Region n, managed the overall study. The participation of the following or-
ganizations is gratefully acknowledged:
• Aquatec, Clayton Environmental Consultants, Environmental
Monitoring Systems, NUS Corporation, and Versar-sample analysis
-------
• Black & Veatch-sample collection and preparation for shipment
• Cambridge Analytical Associates-assistance with revision of
analytical method; sample analysis and laboratory management sup-
port; assistance with statistical design and analysis; and review and
evaluation of analytical data
• DPL & Associates, Analytical Sciences, Inc., and Princeton Univer-
sity—assistance with statistical design and analysis
• Ecology and Environment-preparation of field equipment; logisti-
cal support; sample collection
• Horizon Systems Corporation-automated systems support; develop-
ment, installation, and maintenance of software systems
• Wen Laboratory Associates—assistance in revising the analytical
method; assistance in managing the laboratory team throughout the
analytical phase; and review and evaluation of analytical data
XI
-------
LIST OF ACRONYMS AND ABBREVIATIONS
BQC Blind quality control
CDC Centers for Disease Control
DHHS U.S. Department of Health and Human Services
EDA Emergency Declaration Area
FHB Field handling blank
EPA U.S. Environmental Protection Agency
EMLS-LV U.S. EPA Environmental Monitoring Systems Laboratory
in Las Vegas
GC/MS Gas chromatograph/mass spectrometer
LCIC Love Canal Indicator Chemical
LEMSCo Lockheed Engineering and Management Services Company
MHB Method/holding blank
MS/MSD Matrix spike/matrix spike duplicate
NEIC U.S. EPA National Enforcement Investigation Center
NYSDEC New York State Department of Environmental
Conservation
NYSDOH New York State Department of Health
OTA U.S. Congressional Office of Technology Assessment
ppb Parts per billion
PHB Preparation handling blank
QAPP Quality Assurance Project Plan
QA Quality assurance
QC Quality control
SAS Statistical Analysis System
Xll
-------
LIST OF ACRONYMS AND ABBREVIATIONS
(continued)
SIM Selected ion monitoring
S S B S hipping and storage blank
TCDD 2,3,7,8-Tetrachlorodibenzo-p-dioxin
TRC Technical Review Committee
Xlll
-------
1.0 SUMMARY
The Love Canal Emergency Declaration Area; Proposed Habitability
Criteria document (NYSDOH and DHHS/CDC, 1986) calls for the perfor-
mance of three environmental studies. The findings of one of these studies,
the soil assessment for indicator chemicals, are reported here. The findings
of the air assessment for indicator chemicals and the soil assessment for
2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) are documented in separate
reports. Other reports in this series describe the decision-making process
during the development of the habitability study and the subsequent peer
review of its three environmental assessments.
The purpose of this study was to provide the Commissioner of Health for the
State of New York with information for use in making a decision on the
habitability of the Emergency Declaration Area (EDA). As directed by the
Technical Review Committee (TRC), which oversees the habitability study,
the purpose of this assessment is to present data and is specifically not to
develop conclusions or interpretations of the data.
As called for in the habitability criteria, the specific goal of this assessment
was to compare the concentrations of Love Canal Indicator Chemicals
(LCICs) in the soils of the 13 EDA neighborhoods to the concentrations of
those chemicals in the soils of three selected comparison areas. The com-
parison areas were western New York communities that are similar to the
EDA but are not close to a known chemical landfill. The original LCICs
called for in the habitability criteria were four specific chlorinated benzene
compounds, two BHC isomers, total BHC, and 2-chloronaphthalene.
Chlorobenzene was dropped from the soil assessment based on poor perfor-
mance in volatile analysis during a pilot study. The three other chlorinated
benzene compounds were retained. Improvements to the semivolatile
analytical method allowed two additional BHC isomers to be added which
provided a more accurate measure of total BHC. The final list of LCICs used
for the soil assessment included 1,2-dichlorobenzene; 1,2,4-trichloroben-
zene; 1,2,3,4-tetrachlorobenzene; 2-chloronaphthalene; alpha-BHC; delta-
BHC; beta-BHC; and gamma-BHC. (Use of the four BHC isomers replaced
the need for total BHC.)
l-l
-------
The study was designed to perform a statistical comparison between EDA
sampling areas and comparison areas. The EDA comparison areas were
formed from neighborhoods on the basis of recommendations from a peer
review of a previously conducted pilot study. The comparison method as-
sumes that data collected from the comparison areas correspond to "normal"
or habitable conditions, and that comparing these data to similar measure-
ments from the EDA will indicate whether differences exist.
The analytical chemistry method used in the soil assessment for LCICs was
developed over the course of the pilot study and was further modified before
the soil samples were collected. The results of a previous sampling study
conducted in the EDA in the summer of 1980 (U.S. EPA, 1982) indicated
that a large percentage of contaminant levels were recorded as non-detects;
therefore, low detection limits were needed to provide the data required for
the statistical comparisons. A method utilizing a gas chromatograph/mass
spectrometer (GC/MS) operating in the selected ion monitoring (SIM) mode
was developed specifically for this study to provide typical operational detec-
tion limits in the 0.2- to 0.3-part-per-billion (ppb) range for all eight LCICs.
The actual detection limits for a particular sample depended on the individual
LCIC, sample interferences, and the GC/MS instrument. For many samples
with limited interferences, LCICs were detected at levels below 0.2 ppb.
An information management system was developed to expedite the schedule
of the study, minimize transcription errors, and facilitate quality as-
surance/quality control (QA/QC) and statistical manipulation of the analyti-
cal results.
To monitor and maintain quality in the collection, management, and use of
environmental data, QA/QC was a key element in the design of the study.
Thus, strict QA/QC requirements and protocols were implemented in the
field sampling, chemical analysis, statistical analysis, and data management
efforts.
Sample collection was conducted during October and November 1987. Ac-
tual implementation of the planned sample allocations among sampling
areas, sampling crews, and laboratories proceeded with no significant devia-
tions from the plan. A total of 879 prepared field samples were sent to seven
laboratories for analysis. Seven hundred eighty-one of these samples were
successfully analyzed for each of the eight LCICs, resulting in
1-2
-------
6,248 analyses. A total of 5,320 valid analytical results (85 percent) were
obtained for use in the statistical comparisons.
Specific quantitative criteria, or control limits, were set for several QC
measures. Certain of these criteria were mandatory, while others were ad-
visory only. For each such measure, a goal was established that either 90 or
100 percent of the analytical results for each LCIC should meet the criteria.
These goals were achieved for all of the QC measures, with the minor ex-
ception of one type of QC sample analysis that did not meet advisory con-
trol limits. The validated results, which met mandatory control limits, were
used in the statistical analysis.
A "nonparametric" statistical method (the Wilcoxon rank-sum test) was
selected for assessing differences in LCIC concentrations because it required
generally less restrictive assumptions than alternative methods, while main-
taining high "power" or ability to detect differences between neighborhoods.
This method also was well-suited to data with relatively large fractions of
nondetectable values. A related test was used to conduct simultaneous (mul-
tivariate) comparisons for all LCICs.
The comparisons between LCIC concentration distributions in the EDA sam-
pling areas and comparison areas indicated some significant differences at
the 0.05 and 0.01 significance levels. (A 0.01 significance level means that
there is a 1-in-100 chance that a difference as large as that observed could
occur by chance alone. A 0.05 significance level corresponds to a l-in-20
chance.) The results, as they relate to the habitability of the EDA, have not
been interpreted in this report. These findings will be reviewed by a panel
of scientists. In accordance with the habitability criteria document, the
results will be presented to the Commissioner of Health for the State of New
York for use in a determination of habitability.
1-3
-------
2.0 INTRODUCTION
This report is Volume HI of a five-volume series. Volume I provides an in-
troduction and documentation of the decision-making during the develop-
ment of the Love Canal EDA Habitability Study. For readers not familiar
with the Love Canal EDA Habitability Study, it is recommended that
Volume I be read before this volume. Volume II reports on the air assess-
ment for LCICs. Volume IV presents the results of the soil assessment for
2,3,7,8-TCDD, and Volume V will summarize the subsequent peer review
of Volumes n, III, and IV and the responses to that peer review.
This document summarizes the design and results of the soil assessment for
LCICs. It is presented in six main sections and contains 11 appendices. This
section gives an overview of the history of the site and a summary of pre-
vious environmental studies conducted in the EDA. Sections 3.0, 4.0,
and 5.0 discuss the study goals, design, and methods, respectively. Sec-
tion 6.0 presents the study results. The appendices contain more detailed,
technical information on the study methods, results, field and laboratory
audits, and analytical data validation.
2.1 BACKGROUND
A detailed history of Love Canal is presented in Volume I. This section
provides a brief summary to assist the reader in understanding the overall in-
tent of the Love Canal EDA Habitability Study and the role of the soil as-
sessment for LCICs.
The former Love Canal landfill was a rectangular, 16-acre tract of land lo-
cated in the southeastern end of the City of Niagara Falls in Niagara Coun-
ty on the western edge of New York State. The landfill takes its name from
William T. Love, whose plan in the 1890s was to dig a canal between the
upper and lower Niagara River to provide inexpensive hydroelectric power
for a proposed model industrial city. The model city project and the partial-
ly dug canal were abandoned before the turn of the century. The abandoned
canal was used as a chemical and municipal waste dump from 1942 to 1954.
In 1954, the site was closed. Subsequently, home-building accelerated
around Love Canal, and the 99th Street School was built adjacent to it. Soil
2-1
-------
from the canal was moved to the 93rd Street School site where it was used
as fill. Anecdotal information indicates that fill material from the canal area
may also have been moved to other locations within the EDA.
In the spring of 1978, following a series of complaints by local residents,
studies were initiated to investigate the health and environmental problems
at Love Canal. Under the direction of the New York State Department of
Health (NYSDOH), basement sumps were sampled, and, under the direction
of the U.S. EPA, air samples were taken at homes abutting the canal. Sample
results revealed significant contamination, and in August 1978, the first of
two states of emergency was declared by President Carter. Homes were sold
on a voluntary basis and remedial efforts were under way by October 1978.
Initial remediation of Love Canal concentrated on site containment. A clay
cap was installed over the canal area, and perimeter/barrier drains were in-
stalled during 1979. A leachate treatment plant was constructed during late
1979 to handle leachate from the collection system. Since that time, an ex-
tensive program for removing contaminated sediments from area sewers and
creeks has been under way, along with improvements to the final cover sys-
tem completed in 1984. (Volume I of this report provides additional details
on Love Canal remediation efforts.)
Meanwhile, concern persisted about the habitability of the residential area
surrounding Love Canal. In May 1980, President Carter issued a second
emergency declaration for Love Canal establishing the EDA as shown in
Figure 2-1. Eligible properties were purchased on a voluntary basis from
residents by the Love Canal Area Revitalization Agency (established by the
New York State Legislature).
In the summer of 1982, the EPA released a report assessing the extent of air,
water, and soil contamination in the EDA. The report was intended for use
as a basis for making recommendations regarding future use of the area (U.S.
EPA, 1982). Subsequently, the U.S. Department of Health and Human Ser-
vices (DHHS) was requested to review the EPA study and other data to deter-
mine whether the EDA was habitable. In July 1982, after considering
comments by the National Bureau of Standards on the procedures used in
the EPA study and after further consultation with the EPA, DHHS affirmed
an earlier provisional decision that the EDA was as habitable as the areas
with which it had been compared. This decision was contingent on the
provision that the storm sewers and their drainage tracts be cleaned and that
2-2
-------
•KOSTRECT
SOHOOt S/TE
IUMOEA awMwrt snton
VICINITY MAP
SOURCE:
EDA BOUNDARIES TAKEN FROM NEW YORK STATE
REAL PROPERTY TAX LAW ARTICLE 17. SECTION 1702
Figure 2-1
SOIL ASSESSMENT—INDICATOR CHEMICALS
LOVE CANAL EMERGENCY DECLARATION AREA
Niagara Falls, New York
SCALE: 1"--750'
LEGEND
—^— EMERGENCY DECLARATION AREA (EDA) BOUNDARY
-------
the remedial actions "...be re-evaluated to guarantee permanent containment
of chemicals in the dump" (NYSDOH andDHHS/CDC, 1986).
In December 1982, the U.S. Congressional Office of Technology Assess-
ment (OTA) was requested to examine the technical basis for and validity of
the habitability decision for the EDA. Questions arose pertaining to the
quality and the limits of detection of some of the previously collected en-
vironmental data. In June 1983, the OTA reported that, with the informa-
tion available, it was impossible to conclude whether or not unsafe levels of
toxic contamination existed in the EDA and that the analysis of available
data did not support the DHHS decision that the EDA was as habitable as
the areas with which it had been compared (OTA, 1983).
In August 1983, in response to the OTA report, the EPA established a TRC
composed of representatives from the EPA, NYSDOH, DHHS/Centers for
Disease Control (CDC), and the New York State Department of Environ-
mental Conservation (NYSDEC) to coordinate and oversee the habitability
study and remedial program at Love Canal. The member agencies of the
TRC asked NYSDOH and DHHS/CDC to develop criteria that would be
considered by the New York State Commissioner of Health in deciding
whether or not the EDA is habitable. The habitability criteria were developed
by NYSDOH and DHHS/CDC with extensive input from the public and an
advisory panel of scientists. The habitability criteria underwent peer review
(Life Systems, 1986) and in 1986, the Love Canal Emergency Declaration
Area; Proposed Habitability Criteria document (NYSDOH and
DHHS/CDC, 1986), which reflects the concerns of the peer review, was
issued.
2.2 HABITABILITY CRITERIA
The habitability criteria stipulate that for chemicals identified in environmen-
tal media to which current and potential residents may have significant ex-
posure, relevant Federal and New York State standards, criteria, and
guidelines will be used to assess the habitability of the EDA. The only known
applicable guideline, however, is the CDC level of concern of 1.0 ppb for
2,3,7,8-TCDD in residential surface soil. Thus, the habitability criteria re-
quire environmental sampling of residential surface soil in the EDA for
2,3,7,8-TCDD.
2-4
-------
Because no known standards, criteria, or guidelines are available for the other
chemicals known or suspected to have been deposited in Love Canal, the
criteria call for a comparison of the results of environmental sampling in
neighborhoods (soil) and residences (air) in the EDA with the results of sam-
pling in similar inhabited communities in western New York. (As stated ear-
lier, the results of the air assessment were presented in Volume n of this
series.)
The habitability criteria require that the comparisons be made on the con-
centrations of LCICs in soil in each neighborhood in the EDA and in the
comparison areas. Eight LCICs were initially selected for the soil assess-
ment:
• Total BHC
• Beta BHC
• Gamma BHC
• Chlorobenzene
• 1,2-dichlorobenzene
• 1,2,4-trichlorobenzene
• 1,2,3,4-tetrachlorobenzene
• 2-chloronaphthalene
These chemicals were selected to represent a larger number of chemicals
potentially originating from Love Canal and to assess the potential chemical
contamination of the EDA by Love Canal (NYSDOH and DHHS/CDC,
1986, Appendix 9). (The list of LCICs used in the soil assessment was later
revised, as discussed in Sections 2.3 and 5.2. of this report.)
The comparison method assumes that data collected from similar western
New York communities not near a chemical landfill will correspond to "nor-
mal" or habitable conditions and that comparing these data with similar
measurements in the EDA will provide an indication of whether EDA neigh-
borhoods are different from the selected comparison areas. The habitability
criteria document states that one of the conditions for a neighborhood in the
EDA to be considered habitable is the following:
.. .the chosen aggregate values (e.g., mean, median, percentiles, etc.) of
each non-TCDD LCIC evaluated both individually (univariate) and col-
lectively (multivariate) are not significantly different than the values
from the comparison areas.. .(NYSDOH and DHHS/CDC, 1986)
2-5
-------
Comparison areas were selected based on the following general characteris-
tics. The comparison areas should:
• Be inhabited residential census tracts in western New York State
• Contain soil types and hydrogeological conditions similar to those
found in the EDA
• Have borders located as far from known chemical landfills as pos-
sible, but no less than one-half mile
The comparison area initially selected was an aggregate (combination) of
neighborhoods in Tonawanda and Cheektowaga, New York. (Additional
comparison areas were selected after peer re view of the pilot study, discussed
in Section 2.5.)
Neighborhoods within the EDA were delineated during the development of
the habitability criteria to facilitate a habitability decision on a neighborhood
basis. The EDA neighborhoods are shown in Figure 2-2.
2.3 PILOT STUDY
In preparation for the soil assessment for LCICs, a pilot study was conducted
to demonstrate the feasibility of implementing the habitability criteria as
proposed (CH2M HELL, 1987c). Specific objectives were to:
• Determine if specially developed techniques for soil sample collec-
tion and preparation were suitable for use in the soil assessment for
LCICs
• Develop and evaluate new analytical procedures for measuring
volatile and semi volatile LCICs in soil at low-level (i.e., 1 to
10 ppb) concentrations
• Provide information on the sources and magnitude of in-
tralaboratory and interlaboratory variability in the analytical data
• Develop methods for estimating LCIC detection limits
2-6
-------
.->:.
Figure 2-2
SOIL ASSESSMENT—INDICATOR CHEMICALS
LOVE CANAL EDA NEIGHBORHOODS
Niagara Falls, New York
SCALE: r-6501
LEGEND
^— NEIGHBORHOOD BOUNDARY
FENCE LINE AROUND LOVE CANAL REMEDIATION SITE
UWDFILL
SOURCE: NEIGHBORHOOD BOUNDARIES ADAPTED FROM THE PROPOSED
HA8ITABILITY CRITERIA DOCUMENT (NYSDOH AND DHHS/CDC. 1986).
-------
• Provide data of known and defensible quality to estimate the statisti-
cal distribution of LCIC concentration data in the EDA and com-
parison areas
• Provide a basis for determining the number of samples that need to
be taken to produce statistically valid results in the soil assessment
forLCICs
A total of 90 soil samples were collected in the EDA, Cheektowaga, and
Tonawanda during July 1986. Samples were collected using a Shelby tube
that was hydraulically pushed into the soil by a Porta Sampler. This provided
an undisturbed soil sample to a depth of about 1 foot. Samples were analyzed
using a GC/MS operating in the SIM mode. The average detection limits for
each of the eight LCICs were estimated to range from 0.5 to 2.5 ppb during
the pilot study. The detection limits provided by the analytical method were
significantly lower than those provided in the 1980 EPA study, generally by
a factor of more than 100.
The following are the conclusions and recommendations of the pilot study
that were incorporated in the soil assessment for LCICs.
• The sample collection method worked well in obtaining a relatively
undisturbed soil sample. However, the full 13-inch depth of the
sample was difficult to impossible to obtain in areas of fill and rub-
ble. It was recommended that the sample depth requirement be
reduced to generate the minimum amount of soil required for the
analytical method.
• The specially developed analytical method was generally capable of
reliably identifying and quantifying LCIC concentrations at the 1-
ppb level with the exception of the BHCs, which had higher detec-
tion limits. Some low-level contamination of laboratory blank
samples was encountered. The analytical method was revised to
reduce the likelihood of laboratory contamination. Variability in the
results suggested that the semivolatile sample extraction procedures
required modifications and that the laboratories should be trained to
obtain a more uniform application of the extraction procedure.
• 1,2-dichlorobenzene was analyzed by both volatile and semivolatile
analysis and exhibited poor chromatographic performance in the
volatile analysis. Higher concentrations were observed in the
2-8
-------
semivolatile analysis. Based on these results, it was recommended
that only the semivolatile analysis be retained for the analytical
process and that chlorobenzene be deleted from the list of indicator
chemicals for the soil assessment, because chlorobenzene was the
only LCIC besides 1,2-dichlorobenzene that was analyzed by the
volatile analysis and it was not detected in any of the samples.
(Chlorobenzene would still be in the habitability study as an in-
dicator chemical for the air assessment.)
• Interlaboratory variability was found to be significant compared to
between-sampling-area variability. It was recommended that
sample allocation among the laboratories be designed to minimize
the effect of interlaboratory variability on the results of the com-
parison in this study.
• An estimator of the method detection limits was developed. The es-
timator was complicated by the sample-specific interferences and
the qualitative nature of the analyte identification criteria.
• No clear evidence of patterns or areas of localized contamination
within neighborhoods existed. This indicated that a randomized
method of selecting sample sites would be appropriate.
• Samples should be allocated among neighborhoods in proportion to
the area of the neighborhood in order to preserve a more or less
uniform spatial density of the sampling points.
• Based on the pilot study data, a minimum number of samples per
neighborhood was set to preserve the power (see Appendix A) of
detecting differences between small neighborhoods and comparison
areas. (Neighborhoods were later combined to form sampling areas
based on peer review recommendations as discussed in Section 2.5.)
Further discussion of the pilot study findings can be found in the pilot study
report.
2.4 NYSDOH SOIL STUDY
A higher percentage of detects of 1,2,4-trichlorobenzene and 1,2,3,4-
tetrachlorobenzene was found in the EDA than in the comparison areas
2-9
-------
during the pilot study. There was reason to believe that the source of
trichlorobenzene and tetrachlorobenzene might be other than Love Canal. In
an attempt to determine whether the 1,2,4-trichlorobenzene and 1,2,3,4-
tetrachlorobenzene soil contamination was widespread in the Niagara Falls
area or restricted to the EDA and to identify the source of contamination,
NYSDOH conducted a sampling program during November 1986. Soil
samples were collected in the EDA, Cheektowaga, and three new areas in or
near the City of Niagara Falls. The areas sampled were selected to account
for significant differences among areas that might be the result of an upwind
industrial source, the City of Niagara Falls water supply (used for watering
lawns, gardens, etc.), or other unknown sources.
The results of this study indicated that similar concentrations of 1,2,4-
trichlorobenzene and 1,2,3,4-tetrachlorobenzene occurred throughout the
areas sampled in the greater Niagara Falls area and were not elevated specifi-
cally within the EDA. However, the source of the contamination was not
conclusively identified.
Using a different analytical method, the NYSDOH study had attained detec-
tion limits of around 0.2 ppb for 1,2,4-trichlorobenzene and 1,2,3,4-
tetrachlorobenzene and found significant numbers of samples with
concentrations above this level but less than 1 ppb. Thus, NYSDOH re-
quested that the analytical method used in the pilot study be further modified
for this soil assessment to provide detection limits as low as possible with a
target of 0.1 to 0.2 ppb for the majority of the LCICs.
2.5 PEER REVIEW
A scientific peer review of the pilot study was conducted in March 1987 (Life
Systems, 1987). The following are the recommendations of the peer
reviewers that were incorporated into the study.
• Based on the results of the independent study by NYSDOH, addi-
tional comparison areas should be added in Niagara Falls.
• Sample depth should be at least 7 and not more than 12 inches,
depending on the ability to collect the samples. Further details on
the recommended sampling protocol can be found in the document
2-10
-------
summarizing responses to the peer review of the pilot study
(CH2MHILL,1987d).
• Only the semivolatile analysis should be retained, thus eliminating
chlorobenzene as an LCIC for the soil assessment
• Specified EDA neighborhoods should be combined to form seven
sampling areas for purposes of sample allocation and statistical com-
parisons. (This was recommended to make the sampling density
similar for each area compared and reduce the number of com-
parisons.) The neighborhoods should be combined as shown in
Figure 2-3. Neighborhood 1 should not be combined, even though
it is small relative to the other neighborhoods, because of the high
number of LCIC detects in that area. This neighborhood may be
more densely sampled than other neighborhoods in the EDA.
• The random sampling method should be modified by using a grid to
create a more uniform spatial distribution of samples.
In response to the peer reviewers' comments, and based on the pilot study
report, the TRC selected two additional comparison areas. Specifically, por-
tions of Census Tracts 221 and 225 in the City of Niagara Falls were selected
based on the habitability criteria guidelines. The relative locations of the
comparison areas and the EDA are shown in Figure 2-4.
Sections 3.0,4.0, and 5.0 describe how these peer review recommendations
were incorporated into the study goals, design, and methods.
2-11
-------
Rgire 2-3
7
SOIL ASSESSMENT—INDICATOR CHEMICALS
EDA SAMPLING AREAS
SCALE: r'6501
LEGEND
13 EMERGENCY DECLARATION AREA (EDA) NEIGHBORHOOD BOUNDARY
— FENCE LINE AROUND LOVE CANAL REMEDIATION SITE
EDA SAMPLING AREA BOUNDARY LA/vD^.a
SOURCE; NEIGHBORHOOD BOUNDARIES ADdPTED FROM THE PROPOSED
HABITABILITY CRITERIA DOCUMENT (NYSDOH AND DHHS/CDC, 19861
-------
Census Tract
225
Comparison
Area
Census Tract 221
Comparison Area
TONAWANDA
GRAND ISLAND
TONAWANDA
Tonawanda
Compari
Cheektowaga
Comparison
Area
CHEEKTOWAGA
Scale: 1" = 3 Mile*
(Approx.)
Legend
Comparison Area
EDA
Figure 2-4
SOIL ASSESSMENT-INDICATOR CHEMICALS
LOCATION OF EDA AND COMPARISON AREAS
IN NEW YORK STATE
-------
3.0 GOALS
The purpose of this study, in accordance with the proposed habitability
criteria (NYSDOH and DHHS/CDC, 1986), was to provide the New York
State Commissioner of Health with information for use in making a decision
on the habitability of the EDA. As called for in the habitability criteria docu-
ment, the specific goal of this assessment was to provide a statistical com-
parison of the concentrations of LCICs in the EDA neighborhoods with
concentrations in each of the selected comparison areas on a univariate (in-
dividual) and multivariate (collective) LCIC basis.
To conduct the comparisons and to provide data of known and defensible
quality, the following specific objectives were developed:
1. Meet the overall QA/QC objectives detailed in the sample collection and
preparation Quality Assurance Project Plan (QAPP [CH2M HILL,
1987a]) and in the sample analysis QAPP (CH2M HILL, 1987b).
2. Use a sufficient number of samples so that the statistical comparisons
can achieve a 90 percent probability of being able to detect, with 95 per-
cent confidence, an order-of-magnitude difference in median concentra-
tion of LCICs between EDA sampling areas and the comparison areas.
The statistical considerations used to develop this objective are dis-
cussed in Appendix 10 of the habitability criteria document.
3. Achieve the lowest technically feasible analytical detection limits.
3-1
-------
4.0 DESIGN
The overall design of the soil assessment for LCICs was based on the
proposed habitability criteria (NYSDOH and DHHS/CDC, 1986), the results
of the pilot study (CH2M HILL, 1987c), and TRC directives. This design
involves three main elements: project organization, study design, and
QA/QC design. Each of these three elements is discussed below.
4.1 PROJECT ORGANIZATION
Project organization and responsibilities of key personnel are detailed in the
soil sample collection and preparation QAPP (CH2M HILL, 1987a) and the
soil sample analysis QAPP (CH2M HILL, 1987b). Organizational respon-
sibilities are illustrated in Figure 4-1.
EPA Region n directed the project and provided technical guidance as a
member of the TRC. Technical guidance was also provided by the other
members of the TRC: DHHS/CDC, NYSDEC, and NYSDOH. The data
that were collected and analyzed were subjected to independent review by
Environmental Monitoring Systems Laboratory in Las Vegas/Lockheed En-
gineering and Management Services Company (EMSL-LV/LEMSCo.)
CH2M HILL served as the prime contractor.
Individual teams were organized to execute the five major components of
the project, as shown on Figure 4-1: sample collection, sample preparation,
sample analysis, real-time laboratory QC, and statistical design and analysis.
Additional teams provided automated systems support and laboratory-re-
lated assistance throughout the project.
4.2 STUDY DESIGN
The overall design addressed all phases of the study, including sampling
design, analytical design, and data base development.
The statistical design affects most aspects of the data collection and analysis
and accordingly is discussed first here. Although the technical detail
4-1
-------
W63394.T1(SA)
TRC
NYSDEC
NYSDOH
U.S. DHHS/CDC
U.S. EPA
Community Relations
NYSDOH
NYSDEC
EPA
Region II
Project Management
EMSL-LV/LEMSCO
QA/QC
Data Validation
Method QA Oversight
Independent Real-Time QC
Overall
Laboratory Related Support
Cambridge Analytical Assoc.
CH2M HILL
(Prime Contractor)
Project Management
QA/QC
Automatic
Data
Processing
Horizons Systems
Corp.
Sample
Collection
Black & Veatch
Ecology & Environment
Sample
Preparation
Black & Veatch
Cambridge Analytical Assoc.
Sample
Analysis
Aquatec
Cambridge Analytical
Associates
CH2M HILL
Clayton Environmental
Consultants
Environmental Monitoring
Systems, Inc.
NUS
Versar
Real-Time QC
Data Validation
Cambridge Analytical Assoc.
CH2M HILL
Wen Laboratories Assoc.
Statistical Design
and Analysis
Analytical Sciences, Inc.
CH2M HILL
DPL & Assoc.
Princeton Univ.
Figure 4-1
SOIL ASSESSMENT - - INDICATOR CHEMICALS
PROJECT ORGANIZATION
-------
included in this section is purposely kept to a minimum, the reader unfamiliar
with basic statistical principles may find it helpful to read Appendix A, Statis-
tical Concepts, as a supplement both to this section and to Section 5.5, Statis-
tical Analysis. Appendix B provides an in-depth and more technically
oriented discussion of the statistical design and methods used in this study.
The statistical design of any sampling program should be tied to the charac-
teristics of the data that are to be analyzed. In this study, the data are the soil
concentrations at specific locations in the EDA and comparison areas.
Theoretically, the concentration in any sample can be considered to reflect
four effects, or factors:
1. Regional characteristics common to all EDA sampling areas and com-
parison areas (for instance, regional geologic characteristics)
2. Local effects specific to the given EDA sampling or comparison area
3. Laboratory characteristics specific to the laboratory at which the chemi-
cal analysis was conducted
4. All other effects, including spatial variations within a sampling area,
intra-laboratory measurement variability, and field data collection and
sample handling variability
The primary motivation in the sampling design was to isolate and measure
the effects of factor 2, differences between the EDA sampling areas and the
comparison areas. In this respect, factor 1 is irrelevant, because it affects all
samples, regardless of area. Factors 3 and 4 can be considered to represent
external effects that must be carefully controlled to avoid biasing or invalidat-
ing the results. For instance, assume that all samples from Comparison
Area 1 were collected by sampling crew A, and all samples from EDA Sam-
pling Area 2 were collected by sampling crew B. Then if samples collected
by sampling crew A tended to have consistently higher concentrations than
those collected by sampling crew B (for instance, because of improper
sample handling, resulting in contamination of the samples), a conclusion
that soil concentrations in Comparison Area 1 and EDA Sampling Area 1
were statistically different might well be invalid.
Randomization is one way to prevent external effects from biasing results.
The idea behind randomization is that if an external effect is properly dis-
4-3
-------
tributed over the known factors, then statistically it will affect the known fac-
tors equally and will not bias the results. However, randomizing the un-
wanted external effects instead of estimating them explicitly may obscure
some of the true differences, if any, between EDA sampling areas and com-
parison areas by increasing the variability of the data distribution. Increased
variability makes a larger number of samples necessary to detect differences
between areas. For this reason, it is desirable to minimize the external ef-
fects as much as possible through rigorous oversight and QA/QC procedures.
Randomization was insufficient to control for the effect of inter-laboratory
variability, which was found in the pilot study to be larger than other exter-
nal effects. Accordingly, "blocking," another statistical procedure to control
external effects, also was used. To block inter-laboratory effects, all statis-
tics about an area are accumulated using only data analyzed by a given
laboratory. Then these statistics (one for each laboratory) are combined to
form a grand statistic that is used to perform the comparisons.
In summary, the statistical design that was adopted treated external effects
as follows:
• The effects of sampling and laboratory analysis protocol, including
the assignment of sampling teams to neighborhoods, the assignment
of samples collected by each team to laboratories, the order of
sample collection, and the order of sample analysis by each
laboratory, were randomized. In addition, the QA/QC protocol
acted to keep the magnitude of these external factors to a minimum.
• The effects of inter-laboratory variations were accounted for by
blocking.
Minimum sample sizes for each sampling area were determined from simula-
tions of the power of the proposed statistical tests, using the distribution from
the pilot study results and pilot study analytical detection limits. The mini-
mum sample size per sampling area was set at 75, with the exception of EDA
Sampling Area 1, for which the minimum sample size was 50. (Sampling
Area 1 is significantly smaller than the other sampling areas; thus,
50 samples for this sampling area represented a compromise between main-
taining approximately equal spatial sampling densities and maintaining ap-
proximately equal sample sizes for each sampling area. Sampling density
4-4
-------
was higher in Sampling Area 1.) The selection of minimum sample size is
discussed in more detail in Appendix B and the pilot study report.
Specific sampling sites within each sampling area were identified through a
multi-step process. First, potential sites were identified by overlaying a map
of the EDA sampling area or comparison area with a grid. The grid con-
tained three to four times as many nodes as the number of samples allocated
to that area. The nodes were then randomly numbered and chosen in numeri-
cal order. A determination was then made regarding whether or not it was
feasible to collect a sample at the proposed site. Feasibility included physi-
cal feasibility (e.g., a pervious surface) and legal feasibility (e.g., appropriate
landowner permissions granted).
The sampling design took into consideration several logistical constraints.
Because of the large number of soil samples to be collected, four sampling
teams were used so that all field sampling could be completed in a 4-week
period. Similarly, seven laboratories were used so that the laboratory
analysis could be completed quickly (within 30 days) to meet the study
schedule and administrative requirements for sample holding time (see Sec-
tion 5.2).
To avoid unknown biases that might result from the assignment in the field
of specific sampling areas or subareas to specific sampling teams,
laboratories, or sampling days, the samples to be collected were randomly
allocated prior to field work with the following constraints:
• Samples assigned to a sampling team on a given day should be near-
ly contiguous (in the same general area) to minimize team travel.
• On each day three sampling teams should be in the EDA and one in
a comparison area to optimize the use of surveying crews. (Com-
parison area sampling points were not surveyed.)
• Each of the four sampling teams should collect no more than a total
of 12 field samples each day to allow adequate time for quality as-
surance documentation.
• Each laboratory should receive approximately equal numbers of
samples from each sampling area to balance the effects of inter-
laboratory variability.
4-5
-------
• Each laboratory should receive no more than its capacity (six to
eight samples) each day to minimize its backlog of samples.
The second major issue in the study design was the development of an
analytical method using a GC/MS operating in the SIM mode. This method
provided typical operational detection limits in the 0.2- to 0.3-ppb range for
the LCICs. Further discussion of the analytical method developed for this
study is presented in Section 5.2.
The third element of the study design was the development of an informa-
tion management system to handle the large volume of data generated. The
extensive system designed for this purpose, described further in Section 5.3,
ultimately produced a final data base containing the data collected during the
study.
4.3 DATA QA/QC DESIGN
As discussed in Section 3.0, Goals, the study was designed to provide data
of known, defensible quality. To achieve this goal, several data QA/QC ob-
jectives were selected; these objectives relate to the data's precision, ac-
curacy, representativeness, completeness, and comparability. The following
paragraphs define the data QA/QC objectives and briefly summarize how
each is achieved and measured. The sample collection and analysis QAPPs
provide more detailed discussions of the QA/QC measures. Section 5.4
describes the overall study QA/QC program, and Section 6.2 provides an
overview of the QA/QC data assessment.
Precision
Precision is a measure of the reproducibility of measurements under a given
set of conditions. Specifically, it is a quantitative measure of the variability
of a group of measurements compared to their average value. The study
design provided for laboratory precision to be measured by splitting the field
samples and sending the splits to the same or to different laboratories. The
design provided for laboratory precision also to be measured by splitting
samples with known "spiked" contaminant concentrations and analyzing
both splits at the same laboratory.
4-6
-------
Accuracy
Accuracy measures the bias in a measurement system. Bias in this study may
result from procedures or contamination associated with sample collection,
shipment, preparation, or analysis. The study was designed to measure con-
tamination bias in these processes by introducing "blank" samples and
evaluating the results. The design also provided for the measurement of
potential bias in the sample analysis through the use of QC samples and
spiked samples.
Representativeness
Representativeness is a qualitative measure of how well the soil sample rep-
resents the surrounding soil and is a measure of how well the measured results
reflect the actual concentration or distribution of the chemical compounds in
the sampled soil. Sampling protocols were designed to obtain soil samples
that were representative of the areas sampled, and sample handling protocols
were developed to preserve the integrity, and thus the representativeness, of
the collected samples. Analytical protocols and analyses of spiked samples
allowed a qualitative assessment of the representativeness of measured
results with respect to actual concentrations in the soil sample.
Completeness
Completeness is a measure of the amount of valid data obtained from the
analytical measurement system. It is defined as the total number of samples
for which acceptable analytical data are generated, divided by the total num-
ber of samples collected, and multiplied by 100. The completeness objec-
tive for this study was to obtain valid analytical data for at least 90 percent
of the samples collected during the study.
Comparability
Comparability is a parameter expressing the confidence with which one data
set can be compared with another. Sample data should be comparable with
other measurement data for similar samples and sample conditions. Com-
parability for this study was designed to be achieved by the use of a narrow-
ly defined laboratory statement of work; training of laboratory personnel;
practice and performance evaluation on the analytical method; use of
4-7
-------
standard glassware and the same batches of solvents and reagents by all the
laboratories; strict quality control criteria for sample analysis; onsite inspec-
tions; and close daily monitoring of laboratory work quality through
electronic data transfer and telephone contact.
The design criteria and issues described in this section formed the basis for
the study methods selected and developed. Section 5.0 provides a discus-
sion of these methods.
4-8
-------
5.0 METHODS
To fulfill the project goals and design criteria described in Sections 3.0
and 4.0, methods were developed for sample collection and preparation,
sample analysis, data management, QA/QC, and statistical analysis. A
description of these methods follows.
5.1 SAMPLE COLLECTION AND PREPARATION
Field sample locations were identified as discussed in Section 4.2, Study
Design. The location of each EDA field sample was established relative to
the New York State Planar Coordinate System (using a permanent
benchmark) prior to sampling. Surveying these locations in the EDA was
essential in the event that there was a need to locate the sample site for
remediation. Comparison area field sample locations were not surveyed, but
their locations relative to permanent structures were measured and docu-
mented.
At each field location, leaf litter and sod were removed before the sample
was collected with a Zero Contamination Sampler (a stainless steel hand
sampler with a removable, 12-inch-long, 1-inch-internal-diameter, stainless
steel liner). To satisfy the minimum soil volume requirements for the analyti-
cal method, the minimum length of a collected sample was established as
7 inches. The Zero Contamination Samplers were decontaminated between
sample locations.
The removable liners housing the samples were individually packaged,
placed in coolers, and surrounded with refrigerant packs for shipping at a
temperature of approximately 4°C. The coolers were then sent via overnight
delivery to the preparation laboratory where the soil was extruded from the
liner and mixed prior to shipment to the analytical laboratories. Further
details of the field sampling methods can be found in the sample collection
and preparation QAPP (CH2M HILL, 1987a).
As each sample was collected, a sample collection form preprinted with
sample station number and property address was completed by the sampling
team. Chain-of-custody forms were completed for each sample, and traffic
5-1
-------
reports were completed daily for each cooler of samples shipped. Data from
the sample collection forms were cross-checked for discrepancies, and the
discrepancies were investigated and resolved. The sample collection form
data were entered daily into the sample collection data base using double data
entry with a computerized cross-check of the entered data.
After the coolers with the samples were received at the preparation laboratory
and their contents verified, the liners were placed in refrigerated (4°C) storage
until preparation later that day. The preparation process included extruding
the soil from the liner using a stainless steel hydraulic extruder, hand-mixing
the soil in a pan for 90 seconds using a spoon, and filling a glass jar with the
mixed soil sample. The head space in the jar was to be minimal, so any head
space was measured and noted. The extruder head was decontaminated after
each sample was prepared. The jars were packaged, placed in coolers, sur-
rounded with refrigerant packs, and shipped at approximately 4°C to the
analytical laboratories.
Data from the preparation laboratory activities were recorded using data sys-
tems developed specifically for this program (see Section 5.3). The sample
preparation forms generated by the data system provided information includ-
ing the sample number and the analytical laboratory to which the sample was
to be sent. Sample preparers entered the station number of the sample, along
with the date and time the sample was prepared, and signed the form. The
system then generated corresponding chain-of-custody forms and traffic
reports for shipment of the samples.
5.2 SAMPLE ANALYSIS
The analytical method used for sample analysis in this study was developed
to meet the study objective of achieving the lowest technically feasible detec-
tion limits. (The typical operational detection limit was 0.2 to 0.3 ppb for
the LCICs.) To allow the sample collection schedule to be somewhat inde-
pendent from the sample analysis schedule, the maximum allowable sample
holding time for extraction (an administrative requirement limiting the time
from sample collection to sample extraction) was increased from the 10-day
period used during the pilot study to 30 days, as discussed later in this sec-
tion. An overview of the analytical method and its development is provided
below.
5-2
-------
Analytical Method
A protocol using GC/MS/SIM for the analysis of soil samples for LCIC con-
centrations was developed, validated, and successfully used for the pilot
study. GC/MS operating in the SIM mode was chosen as an instrumentation
technique that could provide suitable detection limits and selectivity for
LCIC analysis. During the pilot study, LCICs were typically identified using
this technique at concentrations as low as 1 ppb.
Based on the experience gained from the pilot study, a number of modifica-
tions to the method were made prior to this soil assessment to achieve lower
detection limits and reduce chemical interferences for the BHC isomers.
These changes lowered the typical operational detection limits by a factor of
five, from about 1 ppb to 0.2 ppb for most of the LCICs. The actual, but un-
measurable, detection limits for a particular sample depended on the in-
dividual LCIC, sample interferences, and the GC/MS/SIM instrument.
Modifications made to reduce the hydrocarbon interferences allowed two ad-
ditional BHC isomers, alpha- and delta-BHC, to be analyzed with the
method, thus replacing the LCIC "total BHC" with its component isomers,
alpha-, beta-, gamma-, and delta-BHC. Other modifications were made to
increase laboratory throughput.
In general, the analytical method consisted of solvent extraction, extensive
cleanup, and analysis by GC/MS/SIM. To obtain low detection limits, reli-
able LCIC identification, and accurate quantification, an analytical protocol
was developed that detailed extraction procedures as well as calibration pro-
cedures and specified instrument parameters.
The initial development of the analytical method is described in the pilot
study report (CH2M HILL, 1987c). Modifications made to the analytical
method after the pilot study are summarized in more detail in Appendix C
of this document. A detailed description of the analytical method can be
found in the soil sample laboratory analysis QAPP (CH2M HILL, 1987b).
Holding Time Study
The administratively defined laboratory holding time for sample extraction
was increased from 10 days to 30 days. During the pilot study, the holding
time from sample collection to extraction was 10 days (the time typically
5-3
-------
allowed by the EPA's Contract Laboratory Program), and the holding time
from collection to analysis was 50 days. In order to allow decoupling of the
field and laboratory schedules, the holding time from sample collection to
extraction was increased while retaining the overall holding time from col-
lection to analysis. (Decoupling of the field and analytical schedules became
necessary when the field sampling schedule was accelerated to allow sam-
pling to be completed before the onset of inclement winter weather.) There
was no evidence during the pilot study that an increased extraction holding
time would affect the concentration of LCICs in the soil samples.
A holding time experiment was initiated prior to the start of sample analysis
to study the effect of extending extraction holding times; this experiment
proceeded concurrently with the soil assessment. The results of the holding
time study verified that extending the 10-day holding time during this study
did not have a significant effect on the concentrations of LCICs in the sample.
It should be noted that holding time effects, if any, would not affect the results
of the designed statistical comparisons, because the sequence of sample col-
lection (and subsequent analysis) was randomized. A more complete discus-
sion of the holding time study design and results is provided in Appendix D.
5.3 DATA AND INFORMATION SYSTEMS
DEVELOPMENT
A large volume of data was generated throughout the study, from the ran-
dom selection of field sample sites, through sample collection and prepara-
tion, to the analytical results and their validation. An information
management system was developed to capture, manage, and maintain the
data, and to provide a final data base containing all the data collected during
the study. The system expedited the study schedule by tracking samples,
maintaining shipment information, facilitating QA/QC of the analytical
results, and minimizing transcription errors.
On the basis of the experience gained from using software systems during
the pilot study and soil assessment for dioxin, software requirements and a
data flow diagram of the project's information environment were developed.
(More details are provided in Appendix E.) The subsequent design of the
overall information management system was influenced by two primary con-
siderations: the diverse locations of study activities and the complexity of
the data collection and storage. Study activities occurred concurrently in
5-4
-------
separate geographic locations across the country. Processing capabilities and
functions were therefore designed for personal computers wherever possible.
In addition, a project communications link was established between the
various project systems and personnel.
Virtually all data generated during each phase of the study were stored
through the use of software systems. The systems tracked the chain-of-cus-
tody, linked the analytical results and the sample locations, and supported
the complex analytical requirements of the study. The analytical data were
generated and captured electronically on each analytical instrument's
microprocessor.
Based on the software requirements for the study and the existing software
from both the pilot study and the soil assessment for dioxin, separate software
systems were designed. The 10 major software systems addressed site selec-
tion, sample collection, tracking sample preparation, sample analysis, real-
time (analytical) QC, data validation, double data-entry, data base modeling,
data base integration, and project-wide communications.
Eight of these systems were implemented on personal computers, while two
were implemented on the EPA mainframe system. The systems were
designed to be user-friendly and to provide editing capabilities. A Statisti-
cal Analysis System (S AS) relational data base was used for the resulting in-
tegrated data base, which includes 1.8 million cells of data. Appendix E
contains additional details on these systems.
5.4 QA/QC PROGRAM
As discussed in Section 4.3, the QA/QC program was designed to provide
data that are precise, accurate, representative, complete, and comparable.
This section summarizes the QA/QC methods used in the study, while Ap-
pendix F provides a more detailed program description. Specifically,
QA/QC methods were incorporated into each phase of the study to achieve
the following:
• Provide accurate field sample location information--1!^ was
achieved using several procedures. The EDA sample location was
surveyed, and the documentation was filled out by the survey crew
and the sampling crew. The documentation was checked for
5-5
-------
completeness and cross-checked for discrepancies. Missing infor-
mation and discrepancies were resolved and corrected. After sam-
pling, a map was generated showing the points sampled. The map
was then cross-checked against the survey documentation, and dis-
crepancies were resolved and corrected.
• Minimize and assess cross-contamination of samples-This was ac-
complished by implementing QA/QC procedures in the sample col-
lection and preparation processes. Blank samples were interspersed
at various stages in the sample collection, preparation, and analyti-
cal sequence to allow for a measure of cross-contamination, decon-
tamination efficiency, and potential errors that might be introduced
from sources other than the sample.
• Maintain correct link between analytical results and sample loca-
tions-Documentation was generated electronically, assigning the
sample identification number in the sample preparation laboratory.
The sample retained this identifier through the entire laboratory
analytical process. Documentation of sample shipment and chain-
of-custody was prepared, the documentation was cross-checked for
discrepancies, and discrepancies were resolved. All data entered in
the electronic data management system were reviewed and the
functioning of the data-storage software was checked to verify that
the system was operating correctly. The data entered in the system
were merged, and the merge was checked to provide the correct link
between analytical results and sample locations.
• Minimize variability among laboratory performances-Using the
newly developed analytical procedure detailed in the sample
analysis QAPP, a training session was conducted for the staff of
each of the analytical laboratories and evidentiary and technical
audits were conducted at each laboratory to assess whether the
laboratory had followed the QAPP. The evidentiary audits are docu-
mented in Appendix G. A real-time QC system also was used to
monitor the laboratory performance. Standard glassware was used
by all the laboratories to minimize variability introduced by using
slightly different glassware.
• Minimize potential for laboratory contamination and assess
laboratory contamination—On a project-wide basis, all solvents and
materials were analyzed. Batches with the lowest levels of
5-6
-------
contamination were identified, and a single source of solvents and
reagents was used for all the laboratories to minimize this potential
source of contamination. In addition, strict cleanup protocols were
incorporated into the analytical method. QC samples were used to
assess whether contamination had occurred, and appropriate actions
were taken to eliminate discovered sources of contamination and to
reanalyze samples discovered to be potentially affected.
• Assess ongoing laboratory performance--!^ analytical process in-
cluded the analysis of QC samples designed specifically to assess
the accuracy and overall performance of the analytical process at
each laboratory. Each laboratory internally reviewed and assessed
the results of its analytical process. Blind QC (BQC) samples were
sent to the laboratory, and the results were reviewed by an inde-
pendent agency review team (EMSL-LV/LEMSCo). In addition, a
computerized system allowed online monitoring of the laboratory
performance and a review of the results by the prime contractor and
the independent agency review team. Problems in laboratory perfor-
mance were then expeditiously corrected.
• Assess validity of data—The final analytical data were assessed and
validated by the independent agency review team. The team
flagged the analytical results with data qualifiers and data usability
flags. The data usability flags indicated the results that were recom-
mended for use in the statistical comparisons. The data validation
efforts are reported in Appendix H.
• Conduct evidentiary aitfftw-Evidentiary audits were conducted on
the field sampling, sample preparation, and analytical activities.
The results of these audits are reported in Appendix G.
The specific QA/QC activities and measures in the field sampling, laboratory
analysis, information management, and statistical comparison phases are
described in more detail in Appendix F. The sample collection and prepara-
tion QAPP and the sample analysis QAPP fully document the design of the
QA/QC measures for these phases of the project.
5-7
-------
5.5 STATISTICAL ANALYSIS
As described in Section 3.0, the goal of the sampling program was to be able
to compare each EDA sampling area with each comparison area so that there
was a 90 percent probability of detecting an order-of-magnitude difference
in median concentrations with 95 percent confidence. The motivation for
the conceptual framework used in the statistical sampling design reported in
Section 4.2 was to minimize the influence of external factors in the statisti-
cal comparisons, while maximizing the ability to detect differences between
EDA sampling areas and comparison areas. This section describes the choice
of statistical methods used for the comparisons and the basis for their selec-
tion. The results of the statistical analysis of the data are presented in Sec-
tion 6.4.
As in Section 4.3, some of the issues discussed here require the use of statis-
tical terminology. An attempt is made to minimize the use of technical ter-
minology; however, the reader unfamiliar with basic statistical concepts may
find it desirable to read Appendix A before proceeding. Appendix B
provides a more in-depth and technically oriented discussion of the study's
statistical design and methods.
As discussed in Section 4.2, the statistical design and the statistical analysis
methods are closely linked. The objective of the statistical comparison
analysis was to produce a high level of confidence that the analysis would
detect true differences in concentrations between sampling areas if such dif-
ferences exist. To accomplish this, the statistical analysis methods had to be
able to deal with the following characteristics of the data set:
• The concentration estimates are never negative. They are usually
small (or non-detect) and also have a few high values.
• The number of non-detectable concentrations for some LCICs is
quite high.
In addition, the habitability criteria required a version of the statistical test
that could analyze for differences between areas using all the LCICs together
as one data group (the multivariate statistical test).
Many statistical methods that are available are particularly applicable to the
comparison approach. These methods can be broadly divided into two
5-8
-------
groups: parametric and nonparametric. Parametric methods require
knowledge (or an assumption) about the statistical distribution of the data;
nonparametric methods do not. If one applies any statistical test to data that
are not consistent with the underlying assumptions, then the resulting statis-
tical comparisons may be invalid. The degree to which a statistical test will
provide acceptable performance if the assumptions are violated is known as
"robustness." In general, nonparametric tests are more robust than
parametric tests. Extensive computer simulations were conducted to
evaluate the robustness of several candidate parametric and nonparametric
statistical tests. The results of these studies are reported in Appendix B.
In the computer simulations, the test found to have the best overall perfor-
mance was the nonparametric Wilcoxon rank-sum test (Hollander and
Wolfe, 1973). The Wilcoxon rank-sum test is so named because it uses a
test statistic that is based on the sum of the ranks. For example, to compare
two areas, all of the concentrations from both areas are ranked together from
lowest to highest. Nondetectable concentrations are all given the same rank.
All of the ranks from each area are then summed and standardized to form
the test statistic. The test statistic is expected to be small if two areas being
compared are similar. If the test statistic is larger than a critical value, the
null hypothesis (the hypothesis that is accepted unless there is sufficient
evidence to the contrary) that the areas are similar is rejected. Critical values
of the test statistic can be found in standard statistical references (e.g., Hol-
lander and Wolfe, 1973).
Robustness and sensitivity are related yet distinct issues that have important
implications for the choice of a statistical method. As described above,
robustness refers to the performance of a statistical test when the test assump-
tions are not met; robustness is best assessed using computer-simulated data
with characteristics as similar as possible to the real data to which the test
would ultimately be applied.
In contrast, the term "sensitivity" in statistical analysis refers to an evalua-
tion of specific changes in results (that is, the conclusion of whether two sam-
pling areas are or are not different at a given significance level) that would
occur if specific data values were changed. The Wilcoxon rank-sum test was
evaluated for its sensitivity to the soil assessment data by using three varia-
tions of the basic data set. The first variation consisted of all of the validated
data deemed usable for the statistical comparisons. These data were desig-
nated as "Good." The second consisted of the Good data plus the data that
5-9
-------
were of somewhat uncertain validity (designated "Uncertain") according to
QC criteria. (Assessment of data validity is discussed further in Section 6.1.)
The third variation was identical to the first, except that all data values below
1 part per billion (ppb) were treated as non-detectable (the actual operation-
al detection limit was sample-dependent, but was estimated to be typically
0.2 to 0.3 ppb).
The three alternate data sets served to provide a qualitative measure of the
sensitivity of the statistical analyses. The inclusion of Uncertain data as well
as Good data in the analysis was intended to allow evaluation of the sen-
sitivity of the statistical analysis to variations in laboratory analytical quality.
Artificially increasing the detection limit to 1 ppb and analyzing the result-
ing data set made it possible to assess the sensitivity of the statistical analysis
to the laboratory detection limits. If the results had changed dramatically
with the increased detection limits, then it could be concluded that the detec-
tion limits were exerting a strong influence on the results.
5-10
-------
6.0 RESULTS
The results of the soil assessment for LCICs include the samples actually col-
lected and analyzed, study performance in relation to the QA/QC objectives,
and the statistical comparisons between EDA sampling areas and comparison
areas.
The sampling and analytical activities are summarized in Section 6.1. This
includes an accounting of the number of field samples and quality control
samples involved in the study. Section 6.2 provides an overview of the
QA/QC data assessment. The analytical results are reported in Section 6.3.
Section 6.4 reports the results of the statistical comparisons. As directed by
the TRC, the purpose of this assessment is to present data and is specifical-
ly not to develop conclusions or interpretations of the data.
6.1 SAMPLING AND ANALYTICAL ACTIVITIES
Collection of soil samples began in October 1987 and ended in November
1987. Figures 1-1 through I-11 of Appendix I show the locations of the
samples collected in the seven sampling areas of the EDA and the three com-
parison areas (combined Cheektowaga and Tonawanda, Census Tract 221,
and Census Tract 225), respectively. Table 6-1 shows the number of
samples collected in each of the areas mentioned above. As indicated by
Table 6-1, a total of 887 field samples were collected out of the 894 planned.
Five samples were not collected because of field conditions such as gravel
or standing water, and two samples were not collected because permission
to sample was revoked. The sampling design specified that the collection of
samples be randomly allocated among collection teams and neighborhoods.
Table 6-2 shows the number of samples actually collected during the field
sampling by the four collection teams in each neighborhood.
Of the 887 field samples collected, 866 were extruded, prepared, and sent
for analysis (Table 6-1). Twenty-one regular field samples could not be ex-
truded. Sixteen unextrudable samples were replaced with 16 other samples
taken approximately 5 feet away from the original locations. These samples
are referred to as "replacement samples" in Table 6-1 and are identified with
an "R" suffix (e.g., 444R) on the maps (Figures 1-1 through Ml). Three of
6-1
-------
to
Table 6-1
NUMBER AND TYPE OF FIELD SAMPLES COLLECTED AND SENT FOR ANALYSIS
Samples
EDA Planned
Sampling EDA for
Areas3 Neighborhood Collection
1
2
3
4
5
6
7
Total EDA
Comparison Areas:
Census Tract 221
Census Tract 225
Tonawanda
Cheektowaga
TOTAL STUDY
1
2
3
4
5
6
7
8
9
11
12
10
13
50
48
52
100
70
35
105
85
35
35
35
105
35
73
108
35
81
116
669
75
75
37
38
75
894
Samples
Actually
Collected
47
48
52
100
69
35
104
84
36
35
35
106
35
72
107
35
82
117
665
73
75
74
887
Unextrudable
Samples
0
2
2
4
2
1
3
0
1
1
2
4
1
0
1
0
0
0
12
8
1
0
21
Replacement
Samples
0
2b
3
1
1
2
0
1
3
1
0
1
0
0
0
9
6
1
0
16
Samples
Sent
for
Analysis
47
47
50
97
68
35
103
84
36
34
34
104
35
72
107
35
82
117
659
71
75
74
879
Contingency
Samples
Available
for Analysis
7
4
4
8
3
3
6
8
6
3
6
15
5
4
9
8
2
10
63
3
7
3
76
8A sampling area consists of one or more neighborhoods as Indicated.
bBoth replacement samples for Neighborhood 3 were unextrudable; one replacement sample for Neighborhood 9 was unextrudable.
-------
Table 6-2
NUMBER OF FIELD SAMPLES COLLECTED BY SAMPLING AREA, NEIGHBORHOOD, AND SAMPLING TEAM
EDA EDA Sampling Team
Sampling Area Neighborhood
1
2
3
4
5
6
7
Total EDA
Comparison Areas:
Census Tract 221
Census Tract 225
1
2
3
4
5
6
7
8
9
11
12
10
13
Tonawanda/Cheektowaga
TOTAL STUDY
1
13
17
16
33
16
12
28
20
10
8
14
32
0
15
15
12
29
41
182
13
12
11
218
2
10
7
12
19
24
0
24
12
1
13
3
17
12
18
30
3
9
12
124
35
26
30
215
3
12
12
12
24
17
12
39
28
13
2
12
27
11
18
29
8
12
20
169
16
25
21
231
4
12
12
12
24
12
11
23
24
12
12
6
30
12
21
33
12
32
44
190
9
12
12
223
Total
47
48
52
100
69
35
104
84
36
35
35
106
35
72
107
35
82
117
665
73
75
74
887
-------
the replacement samples also had extrusion problems and were not themsel-
ves replaced. In addition, 78 contingency samples were collected in the
event additional samples were needed for analysis; two of these were unex-
trudable.
The resulting 879 extruded and prepared samples (Table 6-1) were sent to
seven analytical laboratories for analysis of the LCICs. Table 6-3 shows the
number of samples that each laboratory was planned to have received, the
number actually received, and the number of samples that each laboratory
analyzed. As shown in Table 6-3,781 samples were actually analyzed for
the LCICs. Ninety-seven samples were not analyzed because of problems
at Laboratory 4, as discussed below. One sample was not analyzed at
Laboratory 3. Figures 6-1 through 6-5 show the locations of the
781 samples that were analyzed in the EDA, Cheektowaga, Tonawanda,
Census Tract 221, and Census Tract 225, respectively.
The electronic and hardcopy data provided by the laboratories were subjected
to a detailed data assessment process by EMSL-LV, described in Appen-
dix H. This resulted in an assignment of data qualifiers to each sample and
each LCIC analyte within a sample. From these data qualifiers, an overall
assessment of the usability of the results for the statistical comparison was
made. Each LCIC result was defined as being Good, Uncertain, or Bad. It
should be noted that the definition of usability of the results was not related
to the concentration estimates. Only results classified as Good were used in
the statistical comparisons. (Results initially classified as Uncertain based
only on holding time violations were reclassified as Good for purposes of
statistical analysis.) Table 6-4 presents a summary of the Good data used for
the statistical analysis by LCIC, for the 781 samples analyzed.
Table I-1 in Appendix I shows the final, validated results for all field samples
by LCIC. Analyses that yielded Uncertain or Bad data are designated in the
table as not meeting QC criteria. This table is discussed further in
Section 6.3.
The nominal or desired allocation of samples is discussed in Section 4.2. A
first step in assessing design implementation was to compare the actual al-
location of samples to the planned allocation of samples, as shown in
Tables 6-1 through 6-3. In all cases, the differences were minor. The agree-
ment between the actual and planned sample allocations confirms that the
6-4
-------
Table 6-3
NUMBER OF FIELD SAMPLES RECEIVED AND ANALYZED BY LABORATORY
Analytical
Laboratory8
1
2
3
4
6
7
8
TOTALS
No. of Samples
Planned
122
125
128
135
137
110
137
894
No. of Samples
Received
119
121
125
135
137
109
133
879
No. of Samples
Analyzed
119 + 38b
121
124°
Od
133°
109
137°
781
"Laboratory 5 was not selected for study participation.
thirty-eight samples extracted by Laboratory 4 were sent to Laboratory 1 for analysis.
<*0ne sample was not analyzed.
Laboratory 4 did not successfully analyze any samples.
eFour samples originally designated for Laboratory 6 were accidentally sent to Laboratory 8
for analysis.
6-5
-------
SCHOOL srre
'•' fJHOEfi SfftVWf 8TUWI
:- id I ILL
Rqire 6-1
SOIL ASSESSMENT—INDICATOR
CHEMICALS
LOCATIONS OF ANALYZED
SAMPLING POINTS - EDA
SCALE: 1"'660'
LEGEND
««««" EMERGENCY DECLARATION AREA (EDA) BOUNDARY
— FENCE LINE AROUND LOVE CANAL REMEDIATION SITE
LOCATION OF SAMPLE
SOURCE; EDA BOUNDARIES TAKEN FROM NEW YORK
STATE PROPERTY TAX LAW ARTICLE 17. SECTION 1702
-------
WfMWBSSftt,
///////////MsmntMjfatv //7
. ; ;Ci j • \— :*J- ^ :S; •
Figure 6-2
SOIL ASSESSMENT—INDICATOR CHEMICALS
LOCATIONS OF ANALYZED SAMPLES
CHEEKTOWAQA COMPARISON AREA
LEGEND
— - SAMPLING BOUNDARY
• SAMPLING POINT
AREA UNSUITABLE FOR SAMPLING AS DETERMINED
BY A NYSDOH/NYSDEC/EDA CITIZEN PANEL
SCALE: 1"-6«y
SOURCE: MAP DRAWN FROM NEW YORK
STATE PROPERTY TAX LAW ARTICLE 17. SECTION 1702
-------
P*"T^7"r«^*p»-r •»-»"»••».«r ••-.
f i—-J • i ! i i i h -i
m '•.—M
•-i W1.'V-'''* '"• ••'**—• ; •*—"^ ' i i *• i * p.....tr^j i i- - i -i • ^Zf^.
• T» •" V-' - '' ,**r-—4 • 1 i : ••!-•••—! r»—j ; t......^ i C • ~ j_ : i f-
t=Sfe|;y*S M fS M te Si ii§
; ¥>'. v" •« ; ' ' < ' j >*• 1 •---- ^ • ~~—^ •
; V; %v-'?^~if: ! 1 l---.-i '• < 1 !'"•'> -5 I i- -j j—^ 1 i f j i
V)1, \ ,,'^ xJJ<—n^"- ™!4....."j f~l? ' ?-'." ''* i i i- H ;—-4- < ^ r--;| i
"'' "
W>d*C¥lf .V^M*
I : """"
Figure 6-3
SOIL ASSESSMENT— INDICATOR CHEMICALS
LOCATIONS OF ANALYZED SAMPLES
TONAWANDA COMPARISON AREA
SCALE: 1-.6501
LEGEND
... SAMPLING BOUNDARY
• SAMPLING POINT
^
AREA UNSUITABLE FOR SAMPLING AS DETERMINED
BY A NYSDOH/NYSDEC/EOA CITIZEN PANEL
SOURCE: MAP DRAWN FROM NEW YORK
STATE PROPERTY TAX LAW ARTICLE 17. SECTION 1702
-------
"i^^,-
0
—I I . —!
f eiKAKD .*i'fM/£
u~
.
=J—
•
T"
-
: __
P " ( .........
__ J I :
: I P ......... — — _ ;
_ • * i. ______ : i f "1 -
-iiu- ...... tra .......... 4
":|l:
-3£l
—tJU
t::::
„ .-,4- -
r'"
—j
i-
fc
* i....
n
Rgure 6-4
SOIL ASSESSMENT—INDICATOR CHEMICALS
LOCATIONS OF ANALYZED SAMPLES
CENSUS TRACT 221 COMPARISON AREA
SCALE: r-400'
LEGEND
«— — CENSUS TRACT 221 BOUNDARY
• LOCATION OF SAMPLING
SOURCE: BOUNDARIES TAKEN FROM NEW YORK
STATE PROPERTY TAX LAW ARTICLE 17. SECTION 1702
-------
••/
iy'1 I8!
!/ "= j
i S~
f.
......... i
....... _ i
. i
Figure 6-5
SOIL ASSESSMENT—INDICATOR CHEMICALS
LOCATIONS OF ANALYZED SAMPLES
CENSUS TRACT 225 COMPARISON AREA
SCALE: f-4001
LEGEND
— — CENSUS TRACT 225 BOUNDARY
LOCATION OF SAMPLING
SOURCE: BOUNDARIES TAKEN FROM NEW YORK
STATE FflOPERTY TAX LAW ARTICLE 17, SECTION 1702
-------
Table 6-4
SUMMARY OF LCIC RESULTS USED FOR STATISTICAL ANALYSES
LCIC
1 ,2-Dlchlorobenzene
1 ,2,4-Trlchlorobenzene
1 ,2,3,4-Tetrachlorobenzene
2-Chloronaphthalene
Alpha-BHC
Delta-BHC
Beta-BHC
Gamma-BHC
Total All Analytes
Number of
Samples
Analyzed
781
781
781
781
781
781
781
781
6,248
Number of
Individual LCIC
Results Designated
As Good9
Number
655
685
676
640
685
673
634
681
5,329
%
84
88
86
82
88
86
81
88
85
Number of
Individual
LCIC Results
Designated
As Uncertain8
Number 3
73
48
53
43
52
61
fc
9
6
7
6
7
8
96 12
50
476
6
8
Number of
Individual
LCIC Results
Designated
As Bad3
Number
53
48
52
98
44
47
51
50
443
%
7
6
7
12
5
6
7
6
7
"Only results that were designated as Good were used In the statistical analyses.
-------
NUMBER OF FIELD
Analytical
Laboratory8
FHB
1 14
2 12
3 11
4 12
6 11
7 10
8 6
TOTALS 76
Table 6-5
QC SAMPLES RECEIVED AND ANALYZED BY LABORATORY
Field QC Samples
Received
Split
22
17
6
11b
14
9
4
83
PHB
8
7
10
9
6
7
9
56C
SSB
19
18
19
21
18
17
18
130
FHB
14+3d
12
11
0
11
10
6
67C
Field QC Samples
Analyzed
Split
22
17
6
11
14
9
4
83
PHBf SSB1
___
t_ ^
_ —
t___ __
. ^
— —
FHB = field handling blank
PHB = preparation handling blank
SSB = shipping and storage blank
"Laboratory 5 was not selected for study participation.
One split was not sent to Laboratory 4.
*JA total of 57 PHBs were used; however, one was unextrudable.
Three of Laboratory 4's FHBs were analyzed by Laboratory 1.
''Nine FHBs were not analyzed by Laboratory 4.
PHBs and SSBs were stored but not analyzed.
-------
original randomization was maintained with respect to the sampling crew
and sample analysis.
Four types of field QC samples were used during sample handling and track-
ing, as discussed in Appendix F. They included field handling blanks, splits,
preparation handling blanks, and shipping and storage blanks. Table 6-5
presents information regarding the number of field QC samples received and
subsequently analyzed by each analytical laboratory. Upon QC review of
the field sampling it was discovered that the allocation of split samples to
analytical laboratories deviated from the planned allocation. Only two of the
intra-laboratory split samples were generated. This limited the opportunity
to assess intra-laboratory variability using split samples. Intra-laboratory
variability, however, can also be assessed by other QC samples, e.g., BQC
samples. (The results of these analyses are given in Appendix H.)
Four types of laboratory QC samples were used:
• Method/holding blanks (MHBs)—1 in 10 analyses
• BQCs-1 in each batch of samples (approximately 10)
• Matrix spike/ matrix spike duplicates (MS/MSDs)--l in 20 analyses
• EPA check standards—once every 2 weeks
These QC analyses totaled approximately 500.
Only two problems were encountered during the sample analysis. First, early
in the laboratory analysis phase of the project it was discovered simultaneous-
ly in several laboratories that an interfering chemical was present in the
MHBs. This chemical interfered with measurements of beta- and gamma-
BHC. An investigation was initiated in all laboratories to isolate and
eliminate the source of contamination. It appeared that the interference was
from the detergent commonly used to clean the laboratory glassware. The
laboratories took appropriate corrective actions to resolve the problem. Cor-
rective actions included minimizing the use of detergent, using more effi-
cient rinses after the glassware had been washed, and adjusting the GC/MS
so that the interfering compound was separated from the target LCICs.
6-13
-------
Second, one laboratory experienced considerable difficulty in successfully
analyzing the samples. This laboratory, designated Laboratory 4, en-
countered instrumentation problems early in the study. Once the GC/MS in-
strument was functioning properly, the blank interference problem discussed
above was discovered. Because of these problems, the BQC samples from
Laboratory 4 were analyzed by Laboratory 1. Once these samples were
analyzed, it was found that only five of the remaining batches had accept-
able BQC samples. The field samples and associated QC samples from these
five batches were analyzed by Laboratory 1, and the data from these samples
were subjected to data validation by EMSL-LV.
6.2 QA/QC DATA ASSESSMENT
As discussed in Sections 3.0 and 4.3, the study was designed to achieve the
QA/QC objectives of precision, accuracy, representativeness, comparability,
and completeness as specified in the sample collection and analysis QAPPs
(CH2M HILL, 1987a and 1987b). This section summarizes the results of
the QC sample analysis, while Appendix J provides a more detailed discus-
sion.
Table 6-6 summarizes the results of all QC sample data. QC control limits
were set for the various measures of QC samples. The QAPP-specified goal
was that the laboratories should meet the QC control limits for either 90 per-
cent or 100 percent of the analyses, depending on the type of measure in-
volved. This goal was met for all types of QC measures, with the minor
exception of MS/ MSD analyses. MS and MSD samples are two separate
portions of the same field sample, each of which is spiked with a known
amount of LCICs.
The MS/MSD samples were analyzed to assess the accuracy and precision
of the analytical method. The accuracy is measured by the "recoveries," i.e.,
the laboratory's ability to measure the spiked concentrations. Stringent
QAPP advisory control limits were developed for MS/MSD recoveries,
based on method validation results from the two laboratories used in develop-
ing the method. When the six laboratories used for the soil assessment
analyzed the QC samples, none achieved the goal of meeting the QC criteria
for 90 percent of the analyses. The control limits were not revised because
the time constraints did not permit a second multi-laboratory MS/MSD study.
6-14
-------
Table 6-6
QC DATA SUMMARY
QA Measure
MHB
Surrogate Spike
Recovery
MS Recovery
(Advisory Only)
MSD Precision
(Advisory Only)
BQC Recovery
Laboratory
1
2
3
6
7
8
1
2
3
6
7
8
1
2
3
6
7
8
1
2
3
6
7
8
1
2
3
6
7
8
Data Meeting QC Criteria
Achieved
Goal (%) (Y or N)
90 Y
Y
Y
Y
Y
Y
90 Y
Y
Y
Y
Y
Y
90 N
N
N
N
N
N
90 Y
Y
N
Y
Y
N
90 Y
Y
Y
Y
Y
Y
6-15
-------
QA Measure
EPA Check
Standard
Recovery
Internal Standard
Area and Retention
Time Variation
Performance
Check Standard
(PC1 and PC2)
Initial Calibration
Standard
Continuing
Calibration
Standard
Holding Time
(for original
extraction and
analysis only)
Table 6-6
(continued)
Laboratory
1
2
3
6
7
8
1
2
3
6
7
8
1
2
3
6
7
8
1
2
3
6
7
8
1
2
3
6
7
8
1
2
3
6
7
8
Data Meeting QC Criteria
Achieved
Goal (%) (Y or N)
100 Y
Y
Y
Y
Y
Y
90 Y
Y
Y
Y
Y
Y
100 Y
Y
Y
Y
Y
Y
100 Y
Y
Y
Y
Y
Y
100 Y
Y
Y
Y
Y
Y
100 Y
Y
Y
Y
Y
Y
6-16
-------
The samples were not reanalyzed because these control limits were advisory
only.
Most of the laboratories (four of the six) did meet the 90 percent goal for
MS/MSD precision; that is, the measured values for both MS and MSD
samples were within the specified range. As Table 6-6 shows, the 90 per-
cent or 100 percent goals were met for all other QC measures.
6.3 DETECTION LIMIT AND ANALYTICAL RESULTS
Detection Limit
Previous Love Canal studies and the pilot study have indicated the major im-
portance of achieving an understanding of the detection limits for the LCICs
in soil samples from the Niagara Falls area. Unfortunately, there is no
generally accepted approach for estimating detection limits. For a method
using GC/MS/SIM, at best an estimated detection limit can only indicate the
typical performance of the analytical method in a given laboratory for an
LCIC. The actual detection limit for an individual sample, which cannot be
measured, depends on the LCIC in question, the level of interferences in the
sample, and the performance of the GC/MS instrument at the time of analysis.
For this analysis, there is an additional complicating factor in that LCICs
need to be both detected and identified. As was observed in the pilot study,
the qualitative identification criteria are often more difficult to satisfy than
the detection criteria.
The operational detection limit was estimated based on the large number
(ranging from 62 percent of 2-chloronaphthalene to 12 percent of delta-
BHC) of LCIC results reported for field samples at concentrations below
0.3 ppb. The operational detection limit is estimated to be between 0.2 and
0.3 ppb. This estimate is also supported by the performance of the
laboratories on the BQC samples at this concentration level, as discussed in
Appendix J.
6-17
-------
Analytical Results
Table 1-1 of Appendix I shows the final, validated results for all field samples
and each LCIC within a sample. LCIC concentrations are reported only for
results designated as usable for the statistical comparisons. These data are
designated as Good. Results in which LCICs were not detected are listed as
ND. Analytical results that were classified as Uncertain or Bad are listed as
QCF. Samples that were not taken and were not analyzed (contingency
samples) are designated as NT and NAC, respectively. Samples that were
not analyzed because of Laboratory 4's problems described in Section 6.1
are listed as NA.
Tables 6-7a through 6-7h summarize the analytical results for each of the
seven EDA sampling areas and each of the three comparison areas.
Graphic Representations of Analytical Results
Histograms and box plots were used to visually compare the LCIC concentra-
tion data from the EDA sampling areas to the corresponding data from the
comparison areas. Although these graphic representations are not a sub-
stitute for formal statistical analysis (discussed in Section 6.4), they should
be generally consistent with the statistical analysis and accordingly provide
a qualitative cross-check of the results of the statistical comparisons. The
following paragraphs discuss the histograms and box plots in more detail.
Histograms. Histograms are graphic representations of the concentrations
of each of the LCICs in each of the EDA sampling areas and comparison
areas. Figures 6-6a through 6-6h are histograms illustrating the distribution
of concentrations found for each LCIC in each of the EDA sampling areas
and comparison areas. The vertical axis of each histogram represents the
relative frequency of each concentration range represented on the horizontal
axis. The highest bar represents the most commonly found concentration
range. Gaps in the horizontal axis indicate that no concentrations were ob-
served in that range. Non-detectable concentrations are coded as zero in
these figures.
Box Plots. An example of a box plot is shown in Figure 6-7. The vertical
axis is the concentration of an LCIC in ppb. The box includes the middle
50 percent of the data. The line in the box marks the median, or 50th
6-18
-------
Table 6-7a
SUMMARY OF ANALYTICAL RESULTS FOR
1,2-DICHLOROBENZENE (DCB)
Concentration (ppb)
Sampling
Area
1
2
3
4
5
6
7
Census Tract 221
Census Tract 225
Tonawanda and
Cheektowaga
TOTAL
Number
of Good
Results8
39
68
73
60
80
84
86
50
58
57
655
Percent
Detects
100
98
99
95
100
100
99
100
100
100
Minimum
(Includes
NDsb)
0.20
ND
ND
ND
0.14
0.16
ND
0.19
0.18
0.11
Minimum
(NDs not
Included)
0.20
0.13
0.12
0.13
0.14
0.16
0.16
0.19
0.18
0.11
Maximum
5.65
19.80
2.37
2.23
3.19
1.51
0.94
1.15
1.40
1.38
Median
(Includes
NDs)
1.01
0.39
0.40
0.36
0.39
0.36
0.40
0.39
0.43
0.36
""Good Results" are those that passed all QC criteria.
bND = non-detect.
-------
Table 6-7b
SUMMARY OF ANALYTICAL RESULTS FOR 1
,2,4-TRICHLOROBENZENE (TCB)
Concentration (ppb)
Sampling
Area
1
2
3
4
5
6
7
Census Tract 221
Census Tract 225
Tonawanda and
Cheektowaga
TOTAL
Number
of Good
Results8
40
73
82
61
80
88
88
53
60
60
685
Percent
Detects
100
100
100
98
100
100
100
100
100
100
Minimum
(Includes
NDsb)
0.15
0.17
0.08
ND
0.08
0.04
0.15
0.25
0.12
0.07
Minimum
(NDs not
Included)
0.15
0.17
0.08
0.11
0.08
0.04
0.15
0.25
0.12
0.07
Maximum
45.11
167.33
7.65
12.08
35.65
3.51
4.49
3.12
33.07
0.92
Median
(Includes
NDs)
8.67
0.91
0.89
0.43
0.52
0.37
0.58
0.65
0.60
0.14
""Good Results" are those that passed all QC criteria.
bND = non-detect.
-------
Table 6-7c
SUMMARY OF ANALYTICAL RESULTS FOR 1,2,3,4-TETRACHLOROBENZENE (TeCB)
Concentration (ppb)
Sampling
Area
1
2
3
4
5
6
7
Census Tract 221
Census Tract 225
Tonawanda and
Cheektowaga
TOTAL
Number
of Good
Results8
37
72
82
60
78
87
88
52
59
61
676
Percent
Detects
100
100
99
98
99
98
100
100
100
79
Minimum
(includes
NDsb)
0.11
0.23
ND
ND
ND
ND
0.10
0.22
0.06
ND
Minimum
(NDs not
Included)
0.11
0.23
0.10
0.04
0.09
0.04
0.10
0.22
0.06
0.02
Maximum
67.20
66.66
9.55
182.41
168.64
2.90
3.91
1.33
64.25
0.84
Median
(includes
NDs)
11.48
1.16
1.06
0.38
0.43
0.31
0.54
0.53
0.61
0.05
""Good Results" are those that passed all QC criteria.
bND = non-detect.
-------
N>
SUMMARY OF ANALYTICAL
Table 6-7d
RESULTS FOR 2-CHLORONAPHTHALENE (CNP)
Concentration (ppb)
Sampling
Area
1
2
3
4
5
6
7
Census Tract 221
Census Tract 225
Tonawanda and
Cheektowaga
TOTAL
Number
of Good
Results"
36
70
75
52
72
88
82
51
60
54
640
Percent
Detects
44
61
51
62
67
65
70
67
72
57
Minimum
(includes
NDsb)
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
Minimum
(NDs not
Included)
0.06
0.01
0.04
0.02
0.02
0.02
0.02
0.03
0.04
0.01
Maximum
0.21
0.24
0.15
0.13
0.16
0.19
0.18
0.22
0.32
0.21
Median
(includes
NDs)
ND
0.04
0.04
0.04
0.06
0.06
0.06
0.06
0.07
0.03
a..
Good Results" are those that passed all QC criteria.
"ND = non-detect.
-------
SUMMARY
Table 6-7e
OF ANALYTICAL RESULTS
FOR ALPHA-BHC
(A-BHC)
Concentration (ppb)
Sampling
Area
1
2
3
4
5
6
7
10 Census Tract 221
w
Census Tract 225
Tonawanda and
Cheektowaga
TOTAL
Number
of Good
Results8
39
73
81
61
80
88
89
53
60
61
685
Percent
Detects
97
86
76
70
64
61
78
79
65
5
Minimum
(Includes
NDsb)
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
Minimum
(NDs not
Included)
0.66
0.06
0.01
0.03
0.03
0.03
0.02
0.06
0.02
0.01
Maximum
69.70
100.31
16.05
152.53
83.51
1.99
2.42
1.19
33.97
0.17
Median
(includes
NDs)
8.25
0.40
0.22
0.12
0.14
0.07
0.14
0.18
0.10
ND
""Good Results" are those that passed all QC criteria.
bND = non-detect.
-------
Table 6-7f
SUMMARY OF ANALYTICAL RESULTS FOR DELTA-BHC (D-BHC)
Concentration (ppb)
Sampling
Area
1
2
3
4
5
6
7
7s
K> Census Tract 221
Census Tract 225
Tonawanda and
Cheektowaga
TOTAL
Number
of Good
Results8
39
72
79
61
80
84
86
52
59
61
673
Percent
Detects
69
51
30
20
22
11
14
8
15
0
Minimum
(Includes
NDsb)
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
Minimum
(NDs not
Included)
0.28
0.03
0.05
0.08
0.03
0.03
0.05
0.04
0.07
ND
Maximum
38.83
2.33
79.99
3.04
9.96
3.44
1.25
0.94
5.40
ND
Median
(Includes
NDs)
1.13
0.04
ND
ND
ND
ND
ND
ND
ND
ND
a,,,
Good Results" are those that passed all QC criteria.
°ND = non-detect.
-------
Table 6-7g
SUMMARY OF ANALYTICAL RESULTS FOR BETA-BHC (B-BHC)
Concentration (ppb)
Sampling
Area
1
2
3
4
5
6
7
o\
N> Census Tract 221
Census Tract 225
Tonawanda and
Cheektowaga
TOTAL
Number
of Good
Results"
39
70
77
56
77
79
74
49
54
59
634
Percent
Detects
97
64
54
48
32
21
45
29
46
3
Minimum
(includes
NDsb)
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
Minimum
(NDs not
Included)
1.89
0.07
0.06
0.03
0.01
0.10
0.05
0.07
0.01
0.36
Maximum
101.72
29.97
51.17
4,108.31
663.49
3.15
2.01
1.03
50.57
5.36
Median
(Includes
NDs)
11.58
0.20
0.13
ND
ND
ND
ND
ND
ND
ND
""Good Results" are those that passed all QC criteria.
bND = non-detect.
-------
Table 6-7h
SUMMARY OF ANALYTICAL RESULTS
FOR GAMMA-BHC (G-BHC)
Concentration (ppb)
Sampling
Area
1
2
3
4
5
6
7
to Census Tract 221
Census Tract 225
Tonawanda and
Cheektowaga
TOTAL
Number
of Good
Results8
40
72
80
60
80
88
89
53
60
59
681
Percent
Detects
85
56
45
33
26
17
24
17
30
2
Minimum
(Includes
NDsb)
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
Minimum
(NDs not
Included)
0.36
0.03
0.01
0.03
0.04
0.04
0.04
0.04
0.03
0.04
Maximum
20.99
6.67
12.68
85.60
26.33
0.89
0.61
80.81
15.83
0.04
Median
(includes
NDs)
1.72
0.08
ND
ND
ND
ND
ND
ND
ND
ND
""Good Results" are those that passed all QC criteria.
bND = non-detect.
-------
AREA=EDA 1
1 OO 1 OO
ao ao
tt~ eo H- ao
J :: J ::
Cone . i n RRO
AREA — E D A2
100 100
ao . ao
th eo H- eo
J •« • J "
20 ^H m 20
o -9- T ^ If* T T" T" T ••»•' "r1 °
8t828t828t828iol8i828
Cone . 1 n F»PB
AREA — EDAJ3
1OO i 1 OO
ao ao
1 Io J !o
20 "• "• ^^ 20
Cone. In Pf>B
AREA— EDAA
1OO IOO
ao ao
9h eo it" eo
20 "•*• 20
0 ^JBUIl^ , , j _j_ j j j j j j _ _ _ _ g_ o
Cona . 1 n F»F»B
ARE A — E DAS
IOO IOO
BO BO
ft" ao H- eo
E ^H s:
*P *° •• «i= 4O
20 VH ^H 20
O -Q — ^^ — I — T" — | "f }•' l 3""^~ — 1 ] ~ — 5 — °
Cona. in F>F>B
ARE A = EDA6
Cone . 1 n PPB
ARE A— ED A7
Biflil"]
Cone . 1 n PPB
AREA— CEMSUS TRACT 221
Cone . 1 n PPB
AREA— CEMSUS TRACT 225
Cono . I n PPB
AR E A — CUE E K TOWAG A A NJ D TOtslAWA
Cono. in RPB
Figure 6-6a
1, 2-DICHLOROBENZENE
SOIL ASSESSMENT — INDICATOR
CONCENTRATIONS BY AREA
« 8 J 8
1828
1828
} 8 t 8
MD A
1828
CHEMICALS
-------
ARE A = E DA1
ARE A = EDA6
1OO
ao
eo
»o
A 4 6 »
Cone. In PP(
AREA—E DA2
; o j—JAJj
°46»6»6
Cone. in PPB
ARE A«-E D A 7
8
£
100
ao
eo
20
o
1 OO
ao
_*- eo
I I 2 I t
S I I
6464
Cone. In PPE
AREA —E DA3
1 OO
ao
60
»°
20
o
I,
O 4 O
Cone. In PPB
AREA—CENSUS TRACT 221
»6 • J^4lJJ***11»l*
Cone. In PPB
AREA=CENSUS TRACT 225 .
i8
6.6*6 * * * T' * *
Cone
AREA=CMEEKTOWAGA AND TONAWANDA
< » <• i" ' I * i i 1 i i i
I n f»P(
Cono. In PPB
6 » 6
Figure 6-6b
1, 2, 4-TRICHLOROBENZENE
SOIL ASSESSMENT — INDICATOR CHEMICALS
CONCENTRATIONS BY AREA
-------
1 OO
• o
ft- eo
J *o
20
1 OO
80
If- ao
zo
1 OO i
BO
If* ao
S A_
£ *°
20
1 OO
ao
Jh ao
M *°
20
too
eo
K- ao
J -
ao
o
AREA=EDA 1
TOO
ao
Ih ao
J! *°
20
Cone . 1 n RRO
ARE A— ED A2
1OO
ao
9J- eo
H _ J ~
Cone . 1 n PPB
AREA — E D A 3
100
80
•h eo
II. ' =
Cone. In F»PB
AREA— EDAA
1 OO
ao
B -t" B°
9 j *°
HB 2!
Cone . ! n f»F>B
AREA — E D A 5
1 OO
ao
R »- eo
1 * "
H >o
HB-BL^_->v> — _ ,_ — .-- — rz — .. n .. — r, — .-, — n — r. — r_ — r. — „ — r. — .- '•• o
Cono. In Rf>B
AREA=EDAB
Cone. In Rf*B
AREA — E D A 7
9
in
Cone . 1 n PPB
AREA— CENSUS TRACT 221
|_
1
Cone. In PPB
ARE A— CE MSUS TRACT 225'
B
fl
!•
Cone . 1 n f>RB
AR E A = C ME E K T OWAG A AMD TONJAWAMDA
Cono. in RF»B
Figure 6-6c
1, 2, 3, 4-TETRACHLOROBENZENE
SOIL ASSESSMENT — INDICATOR CHEMICALS
CONCENTRATIONS BY AREA
-------
AREA=EDA
AREA=EDAB
AREA=CEMSUS TRACT 221
1 OO
eo
_l-~ eo
ii A °
20
o
i oo
ao
eo
AO
20
o
i oo
ao
Or eo
J -
2 o
TRACT 225
~5— o A B—~B~—B 5 5 B H B B—"S~
'?$Vi(gSiiJi4iA«iJ4ii4
o3a«8aaa«8alli4o6jB«C
Cone. In PP*Q
AMD TOMAWANDA
o o o o A B B—"B™—B B B B o 5 5~
8:jj£44gfi3*i!o?
82B4oe2S*o«za»o
Figure 6-6d
2-CHLORONAPHTHALENE
SOIL ASSESSMENT — INDICATOR CHEMICALS
CONCENTRATIONS BY AREA
-------
AREA=E DA 1
ARE A = E DAS
i oo
ao
C o n c . in
AREA —E D A2
I 4—i—•—TO—r»—r*—n—n IB n—r*—n •• SB si n—n—ir
Cone. in PPO
AREA — E DA 7
1 OO
ao
_»• eo
e
.= *°
20
o
1 OO
ao
th eo
J -
20
o
I ia—n—r*—rm—1» »a 11—r*—n—»» IB 11
Cone. In PPB
AREA — E DA3
»• 1H
Cone. In PPO
AREA—CENSUS TRACT
2 2 1
1 OO
ao
00
*0
20
o
Cone . I n P
AREA —E DAt
Cone. In PPB
AREA—CEMSUS TRACT
225
Co n a . In R
AREA — E DA 5
1 OO
ao
eo
•°
20
o
Cone. In PP
AREA—CHEEKTOWACA AMD
TONIAWANDA
i oo
ao
eo
Cono. in PPC
C o n o . i n
Figure 6-6e
ALPHA-BHC
SOIL ASSESSMENT — INDICATOR CHEMICALS
CONCENTRATIONS BY AREA
-------
B
1
"e
5
*•
100 ,
ao
eo
oo
20
o
100,
ao
eo
oo
20
o
100,
ao
eo
oo
20
o
1 OO
ao
eo
00
20
o
too.
ao
ao
20
o
ARE A = E DA 1
1 OO
tt~ eo
H <0
| — m __ _ * zo
Cone . 1 n F»f B
AREA — E D A2
1 OO
ao
^^B 9* e o
BB n 2°
^^^P^^ "T r ~1 l~ T °
Cone. In P*F>B
AREA — E D A3
1 OO
• "°
B Jh eo
Hfl C
••• — o o
111 20
Cone . 1 n PPB
AREA — EDA-*
1 OO
Lao
f "
20
Cone . 1 n r>r>B
ARE A — EDA 5
1 OO
IB- "
* ::
C O o o . 1 n f*f*&
AREA=EDA6
Cone. Ir^ ^PO
ARE A=-EDA 7
Cone . 1 n »»PB
AREA— CENSUS TRACT 221
Cone . 1 n »»B
AREA— CENISUS TRACT 225 '
Cone . 1 n *»RB
ARE A— CHE E K TOWAC A AMD TOMAWAMDA
Cono. In f»F>B
Figure 6-6f
DELTA-BHC
SOIL ASSESSMENT — INDICATOR CHEMICALS
CONCENTRATIONS BY AREA
-------
ARE A = E D A 1
AREA=EDA6
m
i oo
ao
eo
20
o
i oo
BO
Cone . t n f>
ARE A— EDA2
ARE A =
in F>F
:D A 7
*• eo
1 _
_»•
5
£
20
o
1 OO
BO
eo
20
o
1 oo
ao
Jf- eo
£ *°
20
O
1 OO
ao
eo
4VO
ao
o
i oo
ao
eo
* o
20
o
Cone . 1 n PPO
ARE A — E DA3
Cone. In PF*B
AREA—CENSUS TRACT 221
Cone. In f>
AREA — E D A A
Cone. In Pt»B
AREA —CENSUS TRACT 225.
C 0 n 0 . t n f»
AREA — E D A 5
_*•
s
1 OO
ao
eo
20
o
t oo
BO
eo
20
o
• i~5—T3—TB—n—31—TT—TS—53—H—51—IT—*t \i 4' 4-
Cone. In Pf»a
AREA—CMEEKTOWACA AND TONAWANDA
Figure 6-6g
BETA-BHC
SOIL ASSESSMENT — INDICATOR CHEMICALS
CONCENTRATIONS BY AREA
-------
AREA = E DA1
ARE A = E DAS
1 oo
• o
J
100
BO
«- .o
J *°
20
1 OO
eo
20
O
too
ao
_*• eo
J -
20
1 OO
ao
h oo
J -
Cone. In RF*e
AREA—E DA2
^^"^^"^^•^P I !
Cone. In P»P
ARE A —E DA3
^^~^^~^T"""
Cone. In
AREA—
Co n o . In PO
AREA —E OA5
20
O
1OO
ao
eo
20
o
a 4 6 »
Cone- In f*F*O
AREA —CE.tsJSUS TRACT 221
6 » 6 » O » 6
Cone . I n t>PB
AREA—CEMSUS TRACT 225
AR E A —CHE E K TOWAG A AMD T OM A\A/A Isl D A
C o n o . in F»f»B
Figure 6-6h
GAMMA-BHC
SOIL ASSESSMENT — INDICATOR CHEMICALS
CONCENTRATIONS BY AREA
-------
BOX PLOTS
value
90
80
70
60
50
40
30
20
10
The top of the box
is the 75th percentile
The bar shows the
median
---- >
"Extreme Values"
< larger but not extreme values
The bottom of the box is the
25th percentile.
This axis would list categories of plots, e.g. labs or areas
Figure 6-7
SOIL ASSESSMENT — INDICATOR CHEMICALS
BOX PLOT EXAMPLE
-------
percentile; 25 percent of the concentrations are included in the part of the
box above the median line, and 25 percent are in the part of the box below
the median line.
The upper "whisker" (vertical line above the box) extends to the largest non-
extreme value (the interpretation of "extreme" is based on a relationship with
the normal [Gaussian] distribution). The extreme values are plotted with
asterisks and the more extreme with circles. Those extreme values that lie
off the scale are printed at the top of the figure.
Figures 6-8a to 6-8h present the results of total EDA, EDA sampling areas,
and comparison areas using box plots. The "boxes" in these plots are reduced
to lines when more than 75 percent of the data is non-detect. In these plots,
non-detect concentrations are coded with a value of zero.
6.4 STATISTICAL COMPARISONS OF SAMPLED
AREAS
This section gives the results of applying the statistical techniques described
in Section 5.5, Statistical Analysis, and Appendix B to the soil concentration
data. One of the objectives of the study was to produce precise data of known
quality suitable for making reliable comparisons between EDA sampling
areas and comparison areas. Therefore, the statistical comparisons reported
upon in this section use only the Good data (observations) as described in
Section 6.2 of this report. There is no evidence that the usability of the data
is related to sampling area or concentration in any way that would bias the
comparisons.
Because the design of the sampling program explicitly randomized certain
external factors, statistical inferences for other data groupings would require
analysis to check that an unbiased allocation of factors occurred for the new
groupings. If such an allocation were lacking, the external factors not ran-
domized could bias the comparison results.
The statistical method outlined in Section 5.5, based on the Wilcoxon rank-
sum test, was applied to the data described in Section 6.3. Statistical com-
parisons were made, first, between the seven EDA sampling areas and the
three comparison areas, and second, among the three comparison areas.
Table 6-8 shows these comparisons for each LCIC (univariate comparisons),
6-36
-------
5.6
o
19.8
5.6
o
o
o
o
EDA1
EDA2
EDA3
EDA4
EOA5
Note:
EDA (Column 8) - Composite of Columns
1 through 7
C 8 T = Cheektowaga and Tonawanda
EDA6
Area ID
EDA7
EDA
C1221
C1225
C&T
Figure 6-8a
LCIC = 1, 2-DICHLOROBENZENE
SOIL ASSESSMENT — INDICATOR CHEMICALS
CONCENTRATIONS BY SAMPLED AREAS
-------
30
C
n 20
c
e
n
t
r
0
1
i
o
n
i
n
P 10 i
C
0-
o
45
4
31
167.3
52.6
45. 1
. 1 41.7
.7 167.3 35.6
.7 52.6 35.6 31.7 33.1
o
o o
o
o
8
&
e 8
0 0 |
8
8'
0 o 8
o Q
0 I
o o
* ° i
; ° : o § i :
I J e :
1*1 1 I
1 1 I | g T 1 |
" col 1 — ( t ol 1 <» T
-------
40
30
C
o
rt
C
e
n
I
t
o
I
0
n
20-
10
67.2
30.8
50.2
44.2
66.7
44.8
182.4
168.6
52.0
182.4
168.6
67.2
66.7
52.0
50.8
50.2
44.8
44.2
64.3
EDA1
EOA2
EDA3
EDA4
EDA5
Note:
EDA (Column 8) = Composite of Columns
1 through 7
C 4 T = Cheektowaga and Tonawanda
EOA6
Areo ID
EDA7
EDA
CT221
CT225
C&T
Figure 6-8c
LCIC = 1, 2, 3, 4-TETRACHLOROBENZENE
SOIL ASSESSMENT — INDICATOR CHEMICALS
CONCENTRATIONS BY SAMPLED AREAS
-------
0.32
0.31
0.30-
0.29
0.28
0.27-
0.26-
0.25-
0.24
0.23
o 0.22-
n 0 21 -
e 0.20-
n 0. 19
J 0.18
o 0.17-
1 0. 16-
1
o 0. 15
n 0. 14
i 0. 13
n 0. 12
p 0.11
P 0. 10
0 09
\j . v y
0.08
0.07
Of\ C
. Ub
0.05
Of\A
\J"
0.03
0.02
0.01
0 00
u>
0
«
— 1 —
«n
™
I ~1
r^
CD
CD
£
*
11
j , | ,
«/•»
*
irt
*
s z
EDA1
EDA2
EDA3
EDA4
EOA5
Note:
EDA (Column 8) = Composite of Columns
1 through 7
C S T = Cheektowaga and Tonawanda
EDA6
Area ID
EDA7
EDA
CT221
CT225
C&T
Figure 6-8d
LCIC = 2-CHLORONAPHTHALENE
SOIL ASSESSMENT — INDICATOR CHEMICALS
CONCENTRATIONS BY SAMPLED AREAS
-------
40
30-
C
o
n
C
e
n
1
i
0
J 20-
0
n
i
n
P
C
10
0
o>
*o
69
152.5
129.9
100.3
83.5
152.5 83.5 69.7
7 100.3 129.9 57.6 57.6
8
o
0
o
o
o
e
§
o
o
o
o
0 0 ° S °
0 0
O D
8 „ o g
Boo" |
oo 6 g
* « o . „ o • 9
* *\ S o ° T
| " ' I • | • ! | • | r | • | i | . | . | . | | . | • |
EDA1 EOA2 EDA3 EDA4 EDA5 EDA6 EOA7 EDA CT221 CT225 C&T
Note: Are° ID
EDA (Column 8) = Composite ol Columns eijin«» C Oa
1 through 7 figure o-ue
C * T = Cheektowaga and Tonawanda LCIC = ALPH A-BHC
SOIL ASSESSMENT — INDICATOR CHEMICALS
CONCENTRATIONS BY SAMPLED AREAS
-------
80.0
38.8
5
4
C
O
n
c
p
n 3
1
r
0
1
i
o
n
i 2
n
P
C
1
0
38.8 10.0
9.8 9.8
5.8 80.0 1Q,0 508 5.4
o>
o
0 g
o
o
0 0
0 0
0 8
o
0 0
1
o
o 8
o
8
Q
o o o O 0
* o |
o o B
0 8
«• e
0 M
1 § 0 0 ° i ° °
| 1^8l e
CM o.| | - 1 o jj < | a> „ op—1 1 CM Q 0. 0 -
^ , , , , , , , , , , , 1 1 , , 1 , , . | . , . i
EDA1 EDA2 EDA3 EDA4 EOA5 EDA6 EDA7 EDA CT221 CT225 C&T
Areo ID
Note:
EDA (Column 8) = Composite of Columns
1 through 7
C 8 T = Cheektowaga and Tonawanda
Figure 6-81
LCIC = DELTA-BHC
SOIL ASSESSMENT — INDICATOR CHEMICALS
CONCENTRATIONS BY SAMPLED AREAS
-------
60
50-
P 20
10
101.7
72.2
4108.3
1729.1
241.3
82.5
663.5
89.5
EOA1
EDA2
EDA3
EDA4
EDA5
Note!
EDA (Column 8) - Composite of Columns
1 through 7
C 4 T = Cheektowaga and Tonawanda
EDA6
Areo ID
1 •
EOA7
4108.3
1729. 1
663.5
241.3
101.7
89.5
82.5
7^.2
,
EDA
CT221
CT225
C&T
.-• c 0_
Figure 6-8g
LCIC = BETA-BHC
SOIL ASSESSMENT — INDICATOR CHEMICALS
CONCENTRATIONS BY SAMPLED AREAS
-------
85.6
47.0
26.3
21 .0
16.3
14.0
13.3
10 -
9
8
C 7
o
n
t
e
n 6
t
r
0
! 5
0
n
i 4
n
P
E 3
2
1
0
21 0 85.6 26.3 12.7 ,5
12,2 47.0 14.0 12^2
1.4 12.7 16.3 13.3 11.4
0 °
0 °
o
o °
8
o
o
a
o
o
o ®
o
o 8
n
o o i
0 0
e §
o 2
2 * ° o ° 8
1 8 ° | •
rJTL ~\ o( 1 ^. i c^ 0 eo B °* 8 C)| 1 *° B Si
.8
o
o
o
e
* o>
1 1 1 1 r
EDA1 EDA2 EDA 3 EDA 4 EDA 5 EDA6 LDA7 EDA CT221 CT225 C&T
Areo ID
Note: Figure 6-8h
EDA (Column 8) = Composite of Columns LCIC = GAMMA-BHC
1 tt"ough 7 A SOIL ASSESSMENT -
C » T = Cheektowaga and Tonawanda CONCENTRATIONS BY
INDICATOR CHEMICALS
SAMPLED AREAS
-------
Table 6-8
RESULTS OF NONPARAMETRIC UNIVARIATE STATISTICAL COMPARISONS WITH OBSERVATIONS CLASSIFIED
AS GOODa>b'c
Ol
Median
Comparison Cone.
EDA Sampling Areas
Comparison Areas
LCIC Area
1 ,2-dichlorobenzene 221
225,
C&T*
1,2,4-trichlorobenzene 221
225
C&T
1,2,3,4-tetrachlorobenzene 221
225
C&T
2-chloronaphthalene 221
225
C&T
alpha-BHC 221
225
C&T
(ppb)
0.39
0.43
0.36
0.65
0.60
0.13
0.53
0.61
0.05
0.06
0.07
0.03
0.18
0.11
ND
1
1.01
++
++
++
8.67
++
++
++
11.48
++
++
++
ND
0
-
O
8.25
++
++
++
2
0.39
O
O
0.91
++
++
++
1.17
++
++
++
0.04
0
-
O
0.40
++
++
++
3
0.40
0
O
++
0.89
+
+
++
1.06
++
++
++
0.04
O
~
o
0.22
0
0
++
4
0.36
—
.
O
0.43
-
O
•n-
0.38
-
0
++
0.04
0
O
O
0.12
0
0
++
5
0.39
0
0
o
0.52
0
0
++
0.43
0
O
++
0.06
o
-
o
0.13
0
0
++
6
0.36
—
0
o
0.37
—
~
•n-
0.31
—
—
++
0.05
O
O
0
0.07
-
O
+4-
7
0.41
0
0
0.58
0
0
++
0.54
O
O
•M-
0.06
O
O
+
0.14
O
O
++
221
0.39
O
0.65
0
•M-
0.53
0
++
0.06
o
o
0.18
0
++
225
0.43
0
O
0.60
0
++
0.61
0
++
0.07
o
•+•
0.11
O
++
C&T
0.36
~
o
0.13
—
—
0.05
—
—
0.03
O
-
ND
—
—
a. ++: EDA > Comparison area at 0.01 significance level
+: EDA > Comparison area at 0.05
significance
o: No significant difference at 0.05 significance
level
level
-: EDA < Comparison area at 0.05 significance level
--: EDA < Comparison area at 0.01 significance level
b. All test results reported are based on two-sided p-values.
c. The first entry In each column is median concentration. ND indicates non-detect.
d. C&T = Cheektowaga and Tonawanda
-------
Median
Table 6-8
(continued)
Comparison Cone.
LCIC
delta-BHC
beta-BHC
gamma-BHC
TOTALS
Area
221
225
C&T
221
225
C&T
221
225
C&T
++
+
O
-
—
(ppb)
ND
ND
ND
ND
ND
ND
ND
ND
ND
1
1.13
++
++
++
11.58
++
++
++
1.73
++
++
++
21
0
2
1
0
2
0.04
++
++
++
0.20
++
O
++
0.09
++
+
++
17
1
5
1
0
EDA Sampling
3
ND
++
O
++
0.13
++
O
++
ND
++
O
++
12
2
9
0
1
4
ND
•«•
O
++
ND
+
O
++
ND
+
O
++
6
3
11
3
1
Areas
5
ND
+
o
++
ND
O
o
++
ND
O
o
++
6
1
16
1
0
Comparison Areas
6
ND
o
o
++
ND
O
~
++
ND
O
-
++
6
0
10
2
6
7
ND
o
o
•M-
ND
0
O
•H-
ND
O
O
•M-
6
2
16
0
0
221
ND
0
o
ND
-
•M-
ND
O
++
6
0
9
1
0
225
ND
0
++
ND
+
•M-
ND
0
++
6
2
8
0
0
C&T
ND
O
..
ND
—
..
ND
~
~
~0
0
3
1
12
All Symbols
24
24
24
24
24
24
24
16
16
16
-------
and Table 6-9 shows the comparisons for the LCICs as a group (multivariate
comparisons). Table 6-10 provides a summary of the number of univariate
comparisons that were significant. (Tables of the actual "P" values are in
Appendix K.)
Five symbols are shown in the following tables: ++, +, o, -, and --. The ++
symbol indicates that the concentration of an LCIC (or the grouped LCICs)
is significantly higher than in the comparison area at the 0.01 significance
level (two-sided); the + symbol shows significance at the 0.05 level (two-
sided); o indicates no significant difference at the 0.01 or 0.05 level; - indi-
cates that the LCIC (or grouped LCICs) concentration in the EDA is
significantly lower than in the comparison area at the 0.05 level; and — indi-
cates significantly lower at the 0.01 level. (A 0.01 significance level means
that there is a 1-in-100 chance that a difference as large as that observed could
occur by chance alone. The 0.05 significance level corresponds to a 1-in-20
chance.) The median concentration values are also given to show the mag-
nitude of the concentrations being compared.
Figure 6-9 graphically displays the results of these comparisons between the
EDA sampling areas and the comparison areas for each LCIC; Figure 6-10
shows the results for the LCICs as a group; and Figure 6-11 shows the results
for the comparison area analyses. In these figures, the "up" arrows represent
LCIC concentrations that are significantly higher at the 0.05 level (two-
sided) than those in the comparison area, while the "down" arrows represent
LCIC concentrations that are significantly lower than comparison area con-
centrations.
Some additional comments are needed with respect to Tables 6-8 to 6-10.
No attempt was made to adjust the tables for two possible effects that would
influence the comparison results. The first effect is the number of "false posi-
tive" results arising from conducting a large number of comparisons. The
significance level represents the probability that the null hypothesis (no dif-
ference between two sampling areas) would be rejected even if there were in
fact no true difference. (In other words, the probability that the comparison
could mistakenly find a difference between two areas.) When many tests are
conducted, some rejections of the null hypothesis are expected to occur by
chance ("false positives"). This effect can be corrected under certain assump-
tions, but there is no universally valid method of making such a correction.
6-47
-------
Table 6-9
RESULTS OF NONPARAMETRIC MULTIVARIATE STATISTICAL COMPARISONS WITH OBSERVATIONS
CLASSIFIED AS GOODa'b
Comparison
Area 1
221 ++
225 ++
C&T° ++
EDA Sampling Areas
23456
++ -M- — +
++ + O • -
++ -M- ++ ++ ++
Comparison Areas
7 221 225 C&T
++ + —
o -
++ -M- ++
a. ++: EDA > Comparison area at 0.01 significance level
+: EDA > Comparison area at 0.05 significance level
o: No significant difference at 0.05 significance level
oo
-: EDA < Comparison area at 0.05 significance level
--: EDA < Comparison area at 0.01 significance level
b. The direction of the difference between the EDA and Comparison Areas is based on the sign of the sum of the elements of the 8X1
vector of rank-sums (Appendix B)
c. C&T = Cheektowaga and Tonawanda
-------
Table 6-10
SUMMARY OF NONPARAMETRIC UNIVARIATE STATISTICAL COMPARISONS WITH OBSERVATIONS
01 AQQIPIPH AQ nnr>nai b
CLASSIFIED AS GOOD
Totals over all LCICs
Comparison
Area
CT221
CT225
ox C&T°
A-
\0
1
++ 7
+ 0
0 1
0
0
++ 7
+ 0
0 0
1
0
++ 7
+ 0
0 1
0
0
2
6
0
2
0
0
4
1
2
1
0
7
0
1
0
0
EDA Sampling Areas
3
4
1
3
0
0
1
1
5
- 0
1
7
0
1
0
0
4
0
3
2
2
1
0
0
7
1
0
6
0
2
0
0
5
0
1
7
0
0
0
0
7
1
0
6
0
2
0
0
6
0
0
4
1
3
0
0
4
1
3
6
0
2
0
0
7
0
0
8
0
0
0
0
8
0
0
6
2
0
0
0
Comparison Areas
221
0
0
7
1
0
6
0
2
0
0
225
0
1
7
0
0
6
1
1
0
0
C&T
0
0
2
0
6
0
0
1
1
6
a. ++:
o:
EDA > Comparison area at 0.01 significance level
EDA > Comparison area at 0.05 significance level
No significant difference at 0.05 significance level
-: EDA < Comparison area at 0.05 significance level
-: EDA < Comparison area at 0.01 significance level
b. All test results reported are based on two-sided p-values.
c. C&T = Cheektowaga and Tonawanda
-------
II
;j.
II
oM
. SAMPLING
AREA
BOUNDARY
LOVE CANAL
0 8000000
MMMM
LEGEND:
A LCIC concentration in this EDA
1 sampling area is greater than that in the
comparison area at the 0.05 level of
significance (two-sided)
O No significant difference
iLCIC concentration in this EDA
sampling area is less than that in the
comparison area at the 0.05 level of
significance (two-sided)
Top Line = Census Tract 221
Middle Line = Census Tract 225
Bottom Line = Cheektowaga and Tonawanda
LCICs are in the following order on the line:
7 7ill 77/
o
oil
of!
Mi
i
oil
oil
Ml
N.
ooMI
1 M M
c
7
IF:
-I
2
I
0MM
**0*
oMM
sL 1
^N;
5-
Ml.MM
Ml MM
M*!MM
SOURCE: Neighborhood boundaries adopted from
the Proposed Habitability Criteria document
(NYSDOH and DHHS/CDC, 1986)
////ftft
•V ^" "V
"•'
-------
LEGEND:
t
Q No significant difference
LCIC concentrations in this EDA
sampling area are generally greater
than those in the comparison area
at the 0.05 level of significance
(two-sided)
I
LCIC concentrations in this EDA
sampling area are generally less
than those in the comparison area
at the 0.05 level of significance
(two-sided)
Comparison area symbols are in the
following order:
CT221 CT225 CK&TON
NOTE: The direction of the difference between
the EDA and Comparison areas is based on the
sign of the sum of the elements of the
8x1 vector of rank-sums (Appendix B).
SOURCE: Neighborhood boundaries adopted from
the Proposed Habitability Criteria document
(NYSDOH and DHHS/CDC, 1986)
Figure 6-10
SOIL ASSESSMENT — INDICATOR CHEMICALS
SUMMARY OF MULTIVARIATE COMPARISONS
(LCICs AS A GROUP)
EDA SAMPLING AREA TO COMPARISON AREA
-------
f I
H I
CENSUS TRACT 221
i
I
I
CENSUS TRACT 225
r.rt
UN I VARIATE
t t
1
r t n t
MULTI -
VARIATE
b 1
J 1
UMIVARIATE
.000.'
t t M t 1
MULTI-
VARIATE
^ 1
t 1
LEGEND:
LCIC concentration in this
i comparison area is greater at the
T 0.05 level of significance (two-
I sided) than that in the comparison
area indicated on the line.
O No significant difference
iLCIC concentration in this
comparison area is less at the
0.05 level of significance (two-
sided) than that in the comparison
area indicated on the line.
LCICs are in the following order on the line:
NOTE: The direction of the difference between
the EOA and Comparison areas for the
Multivariate comparison is based on
the sign of the sum of the elements of
the 8 x i vector of rank-sums (Appendix B).
I—J
CT22S
UNIVARIATE
1 U ° I ° »
°miii
MULTI.
VARIATE
1 T
1 T
Figure 6-11
SOIL ASSESSMENT —INDICATOR CHEMICALS
SUMMARY OF STATISTICAL COMPARISONS:
COMPARISON AREA TO COMPARISON AREA,
UNIVARIATE AND MULTIVARIATE
-------
A second effect that influences only the univariate test results is that of cor-
relation among the LCICs. Because some of the LCICs are correlated (see
Appendix B), there is some redundancy in the information contributed by
different LCICs about the sampled areas. For instance, the BHC isomers
tended to correlate with other isomers of BHC. Similarly, for example, if a
high value of 1,2,4-trichlorobenzene was found in a given sample, a high
concentration of 1,2,3,4-tetrachlorobenzene was also likely to occur. This
means that if one LCIC is found to be significantly different in two areas, the
LCICs highly correlated with it will also tend to be significantly different.
This effect also occurs even if there are no differences between sampling
areas (null hypothesis is true), resulting in an increase in the apparent rejec-
tions that would be due to chance ("false positives"). The effect of correla-
tion among the LCICs is explicitly accounted for in the multivariate test. The
multivariate test results give an overall picture of differences between the
EDA sampling areas and the comparison areas with respect to all the LCICs.
In Tables 6-8 to 6-10, there were very few changes in the comparison results
between those carried out at the 0.01 and 0.05 significance levels, indicating
that most of the differences were more than marginally significant.
Additional analyses of a supplementary nature were carried out and are
reported in Appendix K. Two kinds of data sensitivity analyses were per-
formed: First, the effect of relaxing the data quality control to include "un-
certain" data was considered. Second, the effect of classifying all data below
1 ppb as non-detects was considered. It was concluded that these changes
had only minimal effect on the comparison results. In addition, it was deter-
mined that the Wilcoxon rank-sum test, applied to laboratory blocked data
(described in Section 5.5), met the statistical goal for the statistical com-
parisons as set forth in Section 3.0.
As directed by the TRC, the purpose of this assessment is to present data and
is specifically not to develop conclusions or interpretations of the data. The
findings of this study will be reviewed by a panel of scientists. In accord-
ance with the habitability criteria document (NYSDOH and DHHS/CDC,
1986), the results will then be presented to the Commissioner of Health for
the State of New York for use in a determination of habitability.
6-53
-------
REFERENCES
CH2M HILL. 1987a. Love Canal Habitability Study-Soil Sample Collec-
tion and Preparation Quality Assurance Project Plan (Final Revised
Version).
CH2M HILL. 19875. Love Canal Habitability Study-Soil Sample
Laboratory Analysis Quality Assurance Project Plan.
CH2M HILL. 1987c. Pilot Study for the Love Canal EDA Habitability
Study. Volumes I and II.
CH2MHILL. 1987d. Summary of Responses to the Peer Review of the Pilot
Study for the Love Canal EDA Habitability Study. Volumes I and n.
Hollander, M. and Wolfe, D. A. 1973. Nonparametric Statistical Methods.
]. Wiley, New York.
Life Systems. 1986. Peer Review of the Proposed Habitability Criteria for
the Love Canal Emergency Declaration Area.
Life Systems. 1987. Peer Review of the Love Canal Full-Scale Sampling
Plan.
Life Systems. 1988. Love Canal Emergency Declaration Area, Habitability
Study-Final Report, Volume I, Introduction and Decision Making
Documentation Report.
NYSDOH and DHHS/CDC. 1986. Love Canal Emergency Declaration
Area; Proposed Habitability Criteria.
OTA. 1983. Habitability of the Love Canal Area, An Analysis of the Tech-
nical Basis for the Decision on the Habitability of the Emergency
Declaration Area-A Technical Memorandum.
U.S. EPA, Office of Research and Development. 1982. Environmental
Monitoring at Love Canal. Volumes 1, 2, and 3.
R-l
-------
APPENDIX A
Statistical Concepts
-------
Appendix A
STATISTICAL CONCEPTS
The soil habitability study involved the use of a comparison
approach to determine whether an EDA sampling area, taken as
a whole, had different (higher or lower) LCIC concentrations
than a specified comparison area. Both the design of the
sampling protocol and the analysis of the data were con-
ducted in a statistical framework. This appendix reviews
some basic statistical concepts that are needed to under-
stand the statistical sampling design (Section 4.2), the
statistical methods (Section 5.5), the statistical results
(Section 6.4), and the supporting material presented in
Appendices B, I, J, and K.
INTRODUCTION
What do we mean by "different" and how do we measure this?
Comparing two single numbers is easy—10 is different (and
larger) than 6. Comparisons between EDA sampling and com-
parison areas are harder. Are the 50 or more LCIC sample
concentrations (from 50 or more soil samples for example)
different than the 50 or so LCIC sample concentrations from
the comparison area? While some values may be larger and
some smaller, the values are variable both within and be-
tween the two areas, regardless of whether or not contamina-
tion is present, because of natural spatial variability,
imprecise measurement procedures, and for other reasons dis-
cussed in Section 4.2.
Statistical hypothesis testing is a formal, mathematical
procedure for taking sets of imprecise (variable) data and
making comparisons as to whether the data sets are
different.
The sets of data can be represented by probability distri-
butions (also referred to as probability density functions).
The area under the probability density function in any given
interval is the probability that a sample drawn from that
probability distribution will have a value (e.g., concen-
tration) in that interval (see Figure A-l).
Figure A-2 illustrates the essential elements of a statis-
tical comparison. Usually some measure of the data, a test
statistic, is used for comparison. For the illustrations
here, assume that the median concentration can be used to
represent the data. If an unlimited amount of data were
available, the true median could be calculated exactly.
However, in practice sample sizes are finite, the true
median cannot be known, and the sample median must be con-
sidered to be an estimate of the true median, with some
A-l
-------
associated uncertainty. Different sets of samples would
give different values for the sample median. Figure A-2(b)
depicts the probability distribution of the median. In some
cases, probability theory will permit an estimate of this
probability distribution from the probability distribution
of the underlying process, i.e., the concentrations.
Probability
Density
Area = Probability G,< C <
Concentration (C)
Figure A-1
PROBABILITY DENSITY FUNCTION
To make a comparison, a working hypothesis is needed. The
usual hypothesis in statistical testing of the type de-
scribed here is that (the probability distributions of) con-
centrations from the EDA sampling and comparison areas are
the same. In the hypothetical example, this is equivalent
to a statement that the medians are equal or the difference
in medians is zero. This primary hypothesis is defined as
the null hypothesis. It is the hypothesis that is addressed
by statistical comparison tests.
The next step, shown in Figure A-2(c), is to determine the
probability distribution (the density function) of the test
statistic under the null hypothesis. The null hypothesis
assumes that the difference in the medians is zero, but
since the sample medians are uncertain their difference is
also uncertain. This uncertainty can be represented by a
probability density function.
A-2
-------
(a) Probability
Density
Area 1 Data
mean
Probability
Density
Concentration
Area 2 Data
Concentration
mean
(b) Probability
Density
Sample
mean-)
Probability
Density
mean
concentration
Sample
mean
concentration
Null Hypothesis Ho: mean1 -mean2 = 0
I
It null hypothesis true
(c)
sample
difference
difference in means
Figure A-2
STATISTICAL COMPARISON STEPS
-------
The final step is to test the value of the test statistic
(in the example, the difference of the sample medians,
delta) as to whether it is consistent statistically with the
probability distribution under the null hypothesis. Under
the null hypothesis, the expected value of delta, the dif-
ference in the sample medians, is zero. Therefore, delta is
tested to see if it is significantly different from zero.
The significance level is the probability of rejecting the
null hypothesis (H ) when it is in fact true. This proba-
bility, often described by the symbol alpha, represents the
probability of making the error of rejecting H when it
should be accepted; this error is referred to as a Type I
error.
If the only concern is positive deltas (i.e., whether the
EDA has a higher median than the comparison area), then
specifying alpha allows one to calculate a critical value D
as shown in Figure A-3. Any sample value of delta equal to
or greater than D results in rejecting the null hypothesis
that there is no Sifference in the two data sets.
Probability
Density
Probability Density for l-fc true
Area = a (Alpha)
difference
Critical region
2)
If the samples give m1-m2 > Dc, reject HO
Figure A-3
LOCATION OF THE CRITICAL REGION
There are several points to keep in mind:
o The statistical test is carried out on a test sta-
tistic (in this example the difference in the sam-
ple medians), where the probability density is
derived from the probability distribution of the
se8855/049/4
A-4
-------
concentrations and the number of samples. There-
fore, taking more samples reduces the variability
in the test statistics. A larger variability in
the underlying process (the concentration data)
increases the variability in the test statistics.
Figure A-4 illustrates this.
Probability
Density
of
concentration
High Variability
Probability
Density
of
concentration
Mean
Concentration
Low Variability
Mean
Concentration
Probability
Density
for Mean
Taking N samples
sample
Mean
Probability
Density
for
Mean
Mean
value
Taking N samples
sample
Mean
Mean
value
Probability
Density
for
Mean
Taking 2N Samples
More samples needed
for same variability
of estimate of mean.
If data variability higher.
sample
Mean
Mean
value
Figure A-4
INFLUENCE OF SAMPLE SIZE ON VARIABILITY
-------
o In the comparison example above, the null hypothe-
sis H is accepted or rejected. No reference is
made to what is accepted if H is rejected, beyond
the statement that the two comparison parameters
are statistically different. It is not explicitly
determined "how different" one area is. What can
be analyzed is the probability of detecting a
given difference, the minimum unacceptable differ-
ence. The minimum unacceptable difference is the
smallest difference that will imply that the EDA
is not habitable. This is usually called the al-
ternative hypothesis, H . The probability of de-
tection at this given difference is referred to as
the power of the statistical test. Figure A-5
illustrates the relationship between the probabil-
ity determination for H and H and the power
curve. ° a
Computationally, power is the area under the H probability
density curve to the right of D (see Figure A-5), the cri-
tical value defined earlier. Tne area under the H prob-
ability density curve to the left of D (i.e., foravalues of
the sample difference where H would nSt be rejected) is
known as beta (one minus power) and is referred to as the
Type II error. It is sometimes also referred to as the
"false negative" probability. A Type II error occurs when
H is accepted when in fact it is not true.
SOURCES OF VARIABILITY
A set of concentration data is the end result of collecting
and analyzing samples. A data value can be thought of as
being comprised of (the sum of) a "true" value and a random
"error" component. An important component of the random
error is natural spatial variability, for instance, within a
comparison area. In addition to this spatial variability,
there are a variety of sources of error that are related to
the data collection and sample analysis procedures. Some of
these sources are:
o Sample collection variability. It is possible
that sample collection procedures may introduce
variability in the measured concentrated values.
o Sample handling variability. Some of the contami-
nants of interest are volatile organics. If they
are not properly handled and carefully stored,
their concentrations may not correctly reflect
field concentrations.
A-6
-------
Null Hypothesis HO A = 6
Alternate Hypothesis H0 A = A mtn
H
Area = beta
Probability
Density
Probability
Density
Area =
Power (1-Beta)
a)forHa:^mjnsmaii
difference
of means
b) for HaAmln large
difference
of means
c) Power as function of
difference
of means
Figure A-5
INFLUENCE OF A rtn (MINIMUM ACCEPTABLE DIFFERENCE)
ON THE PROBABILITY OF DETECTION (POWER)
-------
o In the comparison example above, the null hypothe-
sis H is accepted or rejected. No reference is
made ?o what is accepted if H is rejected, beyond
the statement that the two comparison parameters
are statistically different. It is not explicitly
determined "how different" one area is. What can
be analyzed is the probability of detecting a
given difference, the minimum unacceptable differ-
ence. The minimum unacceptable difference is the
smallest difference that will imply that the EDA
is not habitable. This is usually called the al-
ternative hypothesis, H . The probability of de-
tection at this given difference is referred to as
the power of the statistical test. Figure A-5
illustrates the relationship between the probabil-
ity determination for H and H and the power
O 3.
curve.
Computationally, power is the area under the H probability
density curve to the right of D (see Figure A-5), the cri-
tical value defined earlier. Tne area under the H prob-
ability density curve to the left of D (i.e., foravalues of
the sample difference where H would nSt be rejected) is
known as beta (one minus power) and is referred to as the
Type II error. It is sometimes also referred to as the
"false negative" probability. A Type II error occurs when
H is accepted when in fact it is not true.
SOURCES OF VARIABILITY
A set of concentration data is the end result of collecting
and analyzing samples. A data value can be thought of as
being comprised of (the sum of) a "true" value and a random
"error" component. An important component of the random
error is natural spatial variability, for instance, within a
comparison area. In addition to this spatial variability,
there are a variety of sources of error that are related to
the data collection and sample analysis procedures. Some of
these sources are:
o Sample collection variability. It is possible
that sample collection procedures may introduce
variability in the measured concentrated values.
o Sample handling variability. Some of the contami-
nants of interest are volatile organics. If they
are not properly handled and carefully stored,
their concentrations may not correctly reflect
field concentrations.
A-8
-------
Laboratory variability. There is inevitable im-
precision in all laboratory results. Laboratory
variability can be represented by the probability
distribution of the results of analyzing (splits
of) a single sample many times.
Inter-laboratory variability. It is possible that
a laboratory could analyze many splits and show
relatively low variability, but that all the
analyses would yield concentration estimates that
were different from those performed by another
laboratory.
MULTIPLE COMPARISONS
In the statistical comparisons discussed up to now, it has
been assumed that the test statistic is based on two data
sets: one from an EDA sampling area and one from a compari-
son area. On the basis of the test statistic and the speci-
fied alpha level (Type I error probability), the given EDA
sampling area can be classified as having, or not having,
higher concentrations of the LCICs than the comparison area.
If such a classification is made sampling area by sampling
area for each LCIC, then the probability of incorrectly
classifying a sampling area increases substantially for the
study as a whole with the number of LCICs compared. A simi-
lar problem exists for evaluation of the classification of
multiple sampling areas, or for classification of a given
sampling area with respect to multiple comparison areas.
For a given LCIC and sampling area-comparison area pair, an
alpha of 0.05 implies a 5 percent probability of rejecting
the null hypothesis that there is no difference between the
EDA neighborhood and the comparison area. This Type I error
probability refers to a single LCIC comparison. Now if
eight LCIC comparisons are made separately for each EDA
neighborhood, and the LCICs are statistically independent,
the probability of misclassifying the EDA neighborhood for
at least one LCIC increases to almost 34 percent. For a
given LCIC with comparisons between seven sampling areas and
three comparison areas (21 total comparisons), the probabil-
ity of at least one misclassification is 66 percent.
There are a number of alternatives that can address this
problem. Two of these are:
o Adjust the significance level. The significance
level (alpha) for the single comparisons could be
adjusted to result in an alpha level that is con-
sistent with respect to the total number of com-
parisons. If the individual comparisons are
A-9
-------
between statistically independent data, one ap-
proach to doing this is to divide the nominal
alpha by the number of comparisons to be made.
This is referred to as the Bonferroni correction.
The approach is extremely conservative and can
result in almost no comparison being significant
if a large number of comparisons are made.
Use a multivariate comparison approach. With a
multivariate approach, all the LCIC data for an
EDA neighborhood are analyzed together so that the
number of comparisons is reduced. This also has
the advantage that the correlation between vari-
ables is accounted for explicitly. The only com-
plication with the multivariate approach is
interpretation. It is possible, for instance, for
the multivariate test to indicate a statistically
significant difference for the eight LCICs as a
group, even though there are no significant dif-
ferences among individual LCICs. In addition, it
may be difficult to interpret the direction of
change. For example, the multivariate test may
indicate a statistically significant difference
between sampling areas when the univariate test on
the individual LCICs indicates that four are
greater in the EDA sampling area than in the com-
parison area, and four are less.
A-10
-------
APPENDIX B
Development of a Statistical Approach
for the Sampling Area Comparisons
-------
Appendix B
DEVELOPMENT OF A STATISTICAL APPROACH FOR
THE SAMPLING AREA COMPARISONS
This appendix describes in detail the statistical methodology
that was used to design the LCIC soil assessment sampling
program and to analyze the resulting data. The appendix is
divided into five sections. Section 1 discusses the statis-
tical problems that were encountered in the 1982 habitability
study and emphasizes the comparison approach basis of the
present study. Section 2 describes the process that was
followed to determine an appropriate statistical model for
the LCIC data and to determine an appropriate statistical
method (based on this model) to perform the comparisons.
Section 3 briefly describes how the sampling design was cre-
ated based on the chosen statistical model and method. Sec-
tion 4 describes the slightly modified statistical model and
method that were used to analyze the resulting LCIC data.
Finally, Section 5 gives a retrospective power analysis,
based on the observed distributions of the LCICs in the com-
parison areas.
B.1 BACKGROUND
The assessment of EDA habitability with respect to soil
chemistry was initially based on data resulting from an en-
vironmental study released in 1982 (U.S. EPA, 1982) . The
primary objective of the 1982 study was to characterize soil
chemistry in the Love Canal area; comparison of the EDA soil
chemistry with comparison areas was not an original objective
of the study. However, the 1982 study results were analyzed
using a comparison approach; that is, chemical concentra-
tions estimated from soil samples collected in Love Canal
neighborhoods were compared to samples from a control neigh-
borhood. During subsequent review, the conclusions of the
study regarding habitability were called into question (OTA,
1983). Specifically, questions arose pertaining to the qual-
ity and limits of detection of some of the soil (and other
media) samples and the appropriateness of the statistical
design.
Because the 1982 study was not originally intended to .com-
pare EDA areas with control areas, the design was highly
imbalanced: EDA sample sizes were much larger than those
from the control area. In addition, the fraction of non-
detectable concentrations was extremely high (more than
95 percent for most of the chemicals that were analyzed).
The 1982 study conclusions regarding habitability were set
aside following external review, largely because the small
sample sizes in the control area precluded any statistical
B-l
-------
comparison from achieving an adequate power to detect possi-
ble differences between the Love Canal and control area, and
because laboratory analyses were of unknown quality. The
present study was initiated largely to provide a data col-
lection design and results that would avoid the problems
encountered in the 1982 study and to enable a valid statis-
tical comparison approach to be implemented. The problems
encountered in the 1982 study highlight the importance of
using a proper experimental design to aid in reaching a
decision regarding the habitability of the EDA.
The current study was based on a statistical comparison
approach, as directed by the TRC (NYSDOH and DHHS/CDC,
1986). The EDA neighborhoods were to be compared to one or
more "similar" neighborhoods that were known to be distant
(>l/2 mile) from any hazardous waste site. Early in the
study, the EDA was subdivided into 13 neighborhoods, which
were later aggregated into seven EDA sampling areas. One
aggregate comparison area was selected that was later aug-
mented with two additional comparison areas. The EDA sam-
pling areas and comparison areas were selected on the basis
of socioeconomic, geographic, and other factors (NYSDOH and
DHHS/CDC, 1986).
In the design of a statistical assessment procedure to im-
plement the comparison method, several questions were
addressed:
1. What are the most appropriate statistical methods?
2. How should samples be allocated to the EDA and compari-
son areas to assure that the comparison procedure is
sufficiently sensitive to differences, if any, between
the EDA and comparison areas?
3. How can the experimental design account for unavoidable
within-laboratory variability and interlaboratory dif-
ferences in analytical methods? The decisions made
regarding each of these issues, and the analyses that
were conducted to support the decisions, are summarized
in this appendix. The proposed statistical analysis
methods and field sampling design were reviewed by a
peer review panel, which met in Niagara Falls on
March 24-26, 1987. The peer reviewers supported the
concept of the comparison approach and its implementa-
tion through the field sampling design and the statis-
tical analysis procedure. The peer reviewers' comments
are summarized in Peer Review of the Love Canal Full-
Scale Sampling Plan (Life Systems, 1987). The substan-
tive comments of the peer reviewers are addressed in
this appendix.
B-2
-------
B.2 INITIAL STATISTICAL MODEL AND EVALUATION OF STATISTICAL
METHODS
On the advice of the TRC, a decision was made that the soil
chemical characteristics of the comparison areas were to be
considered fixed. That is, although samples collected in
each comparison area were intended to provide a statistically
representative sample of conditions in those particular com-
parison areas, neither the comparison area samples nor the
comparison areas themselves could necessarily be considered
to be statistically representative of the soil chemistry in
the Niagara Falls-Buffalo area taken as a whole. The com-
parison areas were, however, thought to be typical of condi-
tions in the larger area.
In a statistical sense, the above is equivalent to the
statement that the comparison areas were treated as fixed,
rather than random, effects in the comparisons. This dis-
tinction is important because it essentially means that the
comparison between any given EDA sampling area and compari-
son area should be interpreted individually.
B.2.1 Initial Model and Candidate Statistical Tests
The decision to design a comparison study leads directly to
the consideration of two-sample statistical tests. When
work was initiated on the statistical design, the extent of
interlaboratory differences was unknown, and the basic model
for a single LCIC that was considered was:
where:
C, = the concentration of the the I ' th observation in
sampling area k
M, = the mean or median concentration in sampling
area k (FIXED)
e, = random error in the £'th observation in sampling
area k (RANDOM)
k = 1,2 (sampling area index; 1 = EDA, 2 = comparison
area)
£ = l,...,n, (sampling1 area replicate index)
The e, terms were considered to be independent random vari-
ables with mean zero that represent four sources of vari-
ability: (1) between-laboratory variability, (2) within-
laboratory variability, (3) variability between sampling
B-3
-------
sites within a given sampling area, and (4) variability as-
sociated with other factors that were not explicitly modeled
(e.g;, collection procedures, storage, transportation, etc).
In terms of the model of Eq. B.I, the habitability criteria
specify a comparison of the magnitude and direction of the
M, terms. To perform these comparisons, four classes of
univariate tests were considered:
1. Two-sample location tests including Student's t-test
(Zar, 1984) and the Wilcoxon rank-sum test (Hollander
and Wolfe, 1973)
2. Two-sample dispersion tests including Shorack's APF-
test (Shorack, 1969)
3. Two-sample proportion tests including Fisher's exact
test (Zar, 1984)
4. Two-sample distribution tests including the Kolmogorov-
Smirnoff test and the Mood quantile test (Zar, 1984)
In addition, two multivariate tests were considered:
1. The multivariate Wilcoxon rank-sum test (Puri and Sen,
1971; Hettmansperger, 1984)
2. An ad hoc adaptation of the univariate Wilcoxon
rank-sum test based on average (over LCICs) ranks
The method of evaluating differences in multiple LCICs is
important because of what is known in statistics as the mul-
tiple comparison problem, which is discussed in Appendix A.
Multivariate tests circumvent this problem by simultaneously
testing differences in multiple (LCIC) chemicals.
Each of the above univariate and multivariate tests, with
the exception of the multivariate Wilcoxon rank-sum test, is
described in Attachment G-2 of Volume II of the pilot study
report (CH2M HILL, 1987).
B.2.2 Description of the Monte Carlo Study to Evaluate the
Candidate Statistical Tests
Based on the model of Eq. B.I, the relative performances of
the candidate statistical tests were assessed using Monte
Carlo simulation. Monte Carlo simulation involves the re-
peated generation of synthetic samples with known character-
istics from a (specified) random population. Because the
true characteristics of the statistical populations are
known (such as the difference in means or medians between a
hypothetical EDA and comparison area), the power and Type I
B-4
-------
error level of the statistical tests, and the effect of any
deviations from the test assumptions, can be evaluated.
The performances of the candidate statistical tests were
assessed with respect to their power as a function of:
o Sample size
o Coefficient of variation of the error terms
o Distribution of the error terms
o The ratio of within-laboratory variability to the
remaining sources of variability
o The fraction of analyses that resulted in
non-detects
Six distributions were used in the Monte Carlo study:
o Normal (with negative values truncated)
o Lognormal
o Gamma
o Mixed exponential
o Generalized extreme value
o Lognormal mixture
According to the null hypothesis, the distribution of LCIC
concentrations in the EDA sampling area is the same as the
distribution in the comparison area. According to the
alternative hypothesis, the distribution of LCIC concentra-
tions in the EDA sampling area is "shifted" relative to the
distribution in the comparison area. Four different kinds
of shifts in distribution were used in the Monte Carlo
s tudy:
o Additive shift, written as f,(x-A)"= f2(x), where
f, (y) and f~(y) denote the probability density
function (paf) of the EDA and comparison area sam-
ples, respectively
o Scale shift, written as f^{ [(x-M)/c] + M }/c =
f2 (x), where M denotes the mean of both the EDA
and comparison areas
o Multiplicative shift, written as f,(x/c)/c = f~(x)
o Upper quantile shift, written as f, (x) = p f2(x) +
(l-p)g(x), where p is a mixing fraction between
zero and 1 (but generally near 1), and g(y) is an
alternative distribution with a larger variance
than that of f0(y)
B-5
-------
The first alternative hypothesis represents a simple shift
of the EDA area LCIC concentrations relative to those of the
comparison area by a constant, that is, all moments of the
populations are the same except the mean. Under the scale
shift, the EDA population of LCIC concentrations has the
same mean as the comparison area, but its standard deviation
is larger by the multiplicative constant c. The multiplica-
tive shift results in an EDA area population that has the
same coefficient of variation as the comparison area, but
has a mean and standard deviation that are larger by the
multiplicative constant c. Under the fourth alternative
hypothesis, the EDA population is a mixture of the compari-
son area population and a second, more variable population.
The upper quantile shift was of particular interest because
it represents a situation in which most of the EDA and com-
parison area samples are drawn from the same distribution,
but a few larger values (e.g., areas of localized contamina-
tion) occur in the EDA neighborhood.
Because the form of the difference (if any) between the EDA
and comparison area populations was unknown, it was important
that the test or tests that were selected should perform
well over all of the alternative shift hypotheses, even
though each test is usually considered optimal for only one
of these types of shifts. However, attention was ultimately
focused on the multiplicative and upper quantile shifts,
which were felt to be the most reasonable. The additive and'
scale shifts are not compatible with the constraint that the
lower bound on LCIC concentration be zero for both the EDA
and comparison areas: an additive shift implies that the
lower bound (zero) of the EDA population is shifted to the
right by the amount A (assuming A >0), while a scale shift
implies that the lower bound is shifted to the left by a
fraction of the mean equal to c - 1 (assuming c >1). On the
other hand, multiplicative and upper quantile shifts pre-
serve the lower bound of both populations at zero.
B.2.3 Results of the Monte Carlo Study
The performance of each of the tests was evaluated for a
range of alternative hypotheses, probability distributions,
and statistical parameters. The three univariate tests that
performed the best over all the simulations were the Wilcoxon
rank-sum test, Shorack's test, and Mood's test. Selected
results are given in Table B-l. The ordering of the tests
shown in this table is based on the observed power of the
tests over a wide range of sample sizes for an order-of-
magnitude shift. These orderings, however, are based on the
assumption that the nominal alpha-level is maintained under
the null hypothesis. Further evaluation of the tests showed
that, in some cases, Shorack's test was liberal with respect
to Type I error probability; that is, the true alpha-level
was greater than the stated value. After attempts were made
B-6
-------
Table B-l
SUMMARY OF SIMULATION -RESULTS FOR WILCOXON,
SHORACK, AND MOOD TESTS
Shift Percent
Type5 Cv Nondetect Order
MLT 2.0 0.85 W,S,M
MLT 2.0 0.95 W,S,M
UPQ 2.0 0.85 S,W,M
UPQ 2.0 0.95 S,W,M
MLT 1.414 0.85 S,W,M
MLT 1.414 0.95 S,W,M
aMLT = multiplicative shift, UPQ = upper quantile shift with
mixing fraction p = 0.3. Shift magnitude under the alter-
native hypothesis was 10 (one order of magnitude) in all
cases. Results are for a = 0.05, but are generally con-
sistent at other a-levels.
W = Wilcoxon, S = Shorack, M = Mood. Ordering reflects
dominant performance over a range of sample sizes from 10
to 100, and generally was consistent for sample sizes
larger than 30. For description of tests, see pilot study
report, Volume II (CH2M HILL, 1987).
8856A/013/1
B-7
-------
to modify Shorack's test to maintain the nominal alpha-
level, the power of the Shorack test generally dropped below
that of the Wilcoxon. Additional analysis of the results
also showed that the Mood test was almost always dominated
by the Wilcoxon test.
The high power of the Wilcoxon test against the additive
shift is generally recognized (Lehmann, 1975). The Wilcoxon
test is not generally mentioned in the literature as an ef-
fective test against a multiplicative or upper quantile
shift (the type of shifts that hold the greatest interest
for this study). However, as long as the null hypothesis is
that the EDA and comparison area populations are indepen-
dently identically distributed, it appears that the Wilcoxon
test can be expected to perform well relative to alternative
candidate tests.
Table B-2 summarizes a comparison of the two multivariate
tests. The first test is the nonparametric analogue of
Hotelling's two-sample T test (Johnson and Wichern, 1982).
It uses an orthogonal rotation of the squares of the uni-
variate Wilcoxon statistics, which can be shown to have
(approximately) a chi-squared distribution. This test is
described in Hettmansperger (1984) and Puri and Sen (1971).
The second test is an ad hoc procedure that applies a uni-
variate Wilcoxon test to a set of ranks that are computed by
averaging ranks, computed within each LCIC, over all LCICs.
The Monte Carlo simulations were performed in a manner simi-
lar to that used for the univariate tests. Results in
Table B-2 include cases for which distributional shifts were
applied to only one, as well as five, of five chemicals.
The results show that although the observed alpha-level of
the average rank test is very far below the nominal alpha-
level, it performed nearly as well as the usual multivariate
Wilcoxon test in terms of power when all five chemicals were
shifted. This ad hoc test, however, did not perform as well
when only one chemical was shifted.
The general conclusions from the Monte Carlo simulations
were:
o The univariate Wilcoxon test performed well re-
gardless of the fraction of nondetectable values.
The Wilcoxon test procedure treats nondetectable
values as ties, therefore, no change in the formal
procedure is required to handle nondetectable
values.
o The best proportional allocation of samples to the
EDA and comparison area(s) generally has equal
sample sizes in each EDA neighborhood and compari-
son area (balanced design). However, modest devia-
tions from a balanced design do not greatly affect
the power of the tests with the best performance.
B-8
-------
Table B-2
SUMMARY OF SIMULATION RESULTS FOR THE MULTIVARIATE
RANK SUM AND AD HOC AVERAGE RANK TESTa
Sample
Size
10
10
20
20
30
30
50
50
Number of
Chemicals
Shifted
1
5
1
5
1
5
1
5
Nominal
g-level
.05
.05
.10
.10
.10
.10
.10
.10
Observed
a-level
Power
Avg
Rank
Mv
Rank
Avg
Rank
Mv
Rank
.002
.002
.006
.006
.002
.002
.002
.002
.066
.066
.102
.102
.098
.098
.106
.106
.014
.412
.028
.816
.024
.940
.064
.998
.134
.458
.302
.852
.380
.970
.592
1.000
Observations were distributed lognormally with mean 1 and a
coefficient of variation 1.414 under the null hypothesis;
power results are for an upper quantile shift (contaminated
lognormal distribution) with mixing fraction p = 0.3, and
composite mean 10 (order of magnitude shift) and coeffi-
cient of variation 2.0.
8856A/013/2
B-9
-------
o Using laboratory replicates to reduce the effect
of laboratory errors was unwarranted. The cost of
laboratory analyses was large relative to the sam-
ple collection cost, and resources would be better
allocated to increasing the overall sample size as
opposed to conducting multiple laboratory analyses
on a smaller sample set. This conclusion does not
apply to laboratory replication that was conducted
as part of the QA/QC program, which had different
objectives.
o The usual multivariate Wilcoxon test performed the
best for multivariate comparisons of the data.
B.3 CREATION OF THE SAMPLING DESIGN
Based on an assumed difference in median concentration be-
tween the EDA and comparison areas of an order of magnitude
(as specified in the habitability criteria), initial esti-
mates of sample sizes were required that would allow the
design to meet the habitability study objectives of 90 per-
cent power at 95 percent confidence. The appropriate sample
sizes were found from the Monte Carlo simulations to be
approximately 50 to 100 samples in each EDA neighborhood and
comparison area. However, it was recognized that some field
data would be required to finalize the sample size require-
ments and test analytical and field sampling methods. There-
fore, a pilot study was undertaken.
A pilot study was conducted in June 1986 to finalize esti-
mates of sample size requirements, test the analytical
methods in the laboratories, test field sampling techniques,
and develop the QA/QC procedures that would be used in the
full sampling study. During the pilot study, 45 samples
were collected from random locations throughout the EDA and
another 45 were collected in a composite comparison area.
The samples were assigned randomly to three laboratories.
All samples were split and the second half of the split was
assigned at random to one of the two remaining laboratories.
The most important results of the pilot study were:
o The EDA samples appeared to be well-fit by a log-
normal mixture distribution. Most samples had
concentrations near detection limits, but a few
relatively high values were found in the EDA.
o Analysis of the sample splits showed that there
were substantial differences between the labora-
tories and, in some cases, the interlaboratory
variability was larger than the natural (between-
sample) variability in the LCIC concentrations.
B-10
-------
Complete results of the pilot study are reported in the Soil
Pilot Study Report.
B.4 MODIFIED STATISTICAL ANALYSIS APPROACH
Because of the substantial between-laboratory variability,
the statistical design and analysis that was used in the
final study included blocking by laboratories. Heuristically,
this simply amounts to conducting the Wilcoxon test (univari-
ate or multivariate) for each set of observations analyzed
by one laboratory, and then computing an overall statistic
by adding the statistics associated with each laboratory.
Because the univariate Wilcoxon statistics have (asymptoti-
cally) a normal distribution, the sum of the Wilcoxon sta-
tistics is also asymptotically normal. Likewise, the
individual multivariate statistics for a single laboratory
are approximately Chi-squared, so the sum (over labora-
tories) also is distributed as Chi-squared. The next sec-
tion describes the model for the LCIC data that is modified
to incorporate blocking by laboratories. The section fol-
lowing that describes in detail the statistical methods that
were used to analyze the field study data.
B.4.1 The Statistical Model Modified for Blocking
The sampling design that was used for the Love Canal Habita-
bility Study can be modeled as a randomized complete block
design (Anderson and McLean) with laboratories classified as
blocks and sampling areas classified as treatments. The
comparisons between EDA sampling areas and comparison areas
can be made separately for each LCIC (a univariate approach),
or the information from all LCICs can be combined and used
in a multivariate approach. In either case, the general
layout of the data can be represented as:
Sampling Area (Treatment)
1
EDAl
2
EDA2
7
EDA7
8
CT221
9
CT225
10
C&T
Lab
(Block)
1
2
3
4
5
6
7
n
11
n
71
n
12
n
72
n
17
n
77
n
18
n
78
n
19
n
79
n
1,10
n
7,10
where n., represents the number of soil samples from sampling
area k iK = 1,...,10) that were analyzed by laboratory j
(j = 1,...,7). Note that for ease of notation, laboratories
B-ll
-------
have been arbitrarily labeled sequentially from 1 to 7.
This labeling system is totally distinct from the labeling
system used in the main body of this report.
A conceptual model for the LCIC measurements that includes
the effects of interlaboratory variability is:
Cjk£ = M + Lj + Ak + 6jk£ (B'2)
where:
C-,. = the concentration for the £'th observation in
-1 sampling area k that was analyzed by labora-
tory j
M = the overall mean concentration, over areas and
laboratories (FIXED)
L. = block effect because of laboratory j (FIXED)
A, = treatment effect because of sampling area k
(FIXED)
e., = random error in the I'th observation in sampling
-1 area k that was analyzed by laboratory j (RANDOM)
j = 1,...,7 (laboratory index)
k = 1,...,10 (sampling area index)
£ = l,...,n., (laboratory by area replicate index)
JK
7 10
E L . = E A. = 0
j=l D k=l k
Equation B. 2 is essentially the same as Eq. B.I, except that
the effect of laboratory is treated explicitly in Eq. B.2 as
a fixed effect, whereas in Eq. B.I it is included in the
overall error term e, .
A more complex model could be postulated that includes a
laboratory by sampling area interaction term. However,
there is no theoretical justification for such a term, and
visual inspection of boxplots of LCIC concentrations by sam-
pling area and laboratory gave no evidence that such an inter-
action term is necessary. Thus, interaction terms were not
used.
The e.. terms are assumed to be independent and identically
distributed (iid) random variables with mean zero under the
null hypothesis. These error terms reflect the last three
of the four sources of variability represented by the error
B-12
-------
terms in Eq. B.I; that is: (1) within-laboratory variabil-
ity, (2) between-sampling-site variability, and (3) vari-
ability associated with factors not explicitly modeled (for
example, data collection procedures, storage, transporta-
tion, etc.). The other source of variability accounted for
by the error terms in Eq. B.I, between-laboratory variabil-
ity, has been explicitly modelled in Eq. B.2. A multi-
variate extension to Eq. B.3 replaces C., with £'.:],£ where
X' denotes the transpose of a matrix X, or •*
'" "C8jk£) ' (B>3)
and
C.... = The concentration of the i'th LCIC for the
-' £' th observation in sampling area k that was
analyzed by laboratory j, with i = 1,...,8
(LCIC index)
Similarly, M, L. , A,, and e., are all replaced with their
multivariate analogues.
In terms of these models, the habitability criteria specify
a comparison of the magnitudes and directions of the effects
A,,...,A_ with the effects AR, Aq, and A,n yielding 7 x 3 = 21
pairwise EDA sampling area--comparison area comparisons. In
addition, three pairwise comparison area--comparison area
comparisons were performed. Thus, a total of 24 (pairs) x 8
(LCICs) = 192 univariate comparisons were performed, along
with 24 multivariate comparisons.
B.4.2 Statistical Methods Modified for Blocking
For univariate comparisons of data that can be described by
Eq. B.I, the Wilcoxon rank-sum test is appropriate to com-
pare one EDA sampling area with one comparison area. An
obvious heuristic extension to the Wilcoxon test that allows
for blocking by laboratory is to compute the Wilcoxon rank
sums for within-laboratory-between-sampling-area compari-
sons, and then to compute a grand statistic by summing over
laboratories. The variance of the summed statistic can be
computed in a straightforward manner by assuming that no
between-laboratory correlation exists.
The extension of the univariate Wilcoxon rank-sum test to
account for blocking is presented in various forms by Benard
and Van Elteren (1953) , Prentice (1979) , and Mack and Skill-
ings (1980). A confusing side issue is that these authors
refer to their tests as extensions of Friedman's test (Con-
over, 1980), rather than the Wilcoxon rank-sum test.
Table B-3 illustrates why. In the case of only one block
and two treatments, however, all three tests reduce to the
B-13
-------
Table B-3
RELATIONSHIPS BETWEEN SOME NONPARAMETRIC TESTS
Number
of
Blocks
Number of Treatments
2 K
Wilcoxon Rank-Sum Test
n , n general
Sign Test
j = 1,. . . ,J
Kruskall-Wallis Test
n ,...,n general
1 K
Friedman's Test
n = ... = n = l;j = 1,...,J
J1 JK-
Benard & Van Elferen (1953),
Prentice (1979),
Mack & Skillings (1982)
n ,...,n general
B-14
-------
Wilcoxon rank-sum test. Prentice's (1979) test will now be
described in the context of J blocks and two treatments.
In the following, it will be assumed that only two sampling
areas are being compared. Thus, n... will denote the sample
size (within laboratory j) in the EDA sampling area that is
being considered, while n.- will denote the sample size
(within laboratory j) in tine comparison area that is being
considered. A dot (.) substituted for a subscript denotes
the summation over that subscript. For example:
2
n- = I n-. The sample size within block
J* k=l -1 (laboratory) j over both treatments
sampling areas)
7 2
n = I I. n.. The total sample size over treat-
' ' j=l k=l ^ ments and laboratories
The blocked Wilcoxon test is presented in terms of a scaled
rank score r., /{n. +1), rather than the rank sum, where
JKJ6 J .
r., = The rank of C., , where ranks are assigned
11 from 1 to n..3within block (laboratory) j,
and n. denotes the number of observations
within block (laboratory) j.
It can be shown that under the null hypothesis of no differ-
ences between areas, the expected value of r-vo/(n- +D is
, -vo
1/2. The centered scaled rank is then defines
as
To test the null hypothesis of no difference in LCIC concen-
trations between sampling areas for a single LCIC, the fol-
lowing steps are performed:
1. Compute W. = X R . , for j = 1,...,7.
D £=1 m*
R . .. are the ranks for the data from the EDA sampling
area for the sample analyzed by laboratory j. W. is
the sum of the centered scaled ranks within laboratory j
for the EDA sampling area.
2. Compute V. = n., n.- /[12(n. +1)] for j
J J *• j ^ j •
V. is the variance of W. under the nul
no differences in LCIC concentrations
sampling area and the comparison area.
B-15
-------
7 7
3. Compute Z = W.//V. , where W. = Z W. and V. = Z V.
Under the null hypothesis of no difference in LCIC con-
centrations between the EDA sampling area and the com-
parison area, Z is approximately distributed as a
standard normal variable.
4. Compute the two-sided p-value as p = 2[1 - ( | Z | ) ] ,
where $ (y) denotes the cumulative distribution function
of the standard normal distribution and |z| denotes the
absolute value of Z. The essence of the blocked Wil-
coxon test is thus to compute the usual Wilcoxon sta-
tistic for each block (laboratory) and then to sum
these statistics across laboratories by weighting by
(block size + 1) to account for the difference in sam-
ple sizes between blocks. Because observations between
laboratories are independent, the variance of the sum
of these statistics is the sum of the variances of the
statistics for each laboratory.
Some variations of Prentice's (1979) extension have been
explored. Benard and Van Elteren (1953) base their test on
r., , while Mack and Skillings (1980) use r., /n. . Both
Prentice (1979) and Mack and Skillings show3 tfiat-1 "their ap-
proach is more efficient than that of Benard and Van Elteren
(1953). Prentice's approach, however, is the most computa-
tionally convenient. Some exploratory comparisons of the
three variations were conducted for Love Canal-like data and
the differences in performance of the three variations of
the test were found to be negligible. In addition, Prentice's
approach for both the univariate and multivariate case is
available on the statistical package SAS under the procedure
MRANK (Sarle, 1986) .
The blocked Wilcoxon test is easily adapted to the multi-
variate case. In essence, a multivariate generalization of
the Wilcoxon statistic (Puri and Sen, 1971, page 186; 1985,
page 185) is computed for each laboratory, along with an
estimate of the variance-covariance matrix of this statistic.
The multivariate Wilcoxon statistics are then summed over
labs, and a chi-squared statistic based on the summed (over
labs) variance-covariance matrices is formed.
In order to describe the multivariate test that was used to
compare sampling areas, it will be convenient to refer to
Eq. B.3 and to introduce the following notation:
-' jkt = (rljkfc' ' " 'r8jk£)
B-16
-------
r. ., . = The rank of C. ••^9, where ranks are assigned
1D from 1 to nj. Within block (laboratory) j;
ranking is performed separately for each of the
eight LCICs
-'jk£ = (Rljk£'- ""R8jk£)
-j = [R-jll'"" -jln^' £j21""f -J2n.2]
R. is simply the 8 x n.. matrix of centered scaled ranks for
block (laboratory) j. -1
To test the null hypothesis of no difference in LCIC concen-
trations between sampling areas for all of the LCICs, the
following steps are performed:
.J
1. Compute W. = E R-i. for j = 1,...,7.
W. is the sum of the centered scaled rank vectors
within laboratory j for the EDA sampling area.
2. Compute V. = (n.,n.,/n. )[(R1.R.)/(n. -1)] for j = 1,...,7
j J ••• jz J • j j J •
V. is the estimated variance-covariance matrix of W.
under the null hypothesis of no differences in LCIC-1
concentrations between the EDA sampling area and the
comparison area.
2-1 7 7
3. Compute X = W ' (V )W , where W = E W., and V = E V.
Under the null hypothesis of no difference in LCIC con-
centrations between sampling areas for any of the
LCICs, the statistic X is approximately distributed as
chi-squared with 8 degrees of freedom.
2
4. Compute the two-sided p-value as p = 1 - Fp(X ), where
(y) denotes the cumulative distribution function of
e chi-squared distribution with 8 degrees of freedom.
It should be noted that the samples sizes that were used in
the sampling design in order to achieve the specified power
requirements were based on the model of Eg. B.I, in which
between-laboratory variability is incorporated into the
B-17
-------
error term. Because between-laboratory variability is sub-
stantial, however, and because the above tests explicitly
account for blocking, the actual power of these tests should
have exceeded the 90 percent power requirement set by the
habitability criteria. This was indeed the case as reported
in the next section.
B.5 RETROSPECTIVE POWER OF THE SAMPLING DESIGN
One of the study goals was to provide 90 percent power of
detecting an order-of-magnitude difference between each EDA
sampling area and comparison area at 95 percent confidence.
Power is primarily a design issue to be addressed prior to
data collection. However, it is possible to perform a
retrospective analysis to determine what the power would
have been given an underlying probability distribution with
parameters estimated from the observed field data collected
in the comparison areas. (Distribution parameters estimated
from the EDAs are not appropriate because these areas are
presumed to (possibly) be characterized by a distribution
other than that corresponding to the null hypothesis.)
Monte Carlo simulations were conducted using the blocked
univariate and multivariate Wilcoxon procedures described in
the previous section, using coefficients of variation, non-
detect fractions, and (in the case of the multivariate tests)
inter-variable correlations estimated from the soil chemistry
data. The following assumptions were made:
1. The data are lognormally distributed.
2. The alternative hypothesis is a multiplicative (con-
stant coefficient of variation) shift.
3. For each chemical, the log-space standard deviation of
the "natural" process (not including interlaboratory
variability) was taken to be the average sample value
over all laboratories and neighborhoods (note that the
log space standard deviation is a function of the real
space coefficient of variation only, which is unaffected
by a multiplicative shift, if one exists).
4. The percentage of nondetects for each chemical was taken
as the average over the three control areas for all
laboratories.
5. The interlaboratory variance was taken to be 0.5 of the
total variance (natural variability plus interlabora-
tory variability).
The results are summarized in Tables B-4 through B-7. The
results show that for all but a few of the LCICs, the sta-
tistical test had the desired power (90 percent for alpha
= 0.05) for a delta of 1.0 (a 100 percent shift), and in all
B-18
-------
cases the power of the test was essentially 1.0 for an order-
of-magnitude shift. The results give confidence that the
design had at least the required sensitivity to concentration
differences between the EDA and comparison areas in all cases.
B-19
-------
Table B-4
SUMMARY STATISTICS USED IN THE MONTE CARLO SIMULATION FOR
THE RETROSPECTIVE POWER ANALYSIS3
LCIC C\T ND Vt XT
chlorobenzene (1) .45 0.0 .583 0.0
dichlorobenzene (2) .89 0.0 .974 0.0
trichlorobenzene (3) 1.02 7.5 1.06 .124
chloronaphthalene (4) .34 35.5 .456 .761
alpha-BHC (5) 3.25 51.2 1.76 .228
delta-BHC (6) .95 93.0 1.02 2.68
beta-BHC (7) 1.55 75.1 1.33 1.02
gamma-BHC (8) 1.14 84.9 1.13 1.69
Inter-laboratory variance assumed to be 50 percent of
total; essentially identical results were achieved when
interlaboratory variance was assumed to be 25 percent of
total.
Coefficient of variation derived from average (overlabora-
tories and neighborhoods) log standard deviation.
c
Estimated non-detect fraction (average over comparison
areas) in percent.
Log space standard deviation (square root of sum of "na-
tural" variance and inter-laboratory variance).
Q
Inferred non-detect level, as fraction of mean.
8856A/013/3
B-20
-------
Table B-5
POWER (AT a = 0.05) FOR UNIVARIATE TESTS BASED ON PARAMETERS
ESTIMATED FROM "GOOD" DATA GIVEN IN TABLE B-4,
FOR INTER-LABORATORY VARIANCE FRACTION =50%
Delta
0.00
0.05
0.10
0.20
0.50
1.00
2.00
5.00
10.0
LCIC
1
.050
.122
.244
.616
.988
1.000
1.000
2
.050
.088
.144
.286
.808
.990
1.000
•3
.050
.086
.130
.254
.756
.982
1.000
4
.044
.140
.322
.756
1.000
1.000
1.000
5
.048
.062
.096
.136
.338
.676
.958
6
.056
.076
.088
.124
.354
.740
.984
7
.058
.074
.104
.152
.412
.808
.994
8
.042
.070
.084
.156
.452
.832
.998
1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000
1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000
LCIC concentrations in EDA sampling areas are drawn from a
statistical population related to that of the comparison
area by the factor 1+delta; delta = 10.0 corresponds to an
order of magnitude shift.
8856A/013/4
B-21
-------
Table B-6
POWER FOR MULTIVARIATE TEST FOR PARAMETERS GIVEN IN
TABLE B-4 AND INTERVARIABLE CORRELATION GIVEN IN TABLE B-7,
FOR INTERLABORATORY VARIANCE FRACTION 0.5
Alpha
Delta3 .0100 .0250 .0500 .1000 .2000
.00 .000 .006 .022 .066 .176
.05 .000 .000 .022 .072 .192
.10 .000 .004 .034 .102 .216
.20 .014 .048 .096 .174 .340
.50 .256 .402 .528 .656 .808
1.00 .920 .968 .990 .994 .998
2.00 1.000 1.000 1.000 1.000 1.000
5.00 1.000 1.000 1.000 1.000 1.000
10.00 1.000 1.000 1.000 1.000 1.000
LCIC concentrations in EDA sampling areas are drawn from a
statistical population related to that of the comparison
area by the factor 1+delta; delta = 10.0 corresponds to an
order of magnitude shift.
8856A/013/5
B-22
-------
Table B-7
INTERVARIABLE CORRELATIONS USED IN SIMULATIONS
(BASED ON AVERAGE RANK CORRELATION FOR "GOOD" DATA)
LCIC
LCIC 1234567
1
2
3
4
5
6
7
8
8856A/013/6
1
.000
.591
.495
.333
.444
.410
.412
.344
.591
1.000
.957
.119
.750
.590
.591
.573
.495
.957
1.000
.095
.753
.610
.626
.601
.333
.119
.095
1.000
.160
.240
.080
.160
.444
.750
.753
.160
1.000
.670
.760
.687
.410
.590
.610
.240
.670
1.000
.670
.670
.412
.591
.626
.080
.760
.670
1.000
.600
.344
.573
.601
.160
.687
.670
.600
1.000
B-23
-------
REFERENCES
Anderson, V.L., and McLean, R.A. (1974) Design of Experi-
ments: A Realistic Approach. Marcel Dekker, New York.
418 pp.
Benard, A., and Van Elteren, P. (1953) "A Generalization
of the Method of m Rankings." Indagationes Mathematicae,
15, 358-369.
CH2M HILL. 1987. Pilot Study for the Love Canal EDA Habit-
ability Study. Volumes I and II.
Conover, W. J. (1980) Practical Nonparametric Statistics.
Second Edition. Wiley, New York. 493 pp.
Hettmansperger, T. P. 1984. Statistical Inference Based on
Ranks. Wiley, New York.
Hollander, M., and Wolfe, D.A. (1973) Nonparametric Sta-
tistical Methods. Wiley, New York. 503 pp.
Johnson, R.A., and Wichern, D.W. (1982) Applied Multi-
variate Statistical Analysis. Prentice-Hall, Englewood
Cliffs, New Jersey. 594 pp.
Lehmann, E. L. (1975) Nonparametrics: Statistical Methods
Based on Ranks. Holden-Day, San Francisco. 457 pp.
Life Systems. 1987. Peer Review of the Love Canal Full-
Scale Sampling Plan.
Mack, G. A. and Skillings, J. H. (1980) "A Friedman-Type
Rank Test for Main Effects in a Two-Factor ANOVA." Journal
of the American Statistical Association, 75, 947-951.
NYSDOH and DHHS/CDC. 1986. Love Canal Emergency Declara-
tion Area; Proposed Habitability Criteria.
Prentice, M. J. (1979) "On the Problem of m Complete Rank-
ings." Biometrika, 66, 167-170.
Puri, M. L., and Sen, P. K. (1971) Nonparametric Methods
in Multivariate Analysis. Wiley, New York. 440 pp.
(1985) Nonparametric Methods in General Linear
Models. Wiley, New York. 399 pp.
Sarle, W. S. (1986) "The MRANK Procedure." In Hastings,
R. P, ed. SUGI Supplemental Library User's Guide. SAS
Institute, Inc. Cary, North Carolina, pp. 361-375.
B-24
-------
U.S. EPA, Office of Research and Development. 1982. Envi-
ronmental Monitoring at Love Canal. Volumes 1, 2, and 3.
8856A/012
B-25
-------
APPENDIX C
Analytical Method Development
-------
Appendix C
ANALYTICAL METHOD DEVELOPMENT
INTRODUCTION
This appendix describes the development of the method used
for analysis of the field samples collected for the Love
Canal LCIC study. Further details regarding the analytical
procedure can be found in the sample laboratory analysis
QAPP (CH2M HILL, 1987). Several soil LCIC analytical tech-
niques were considered before the pilot study was conducted,
including gas chromatography with electron capture detection
(GC/ECD), gas chromatography mass spectrometry (GC/MS), and
isotope dilution GC/MS. Ultimately, however, selected ion
monitoring GC/MS (GC/MS/SIM) was chosen because it can pro-
vide suitable sensitivity and selectivity for LCIC analysis
(see CH2M HILL, 1987 Appendix F). GC/MS/SIM, which is used
routinely for several types of environmental analyses, moni-
tors only certain characteristic ions for each target com-
pound. With only a few masses monitored, the sensitivity
for each mass channel increases. Enough masses of each tar-
get compound must be monitored to confirm the detection of
that compound.
A method using GC/MS/SIM was developed and validated for the
analysis of soil samples for LCICs and was used successfully
for the pilot study. While LCIC concentrations were detected
and identified to 1 ppb using this method, LCIC concentrations
in the pilot study soil samples from the Emergency Declaration
Area (EDA) were mostly less than 1 part per billion (ppb)
and a large percentage were non-detects. For example, only
1 percent of the EDA samples had 2-chloronaphthalene concen-
trations greater than 1 ppb. Chemical interferences, rather
than simply low LCIC concentrations, also caused a signifi-
cant number of non-detects. This was especially true for
the BHC isomers, which exhibited hydrocarbon interferences.
MODIFICATIONS TO THE ANALYTICAL METHOD
On the basis of experience gained during the pilot study and
the NYSDOH study, a number of modifications were made to the
method prior to the habitability study to achieve detection
limits of around 0.2 to 0.3 ppb for most of the LCICs. Some
of these changes were oriented toward achieving lower detec-
tion limits and reducing the degree of chemical interferences,
particularly for the BHC isomers. Other modifications were
made to improve the productivity and effectiveness of the
labs contracted for the study.
In an effort to remove the hydrocarbon interferences, the
sample cleanup procedure was scrutinized, and major
C-l
-------
modifications were made. First, the alumina cleanup column
was replaced with a silica gel column, because silica gel
was found to be much more efficient than alumina in removing
these interferences. Second, the two sulfuric acid cleanup
steps were moved to before the silica gel column separation
step instead of after. This resulted in a cleaner extract
being applied to the silica column, thus improving the per-
formance of the column separation.
In addition, a bicarbonate wash step was added to follow the
sulfuric acid cleanup steps; the bicarbonate neutralizes any
acid that may be transferred along with the extract. The
sulfuric acid cleanup steps were found to cause fairly rapid
hydrogen/deuterium exchange for the deuterated internal stan-
dards, if any sulfuric acid remained in the extract. This
resulted in a severe high bias because of a reduction in the
concentration of the internal standards in the extracts.
The bicarbonate wash eliminated this problem.
Another modification to reduce the hydrocarbon interferences
involved changing the ions monitored for the four BHC iso-
mers: from m/z 181, 183, and 109 to m/z 217, 219, and 183.
The higher mass ions were found to be less prone to hydro-
carbon interferences. In addition, the ion monitored for
the low-level surrogate, 1,2,3,4-tetrabromobenzene, was
changed from m/z 392 to m/z 234 so that it would be more
subject to the types of interferences experienced by the
BHCs.
The changes to reduce the hydrocarbon interferences allowed
two additional BHC isomers to be added as LCICs: alpha- and
delta-BHC.
Other changes to the pilot study method were made to improve
the method detection limits. The volume of the final ex-
tract was reduced to 100 pi from 200 yl. Doubling the con-
centration of the LCICs in the extract effectively lowers
the detection limit by a factor of two if the limiting fac-
tor is the instrument's sensitivity. By using a more effi-
cient sample cleanup, the LCIC detection limit was lowered.
Another major improvement was to provide the laboratories
with solvents and reagents that were subjected to a rigorous
set of QC analyses prior to use to select lots in which any
LCICs present were at the lowest possible level. Since all
laboratories were using the same lots of solvents and re-
agents, a major source of blank contamination was eliminated
and any blank contamination was expected to be laboratory-
specific; the use of common lots also helped in determining
and eliminating remaining sources of contamination. Using
these materials, it was possible to cut in half the allow-
able blank contamination level from the 1.0 ppb used in the
pilot study to 0.50 ppb (0.60 ppb for 1,2-dichlorobenzene).
C-2
-------
In addition to using the same lots of solvents and reagents,
the laboratories were also provided with identical specialty
glassware, consisting of macro- and micro-Snyder columns and
silica gel chromatography columns. The identical glassware
provided increased comparability between laboratories.
A review of the pilot study data revealed interferences for
certain LCICs that made it difficult and, in some cases,
impossible to identify LCICs, thus resulting in false nega-
tives. By examining the method/holding blank results, it
was determined that the solvents and the reagents were causing
some of these interferences. To avoid this during the habit-
ability study, a detailed set of method/holding blank inter-
pretation rules was written into the sample analysis Quality
Assurance Project Plan (QAPP) (CH2M HILL, 1987). The rules
prohibit interference peaks that are close to the predicted
relative retention time (RRT) of an LCIC and are larger than
the peaks corresponding to the blank contamination level.
A detailed set of rules for identifying LCICs, similar to
the rules regarding blank interpretation, was added to the
sample analysis QAPP (CH2M HILL, 1987). These rules, de-
rived from the experience of the pilot study, were intended
to allow low-level LCIC data to be interpreted uniformly by
all project laboratories.
The levels of all calibration standards and quality control
(QC) compounds were adjusted downward from the levels used
in the pilot study. To adjust to the lower method detection
limits that could be achieved using the modified analytical
protocol, the initial calibration range was reduced from
1 through 200 ppb to 0.25 through 10 ppb, and the concen-
tration of the continuing calibration standard was reduced
from 2 ppb to 0.5 ppb. In addition, internal standard con-
centrations were reduced from 20 ppb to 10 ppb; surrogate
standard concentrations were lowered from 25 ppb and 1 ppb
to 10 ppb and 1 ppb; and matrix spike/matrix spike duplicate
levels were reduced from 50 ppb to 5 ppb.
Taken together, the changes described above lowered the
method detection limits by approximately a factor of five
from about 1 ppb to 0.2 ppb to 0.3 ppb for most of the LCICs.
This was as far as the detection limits could feasibly be
lowered using the GC/MS/SIM approach for the LCICs.
Other changes in the analytical protocol were made to
increase laboratory throughput and provide for greater com-
parability among the project laboratories. The performance
check standard and the continuing calibration standard were
combined into the same solution so that fewer QC injections
were needed for each analysis shift. The width of the RRT
window for each LCIC was reduced from ±0.007 to ±0.005 RRT
units, thus reducing the number of peaks that needed to be
C-3
-------
scrutinized in each chromatogram. Since the instruments
were operating in a mode that enabled sub-ppb levels of LCICs
to be detected, field samples that were suspected of having
high concentrations of LCICs were screened by diluting the
extracts prior to analysis. As a control, other field samples
not suspected of having high LCIC concentrations were also
preselected for screening. This was done to limit the maximum
concentrations to which the GC/MS/SIM would be subjected,
thus reducing the potential for instrument malfunction or
down time.
SUMMARY OF THE MODIFIED ANALYTICAL METHOD
The analytical method that was ultimately developed consisted
of solvent extraction, extensive cleanup, and analysis by
GC/MS/SIM. After a sample was received and logged into the
laboratory, a 20-gram aliquot was weighed out and spiked
with a solution containing the three surrogate standards.
The soil was then dried to a sandy texture by mixing it with
sodium sulfate. The sample was extracted three times for
3 minutes using a mixture of methylene chloride and acetone.
A sonicator probe was used to increase the efficiency with
which the LCICs were extracted from the soil matrix. The
combined extracts were centrifuged, if necessary, to remove
soil particles. The extract was dried further by passing it
through a sodium sulfate column. The dried extract was re-
duced in volume by evaporating the solvent using a Kuderna-
Danish column with a three-ball macro-Snyder column, followed
by further concentration using a two-ball micro-Snyder column,
A second two-ball micro-Snyder column was used to exchange
the extract sample into a hexane solvent.
In the first of the two cleanup steps, concentrated sulfuric
acid was added to the extract, causing a reaction in which
many of the polar compounds that could interfere with the
LCIC analysis were removed. The extract was then washed
with a bicarbonate solution to neutralize any residual acid.
In the second cleanup step, the extract was passed through a
silica gel column. The composition of the solvent used to
extract the silica gel column was carefully selected during
method development to yield high recovery of the LCICs while
separating them from hydrocarbon interferences. Following
silica gel column chromatography, the extract was reduced in
volume using macro- and then micro-Snyder columns. This
solution was stored in a refrigerator at 4°C until the time
of analysis.
Just before analysis, a solution containing the five deuter-
ated internal standards was added to the extract. Using a
gentle stream of purified nitrogen gas, the extract window
was concentrated to a final volume of 100 yl. A 1- to 2-yl
aliquot of the final extract was injected into the GC using
C-4
-------
a splitless injector. The LCICs were separated using a fused
silica column coated with polyphenylmethyl silicone. Follow-
ing the temperature-programmed GC separation, the LCICs were
detected using an electron impact MS operated in the SIM
mode. Three characteristic ions were monitored for each
LCIC. The instrument response for these ions was required
to satisfy the following three criteria to confirm the iden-
tity of the LCIC: (1) all three ions must exhibit a peak
within 0.005 RRT unit of the daily performance check; (2)
all three ions must exhibit peak maxima within two mass spec-
tral scans of each other; and (3) the primary and secondary
ions must have a response ratio within 20 percent of the
theoretical ratio. For confirmed peaks, quantification was
performed using the internal standard that eluted closest in
the chromatogram to the LCIC. Further details on the anal-
ytical procedure can be found in the sample analysis QAPP
(CH2M HILL, 1987) .
REFERENCES
CH2M HILL, 1987. Love Canal Habitability Study—Soil Sample
Laboratory Analysis Quality Assurance Project Plan. Pilot
Study Report Vol. I and II.
8854/035
C-5
-------
APPENDIX D
Holding Time Study
-------
Appendix D
HOLDING TIME STUDY
INTRODUCTION
Prior to the start of field sampling for the soil assessment
study, a study was initiated to assess the effects of sample
extraction holding time on LCIC concentration. For adminis-
trative purposes, the EPA Contract Laboratory Program typi-
cally specifies an extraction holding time of not more than
10 days. During the design of the soil assessment study, it
was decided that the maximum allowable extraction holding
time should be increased to 30 days to allow the field sam-
pling to proceed on schedule. This decision allowed the
field sampling to finish prior to adverse winter conditions,
yet gave the laboratories adequate time to extract and ana-
lyze all of the field samples.
The main purpose of the holding time study was to determine
whether LCIC concentrations decrease over time in a soil
sample that is held in refrigerated storage in the labora-
tory before the sample is extracted. The effects of time
from extraction to analysis were not specifically investi-
gated. The holding time study showed that there was no
significant decrease in LCIC concentration over the studied
extraction time.
The field sampling and laboratory methods that were used in
the holding time study, and details of the results, are de-
scribed in this appendix.
METHODS
The holding time study involved collecting 20 soil samples
from EDA Neighborhood 1 and 20 soil samples from Neighbor-
hood 3. Neighborhood 1 was chosen because of indications
during the pilot study that concentrations of LCICs were
elevated compared to other EDA neighborhoods. Neighbor-
hood 3 was taken to be more typical of the remaining EDA
neighborhoods.
The samples from each neighborhood were sent to the soil
preparation laboratory where they were extruded from the
stainless steel liners. All 20 samples from a single neigh-
borhood were mixed for 5 minutes, resulting in two homo-
geneous batches: one from EDA Neighborhood 1 and one from
EDA Neighborhood 3. Each batch of homogenized soil was then
split into 20 sample jars. Ten of the jars from each neigh-
borhood, along with one shipping blank jar containing blank
soil, were then sent to each of two laboratories (1 and 2)
for analysis.
D-l
-------
All of the holding time samples were shipped from the soil
preparation laboratories to the two analytical laboratories
on the same day. Upon receipt of the holding time samples,
the analytical laboratories placed the jars in refrigerated
storage. The extraction holding time for a particular sam-
ple was defined to be the number of days between the day the
analytical laboratory received the sample and the day the
sample was extracted.
On the day both laboratories received the holding time sam-
ples, each laboratory selected two jars at random from each
neighborhood batch and analyzed the soil. The analysis for
each jar was performed in duplicate, i.e., two separate sub-
samples were taken from a single jar, extracted, and ana-
lyzed. These 16 analyses (2 laboratories x 2 neighborhoods
x 2 jars/neighborhood x 2 subsamples/jar) established the
LCIC concentrations on day zero.
After the initial day zero analyses, each laboratory was
instructed to extract and analyze one sample (jar) per
neighborhood (in duplicate) every fifth day. During the
next three analysis dates, samples from both neighborhoods
were to be analyzed. From then on, starting with the fifth
analysis date, samples for a specific neighborhood were ana-
lyzed every TO days. Table D-l shows the planned holding
times for this study. The longest planned extraction hold-
ing time for the holding time study exceeded the expected
holding times associated with the full field sampling study.
Both laboratories analyzed the holding time experiment sam-
ples using the same analytical procedures that were used for
the field sample analyses.
Table D-l
PLANNED HOLDING TIMES FOR THE HOLDING TIME STUDY
Receipt Extraction Neighborhood Holding Time
Date Date Soils Analyzed (days)
10/14/87 10/14/87 1,3 0
10/14/87 10/19/87 1,3 5
10/14/87 10/24/87 1,3 10
10/14/87 10/29/87 1,3 15
10/14/87 11/03/87 3 20
10/14/87 11/08/87 1 25
10/14/87 11/13/87 3 30
10/14/87 11/18/87 1 35
10/14/87 11/23/87 3 40
10/14/87 11/28/87 1 45
10/14/87 12/03/87 3 50
10/14/87 12/08/87 1 55
10/14/87 12/13/87 3 60
10/14/87 12/18/87 1 65
D-2
-------
RESULTS OF THE HOLDING TIME STUDY
Table D-2 shows the actual holding times and extraction
dates that were used in the holding time study along with
the observed concentrations for each LCIC. Note that the
samples were received by the analytical laboratories I day
early, on October 13, 1988, rather than October 14, 1988.
To investigate the effect, if any, of holding time on LCIC
concentration, a regression of LCIC concentration versus
holding time was performed for each LCIC, each neighborhood,
and each laboratory. Only observations that were flagged as
"Good" by EMSL-LV/LEMSCO were used for the regressions. The
number of observations per regression varied from a maximum
possible of 20 to a minimum of 11. The average number of
observations per regression was 16.
In cases where concentrations were reported as non-detect
(because not all of the qualitative identification criteria
were met), the concentration based on the primary ion area
was used. This was considered to be the best estimator for
non-detectable concentrations because the primary ion area
is the basis for quantification when the analytes do in fact
pass all of the qualitative identification criteria. In
addition, for a specific neighborhood, the holding time sam-
ples were from only one batch of soil that was relatively
homogeneous. Hence, if an LCIC was detected in an earlier
sample taken from a different jar, it was probably present
in the current sample as well, but it was registering as a
non-detect because of some type of interference (perhaps
from the jar or from the extraction process for that
particular day). Note that such multiple measurements on
the same soil were not available for the field samples in
the full study. Ignoring non-detects or treating them as if
they reflected a concentration of zero would have seriously
biased the regression results.
To quantify the relationship between LCIC concentration and
extraction holding time, a least-squares regression was per-
formed along with two tests for trend: the usual parametric
test for a non-zero slope coefficient (Zar, 1984, page 271),
and the nonparametric test for trend based on Kendall's tau
(Conover, 1980, page 256). Because the data contained sev-
eral anomalous observations, the more robust nonparametric
tests were used. Table D-3 summarizes the results of the
nonparametric tests for trend. In every case where a sig-
nificant trend was found for one neighborhood and labora-
tory, it was not significant for the other laboratory.
The results of the holding time study indicate that extend-
ing the extraction holding time from 10 days to 30 days did
not have a significant effect on the concentrations of LCICs
during the full field sample study. It should be noted,
D-3
-------
Table D-2
ACTUAL HOLDING TIMES AND OBSERVED CONCENTRATIONS
EDA NEIGHBORHOOD 1
Extraction
Lab Date
1 10/14/87
10/14/87
10/14/87
10/19/87
10/19/87
10/24/87
10/24/87
10/29/87
11/08/87
11/08/87
11/18/87
11/18/87
11/28/87
11/28/87
12/08/87
12/08/87
12/18/87
12/18/87
2 10/14/87
10/14/87
10/14/87
10/14/87
10/19/87
10/19/87
10/24/87
10/24/87
10/29/87
10/29/87
11/18/87
11/18/87
11/18/87
11/18/87
12/08/87
12/08/87
12/13/87
12/13/87
12/18/87
12/18/87
Extraction
Holding
Time
1
1
1
6
6
11
11
16
26
26
36
36
46
46
56
56
66
66
1
1
1
1
6
6
11
11
16
16
36
36
36
36
56
56
61
61
66
66
DCB
Cone
1.30
1.60
1.70
1.74
1.98
1.50
1.87
1.97
1.59
1.21
0.82
0.58
1.42
2.17
1.06
1.21
1.38
1.42
1.27
2.55
1.37
1.44
0.97
1.34
0.94
1.10
0.77
1.20
1.26
2.92
1.26
0.72
0.92
0.93
1.31
0.88
0.92
1.22
DCB
Flag
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
B
B
U
U
TCB
Cone
7.70
10.04
10.34
11.68
12.30
9.51
9.79
9.18
8.66
7.94
6.08
4.05
-7.77
10.08
6.90
11.77
8.69
10.22
10.94
18.64
10.63
11.90
7.40
10.14
7.97
8.41
5.75
8.19
7.86
13.23
6.47
5.59
6.10
6.70
10.18
5.52
4.96
6.53
TCB
Flag
G
G
G
G
G
G
G
G
G
G
G
G
G
G
U
U
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
B
B
U
U
OCB
Cone
8.43
11.29
11.35
12.49
13.15
11.39
10.60
9.21
9.57
10.68
8.61
12.47
8.85
10.21
8.60
13.50
11.02
11.39
13.07
17.86
12.21
12.62
10.16
11.65
10.07
11.10
11.59
11.06
13.49
14.18
8.39
8.18
9.77
9.64
14.77
8.28
6.13
44.80
QCB
Flag
G
G
G
G
G
G
G
G
G
G
G
G
U
U
U
U
G
G
U
U
U
G
U
G
G
G
G
G
G
G
G
G
G
G
B
B
U
U
CNP (
Cone 1
0.08
0.09
0.09
0.09
0.10
0.13
0.13
0.13
0.17
0.19 *
0.10
0.08
0.15
0.12
0.09
0.10 *
0.14
0.17
0.10
0.14
0.11 *
0.09
0.08
0.20 *
0.08 *
0.09 *
0.04 *
0.05 *
0.06 *
0.06 *
0.08 *
0.08 *
0.04 *
0.05
0.04 *
0.08 *
0.06
0.07 *
:NP
:lag
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
U
U
U
U
U
B
U
G
G
G
G
G
G
G
G
G
G
G
B
B
U
U
BHCA
Cone
6.76
8.93
8.07
13.32
13.53
9.77
9.79
8.85
9.29
9.90
7.81
6.60
10.68
10.04
6.83
14.11
10.02
10.58
11.93
15.99
11.15
12.02
12.26
13.44
13.98
12.39
10.56
10.02
8.40
11.99
8.73
11.53
11.86
11.03
15.55
6.89
6.11
13.28
BHCA
Flag
G
G
G
G
G
G
G
G
G
G
G
G
G
G
U
U
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
B
B
U
U
BHCD E
Cone f
0.49 *
1.42
0.94
1.78
1.67
1.16 *
1.51
1.66
1.33
1.23
0.92
1.13
1.59
1.25
0.86
1.50
1.20
1.18
1.45
2.21
1.63
2.16
1.37
1.68 *
1.49
1.60
1.84
1.67 *
0.88
1.39
0.80
1.27
1.06
1.36
2.40
0.93
0.94 *
0.64 *
IHCD
:lag
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
B
G
G
G
G
G
G
B
B
U
U
BHC8 E
Cone f
1.14 *
17.49
15.16
23.19
22.55
1.89
1.87
13.92
17.69
18.30
11.39
11.06
20.94
18.64
13.56
29.10
12.18
1.83 *
17.75
29.40
19.21
16.50
23.10 *
1.09 *
15.91
13.44
19.72
21.91
12.60
21.42
19.21
12.15
18.48
19.19
37.35
17.99
7.59 *
16.66
IHCB
:lag
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
B
G
G
G
G
G
B
G
G
G
G
U
U
G
U
G
G
B
B
U
U
BHCG B
Cone F
10.58
1.78
1.63
2.48
2.69
14.18
15.72
1.74
1.93
1.85
1.65
1.84
1.83
2.05
1.34
2.65
1.65
12.80
2.04
2.43
1.63
2.33
2.27
1.65 *
2.73 *
2.52 *
2.18 *
2.17 *
1.71
2.46
1.72
2.82
2.28
2.50
3.45
1.62 *
1.08 *
2.05
HCG
lag
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
B
G
G
G
G
G
G
G
G
B
G
G
G
G
G
G
G
B
B
U
U
-------
Table D-2
(cont inued)
EDA NEIGHBORHOOD 2
Extraction
Lab Date
1 10/14/87
10/14/87
10/H/87
10/14/87
10/19/87
10/19/87
10/24/87
10/24/87
10/29/87
10/29/87
11/03/87
11/03/87
11/13/87
11/13/87
11/28/87
11/28/87
12/03/87
12/03/87
12/08/87
12/08/87
2
10/14/87
10/14/87
10/19/87
10/19/87
10/19/87
10/24/87
10/24/87
10/29/87
10/29/87
11/03/87
11/03/87
11/13/87
11/13/87
12/03/87
12/03/87
12/13/87
Extraction
Holding
Time
1
1
1
1
6
6
11
11
16
16
21
21
31
31
46
46
51
51
56
56
,
1
1
6
6
6
11
11
16
16
21
21
31
31
51
51
61
DCB
Cone
0.38
0.43
0.42
0.36
0.51
0.48
0.42
0.40
0.56
0.49
0.44
0.46
0.52
0.50
0.61
0.59
0.44
0.51
0.41
0.43
0.31
0.37
0.41
0.37
0.33
0.39
0.30
0.37
0.51
0.47
0.59
0.39
0.26
0.57
0.55
0.46
0.52
0.52
DCB
Flag
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
U
U
G
G
G
G
G
G
G
G
G
G
G
G
G
U
U
G
G
B
TCB
Cone
0.86
0.85
0.72
0.71
0.97
0.99
0.65
0.73
0.78
0.74
0.66
0.71
0.68
0.75
0.79
0.72
0.70
0.77
0.64
0.65
0.78
0.92
1.16
1.00
0.53
1.08
0.60
0.98
0.94
0.79
0.90
0.95
0.64
0.75
0.70
0.91
0.83
0.78
TCB
Flag
G
G
G
G
G
G
G
G
G
G
G
G
G
G
U
U
G
G
U
U
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
B
QCB
Cone
1.02
0.85
0.67
0.67
0.98
0.95
0.66
0.77
0.84
0.76
0.83
0.96
1.08
0.89
0.77
0.72
0.74
0.84
0.71
0.89
0.84
0.98
1.02
0.95
0.50
0.99
0.65
0.98
.08
.02
.09
.04
.09
.04
0.72
.00
.02
0.83
OCB
Flag
G
G
G
G
G
G
G
G
G
G
G
G
G
G
U
U
G
G
U
U
G
G
G
G
G
U
U
G
G
G
G
G
G
G
G
G
G
B
CNP CNP
Cone Flag
0.06
0.06
0.06
0.07
0.07
0.05
0.10
0.10
0.10
0.10
0.10
0.10
0.13
0.13
0.10
0.10
0.09
0.09
0.11
0.11
0.06
0.07
0.06
0.05
0.05
0.03
0.05
0.05
0.08
0.04 *
0.04 *
0.06 *
0.05 *
0.10
0.06
0.05 *
0.05 *
0.03 *
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
U
U
G
G
G
G
G
G
G
G
G
G
B
BHCA BHCA
Cone F I ag
0.55
0.63
0.45
0.38
0.50
0.40
0.46
0.41
0.46
0.49
0.35
0.47
0.31
0.36
0.56
0.53
0.29
0.43
0.32
0.31
0.45
0.45
0.43
0.43
0.27
0.43
0.30
0.50
0.46
0.53
0.44 *
0.45
0.34
0.31
0.46
0.52
0.50
0.42
G
G
G
G
G
G
G
G
G
G
G
G
G
G
U
U
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
B
BHCD BHCD
Cone F I ag
0.15 *
0.60
0.13
0.14 *
0.09 *
0.09 *
0.14 *
0.09 *
0.07
0.19
0.07 *
0.15 *
0.11
0.12 *
0.15 *
0.14
0.11
0.11 *
0.09 *
0.12
0.08 *
0.10 *
0.12 *
0.16 *
0.09 *
0.20 *
0.08 *
0.33 *
0.18 *
0.06 *
0.05 *
0.10 *
0.09 *
0.10 *
0.21 *
0.09 *
0.19 *
0.08 *
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
U
U
G
G
B
BHCB BHCB
Cone f 1 ag
0.51 *
0.51 *
0.42 *
0.37 *
0.53
0.47
0.42
0.36
0.37
0.38
0.30
0.35
0.31
0.31
0.38
0.46
0.27
0.35
0.30
0.19
0.29
0.44
0.32
0.44
0.20
0.32
0.24
0.49
0.34 *
0.67
0.35 *
0.38
0.34
0.31
0.31
0.46
0.48
0.52 *
G
G
G
G
G
G
G
G
G
G
G
G
G
G
U
U
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
B
BHCG BHCG
Cone F 1 ag
0.36 *
0.23 *
0.19 *
0.17 *
0.21 *
0.13 *
0.18 *
0.16
0.09
0.09
0.15
0.15
0.11
0.09
0.11
0.17 *
0.08
0.11
0.09
0.09
0.09 *
0.06 *
0.14
0.11
0.03 *
0.06 *
0.07 *
0.34 *
0.12
0.13 *
0.13 *
0.10
0.03 *
0.07 *
0.22 *
0.19
0.16 *
0.24 *
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
B
-------
Table D-2
(continued)
12/13/87 61 0.58 B 0.88 B 0.95 B 0.06 B 0.36 B 0.19 * B 0.46 B 0.09 * B
* denotes that this concentration was reported as a non-detect. The primary ion concentration is shown.
. denotes a missing value.
-------
Table D-3
HOLDING TIME STUDY TESTS FOR TREND BASED ON KENDALL'S TAU
Result of
LCIC
1 , 2-Dichlorobenzene
1,2, 4-Trichlorobenzene
2-Chloronaphthalene
1,2,3, 4-Tetrachlorobenzene
Alpha-BHC
Beta-BHC
Neighborhood
1
3
1
3
1
3
1
3
1
3
1
3
Laboratory
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
Trend Test
0
-
+
0
0
—
0
0
0
0
+
0
0
0
0
0
0
0
-
0
0
0
_
0
N
18
16
18
13
16
16
16
15
16
11
20
13
14
12
16
13
16
16
18
15
17
12
18
15
D-7
-------
Table D-3
(continued)
Result of
LCIC Neighborhood Laboratory Trend Test _N
Gamma-BHC 1 1 0 17
2 0 15
3 1-20
2 0 15
Delta-BHC 1 1 0 18
2 - 15
3 1 0 20
2 0 13
+ = a positive trend was indicated at p < 0.05 (two-sided)
- = a negative trend was indicated at p < 0.05 (two-sided)
0 = no trend at p < .05 (two-sided)
D-8
-------
however, that holding time effects, even if they had been
present, would not have affected the results of the EDA
sampling area-comparison area comparisons because the se-
quence of sample analysis was randomized by the requirement
that the laboratories analyze the samples in the order that
they were received.
REFERENCES
Conover, W. J. 1980. Practical Nonparametric Statistics,
John Wiley & Sons, New York, New York.
Zar, Jerold H. 1984. Biostatistical Analysis. Prentice
Hall Inc., Englewood, Cliffs, New Jersey.
8855/001
D-9
-------
APPENDIX E
Data and Information Systems
Development
-------
Appendix E
DATA AND INFORMATION SYSTEMS DEVELOPMENT
The soil assessment for indicator chemicals required that
rigid QA/QC standards be observed within a demanding
sampling schedule. In addition, the data to be collected
for project activities were complex and voluminous. To meet
the QA/QC and data handling requirements, it was necessary
to provide automated support to the project where possible.
Development of the data and information systems built upon
the experience and software developed during the pilot study
and the soil assessment for dioxin. In addition, a func-
tional requirements analysis provided a software require-
ments summary and a data flow diagram (see Figure E-l). The
systems needed to support the study were designed using the
existing soil pilot study software, the dioxin sample tracking
software, and the software requirements summary developed
for this study.
Ten separate software systems were used, eight were imple-
mented on personal computers and two were implemented on the
EPA mainframe system. East of use and the editing capabil-
ities to identify invalid data and prevent illogical func-
tions were high priorities in the development of virtually
all of the systems. A system flow chart depicting nine of
these systems is presented in Figure E-2. The tenth system
(not shown) was a model data base implemented on the
mainframe. Each system is described in Table E-l.
The final integrated data base is maintained as a Statis-
tical Analysis System (SAS) relational data base. SAS was
chosen because it can handle large quantities of data because
it provides extensive data manipulation, merging, reporting,
and statistical analysis capabilities. The data base con-
sists of six files, or tables, each containing the key
fields required to merge or join each table to any related
table. The tables are as follows:
1. Initial Calibrations—Contains one entry for each
five-point initial calibration and EPA performance
check run on the GC/MS/SIM; and includes all quanti-
tation reports and computed data for an initial
calibration.
2. Continuing Calibrations—Contains one entry for each
continuing calibration/performance check analysis and
the mid-analytical run continuing calibration/performance
check analysis for each analytical run and includes all
quantitation reports and computed data for the
calibrations.
se8854/039/l E-l
-------
W63394.T1(SA)
Stainless
Steel Unas
Sample Tracking Data
Quantitafon Results/Performance Check Chromatograms
Perform
On-line
QAQC
QCed Sample Tracking Data Base
QCed Electronic Final Data Package
Generate
Integrated
Data Base
Validated Bectronb Final Data Package
Integrated Data Base
D
Process
Data Flow Between Processes
"Data Stores (Collection of Data Flow)
Source lor Data Flow or Outflow ol Data
/ Conduct
( Statistical
V Analysis
Figure E-1
SOIL ASSESSMENT--
INDICATOR CHEMICALS
INFORMATION ENVIRONMENT
DATA FLOW DIAGRAM
-------
W63394.T1/S.A
PERSONAL COMPUTER SYSTEMS
Sample Collection
System
Sample Collection
Forms
Traffic Reports
Labels
Legend
| EtectronicSystem
Hard Copy
Q Diskette Containing
El
Electronic Into.
Integrated
Electronic Data
Disc
Sample Preparation
System
Sample Preparation
Forms
Traffic Control
Reports
>r
Q
Laboratory
Analytical
Results
Bulletin Board
System
V
Data Validation
System
Validation
Checklists
Real-time
QA/QC System
Audit Shift Lists
init Cai Checklists
Shift Checklists
Sample Checklists
CWQC Results
Control Charts
Blind QC Results
Site Selection
System
Q
Target
Sample Sites
• QAQC Reports
• Verification
Reports
~r >~
, ., ^
Data Base
Integration/
Verification System
"^\
• — -'
Integrated
Data Base
Includes:
1 .Sample Collection Data
2.Sample Preparation Data
3. Laboratory Analytical Data
• Field Sample Results
• Calibration (Initial and Continuing) Results
• QC Sample Results Including
Blanks MS/MSD and Blind QC Samples
4.0ata Validation Results
MAINFRAME SYSTEMS
Figure E-2
SOIL ASSESSMENT-INDICATOR CHEMICALS
INFORMATION MANAGEMENT SYSTEMS
-------
Table E-l
SOFTWARE SYSTEMS DEVELOPED AND USED IN THE
SOIL ASSESSMENT FOR INDICATOR CHEMICALS
System Name
Site Selection
Sample Collection
Description
Sample Preparation
LabData
Online QA/QC
Data Validation
Generates random selection of sample
locations on EDA grid and comparison
area grid maps
Generates forms and labels for field
sampling; collects completed forms
(via double-data entry and comparison);
verifies data; and produces reports on
tracking/collection status
Generates and tracks sample prepara-
tion forms; automatically generates
the project sample identification
number and indicates sample splits and
target matrix spike/matrix spike
duplicate (MS/MSD) samples; edits and
verifies data entered on forms; and
produces labels and reports on
preparation data status
Generates sample analysis data re-
porting forms; performs chromatogram
performance check analyses; provides
automatic interface with project On-
line QA/QC System; generates the
electronic final data package; and
provides automatic interface with the
project Data Validation System
Produces QC reports on samples being
analyzed in the project laboratories
on a real-time basis; and produces
daily control charts for selected QC
data
Interfaces with LabData and auto-
matically displays analytical results
to be reviewed, including shift files
and calibration; provides data entry
for validation checklists; and
provides error and status reports
se8854/040/l
E-4
-------
Table E-l
(continued)
System Name Description
Form I Data Entry Collects and verifies the Form I's
(data reporting form containing LCIC
concentrations) validated by EMSL-LV,
including the data validation and
usability flags assigned; and gener-
ates, a file for uploading to the
mainframe data base
Integrated Model Provides a model SAS data base with
Data Base selected fields filled with data using
distributions from the pilot study
Integrated Data Integrates the data generated from the
Base LabData, Sample Collection, Sample
Preparation, Data Validation, and
Form I Data Entry systems into a
mainframe SAS data base; recalculates
all data reported on the LabData
forms; checks for data consistency
between the systems where the same
data were available from more than one
source; and reports discrepancies or
omissions
Project Bulletin Provides an interface between project
Board systems and project staff
se8854/040/2 E-5
-------
3. Field Samples—Contains one entry for each field sample
and contains all sample collection, preparation, and
analysis data for the sample. The analysis data include
all quantitation report and computed data for the sample
analysis. Pointers, or keys, join each sample's initial
calibration entry, continuing calibration entry,
method/holding blank entry, MS and MSD entries (if any),
Form I entry, and blind QC entry.
4. QC Samples—Contains one entry for each QC sample; con-
tains all sample collection, preparation, and analysis
data for the sample. The analysis data include all
quantitation report and computed data for the sample
analysis.
5. Form I—Contains one entry for each Form I (the data
reporting form containing the LCIC concentrations)
validated by EMSL-LV. Each entry contains all Form I
data plus extract data and all EMSL-LV data validation
and useability flags.
6. Blind QC Spike Levels—Contains one entry for each
blind QC sample sent to the analytical laboratories by
EMSL-LV. Each entry includes the spiking levels for
each analyte spiked into the sample.
The data base included approximately 1.8 million data cells
(the number of fields per table times the number of occur-
rences) . Virtually every piece of data generated during the
project is included.
se8854/039/2 E-6
-------
APPENDIX F
QA/QC Program
-------
Appendix F
QA/QC PROGRAM
INTRODUCTION
The QA/QC measures for this study consisted of a detailed
and extensive system of quality control for samples, proce-
dures, and documentation designed to demonstrate that a high
level of quality had been maintained during every aspect of
the project. Operationally, the study implementation was
divided into four phases: field sampling, laboratory analy-
sis, information management, and statistics. Associated
with each of these phases were two types of QA/QC measures:
quantitative and qualitative. Quantitative QA/QC measures
consisted of specific QC samples and activities designed to
demonstrate that the QA/QC objectives were being met. Qual-
itative QA/QC measures consisted of sufficient documentation
to demonstrate that the qualifications of participants were
acceptable, that all procedures were being followed, and
that the integrity of samples and information was being
maintained.
Figure F-l shows the flow of activities for the four phases
of the project and the QA/QC responsibilities associated
with each study activity. Primary QA/QC responsibility
rested with the CH2M HILL project team. The TRC agencies
provided review and guidance throughout the project. The
EPA Environmental Monitoring Systems Laboratory at Las Vegas
(EMSL/LV) performed an active QA/QC oversight role.
For the laboratory phase of the project, the subcontractor
laboratories assumed major QA/QC responsibility, with over-
sight from both the CH2M HILL team and the EMSL/LV.
FIELD QUALITY ASSURANCE/QUALITY CONTROL PROGRAM
The field QA/QC program involved specific entities for which
acceptability could be assessed quantitatively against a QC
criterion, as well as more qualitative documentation and
procedures designed to show that the data obtained in the
study corresponded correctly to the samples obtained in the
field. Quantitative and qualitative QA/QC are discussed
separately below. Figure F-l shows the flow of activities
in the field and the associated quantitative and qualitative
QA/QC measures.
QUANTITATIVE FIELD QA/QC
A major element in the quantitative field QA/QC procedures
was to interject blank (theoretically zero concentration)
samples at various stages in the sample collection and
F-l
-------
W63394.T1(SA)
PHASE ACTIVITY SUBCONTRACT LABS
i
i
i
\
i
i
t
>
,
o
£3
Q
r
i
i
_i
o.
C/5
Q
CJ
rz
CO
LABANALYS
NFORMATION
MANAGEMENT
STATISTICS
Sampling Plan
)
F
Sample Locations
1
Sample Collection
i
Extrusion/Mixing
)
QAOC RESPONSIBILITY
CH2M HILL TEAM EPA/EMSL-LV
>
Extraction/Cleanup
i
r
GC/MS/SIM Analysis
^
r
LabData Processing
1
Data Package
Electronic Disks
i
r
Personal Computer
Data Base
I
Statistical Analysis
System Data Base
i
r
Statistical
Comparisons
Documentation
Checks
QC Samples
QAQC Measures
EPA Check Standard
Licensed Survey
Decontamination
Procedure
Blank QC Samples
Informal Checklist
Documentation
Checks
QC Samples
Documentation
Checks
Provide Standard
Solvents, Reagents,
and Glassware
Informal Audits
Audit
— i
— Audit
Real-Time QC
Informal Audits
Review of Blank
QC Data
Review
Review and Apply
Data Qualifier Flags
Functional Checks
Final Verification of
Data Linkage
Independent
Verification
Figure F-1
QA/QC ACTIVITIES AND
RESPONSIBILITIES RELATIVE TO
PROJECT PHASES
-------
laboratory analysis sequence. The analysis of blank samples
provided a measure of contamination sources, decontamination
efficiency, and other potential biases that might have been
introduced from sources other than the sample. Blanks for
soil samples consisted of soil that approximated the consis-
tency of Emergency Declaration Area (EDA) soil and had been
prepared and analyzed for the presence of Love Canal Indica-
tor Chemicals (LCICs). (See CH2M HILL, 1987a for more
details on the field QC samples.)
Sample container cleaning blanks (CCBs), sample shipping and
storage blanks (SSBs), field handling blanks (FHBs), and
sample preparation handling blanks (PHBs) were used to check
the integrity of field and preparation laboratory sample
handling. Figure F-2 shows the points in the sampling pro-
cess at which each type of blank was inserted. The sample
collection and preparation QAPP (CH2M HILL, 1987a) identi-
fies the collection frequency of each type of field blank
sample.
In general, at each step in the field sampling and prepara-
tion process where contamination could be introduced, a
blank was inserted into the sample stream. Only the FHBs
were routinely analyzed. If any contamination had been ob-
served in the FHBs, the other types of field blanks would
have been analyzed to isolate the source of contamination.
A sample CCB was taken after each batch of equipment (which
included liners, caps, spoons, and spatulas) was cleaned.
Additional CCBs were taken so that the total number of CCBs
represented at least 5 percent of the items cleaned. If
LCIC contamination was detected, the equipment associated
with the contaminated CCB was recleaned and another CCB was
taken and analyzed.
FBHs were used to detect any contamination from the sampler
or from handling of the liner in the field. A container
that was certified as clean from the CCB procedure was filled
with the blank soil at Cambridge Analytical Associates
(CAA), sealed with custody tape, and shipped to the field.
The FHB was collected by inserting the sampler into the
filled container. It was then handled as a regular sample
going through the sample preparation process. The FHB was
the only blank that was always analyzed. If any LCICs were
detected in an FHB, all other blanks associated with samples
taken after the last "clean" FHB, such as SSBs and PHBs,
also were analyzed.
SSBs were used to indicate if there was any contamination
during shipment from CAA to the field office, during storage
in the field office, during shipment from the field to the
sample preparation laboratory, or during shipment from the
preparation laboratory to the analytical laboratory. At the
F-2
-------
W63394.T1(SA)
QUANTITATIVE QA/QC
ACTIVITY
QUALITATIVE QA/QC
Decontaminate
Sampling Equipment
CCB
Decontaminate
Soil Prep. Equipment
Number Liners & Caps,
Then Wrap
SSB/FHB
Pack Coolers
&
Ship to Field
Survey Sample
Locations
Obtain
Core Samples
I
Ship to
Soil Prep. Lab
PHB
I
Extrude Soil
Core
Splits
Mix Soil
Sample
Ship to
Analysis Lab
Abbreviations
CCB Container Cleaning Blank
FHB Field Handling Blank
PHB Preparation Handling Blank
SSB Shipping and Storage Blank
Survey Logs
Photographs
Field Notebooks
Sample Collection Forms
Traffic Report Forms
Chain-of-Custody Forms
Custody Seals
Sample Prep Forms
Chain of Custody Forms
Traffic Report Forms
Sample Tracking Numbers
Sample ID Labels
Custody Seals
Figure F-2
SOIL ASSESSMENT - - INDICATOR CHEMICALS
SCHEMATIC REPRESENTATION
OF THE FIELD QA/QC PROGRAM
-------
analytical laboratories, the samples were stored at 4°C for
use if the FHB, described above, was contaminated. (If con-
tamination was found in the FHB, and analysis of the cor-
responding SSB indicated contamination, the analytical
results from the field samples corresponding the the SSB
were flagged before use in the data analysis.)
PHBs were used to detect possible contamination in the sample
preparation process. Several liners were filled with refer-
ence soil, capped, sealed, wrapped in aluminum foil, and
then sent to the sample preparation laboratory, where they
were prepared before being sent to the analytical labora-
tories for storage.
The PHB was the last sample prepared each day. It was
shipped with the other samples and blanks to the analytical
laboratories and stored at 4°C. At least one PHB was kept
in the preparation laboratory's storage area for the dura-
tion of the project and sent directly to CAA for analysis
without undergoing any sample preparation other than being
placed in the appropriate vial or jar. This blank indicated
if any contamination occurred in sending the PHBs to the
preparation laboratory or during storage there.
An element of the field QA/QC procedure was to divide field
samples into two parts, called "splits," at the preparation
laboratory. These splits were designed to measure variabil-
ity within and between laboratories.
QUALITATIVE FIELD QA/QC
Documentation of the status of collected samples was main-
tained in a series of report forms, log books, and computer-
ized data management systems. At each step in the process
the correctness of information was verified and discrepancies
were corrected. The daily soil sampling collection report
from the software system was checked against the sample col-
lection form to verify accuracy of the data base in terms of
information supplied. Errors, when detected, were noted on
the daily report for correction. If errors were detected,
corrections were made to the data base, and a new report was
generated.
The surveyor's daily report was checked to determine sam-
pling locations that needed to be moved more than 10 feet.
The reason for moving locations (e.g., a driveway) was noted
in the field team manager's log and in the surveyor's notes.
The street address and location descriptor appearing on the
surveyor's daily report was compared with the information
appearing on the daily sampling schedule. Discrepancies
were resolved, when possible, by examining a map. A possi-
ble discrepancy might include the term "backyard" versus the
F-3
-------
term "north side," for example. Discrepancies concerning
street addresses or whether a sample was on a property boun-
dary required further review. Documentation changes that
could be made at this stage were marked on the daily sam-
pling schedule. Locations sampled were checked against a
copy of the map and plotted using actual coordinates sup-
plied by the surveyors. Discrepancies were resolved using
photographic and/or field verification.
In the soil preparation laboratory, most of the documenta-
tion was generated by the preparation laboratory's personal
computer software. When coolers of samples were received,
the temperatures inside the coolers were recorded. The con-
tents of each cooler were manually verified by checking the
numbers on the liners against the chain-of-custody form. In
addition, the sample tracking numbers on the liners were
checked against the numbers on the traffic report form. As
each liner was extruded, the soil preparation form for that
sample was completed, signed, and dated by the sample pre-
parer. The computerized record of sample preparation was
completed by entering the station number for that sample
into the personal computer software. The computer then gen-
erated the identification number for that sample, the analy-
sis laboratory to which the sample was to be sent, and
whether the sample was to be a split sample. At the end of
each day, the software generated the chain-of-custody forms
and.traffic report forms to be shipped with the samples to
each analytical laboratory. The information on these forms
was verified manually against the contents of the cooler
before the coolers were shipped to the analytical laboratories
In addition to the forms described above, the soil prepara-
tion laboratory personnel kept a sample preparation log book
and refrigerator log books. Details on the preparation lab-
oratory operation are contained in the soil sample collec-
tion and preparation QAPP (CH2M HILL, 1987a).
ANALYTICAL QA/QC PROGRAM
The following description of the analytical QA/QC program
has been organized into three major parts. The first sec-
tion describes the pre-implementation laboratory QA/QC pro-
gram, which consisted of activities conducted before field
sampling began. The second section describes the laboratory
QA/QC program that was implemented during the field samples
analysis; in this section, laboratory QA/QC measures, the
real-time laboratory QA/QC program, and the onsite labora-
tory evaluation/audit are described. The third section
describes the post-implementation QA/QC program, which in-
cluded the data assessment procedures used by the EMSL-LV to
examine the quality of the laboratory results.
F-4
-------
PRE-IMPLEMENTATION LABORATORY QA/QC PROGRAM
Laboratories were selected and qualified for sample analysis
through a sequence of activities designed to result in a
laboratory operation that could meet the demanding QA/QC
requirements and time constraints of the project. A solic-
itation of 20 laboratories participating in EPA's Contract
Laboratory Program (CLP) resulted in 13 laboratories submit-
ting statements of qualifications. These laboratories were
evaluated in four areas: general, Small Business Admin-
istration status, EPA CLP performance evaluation audit, and
price. The seven laboratories identified for further con-
sideration were audited by the contractor and also received
two performance evaluation (PE) samples for analysis. The
laboratories were ranked on the basis of the PE and onsite
audit results. Once contractual matters were negotiated
with the laboratories, five were brought under contract; two
additional laboratories also participated in the study: the
prime contractor's laboratory and a subcontractor laboratory
that had participated in the method development, method
validation, and pilot study.
Several training activities were conducted so that the lab-
oratories would be familiar with the details of the analyti-
cal procedure. A week-long training session was conducted
for the project managers and extraction specialists from the
laboratories. The training included classroom discussion, a
laboratory demonstration of the soil extraction and cleanup
procedure, and a series of laboratory sessions during which
the extraction specialists processed matrix spike and method/
holding blank (MHB) samples. The extracts were analyzed
using the gas chromatograph/mass spectrometer/selected ion
monitoring (GC/MS/SIM) protocol, and the results were exam-
ined to decide whether further practice and evaluation were
required. While the extraction specialists were processing
samples, the project managers met with representatives from
the EMSL-LV and the prime contractor to discuss the state-
ment of work.
At the conclusion of the training session, the laboratories
were provided with a videotape of the sample extraction and
cleanup process to be used for in-house training. The lab-
oratories also were provided with two practice samples for
their in-house training.
For the laboratories to produce comparable results, they
were provided with materials and supplies judged on the
basis of the pilot study experience to be of critical impor-
tance. All LCICs, surrogate, and internal standards used in
the sample analysis were provided by the EMSL-LV. The lab-
oratories also were provided with most of the solvents and
reagents used in the sample preparation and analysis pro-
cedures. These materials were from single-production
F-5
-------
lots that were dedicated to the project and were demonstrated
by extensive QC analysis to have acceptably low levels of
LCICs or LCIC interferences. In addition, specialty glass-
ware for the sample preparation and cleanup was provided.
This included the Snyder distillation columns (micro and
macro) and the silica chromatography columns.
The primary responsibility for record keeping and documenta-
tion rested with the analytical laboratory. This process
began with the laboratories preparing quality assurance
plans (QAP) describing their QA/QC system and the qualifica-
tions and responsibilities of the personnel involved with
the project. Part of each QAP included series of standard
operating procedures (SOPs) that detailed the specific se-
quence of activities to be used on the project. Both the
QAP and the SOPs were submitted to the prime contractor for
review and were made available to auditors during onsite
laboratory audit.
PE samples were used to demonstrate that the laboratories
had become sufficiently proficient with the method to allow
field sample analysis to proceed. This sample analysis pro-
gram is described in greater detail in Appendix B. The PE
samples were prepared by the EMLS-LV using blank homogenized
soil similar to that found at Love Canal. The first set of
PE samples consisted of four samples whose concentrations
were known and one sample whose concentration was unknown to
the laboratories. The laboratories were required to analyze
these five PE samples in duplicate. In order to assess the
analytical proficiency of the laboratories, the data from
the first set of PE samples were used to evaluate the method
detection limit and to establish acceptance windows for the
blind QC (BQC) samples. The statistical analysis on the
first set of PE data produced windows that were excessively
wide. Once the laboratories had taken corrective actions, a
second set of PE samples was analyzed. The results of this
set of analyses were acceptable, indicating that the labora-
tories were ready to analyze field samples.
IMPLEMENTATION LABORATORY QA/QC PROGRAM
A number of laboratory QA/QC requirements were used during
the sample analysis in order to more closely monitor the
quality of the data and to provide assurance that the ana-
lytical system was in control at all times. A schematic
representation of the laboratory QA/QC program is shown in
Figure F-3. This figure relates the specific laboratory
activities to the quantitative and qualitative QA/QC mea-
sures. The laboratory QA/QC program (e.g., laboratory QA/QC
measures, real-time laboratory QA/QC program, and onsite
laboratory evaluation/audit) that was implemented during the
sample analysis is described below.
F-6
-------
W63394.T1(SA)
QUANTITATIVE QA/QC ACTIVITY
Field QC Samples: Splits _
SSB "'" *•
PHB
FHB
Lab DC Sample- MHR ,. fc
Sample Receipt
4
Log In
(In-House & Lab Data)
1
Store
4
Lab QA/QC ^
Measures: surrogates **
MS/MSD
BQC
Lab QA/QC
Measures: 1C *^
EMPC
PC
CC
(
_^
BQC ^
Extract/Cleanup
4
GC/MS/SIM
Analysis
4
Capture Quantitation Repor
& Chromatogram on PC
4
LabData
Processing
4
Forms 1 to X
V
Internal Review
Data Require Reanalysis?
|No
Data Package/Disks
V
External Review
4
Data Validation Report
t
Yes
QUALITATIVE QA/QC
GAP
SOPs
Logbooks
Sample Tracking (In-House)
Chain of Custody
Traffic Report Forms
Abbreviations
BQC Blind Quality Control Sample
CC Continuing Calibration
EMDC EPA Check Standard
FHB Field Handling Blank
GC/MS/SIM Gas Chromatograph/Mass
Spectrometry/Selected Ion
Monitoring
1C Initial Calibration
MHB Method/Holding Blank
MS/MSD Matrix Spike/Matrix Spike
Duplicate
PC Performance Check
PHB Performance Handling Blank
QAP Quality Assurance Plan
P.TQC Real-Time Quality Control
SOPs Standard Operating (Procedures
SSB Shipping and Storage Blank
Figure F-3
SOIL ASSESSMENT - - INDICATOR CHEMICALS
SCHEMATIC REPRESENTATION
OF THE LABORATORY QA/QC PROGRAM
-------
Laboratory QA/QC Measures
Within each laboratory, all QA/QC measures equivalent to
those of the EPA CLP were used, including matrix spikes
(MSs), matrix spike duplicates (MSDs), laboratory MHBs,
initial calibration (1C) standards, EPA check (CC) stan-
dards, PC standards, continuing calibration standards,
surrogate standards, internal standards, and BQC samples.
In addition to these QA/QC measures, a concept of the ana-
lytical run was used for both extraction batches and instru-
ment analysis batches. An extraction batch consisted of up
to 10 field samples or matrix spike samples plus a BQC
sample and a laboratory MHB. An instrument analysis batch
consisted of an initial combined PC/CC standard, a MBH, BQC
sample or check standard, field samples, and an ending PC/CC
standard. According to this concept, the QA/QC samples,
e.g., MHB, MS/MSD, and BQC, that were extracted with field
samples in the same batch would also be analyzed with them.
This would provide a block of analytical data from the
extraction and analysis batches with evidence that the
analytical system was in control.
Another feature of the QA/QC program is the use of either a
12-hour or a 16-hour analysis sequence. The laboratories
had the option of choosing one analytical sequence. The use
of a 16-hour analytical sequence with a PC/CC standard in
the middle of the run was allowed to make the analytical run
compatible with the number of samples in an extraction
batch. The 16-hour run enabled the laboratories to increase
sample throughput with no decrease in the quality of the
data. More detailed information about these QA/QC measures
can be found in the sample analysis QAPP (CH2M HILL, 1987b).
The following describes briefly these QA/QC measures.
Before any sample analysis, the laboratories were required
to analyze a PC standard to demonstrate adequate GC and MS
resolution and sensitivity and mass range calibration. A PC
standard solution contains LCICs, surrogate standards, in-
ternal standards, and 1-chloronaphthane. Also, a reagent
blank (RB) was used to demonstrate that samples could be
brought through the sample preparation and analysis process
without being contaminated by LCICs and LCIC interferences.
After the criteria for the PC standard solution were met, an
1C containing LCICs and surrogate standards was performed
using a minimum of five concentrations to determine the
linearity of response using the LCICs. The analytical lab-
oratories then analyzed the EPA check standard and compared
it to the 1C standards whenever ICs were performed and, in
any event, every 2 weeks. The EPA check standard analysis
was used to validate the accuracy of the 1C curve by com-
parison with a solution of known concentration prepared
F-7
-------
by EMSL-LV (thus eliminating possible bias in the laboratory
preparation of standards).
A continuing calibration PC (CC/PC) standard containing
LCICs and surrogate standards was performed at the beginning
and the end of each analytical run during sample analysis.
The relative response factors from the CC/PC analyzed at the
beginning of the run were compared with the average relative
response factors from the initial calibration for the GC/MS
instrument. The difference was required to be less than
25 percent (30 percent for delta-BHC). If the difference
exceeded the criteria an initial calibration was required.
Both the CC/PCs analyzed at the beginning and the end of the
analytical run were evaluated to demonstrate adequate GC and
MS resolution, sensitivity, and mass range calibration.
All samples were spiked with surrogate standards (compounds
of know concentration that eluted from the GC column at re-
tention times similar to those of LCICs) immediately before
extraction. Compounds used as surrogate standards were
1,4-dibromobenzene, 2,4,6-tribromobiphenyl, and 1,2,4,5-
tetrabromobenzene. A surrogate spike analysis was used to
indicate the recovery of the analytical process (e.g., ex-
traction and analysis).
For each extraction batch (not to exceed 10 or fewer field
samples), a laboratory MHB was analyzed to check for any
background contamination from the laboratory. A laboratory
MHB, which is a blank sand sample provided by EMSL-LV, was
placed in the refrigerator at the same time that a batch of
field samples was received. If contamination was detected
at levels exceeding those specified, the cause was deter-
mined and corrective actions were taken before proceeding
with sample analysis.
BQC samples provided by EMSL-LV were analyzed at a frequency
of one per extraction batch to assess the recovery of the
analytical method. EMSL-LV prepared the BQC samples by add-
ing an amount of LCICs, which was known to EMSL-LV but un-
known to the laboratories, to a sample of specially prepared
soil. The results of this analysis were used to estimate
the bias and variability of the analytical process at each
laboratory.
In addition to the laboratory MHB and BQC sample, an MS/MSD
sample was extracted and analyzed at a frequency of one set
per 20 samples to assess the precision and accuracy of the
analytical method. The MS/MSD sample was prepared by spiking
with 5.0 ppb of LCICs to two separate sample aliquots of one
field sample.
After the samples were extracted, the extracts were spiked
with the internal standards immediately before final
F-8
-------
concentration to final volume of lOOvil. The internal stan-
dards for GC/MS analysis were D4-1,4-dichlorobenzene,
DS-naphthalene, DIO-acenaphthene, DIO-phenanthrene, and
DIO-pyrene. The internal standards were used to compensate
for variation from nominal values in the final concentration.
The laboratories were required to meet the QA/QC frequency
and criteria as described above and in the sample analysis
QAPP (CH2M HILL, 1987b). If the required criteria were not
met, the laboratories were required to take necessary cor-
rective actions to identify and eliminate the problems.
Strict reextraction/reanalysis requirements were implemented
if the laboratories did not meet the required criteria after
the necessary corrective actions.
Real-Time Laboratory QA/QC Program
The information collected for each analytical run was pro-
cessed using the instrument's software in conjunction with
LabData software, which produced a shift results file and a
chromatogram file. The shift results file contained the
quantitation report of chromatographic retention times and
area counts for each of the ions monitored for each chemical
compound (LCICs, surrogate standards, and internal stan-
dards) . The chromatogram file contained those portions of
the PC/CC standards that were needed to assess instrument
performance. In addition to these two files, the instru-
ment's software produced the total ion current (TIC) and
selected ion current profile (SICP) chromatograms for in-
clusion in laboratory's data package.
The shift results file and the chromatogram file were then
electronically transmitted to a personal computer for fur-
ther processing. This personal computer executed the
study-provided LabData software, a customized package for
the soil LCIC study. LabData is described in Exhibit A4 of
Appendix A of the sample laboratory analysis QAPP
(CH2M HILL, 1987b).
LabData processed the shift results and chromatogram files
and then produced draft versions of the data package forms
(Forms I to X). These forms were used by the laboratories
for internal review of the data. The laboratory-related
QA/QC requirements were evaluated by LabData. The draft
forms, in conjunction with the SICPs, were used to determine
if acceptable sample analysis had been achieved. Samples
that were not successfully analyzed were scheduled for re-
injection or reextraction as appropriate.
Once the information from sample receipt, sample prepara-
tion, and analytical run had been processed through LabData,
the QA/QC data from the associated shift file were transmit-
ted (uploaded) to an electronic bulletin board for review.
F-9
-------
This real-time QA/QC report contained information on surro-
gate recoveries, internal standard areas, MS/MSD recoveries,
MHB levels, PC and CC standards, and sample status. The
real-time QA/QC data were then downloaded by the contractor
and EMSL-LV and reviewed using the standard operating pro-
cedures described in Appendix C of the sample laboratory
analysis QAPP.
The real-time review examined surrogate recoveries, MS/MSD
recoveries, internal standard areas and retention times, PC
standards, CC response factors, and extraction and analysis
holding times. (When these criteria are within the QAPP
specified control limits, it is likely that the analytical
system of the laboratory is operating in statistical con-
trol.) In addition to checking that individual analyses met
specifications, the software also accumulated run results
into control charts to allow the QA/QC personnel to spot
trends in instrument performance.
Part of the real-time QA/QC data was the information on the
BQC samples. This information was accessed by EMSL-LV from
the bulletin board and was used to evaluate the laboratory's
performance. Personnel from EMSL-LV used their own software
to access the BQC data and to assess the acceptability of
the BQC samples. Once the BQC data had been reviewed,
EMSL-LV uploaded to the bulletin board a report indicating
the acceptability of the BQC sample. This report was down-
loaded by the laboratory to confirm the status of that BQC
sample. This process is described in more detail in Appen-
dix B of this report.
Other real-time QA/QC procedures performed by the contractor
and EPA involved assessing the overall status of a laboratory,
as well as the laboratory's success at resolving any diffi-
culties encountered during the project. The Laboratory
Technical Manager conducted daily telephone conversations
with all laboratories to assess their current situation.
The discussions addressed any technical issues that had
arisen.
Laboratory Onsite Evaluation/Audit
During the sample analysis, documentation on the status and
processing of samples was included in a series of log books
that were completed by the laboratory personnel and reviewed
periodically by each laboratory's QA staff. The records
consisted of sample log-in, extraction, refrigerator stor-
age, and instrument analysis. These records either were
included in the data packages or have become part of the
laboratory's case file that is sent to the NEIC at the con-
clusion of the project. The case file contains every writ-
ten record of information from the project, including sample
tags, shipment airbills, and correspondence.
F-10
-------
Periodically throughout the study, personnel from the prime
contractor, EMSL-LV, and NEIC conducted onsite laboratory
evaluation and audits that involved a complete examination
of the personnel, facilities, equipment, chain-of-custody
procedures, analytical and QA/QC procedures, and documenta-
tion records being used by the laboratory. These are dis-
cussed in more detail in Appendices G and H.
POST-IMPLEMENTATION LABORATORY QA/QC
PROGRAM (DATA VALIDATION)
During the laboratory phase of this study, the laboratories
prepared final data packages containing the data for all
samples for which complete data had been obtained at that
time. LabData was used to produce the final forms that were
included in each data package. In addition, LabData produced
two sets of floppy disks. One set contained information
used by EMSL-LV for data assessment; the other was sent to
the prime contractor for use in building the final data
base.
EMSL-LV was responsible for providing an independent review
of the validity of the data produced by the laboratories
during the habitability study. EMSL-LV conducted a data
validation and assessment on the final data package. This
validation and assessment consisted of an electronic audit
of QA/QC data in the final data packages and a review of the
hardcopy data packages for quality flaws and verification of
the electronic audit. The EMSL-LV data assessment procedure
is presented in more detail in Appendix H.
DATA BASE QA/QC PROGRAM
The data base QA/QC program consisted of the following major
components:
o QA/QC of project source documents, such as sample
collection forms, sample preparation forms, and
traffic control forms, to verify that the source
documents for the data base were correct.
o QA/QC of software systems that collected, manip-
ulated, and/or generated data that flowed into the
data base to verify that data entry was correct
and that calculations performed by the software
were correct.
o QA/QC of the data loaded into the data base and
the data integration process to verify that the
data were properly loaded into the correct fields
in the data base and that the data linkages, or
F-ll
-------
connections between the various data fields, were
correct.
The data base QA/QC program resulted in a logical approach
to storing correct data in the data base. The components
also reflected the process through which the data flowed
into the data base. To the extent possible, cross-checks
were built into the software and QA/QC processes to provide
more than one way to check the validity of critical data.
The following sections describe each of these components.
SOURCE DOCUMENT QA/QC
The source document QA/QC began with the sample collection
forms and traffic report forms. As these forms were com-
pleted by field personnel, they were reviewed by supervisory
sampling personnel, and the sample collection form was pre-
pared for data entry. Each form was keyed twice by different
data entry clerks, and the resulting files were compared.
Any discrepancies were investigated and resolved. The sam-
ples were then sent to the preparation laboratory.
In the preparation laboratory, the site number of the sample,
taken from the previously reviewed traffic report form, was
entered into the Sample Preparation System. As part of the
sample design, a master table of selected sites (and target
coordinates) was prepared. This table was then used by the
Sample Preparation System to identify samples being prepared
and to automatically generate sample identifiers. The system
also automatically (1) indicated which samples may have
needed to be screened because they were taken from areas in
which high concentrations were found in the pilot study, (2)
identified samples to be split and generated the associated
identifier to be used, and (3) identified samples to be used
as native samples for MS/MSD samples. This automatic gen-
eration of data, as well as the sample preparation form and
associated traffic report form, substantially reduced the
possibility of these data being incorrectly entered on the
forms. In addition, the data used to generate the forms
were validated before they were entered into the system
(i.e., the table driving the system was validated), further
reducing the chance that a form would be inaccurate.
Once the sample collection and preparation process was com-
plete, the resulting files were checked to verify the
following:
o Sites targeted for sampling were sampled.
o Sites were sampled only once (i.e., no duplicate
samples).
F-12
-------
o Field samples sent to the preparation laboratory
were prepared (except those that could not be
extruded) and sent to the analysis laboratory.
o Samples prepared were, in fact, valid samples from
a valid collection site.
o Samples were prepared only once (i.e., duplicate
samples were not sent to the analysis laboratory).
SOFTWARE SYSTEM QA/QC
Before"each software system was placed into operation on an
ongoing basis during the project, the functions of each of
the systems were validated to verify that data were stored
and calculations were performed correctly, and data outputs
were checked for accuracy.
For the site selection system, a pattern analysis was used
to verify that the randomizing algorithm was not biased in
the selection of targeted sites within the boundaries of the
selected neighborhood. In addition, it was verified that
the spacing between the target sites conformed with the sample
design.
For the sample collection system, it was verified that the
data entry programs stored the data in the correct fields in
the sample collection data base. In addition, the program
used in the system to perform the comparison between the two
files was checked to verify that the proper fields were
compared and that any corrections entered were properly
stored in the data base. All reports generated were checked
to verify that the reports accurately reflected the data
contained in the data base.
For the sample preparation system, it was verified that the
system properly retrieved the sample coordinates when given
a valid site identification. In addition, it was verified
that the system randomly generated a project identification
and properly stored the identification with the correct site
identification in the sample preparation data base. It was
also verified that the system was properly identifying sites
for screening. The identification of samples to be split
(i.e., the frequency) and the proper storage of the split's
project identification were verified as well. The system
was checked to verify that it was sending the samples to the
appropriate analytical laboratory since the system automat-
ically generated the traffic report form. In addition, all
reports were verified to determine that the data contained
in the data base were properly reported.
For the LabData system, the verification effort was inten-
sive and complex. This system produced the forms used by
F-13
-------
the analytical laboratory to report the results of the sample
analyses. Data were electronically captured from the GC/MS/
SIM; additional data were entered into the system by labora-
tory personnel. Therefore, QA/QC was performed on every
data element on every form to verify that (I) the data were
being captured correctly by the LabData system, (2) all cal-
culations were being performed correctly, and (3) the results
were being reported properly. This effort was performed at
the beginning of the study using the PE samples provided to
the laboratories and was performed independently by project
personnel and by each analysis laboratory. Periodically,
the same PE sample forms were regenerated to verify that the
software was still producing the same results. Moreover,
calculations were spot-checked to verify their accuracy. In
addition to the validation of the forms generation, it was
also verified that data were being correctly transmitted by
the LabData System to other project software systems. This
included the data transmitted to the real-time QA/QC system,
data sent to the data validation system, and data generated
for upload to the integrated data base.
For the real-time QA/QC system, the software was checked to
verify that the data downloaded from the project bulletin
board (sent there automatically by the LabData system) were
properly stored. The reports and control charts produced by
the system were verified to determine that the data depicted
on the charts and contained on the reports accurately re-
flected the data contained in the system data bases.
For the data validation system, the software was checked to
verify that data were properly loaded onto the system files.
Reports were checked to verify that data contained in the
reports accurately reflected data base contents. In addi-
tion, the summary concentration report for each laboratory
was verified against the Form I's (i.e., matching the re-
ported concentrations) submitted to EMSL-LV for validation.
The software was also checked to verify that the data
validation checklists entered were properly stored.
The Integrated Data Base Model software was checked to verify
that the data being modeled (only a subset of data base
fields were filled with data) were correctly stored and that
the generated data conformed to the distributions identified
in the pilot study.
The Form I Data Entry System software was checked to verify
that the program stored the data in the correct fields in
the Form I system data base. In addition, the program used
to compare the two data entry files was checked to verify
that the proper fields were compared, and that any correc-
tions entered were properly stored in the data base.
F-14
-------
The QA/QC procedures used for the program that loaded the
data generated by the various Love Canal information manage-
ment systems into the integrated data base, as well as the
procedures used to conduct QA/QC procedures and validate the
integrated data base, are described in the following section.
INTEGRATED DATA BASE QA/QC
The data base QA/QC process had two components. First, each
computer program in the system had to be checked to verify
that data were loaded and retrieved correctly and that cal-
culations in the program were being performed correctly.
This established a verified system with which to perform the
second component of the QA/QC process, which was to verify
that the data in the data base were correct and that the
various tables in the data base were linked and linked
correctly.
The software was verified through the normal process of
printing hard copy inputs and outputs from each program and
manually verifying that data were correctly handled and
calculations were properly performed. Additionally, during
the data base load and verification process, the programs
were periodically checked to verify that validation errors
flagged were not actually software bugs.
Inherent in the design of the system was the need to perform
as many of the verification checks as possible using auto-
mated means. Thus, most of the programs in the system
incorporated checks to verify the data being processed.
Figure F-4 depicts a flow chart of the system, and Table F-l
lists the steps in the load and verification process.
STATISTICAL QA/QC
Statistical QA/QC consisted of independent computation and
cross verifications of all major statistical analyses. The
basic statistical comparisons were programmed in SAS, GAUSS,
and FORTRAN in three computer centers by three analysis
groups. The results were cross checked and all inconsis-
tencies resolved.
REFERENCES
CH2M HILL. 1987a. Love Canal Habitability Study—Soil
Sample Collection and Preparation Quality Assurance Project
Plan (Final Revised Version).
CH2M HILL. 1987b. Love Canal Habitability Study—Soil
Sample Laboratory Analysis Quality Assurance Project Plan.
8856A/009
F-15
-------
W63394.T1/S.A.
.,
>
I
' I ' 1
~^
F1STCOMP
H
ICFILE 7
(Initial •< 1
\ •
Sample \ 'Includes data from
Tracking 1 1) Sample Collection System
Data' / 2) Sample Preparation System
LabData \ 'Each analysis shift
System ] 'ram each laboratory is
Data*
Calibrations) \_
Integrated
Data Base
NOTE;
Dashed lines (—) indicate an iterative process, i.e. the
output file becomes the input file for the next iteration or
execution of the program.
Figure F-4
SOIL ASSESSMENT-
INDICATOR CHEMICALS
INTEGRATED DATA BASE
SYSTEM FLOW CHART
-------
Table F-l
STEPS IN THE LOAD AND VERIFICATION PROCESS
Program Name
Description
Validate LabData
(manual process)
Upload Final Data
Packages (manual
process)
MIGRATE
STMERGE
EDUPQC
Using the list prepared by EMSL-LV
during the data validation
process, each laboratory was
visited to verify that its LabData
System data conformed to the list.
(This list, organized by analysis
shift, details all validated
analyses, including calibrations
and field QC analysis.) After
completion of the verification,
final data packages for all valid
analyses shifts were then
generated.
The final data packages, contained
on diskettes, were uploaded to the
EPA mainframe computer. All files
were successfully uploaded.
The MIGRATE program, which
separated the LabData data into
separate files to be loaded into
the integrated data base, was then
executed.
The program that merged the sample
tracking data with the field
sample data was executed. The
validation report from this
program listed any records in
sample tracking that did not match
to a field sample from the LabData
System and vice-versa. All
discrepancies were resolved.
This program processed the QC
samples and recomputed any results
fields, using the original
quantitation report data,
generated by LabData. Only
results fields requiring no data
from other quantitation reports
(e.g., calibration data) were
recalculated. All discrepancies
were resolved in fields used to
join other table entries.
8855/051
F-16
-------
Program Name
Table F-l
(continued)
Description
6.
EDUPIC
7.
EDUPCC
8.
EDUPFS
9.
EDUPFIN
This program processed the 1C
samples and recomputed any results
fields, using the original
quantitation report data,
generated by LabData. All
discrepancies were resolved.
This program processed the CC
samples and recomputed any results
fields, using the original
quantitation report data,
generated by LabData. Only
results fields requiring no data
from other quantitation reports
(e.g., 1C data) were recalculated.
All discrepancies were resolved in
fields used to join other table
entries.
This program processed the field
samples and recomputed selected
results fields, using the original
quantitation report data,
generated by LabData. Only
results fields requiring no data
from other quantitation reports
(e.g., calibration data) were
recalculated. Sample IDs were
verified against the sample
tracking data. Edits were
performed to verify that the
calibration entry for the field
sample was in the data base and
that any QC samples pointed to by
the field sample were also in the
data base. All discrepancies were
resolved in fields used to join
other table entries.
This program processed the field
samples, verifying all fields that
require data from multiple
sources; it recomputed
concentrations, recoveries, etc.,
and all discrepancies were
resolved.
8855/051
F-17
-------
Program Name
Table F-l
(continued)
Description
10.
FISTCOMP
QCSTCOMP
11. SASBUILD
These programs performed a
three-level match between the
Form I data validated by EMSL-LV
and data entered, and the Form I
data contained on LabData. The
initial match was on the Lab Login
identification for the Form I.
This was the unique key used by
the LabData System to track the
sample analysis. Any mismatches
were reported. Then, for those
samples whose identifications
matched, two additional levels
were checked. The sample
identification (the project
identification), analysis date,
and analysis time were matched and
mismatches reported. Then all
LCIC concentrations, percent
moisture and dilution factor were
compared. Discrepancies between
Lab Login IDs were resolved; and
validated Form I's linked to a
valid LabData entry (which
contains all the raw data for the
analysis) and all LabData entries
linked to a validated Form I.
This program loaded the validated
data files into the integrated
data base.
8855/051
F-18
-------
APPENDIX G
NEIC/7echLaw Participation
So/7 Assessment—Indicator Chemicals
Prepared by
U.S. EPA National Enforcement Investigation Center
TechLcav, Inc.
i LaKewood, Colorado
-------
May 4, 1988
SUMMARY OF NEIC AND CEAT
PARTICIPATION IN
LOVE CANAL EMERGENCY DECLARATION AREA
HABITABILITY STUDY (LCHS) SOIL ASSESSMENT
FOR INDICATOR CHEMICALS
In October 1985, the National Enforcement
Investigations Center (NEIC) received a request for evidence
preparation assistance from the Environmental Protection
Agency (EPA) Region II. This request involved work
supporting a Pilot Study and Soil Assessment for Indicator
Chemicals Study related to the Love Canal hazardous waste
site (Superfund Site/Spill Identifier No. 05) in Niagara
Falls, New York. The NEIC assigned the Contract Evidence
Audit Team (CEAT-TechLaw) to participate in these studies by
attending planning meetings, reviewing and commenting on the
Quality Assurance Project Plan (QAPP), and performing
evidence audits of the field and laboratory operations
related to the studies.
The purpose of the CEAT's work effort was to ensure
that the evidence collected during the studies was
consistent with the NEIC Policies and Procedures manual.
The following is a review of the tasks performed by the CEAT
to date and a summary of the CEAT's participation in the
project and related follow-up activity.
SOIL ASSESSMENT FOR INDICATOR CHEMICALS ORGANIZATIONAL
MEETINGS
CEAT personnel attended organizational meetings prior
to both the Pilot Study and the Soil Assessment for
Indicator Chemicals Study. During these meetings, evidence
audit schedules were determined and issues regarding the
proper methods of documentation were discussed.
QAPP REVIEW
CEAT personnel reviewed copies of the QAPP and supplied
comments to the study project managers. The QAPP was
reviewed to determine if the procedures described for sample
collection, sample transfer, and documentation of evidence
related field and laboratory activities met Evidence Audit
Requirements.
FIELD EVIDENCE AUDITS
CEAT personnel conducted two field evidence audits at
Love Canal during the LCHS. The first audit was conducted
during the Pilot Study, July 15-18, 1986. The second audit
was conducted during the Soil Assessment for Indicator
-------
Chemicals Study, October 14-16, 1987. The auditors observed
document control and chain-of-custody procedures followed
during sample collection, sample preparation, and related
data compilation.
Documentation methods were reviewed to determine if the
information recorded was complete, consistent, and
represented an accurate record of the field sample
collection activities. The documents used in the field were
audited to determine if they were accountable. Standard
operating procedures were reviewed to ensure that the
documents are archived at the conclusion of the project.
Sample collection procedures were audited to determine
if the samples are properly identified and if sample
identification is maintained.
CEAT personnel observed the field activities to
determine if the sample is maintained in the custody of the
field collection team and if there are adequate security
measures used in any areas identified as secure areas for
sample storage.
The CEAT auditors recorded their observations on a
checklist during the audit. At the conclusion of the audit,
the audit teams held debriefings with the field teams.
During the debriefings, the auditors discussed their
findings and made recommendations for corrective action.
Implementation of the corrective actions was discussed in
these meetings and was verified at the conclusion of the
Pilot Study during project file evidence audits conducted by
the CEAT. Implementation of corrective action for the Soil
Assessment for Indicator Chemicals Study will be reviewed
when the complete project files are provided to the CEAT in
the next quarter. Written reports of the field evidence
audits were provided to the NEIC and the EPA project manager
for review.
LABORATORY EVIDENCE AUDITS
CEAT personnel conducted evidence audits of all
laboratory operations pertaining to chain-of-custody and
document control procedures for the Pilot Study and the Soil
Assessment for Indicator Chemicals Study. These audits
included a minimum of three audits at each of the three
laboratories participating in the Pilot Study, as well as at
least two audits of each of the seven laboratories which
participated in the Soil Assessment for Indicator Chemicals
Study.
Procedures and documentation related to sample
receiving, sample identification, sample security, sample
storage, sample tracking, and project file organization and
assembly were reviewed for conformance to Evidence Audit
Requirements and the requirements stated in the LCHS
Statement of Work. The CEAT auditors recorded their
observations on a checklist during the audit.
-------
The audits consisted of the following two components:
Procedural audit. The procedural audits consisted of a
review and examination of actual and written standard
operating procedures and accompanying documentation for
the following operations: sample receiving, sample
storage, sample tracking (from receipt to completion of
analysis), and case file organization and assembly.
The CEAT auditors followed the physical path of the
samples and reviewed procedures and documents related
to all activities associated with the samples. The
security of each facility was reviewed.
Evidence audit. The evidence audits consisted
of review and examination of project file
documentation. The case file was reviewed to
determine if the inventory is accurate and
that all the documents related to sample
receipt, sample transfer, sample preparation, and
sample analysis are included in the case file.
At the conclusion of the audits, debriefings were held
with the laboratory personnel. During these debriefings,
the CEAT auditors presented the findings of the audit and
recommendations for corrective action. Implementation of
corrective action was discussed in the debriefings.
Documents in the case files were reviewed at the conclusion
of the Pilot Study to determine if corrective action had
been completed. The auditors also reviewed corrective action
on the items identified during previous laboratory evidence
audits. Implementation of corrective action for the Soil
Assessment for Indicator Chemicals Study will be reviewed
when the complete project files are provided to the CEAT in
the next quarter. Written reports of the laboratory
evidence audits were provided to the NEIC and the EPA
project manager for review.
-------
ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF ENFORCEMENT
NATIONAL ENFORCEMENT INVESTIGATIONS CENTER
BUILDING 53, BOX 25227, DENVER FEDERAL CENTER
DENVER, COLORADO 80225
DATE:
May 5, 1988
MEMORANDUM
To: Doug Garbarini
Emergency & Remedial Response Division, Region II
From: Robert H. Laidlaw
Project Officer
Subject: CH2M Hill Corrective Action Response to LCHS Field
Evidence Audit Report
In reference to your request of March 30, 1988, NEIC tasked
TechLaw to review whether CH2M Hill adequately responded to audit
recommendations made during the October 1987 field sampling
activity. Jeff Worthington has completed this review and his
comments are attached. Chain-of-custody, document control, and
evidence security procedures instituted by CH2M Hill and
Cambridge Analytical Association meet evidence audit
requirements.
Also attached is a summary of NEIC and TechLaw participation
in the soil assessment program. We are pleased to provide this
assistance. If you have any questions or comments please call me
at FTS 776-5122.
Attachments
cc: Michael Travis, Contract Officer, PCMD (PM241-F)
Park Haney, Deputy Team Leader, CEAT
Jeffrey C. Worthington, CEAT
-------
CONTRACT
EVIDENCE
AUDIT
TEAM
MEMO
TO: Rob Laidlaw, NEIC
FROM: Jeffrey C. Worthington, CEA
DATE: April 12, 1988
SUBJECT: CH2M Hill Corrective Action Response to LCHS Field
Evidence Audit Report
The Contract Evidence Audit Team (CEAT-TechLaw) has conducted a
review of CH2M Hill's response to the CEAT recommendations in the
January 11, 1988 LCHS Field Evidence Audit Report. The review
was made in accordance with generally accepted evidence auditing
standards.
Based on the corrective action described in CH2M Hill's response,
the chain-of-custody, document control, and evidence security
procedures followed by CH2M Hill and Cambridge Analytical
Associates for the Love Canal Habitability Study meet or exceed
Evidence Audit Requirements.
JCW:ad
cc: R. Park Haney, CEAT
Betty Malone, CEAT
IF: 111-068
TECHLAW, INC. • 12600 W. COLFAX AVE-, • SUITE C3IO • LAKEWOOD, CO • 80213 • (J03) 233-1248
-------
APPENDIX H
EMSL-LV/LEMSCo Involvement
Soil Assessment—Indicator Chemicals
Prepared by
U.S. EPA Environmental Monitoring Systems Laboratory
Lockheed Engineering and Management Services Company
Las Vegas, Nevada
-------
THE EMSL-LV INVOLVEMENT IN THE
LOVE CANAL
EMERGENCY DECLARATION AREA
HABITABILITY STUDY
SOIL ASSESSMENT — INDICATOR CHEMICALS
by
G.L. Robertson
D.L. Bogen
Environmental Programs
Lockheed Engineering and Management Services Company
Las Vegas, Nevada 89119
Contract No. 68-03-3249
Work Assignment Manager
Dr. J.D. Petty
Quality Assurance Division
Environmental Monitoring Systems Laboratory
Las Vegas, Nevada 89193
ENVIRONMENTAL MONITORING SYSTEMS LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
LAS VEGAS, NEVADA 89193
April 25, 1988
-------
TABLE OF CONTENTS
PAGE
1.0 Introduction 1
2.0 Method Modifications 5
3.0 Performance Evaluation and Blind Quality Control Samples. . 7
4.0 Quality Assurance Requirements 9
5.0 Onsite Evaluations 13
6.0 Statistical Support 15
7.0 Real Time Quality Control 16
8.0 Data Validation 17
9.0 Summary 30
-------
LIST OF TABLES
PAGE
Tables la-g Summary Table of Love Canal Soil BQC Samples
by Spike Level 32
Tables Ih-k Summary Table of Love Canal Sand BQC Samples
be Spike Level 36
Table 2a Love Canal Summary Soil BQC Statistics Table
by Laboratory 38
Table 2b Love Canal Summary Sand BQC Statistics Table
by Laboratory 38
Table 3a Love Canal Summary Soil BQC Statistics Table ... 39
Table 3b Love Canal Summary Sand BQC Statistics Table ... 39
Table 4 Summary of Flags Used to Qualify BQC Sample
Data from Table 1 40
Table 5a Count of Data Valildation Summary Qualifiers
from Field Samples in Form I Database 41
Table 5b Count of Data Validation Individual Qualifiers
from Field Samples in Form I Database 42
Table 6 Sample Percent Recovery Used in the Generation
of the Low Level Prediction Intervals 43
Table 7 Sample Percent Recovery Used in the Generation
of the High Level Prediction Intervals 44
ii
-------
THE EMSL-LV INVOLVEMENT IN THE
LOVE CANAL EDA HABITABILITY STUDY
SOIL ASSESSMENT -- INDICATOR CHEMICALS
1.0 INTRODUCTION
The U.S. EPA Environmental Monitoring Systems Laboratory in Las
Vegas, Nevada (EMSL-LV) provided independent quality assurance
support for the Love Canal Emergency Declaration Area (EDA)
Habitability Study, Soil Assessment -- Indicator Chemicals (LCHSSA).
The purpose of the support was to assist the U.S. EPA, Region 2 in
obtaining defendable data of known and high quality for use in the
decision making process. This report descibes the EMSL-LV
participation in the study and the data validation procedures and
results.
The data validation established that over seventy percent of
the analytical field sample data was useable for the statistical
analysis. This is a high percentage for a complex analytical method
with strict quality assurance criteria.
1.1 Historical
The EMSL-LV involvement in the LCHSSA began in fiscal year
(FY) 1986 when the USEPA Region 2 requested quality assurance
support from the EPA Office of Research and Development who
designated the EMSL-LV to provide this support. The LCHSSA is
designed to provide information regarding the levels of chemical
contamination in the area surrounding the Love Canal hazardous waste
site and to assist the New York State Commissioner of Health in
determining if the neighborhoods surrounding the site are habitable.
The basis of the study is a comparison of the concentration of
chemicals detected in the Love Canal neighborhoods and other similar
residential areas in western New York State that are at least one-
half mile from a chemical waste site. The comparison will be made
using a selected set of eight indicator chemicals. One of the
reasons this approach is being used is because there are no soil
habitability limits available for the selected Love Canal Indicator
Chemicals (LCIC). The LCIC selection process is presented in Love
Canal Emergency Declaration Area / Proposed Habitability Criteria
(NYSDOH and DHHS/CDC, December, 1986) and had been determined prior
to the EMSL-LV involvement in the study.
-------
The data acquired from previous studies indicated that the
concentrations of LCIC expected to be present in the Love Canal area
were significantly lower than existing analytical methods could
detect. Since a statistical comparison of the areas requires
"detects" to be effective, a new analytical method had to be
developed. The decision was made to perform a pilot study to
evaluate the new analytical methodology, the statistical approach,
and to determine the number of samples required for the main study.
A computerized sample tracking and data processing system was also
developed and evaluated.
1.2 Pilot Study Involvement
The EMSL-LV and their contractors became involved in the LCHSSA
at the analytical method development stage in November 1985,
providing advice on the applicability of the proposed methods and
the documentation and quality assurance measures which should be
incorporated. The documentation and quality assurance requirements
of the USEPA Contract Laboratory Program were used as guidelines for
the Pilot Study and the LCHSSA.
The analytical method chosen was Selected Ion Monitoring (SIM)
Gas Chromatography/Mass Spectrometry (GC/MS) with an alumina column
cleanup of the extract prior to analysis. The data from method
optimization work was reviewed and when performance appeared
adequate, performance evaluation samples were prepared at the EMSL-
LV for formal method validation and to evaluate the qualifications
of the laboratories. The results of the analysis of these samples
were reviewed and the analysis of a second set of performance
evaluation samples was recommended. After the evaluation of these
results, the decision was made to proceed with the Pilot Study.
During this same time period, onsite evaluations were performed
at the participating laboratories to examine their capabilities for
performing the analyses with the required levels of quality
assurance and documentation. Additional onsite evaluations were
performed during the Pilot Study to provide assurance that the
analyses were being conducted according to the method and to provide
assistance with unforseen method problems.
Several revisions of the Pilot Study Quality Assurance Project
Plan (QAPP) were reviewed and the data review section, written at
the EMSL-LV, was included in the QAPP.
An electronic bulletin board system was used to make Quality
Assurance (QA) data generated by the laboratories available to the
data reviewers, including EMSL-LV, soon after it was produced. As a
consequence, the QA data could be evaluated, any out of control
analyses identified, and corrective action taken promptly.
-------
The data produced during the Pilot Study was reviewed to
determine the quality of the data, the effectiveness of the required
QA criteria and the performance of the method. Recommendations were
made to Region 2 regarding changes to the method and QA procedures
for incorporation into the plan for the main study.
Independently of the Pilot Study organized by Region 2, the New
York State Department of Health (DOH) collected and analyzed samples
from the Love Canal Emergency Declaration Area (EDA) and other areas
within Niagara Falls, New York using a different method of analysis.
In December 1986, the EMSL-LV contractor personnel visited the DOH
to examine the data to determine its quality and its comparability
with the Pilot Study data. The data were deemed to be of sufficient
quality that the results were considered in the planning of the main
study. The Niagara Falls areas sampled in the DOH study, which were
not sampled during the Pilot Study, provided information on the
distribution of trichlorobenzene and tetrachlorobenzene in Niagara
Falls that was not available from the Pilot Study. This resulted in
control areas within Niagara Falls being chosen for inclusion in the
main study.
1.3 Method Modifications
The DOH requested that the detection limits of the method be
lowered for the Habitability Study because of the large number of
nondetected LCIC in the Pilot Study at the nominal detection limit
of one microgram per kilogram (ug/kg). This required major
modifications to the method cleanup and concentration procedures.
The EMSL-LV participated in the method revision planning meetings
and reviewed the method development data as it was produced. The
modifications to the method resulted in lowering the detection limit
to approximately 0.3 ug/kg for most: of the LCIC.
1.4 Training Session
Prior to the start of the Habitability Study, the EMSL-LV and
contractor personnel participated in a training session that was
held for the laboratories selected to analyze the LCHSSA soil
samples. The training session consisted of a demonstration of the
extraction procedure and the laboratory personnel practicing the
extraction procedure. The extracts were analyzed immediately to
provide feedback as to the trainees' performances. Simultaneously,
presentations were given to the laboratory project managers on the
QA requirements and additional project administrative and logistical
information.
-------
1.5 PE Samples
Following the training session, performance evaluation (PE)
samples were prepared and scored at the EMSL-LV as part of the
method validation and laboratory qualification processes for the
LCHSSA. The analytical standards to be used for the project were
verified as to identity and concentration. Check standards were
also prepared and distributed to the laboratories for comparison
with the laboratory prepared working standards. The periodic
analysis of the check standard gave additional evidence that the
laboratories' calibrations were accurate.
1.6 BQC Samples
Three sets of "Blind Quality Control" (BQC) samples were also
prepared at the EMSL-LV to be used for the study. The soil matrices
were checked for contamination and then the samples were spiked with
varying amounts of the LCIC which were unknown to the laboratories.
One sample was required to be extracted in each batch of 10 or fewer
samples and to be analyzed with that batch of samples. The results
of these analyses were transmitted to the EMSL-LV for scoring within
24 to 48 hours to determine if reextractions were necessary.
1.7 Onsite Evaluations
The EMSL-LV and contractor personnel performed onsite
evaluations of the field sampling teams, the sample preparation
laboratory, and the analytical laboratories during the LCHSSA. The
field sampling teams and preparation laboratory were observed at the
beginning and the midpoint of the study. Each analytical laboratory
was observed twice and laboratories 1, 3, and 4 were visited three
times. A number of recommendations were made during these
evaluations to improve procedures or to eliminate potential
problems; however, no major problems were found.
1.8 Data Validation
The EMSL-LV reviewed the data and supporting documentation
produced by the laboratories during the LCHSSA. Two series of flags
were developed to characterize the quality of the data. The first
was a set of detailed flags indicating QA deficiencies or other
abnormalities of the data. The second was a set of summary flags
which consolidated the detailed flags into a single qualifier for
each analyte, indicating the usefulness of the data for the LCHSSA
statistical analysis. See Tables 5a-b.
-------
2.0 METHOD MODIFICATIONS
The EMSL-LV provided technical assistance in the method
revision and validation process for the LCHSSA. This involved
assisting in the implementation of recommendations from the Pilot
Study and in the modification of the method to attain the lower
detection limits requested by DOH. Attempts were also made to
reduce the interlaboratory variability during the method revision
process.
2.1 Recommendations
The EMSL-LV review of the Pilot Study data produced numerous
recommendations to improve data quality for the LCHSSA. The details
of this may be found in Appendix D of the Pilot Study Love Canal EDA
Habitability Study Report; however, some of the more important
recommendations follow. It was determined that blank interpretation
must include interferences as well as LCIC contamination; the EMSL-
LV proposed a set of blank interpretation rules that incorporated
this concept, which were included in the LCHSSA QAPP. It was also
decided that there must be a low concentration limit below which the
laboratories do not expend great efforts to determine if the LCIC is
present; this limit was set at 0.2 ug/kg. In addition, it was
concluded that the raw data submitted by the laboratories must be
presented in a manner that permits the reconstruction of the
laboratories' results by an independent reviewer; the data package
requirements were developed using the EPA CLP requirements as
guidelines.
The Pilot Study included a volatiles method which was used to
analyze for chlorobenzene and dichlorobenzene. The volatiles
analysis had no chlorobenzene values over 1 ppb, dichlorobenzene was
also analyzed in the semivolatile method, and found at higher
concentrations in the semivolatile method. The volatiles method
added significantly to the workload of the analytical laboratories
without producing data of value. The decision was made to delete
chlorobenzene as an LCIC and drop the volatile method from the
LCHSSA.
The functional detection limits of the Pilot Study were
approximately one part per billion (ppb) for most LCIC (i.e. below
one ppb the interference structure of the chromatograms became so
complex that data interpretation and review time became
prohibitive). This resulted in the reporting of many samples with
no LCIC detected. Because the statistical approach of the study
requires "detects" to establish if a difference exists between the
EDA and the control areas, the DOH requested that an attempt be made
to lower the detection limit of the method for the LCHSSA.
-------
2.2 Implementations
The EMSL-LV helped determine the most likely techniques, to
lower the detection limit while maintaining data quality, and then
participated in the evaluation of the experimental results to deter-
mine if the method modifications met the goals. The method changes
included changing the extract cleanup procedure by using silica gel
instead of alumina and concentrating the extract to a smaller final
volume. This also required changes in surrogate and internal stand-
ard concentrations, a reduction in the calibration range, and other
minor changes. The results of the analyses using the modified
method were reviewed and the EMSL-LV Performance Evaluation samples
were used as the final step in the method validation process.
In order to reduce interlaboratory variability, several actions
were taken. Critical pieces of glassware used in the extraction
process were supplied by the project because some experimental
evidence indicated a possible difference between brands of glass-
ware. The solvents and other reagents used in the analyses were
analyzed, and specific lots numbers with low background were
supplied to the laboratories; this reduced the amount of blank
interferences present in the LCHSSA.
The EMSL-LV reveiwed all method changes, QA requirements,
reporting requirements, and QAPP revisions. The reviews were
performed in an iterative manner and often occurred as participation
in meetings, rather than formal review. This reduced the time
required for revisions, thus shortening the time required for
project completion.
2.3 Training Session
In an effort to further reduce interlaboratory variability and
due to the complexity of the extraction procedure, a training
session was held at Boston University for the extraction chemists
and the laboratory project managers. The extraction chemists
observed a demonstration of the extraction procedure, organized by
project contractor staff, and then performed extractions while at
the training session. The extracts prepared by each laboratory's
representative were analyzed and the results were discussed.
Simultaneously, project QA and administrative requirements were
presented to the laboratory managers by project contractor and EMSL-
LV personnel. Following the training session, the laboratories were
sent practice samples by the prime contractor and then the first set
of Performance Evaluation (PE) samples by EMSL-LV.
2.4 Standards Preparation and Evaluation
The standards used for the LCHSSA were prepared by the U.S. EPA
Quality Assurance Materials Bank at Research Triangle Park, NC. The
-------
specific lot numbers of the standards provided to the laboratories
were verified as to identity and concentration by the EMSL-LV
laboratory. The individual LCIC, surrogate and internal standard
solutions as well as the check standard solution were verified.
The 1-chloronaphthalene standard contained approximately twelve
percent 2-chloronaphthalene. Efforts to obtain pure 1-chloro-
naphthalene commercially were unsuccessful. Pure 1-chloronaph-
thalene was synthesized at EMSL-LV but was not completed in time for
use in the study. The 1-chloronaphthalene is used as part of the
Performance Check/Continuing Calibration standard; the presence of
the 2-chloronaphthalene impurity would tend to cause low bias of the
2-chloronaphthalene results in the field samples. The laboratories
were told not to use a correction factor in the calculations as the
impurity would not affect the goals of the study, which is a
comparison approach, because all values would have the same bias.
3.0 PERFORMANCE EVALUATION AND BLIND QUALITY CONTROL SAMPLES
The EMSL-LV prepared Performance Evaluation (PE) samples and
Blind Quality Control (BQC) samples for use in the LCHSSA. The PE
samples were used to assist in the laboratory selection process and
to demonstrate when the laboratories had acquired sufficient
proficiency with the method to allow field sample analysis to
proceed. They were also used for the final step in the method
validation process and the results were used to establish limits for
various quality assurance measures. The purpose of the BQC samples
was to provide evidence in each extraction and analytical batch that
the system was in control and that the laboratory was producing
acceptable analytical results on a known sample.
3.1 PE and BQC Sample Preparation
The PE and BQC samples were prepared in the EMSL-LV laboratory
using blank homogenized soil, similar to that found at Love Canal,
and one batch of BQC samples was prepared with sand. Prior to use
in the quality control sample, the homogenized soil matrix was
extracted according to the method and evaluated for interferences
and potential problems. Spiking solutions containing the LCIC were
prepared and the concentrations were verified. The spiking
solutions were spiked into the matrix and the jars were labeled
using a coding system produced by a random number generator.
The first set of PE samples consisted of four samples whose
concentrations were known and one sample whose concentration was
unknown to the laboratories. The known samples were spiked at 0.1
ppb, 0.2 ppb, 0.5 ppb, and 2.0 ppb with all the LCIC. The unknown
sample was spiked with 1.0 mL of a spiking solution containing the
LCIC at varying concentrations. The results of these samples were
-------
also part of the method detection limit determination. A second set
of PE samples was analyzed using samples which were prepared, in the
manner described above, for use as the first batch of Blind Quality
Control samples.
The BQC samples sent to the laboratories were from three
different batches. The batch one BQC samples were prepared in the
same soil used for the PE samples and were spiked at four different
levels. The spiking solution prepared for the PE samples, contain-
ing the eight LCIC, was added to the samples at four different
volumes, 0.5 mL, 1 mL, 1.5 mL, and 2 mL. The spike concentrations
and results for these samples are provided in Tables la-d, spike
levels A - D. The batch two BQC samples were composed of extra 0.2
ppb, 0.5 ppb and 2.0 ppb samples and the unknown PE sample from PE
set 1. The spike concentrations and results for these samples are
provided in Tables le-g, spike levels E - G. The laboratories
required additional BQC samples and every effort was made to obtain
more clean soil similar to that found at Love Canal. Three batches
of soil were extracted and analyzed and none were acceptable for use
in the study. Due the immediate need for more samples, it was
necessary to use sand for the third batch of BQC samples. The batch
three samples were prepared exactly as batch one, except using
commercially available sand for the matrix. The spike
concentrations and results for these samples are provided in Tables
Ih-k, spike levels A - D. Each laboratory was sent 16-20 samples
from batch one, 8-12 samples from batch two, and 10 samples from
batch three.
3.2 PE Sample Evaluation
The results of the PE samples were evaluated at the EMSL-LV.
The evaluation was performed by entering the analytical results
into spreadsheets for statistical analysis and examining the
accompanying raw data packages for analytical problems. A computer
analysis, developed by EMSL-LV, was made on the analytical data to
determine tentative acceptance windows for the results of the BQC
samples. The statistical analysis on the first set of PE sample
results produced windows which were excessively wide. This high
variability among the laboratories' data indicated that the
laboratories were not yet ready to analyze field samples. The
analytical problems indicated by the data packages were examined and
corrective actions recommended. These included laboratory specific
actions and refinements to the analytical protocol. After the
completion of these corrective actions, a second set of PE samples
were analyzed by the laboratories. The results of this set of PE
samples were acceptable, indicating that the laboratories chosen by
the prime contractor were capable of performing the analysis. The
results from the second set of PE samples were combined with the
acceptable results from the first set to provide a larger database
for establishing the control criteria for the BQC samples.
-------
It was necessary to generate two sets of acceptance windows,
for low and high concentration samples. The data from PE samples
spiked with less than approximately 1 ppb of the LCIC were used to
calculate the low concentration windows and the samples spiked with
greater than approximately 1 ppb were used to calculate the high
concentration windows. See Tables 6 and 7 for the data used to
generate the windows and Tables la-k for the acceptance windows used
for the individual spike levels.
The BQC samples were evaluated during real time quality
assurance and again during data validation. See the individual
sections of this report for details.
4.0 QUALITY ASSURANCE REQUIREMENTS
A number of Quality Assurance (QA) requirements were added to
the QAPP at the request of the EMSL-LV in order to more closely
monitor the quality of the data and to provide assurance that the
analytical system was in control at all times. The items
recommended and the rational are listed below.
4.1 Blind Quality Control (BQC) Samples
One BQC sample was required to be extracted with each batch of
10 or fewer field samples and the BQC sample associated with a
particular extraction batch was required to be analyzed with all
samples from that batch, except under certain specified conditions.
This frequency requirement was added to provide evidence in each
extraction and analytical batch that the system was in control.
Strict reextraction/reanalysis requirements were also established
for the BQC sample results.
4.2 EPA Check Standard
This solution, prepared by the U.S. EPA Quality Assurance
Materials Bank at Research Triangle Park, NC and verified by EMSL-
LV, contained the LCIC, surrogates and internal standards and was
used to check that the standards used by the laboratories were
accurate. The solution was required to be analyzed after each
initial calibration, every two weeks and with batches of mixed rerun
samples. If the calculated value of the check standard was not
within 80-120 percent of the true value (70-120 percent for delta-
BHC), corrective action was required and a check standard, which met
criteria, had to be analyzed before field sample analysis could be
resumed. This provided assurance that the laboratories' instruments
were calibrated and that the seven laboratories obtained comparable
results. In general, the laboratories had no problems meeting the
check standard criteria.
-------
4.3 Analytical Run
The concept of the analytical run was used for both extraction
batches and instrument analysis batches. An extraction batch
consisted of up to 10 field samples or matrix spike samples plus a
BQC sample and a method/holding blank. An instrument analysis batch
consisted of an initial combined system performance check/continuing
calibration standard, a blank, a BQC sample or check standard, field
samples and an ending system performance check standard. The idea
was that the QC samples that were extracted with field samples would
also be analyzed with them. This would provide evidence in all
extraction and analysis batches that the analytical system was in
control. The laboratories had the option of either a 12 hour or a
16 hour analysis sequence. The use of a 16 hour analytical sequence
with a system performance check and calibration in the middle was
allowed to make the analytical run compatable with the number of
samples in an extraction batch. The 16 hour run enabled the
laboratories to increase sample throughput with no decrease in the
quality of the data, since a full extraction batch could not be
analyzed in a single 12 hour shift but could be analyzed in 16
hours.
4.4 Laboratory Blanks
The laboratories that participated in the Pilot Study had
varying amounts of the LCIC in their blanks due to the use of
different brands or lots of reagents. Another problem in the Pilot
Study was that the only criteria for an acceptable blank was the
presence and concentration of LCIC contamination; interferences that
were not positively identified as a LCIC were not considered.
Several changes were implemented to decrease and control blank
contamination because the LCHSSA required lower detection limits
than the Pilot Study.
Prior to the Habitability Study, the reagents needed for the
Soil Study were analyzed and specific lot numbers, that were
acceptable according to the new blank requirements, were provided to
all the laboratories. This proved to be an effective means of
limiting LCIC contamination in the blanks. During the Habitability
Study, the laboratories found similar levels of LCIC in their
blanks. However, the presence of non-LCIC interferents posed a
different problem. Each laboratory had somewhat different
interferences, and when they were the same, the separation from the
expected retention time of the LCIC tended to be different. The
laboratories had varying degrees of problems with blanks. A common
set of interferences was traced to detergent residues on the
glassware used in sample extraction and cleanup. This was problem
was resolved by a combination of modifying glassware cleaning
procedures and adjusting chromatographic conditions.
10
-------
A set of blank interpretation rules were developed by the EMSL-
LV to allow the use of data which did not meet the exact definition
of an acceptable blank. These rules allowed the use of data which
contained interferences that did not impact the identification and
quantitation of the LCIC. The rules facilitated the analysis of
samples by the laboratories and provided significantly more usable
data for the study. The blank rules and corresponding figures
follow.
BLANK RULES
Rule 1: If more than one interference peak is present in the LCIC
RRT window, the peak causing the most interference must be
used for determining the acceptability of the blank. This
may be the largest peak or the peak closest to the expected
RRT, depending on the situation (see Figure 1).
Rule 2: Interference or LCIC peaks that are present in the blank but
are not present in the samples (indicating a problem only
with the blank) will not require reanalysis of the negative
samples. (For the purposes of this section, a negative
sample is defined as one that contains less than 0.2 ug/kg
of LCIC response.)
Rule 3: If there is baseline resolution between an interference peak
and the. LCIC peak, then the blank is acceptable for that ion
of that sample. If there is no signal above normal noise in
the baseline at the expected RRT of the LCIC, then the blank
is acceptable for that ion of that sample (see Figures 3a-b).
Rule 4a: If the blank contains more than 0.5 ug/kg (0.6 ug/kg for
1,2-dichlorobenzene) of a LCIC or the equivalent
concentration of an interference at both the primary and
secondary ions (the interference may be at different
retention times) within the RRT window of an LCIC, all
samples in the extraction batch containing a response
greater than 0.2 ug/kg for that LCIC must be reextracted and
reanalyzed (see Figure Aa).
Rule 4b: If an interference peak is present within the LCIC RRT
window at only the primary or secondary ion and the sample
contains a peak greater than 0.2 ug/kg equivalent
concentration at the appropriate RRT of the ion with no
interference, the sample must be reextracted and reanalyzed
(see Figure 4b).
NOTE: Rule A applies only if the conditions of Rule 3 are met.
Rule 5: If an interference peak that is present only at the tertiary
ion in the blank is also present in the sample, and if the
primary and secondary ions meet ion ratio criteria and at
least a shoulder is observable on the tertiary ion
11
-------
interference peak at the LCIC RRT, the blank is acceptable
for that ion of that sample. If a shoulder is not present
or of the potential LCIC tertiary ion peak is totally
obscured by the interference peak, the sample must be
reextracted and reanalyzed (see Figure 5a-b).
Rule 6: If an interference peak is present within the LCIC RRT
window at only the primary or secondary ion and the sample
is a nondetect based on the ion without interference,
reextraction and reanalysis are not required (see Figure 6),
2LAKK RULE FIGURES
BLANKS
SAMPLES
a.
b.
-a.
2°
A.
5b.
1"
2°
* Sanrole must be reextracted and reanalyzed
12
-------
5.0 ONSITE EVALUATIONS
Onsite evaluations were performed at the laboratories •
participating in the LCHSSA to determine if the laboratories
possessed the required facilities, personnel, and equipment, if the
appropriate record keeping procedures were in place and being
followed, and if the analytical protocol was being followed.
Assistance was also provided to the laboratories in solving
analytical problems and in interpreting the protocol. The onsite
evaluation reports are available in the LCHSSA case file at the U.S.
EPA, Region 2.
5.1 Analytical Laboratories
All of the laboratories were visited twice, and three of them
three times, by teams consisting of one EMSL-LV EPA chemist and one
EPA contractor chemist. Each team was accompanied by a
representative of the prime contractor, and a representative of
Techlaw, representing the National Enforcement Investigation Center
(NEIC).
The first round of onsite evaluations was held at the beginning
of the LCHSSA, in August 1987. The initial contact with the
laboratories was at odd hours (late afternoon or evening) and
without prior notice in order to observe the operations of the
laboratory when they were not prepared for visitors. The
laboratories had not started analyzing field samples at this time,
so the onsite evaluations consisted of examining the facilities,
procedures, and personnel qualifications. Each laboratory had
appropriate facilities and instrumentation for participation in the
study; however, they had procedural items to correct before sample
analysis. Examples of procedural items that needed implementation
or modification were: temperature logs, standards logbooks, standard
operating procedures, and appropriate sample and standard storage
procedures. Minor changes to clarify requirements were made to the
protocol as a result of these onsite evaluations.
The second round of onsite evaluations was performed at the
midpoint of the study, October 1987. Sample extraction and analysis
procedures were observed, records for sample receipt, extraction,
and analysis were checked, and problems the laboratory was having
were discussed. Suggestions were made for minor improvements;
however, no protocol modifications were made as a result of these
visits. The Sample Preparation Laboratory was also visited during
this round of onsite evaluations. The process of extruding the
samples from the sampling tubes, homogenizing the samples and
aliquoting the samples for shipment to the laboratories was
observed. Minor recommendations were made to improve security and
reduce the chances of sample cross-contamination.
13
-------
The third set of onsite evaluations was performed near the end
of the study, December 1987, when only three laboratories we-re still
analyzing samples. Laboratories 3 and 4 were having analytical
problems and laboratory 1 was preparing to analyze samples extracted
by laboratory 4. The purpose of these onsite evaluations was to
determine that the laboratories with problems were producing data of
improved quality, to assist in problem solving, and to verify that
laboratory 1 was still performing well since they would be receiving
The locations and dates of the onsite evaluations are listed
below:
Laboratory Name and Location
Date
CH2M HILL, Montgomery, AL
Clayton Environmental Consultants, Novi, MI
NUS Laboratory, Pittsbugh, PA
Versar, Inc., Springfield, VA
Cambridge Analytical, Boston, MA
Aquatec, Inc., Burlington, VT
EMSI, Camarillo, CA
Versar, Inc., Springfield, VA
Clayton Environmental Consultants, Novi, MI
NUS Laboratory, Pittsburgh, PA
CH2M HILL, Montgomery, AL
Aquatec, Inc., Burlington, VT
Cambridge Analytical, Boston, MA
EMSI, Camarillo, CA
Versar, Inc., Springfield, VA
NUS Laboratory, Pittsburg, PA
Aquatec, Inc., Burlington, VT
August 26, 1987
August 26, 1987
August 26-27, 1987
August 27-28, 1987
August 27, 1987
August 28, 1987
August 28, 1987
October 19, 1987
October 20, 1987
October 20, 1987
October 21, 1987
October 22, 1987
October 22, 1987
October 23, 1987
December 14, 1987
December 16, 1987
December 18, 1987
5.2 Field Sampling Teams
Onsite evaluations were also performed by the EMSL-LV
contractor field sampling audit team to observe the field sampling
teams and the sample preparation laboratory. The sampling
equipment, sample collection, sample preparation and sample shipping
were evaluated. During the first onsite evaluation, the complete
sampling process was reviewed and only minor procedural problems
were observed, such as a sampler touched the exposed end of the
liner with gloved hands, the sampler was rinsed before loose soil
was brushed off, unlabeled wash bottles were used, etc. The second
visit was to observe if the required QA procedures were still being
followed. In general, the procedures presented in the soil sampling
plan were being followed during the visits.
-------
The locations and dates of the onsite evaluations are listed
below:
Onsite Purpose and Location Date
Love Canal Field Sampling, Niagra Falls, NY October 14, 1987
Sample Preparation Laboratory, Boston, MA October 15, 1987
Sample Preparation Laboratory, Boston, MA October 21, 1987
Love Canal Field Sampling, Niagra Falls, NY October 27, 1987
6.0 STATISTICAL SUPPORT
The EMSL-LV provided statistical support for the Love Canal
Habitability Study. This support consisted of quality assurance and
technical reviews. These efforts were performed by statisticians
with substantial assistance from several chemists and technicians.
6.1 Quality Assurance
The EMSL-LV was responsible for establishing acceptance
criteria for the performance evaluation materials for use in the
LCHSSA. An acceptance window was determined for each of the seven
PE samples based on observed laboratory analytical performance on
these materials. The statistical procedures and software used were
developed previously by the EMSL-LV for the Contract Laboratory
Program. Basically, the procedure involves removing outliers by
Grubb's test and then using the remaining data to estimate
prediction intervals with the desired confidence level. These
intervals were employed as acceptance criteria for the PE samples
and the BQC samples.
6.2 Technical Reviews
The EMSL-LV statistical review function differed from the
chemical review in that the EMSL-LV had little involvement in the
planning and production stages of the statistical analyses. Only
final reports were reviewed whereas, for the chemical analyses, the
EMSL-LV was involved at all stages and provided iterative feedback
as the project progressed. Each major report was reviewed and, at
times, it was found appropriate to suggest changes or alternative
approaches to problems.
The EMSL-LV will review the final statistical report, which is
unavailable at this time, and submit a summary of the review
comments will be submitted at a later date.
15
-------
7.0 REAL TIME QUALITY CONTROL
The seven laboratories that participated in the LCHSSA were
provided with a software package developed for the project called
LabData. This software was used to track the samples in the
laboratory, process the data from the mass spectrometer, and
electronically transmit the results of the analyses to a secure
bulletin board system. The data reviewers were provided with
software to receive and evaluate the files sent by the laboratories.
This system, Real Time Quality Control software (RTQC), made the QA
data generated by the laboratories available to the data reviewers
for examination on a daily basis so that problems could be
identified and corrective action implemented promptly.
Most of the QA parameters required for the method were
transmitted through the RTQC system. The parameters monitored
through the system include: surrogate recovery, internal standard
area and retention time differences, blank contamination, matrix
spike/matrix spike duplicate recovery and relative percent
difference, column performance check criteria, and sample status.
The EMSL-LV evaluated the data as an independent reviewer. The
Bulletin Board was monitored at least daily and if laboratories had
uploaded (sent) data files, the files were downloaded (received) and
immediately logged into the Love Canal Computer Logbook. The files
were evaluated for all the parameters listed above and any anomolies
were recorded and reported to project management. However, the
laboratories did not upload the results to the Bulletin Board daily
as anticipated and some of the information proved more useful than
others in evaluating the problems occuring in the laboratory. For
example, if a surrogate recovery was out of criteria for a
particular sample, it was impossible to tell if the sample was being
reextracted until the results of the reextraction were posted.
However, if the laboratory was only able to analyze a few samples
per shift, a problem was indicated. The RTQC system did not meet
all expectations but demonstrated the future potential for such a
system. Another possible contribution of the RTQC system to the
general high quality of the data was the psychological effect upon
the laboratories of knowing their data was being scrutinized.
The software generated a special database containing the Blind
QC sample results in a format to be merged into a program written at
the EMSL-LV. The program calculated the percent recovery of each
analyte in each sample, compared the recoveries to the appropriate
acceptance windows, and flagged the analytes that did not meet
criteria. A file was then generated that listed the analytes that
were out of acceptance criteria for each sample, whether they were
out high or low, and if the sample was acceptable or not. The
results file was uploaded to the Bulletin Board, coded so that each
laboratory could only able access their own results, with a turn-
16
-------
around time of 12 to 36 hours. Turnaround time was important
because a failed Blind QC sample required reextraction of the entire
associated batch of samples.
A number of problems occurred with the electronic transfer of
the BQC sample data. For instance, one laboratory experienced
problems with their data system and was unable to upload their
results to the Bulletin Board for a few weeks. The EMSL-LV received
the BQC sample results over the telephone and provided the results
of the evaluation to the laboratory also by telephone. On other
occassions, interference was found in the BQC samples that needed to
be evaluated on an individual basis; for these cases, the raw data
was sent to EMSL-LV for evaluation. Overall, the feedback system of
the BQC sample results was very successful as the laboratories were
provided with the results as promptly as their files were uploaded
to the RTQC system.
8.0 DATA VALIDATION
The EMSL-LV was tasked to evaluate and qualify all data
generated for the Love Canal Habitability Soil Study. This task
included evaluating all the sample results, comparing the raw sample
data to the data on the forms, and generating data qualifier and
summary flags for every analyte in every sample. The Standard
Operating Procedure (SOP) used to validate the data is provided at
the end of this report.
8.1 Data Qualifiers
Individual qualifiers were developed to both denote anomolies
in the data and identify data quality deficiencies. The qualifiers
were applied only where necessary and were catagorized as follows:
BQC samples, matrix spike/matrix spike duplicate samples, method/
holding blanks, internal standards, holding times, surrogate
recoveries, analytical sequence deviations, identification and
quantitation, and miscellaneous. These flags were applied either to
individual analytes or samples, depending on the impact of the data
deficiency.
A system of summary data qualification flags was also developed
to assist the statisticians in deciding which data to use for the
statistical analysis. This system grouped the individual data flags
into catagories based on the effect the flag had on the data. The
summary flags are applied on an individual analyte basis to increase
the amount of usable data for the statistical study. A description
of each of the individual and summary data flags follows.
17
-------
Individual Data Validation Qualifiers
G,U,B
QUALIFIER
Blind Quality Control Samples (BQC)
U/B El = If an analyte in a BQC sample is outside the
acceptance criteria but the BQC sample is acceptable,
that analyte in the BQC sample and the associated
samples is flagged with "El" (equivalent to the "E"
flag in the SOW). This flag used in association with
"Z8" is "U".
U E2 = If a BQC sample is reinjected and an analyte(s) is
outside the acceptance window, that analyte(s) in
that reinjected BQC sample and the associated samples
is flagged with "E2".
G/U E3 = If an analyte in a BQC sample is reported as "ND" due
to failed ion ratios and quantitation from the
primary ion meets acceptance criteria, that analyte
in the BQC sample and the associated samples is
flagged with "E3".
U E4 = The results of this BQC sample should be disregarded
due to a possible problem in the preparation of the
BQC sample rather than a laboratory problem; the BQC
sample and the associated samples are flagged with
"EA".
G/U E5 = If an analyte in a BQC sample is reported as "ND" due
to failed scan range and quantitation from the
primary ion meets acceptance criteria, that analyte
in the BQC sample and the associated samples is
flagged with "E5".
B E7 = If analytes in a BQC sample are outside the
acceptance criteria and the BQC is not acceptable,
those analytes in the BQC sample and the associated
samples are flagged with "E7".
G/U E8 = If an analyte in a BQC sample is reported as "ND" due
to failed ion ratios as a result of interference at
the primary ion and quantitation from the secondary
ion meets acceptance criteria, that analyte in the
BQC sample and the associated samples is flagged with
"E8".
18
-------
Matrix Spike/Matrix Spike Duplicate Samples (MS/MSD)
U Ml = If the percent recovery of an analyte in MS or MSB
samples exceeds the advisory limit, that analyte (in
the MS or MSB sample only) is flagged with "Ml".
U M2 = If the percent recovery of an analyte in an MS or MSB
sample is lower than the advisory limit, that analyte
(in the MS or MSB sample only) is flagged with "M2".
U M3 = If the relative percent difference (RPB) of an
analyte in a MS and MSB sample is outside the
advisory limit, that analyte in both the MS and MSB
samples is flagged with "M3".
B MA = If the MS, MSB and native samples have surrogates
that do not meet criteria, the MS and MSB samples
only are flagged with "MA".
Internal Standards (IS)
U 12 = If an internal standard in a sample (except pyrene-
dlO, see 14) exceeds the area percent difference
criteria, the analyte(s) quantitated using that
internal standard is flagged with "12".
B 13 = If an internal standard in a sample is out of the
area percent difference criteria low, the analyte(s)
quantitated using that internal standard is flagged
with "13".
G 14 = If pyrene-dlO exceeds the area percent difference
criteria, the sample is flagged with "TV.
G 15 =: If pyrene-dlO is out of the area percent difference
criteria low, the sample is flagged with "15".
Method Blank Samples
G Bl == If a peak within the relative retention time (RRT)
window of an LCIC is greater than 0.5 ppb (equivalent
concentration) and is NOT reported on the Form IV and
has no impact on the associated samples, that analyte
in the method blank is flagged with "Bl".
G B2 = If a peak within the RRT window of an LCIC is greater
than 0.5 ppb (equivalent concentration) using peak
height or estimated peak height and is NOT reported
on the Form IV and has no impact on the associated
samples, that analyte in the method blank is flagged
with "B2".
19
-------
U/B B3 = If a peak within the RRT window of an LCIC is greater
than 0.5 ppb (equivalent concentration), may or may
not have been reported on the Form IV, and has an
impact on the identification and/or quantitation of
the analyte in the associated samples, that analyte
in the affected samples is flagged with "B3".
G BA = If a peak within the RRT window of an LCIC is greater
than 0.5 ppb (equivalent concentration), may or may
not have been reported on the Form IV, and has NO
impact on the identification and/or quantitation of
the analyte in the associated samples, the analyte in
the blank only is flagged with "BA".
Identification and Quantitation
B Ql = If a laboratory did not report the presence of an
analyte but examination of the data shows the
presence of a peak meeting all criteria at an
equivalent concentration greater than 0.2 ppb, that
analyte is flagged with "Ql".
B Q2 = If a laboratory reported the presence of an analyte
but should have reported "ND" due to failed ion
ratios, that analyte is flagged with "Q2".
U Q3 = If a laboratory reported the presence of an analyte
but should have reported "ND" due to failed scan
range, that analyte is flagged with "Q3".
G QA = If a spiked sample (i.e. BQC, MS/MSD, and EPA Check
Std) has tertiary ion interference which is not found
in the blank and interferes with identification of
the analyte, the affected analyte is flagged with
"QA".
B Q5 = If samples exhibited apparent chromatographic
reversal of B-BHC and G-BHC, the analytes are flagged
with "Q5".
G Q6 = If an analyte does not exceed the calibration range
as a wet weight, but does exceed the calibration
range as a dry weight, the analyte is flagged with
"Q6".
U Q7 = If an analyte(s) exceeds the calibration range,
diluted or undiluted, wet weight and dry weight, the
analyte(s) affected in the sample is flagged with
"Q7".
20
-------
U/G Q8 == If less than 20 + 1 g (SOW required extraction
weight) was extracted, the sample is flagged, with
"Q8".
U QA == If the B-BHC/G-BHC percent valley criteria for PC2 is
not met, B-BHC and/or G-BHC in the positive samples
in the associated shift are flagged with "QA".
Sample Qualifiers
B Zl « If a surrogate(s) is outside acceptance criteria, the
sample is flagged with "Zl".
U/G Z2 - If the corresponding BQC sample was not analyzed with
samples from the associated extraction batch, the
samples that are not analyzed in the same batch are
flagged with "Z2". If the BQC sample was analyzed
previously and was acceptable, and a BQC sample or an
EMPC was analyzed in the analytical shift, there is
no impact on the data quality (G).
G Z3 ~ If a native sample and MS/MSD samples were extracted
on different dates, the MS/MSD samples are flagged
with "Z.3".
G Z4 = If a native sample and MS/MSD samples were extracted
on the same date, but analyzed on different dates,
the MS/MSD samples are flagged with "ZA".
U Z5 = If the BQC sample extracted with holding time (HT)
samples is unacceptable, the BQC sample and all HT
samples from that extraction batch are flagged with
"Z5". Holding time samples can not be reextracted.
G Z6 = If the surrogate recovery is acceptable for the first
injection of a sample, but unacceptable for the rein-
jection, the reinjected sample is flagged with "Z6".
U/G Z7 = If the associated method blank was not analyzed with
samples from the same extraction batch, the
associated samples not injected with the blank are
flagged with "Z7". If the method blank was analyzed
previously and was acceptable, and a method blank or
an instrument blank was analyzed in the analytical
shift, there is no impact on the data quality (G).
G Z8 = If the BQC sample extracted and analyzed with the
samples is a very low level spike (0.2 ppb) with
advisory limits, the samples and BQC sample are
flagged with "Z8".
21
-------
B Z9 = If the BQC sample extracted and analyzed with the
sample is unacceptable, the samples and the BQC
sample are flagged with "Z9".
G ZA = If the BQC sample extracted and analyzed with the
sample was a sand matrix, the samples and the BQC
sample are flagged with "ZA".
G ZB = If the method blank from the appropriate extraction
batch was not reinjected with a reinjected sample,
the reinjected sample is flagged with "ZB".
G ZC = If the BQC sample from the appropriate extraction
batch was not reinjected with a reinjected sample,
the reinjected sample is flagged with "ZC".
G ZD = If the ion ratio of pyrene-dlO is outside the
criteria for PC2, the samples in the shift are
flagged with "ZD".
G ZE = If the lab analyzed the method blank associated with
rerun samples and an EMPC instead of the associated
BQC sample, the samples are flagged with "ZE".
G ZF = If the native and/or MS/MSD samples were extracted
and analyzed on different dates, the MS and MSB
samples are flagged with "ZF".
G ZG = If a BQC sample, previously analyzed and acceptable,
went dry and could not be reinjected with reinjected
samples, but an EMPC was analyzed with the reinjected
samples, the reinjected samples are flagged with
"ZG".
G ZH = If the wrong peak was evaluated for pyrene-dlO at m/z
213, the affected sample is flagged with "ZH".
G ZI = If the wrong peak was evaluated for pyrene-dlO at m/z
213 in the PC2, the samples in the analytical shift
are flagged with "ZI".
U ZJ = If the concentrations on a Form I are wet weight and
not dry weight because percent moisture is not
indicated on the form, the sample is flagged with
"ZJ".
Holding Time Qualifiers
U HI = Samples that exceed the analysis holding time, 50
days from verified time of sample receipt, are
flagged with "HI".
22
-------
U H2 = Samples that exceed the extraction holding time, 30
days from verified time of sample receipt, are
flagged with "H2".
U H3 = Samples that exceed the extraction and/or analysis
holding times due to reruns requested by EMSL-LV
after the data package was submitted are flagged with
"H3".
U H4 = Samples that exceed both the extraction and analysis
holding times are flagged with "HA".
Resubmission Qualifiers
G K2 = If a small discrepancy exists between the concentra-
tions on the Form I and the calculated concentrations
from the quantitation report which was not resolved
by resubmissions, the sample is flagged with "K2".
Summary Qualifiers
G This flag is used on analytes which have either no data quality
defects or have technical violations of the QAPP which have no
effect on data quality. Data with this flag may also have
informational flags.
U This flag is used on analytes which have significant data
quality defects but only low probability of this having an
effect on the quality of the data. This includes holding time
violations and exceeding the calibration range etc. Data with
this qualifier is usable for some purposes and most of the
defects which make up this catagory would result in reported
values which are lower than expected. This data should only be
used by persons knowledgeable in the effects a particular
deficiency would have on the data.
B This flag is used on analytes which have data quality defects
that may, but do not necessarily have an effect on the data.
Examples of this are BQC samples out of criteria, surrogates
out of criteria, method/holding blanks out of criteria, etc.
Data with this flag should be used only with extreme care and
with full knowledge of the limitations of the data quality and
the consequences of using flawed data.
These flags are applied to the data to reflect compliance with
the QA criteria in the QAPP and Good Laboratory Practice. They do
not indicate compliance with nontechnical requirements of the QAPP
and are not intended to do so.
23
-------
8.2 Count of Data Validation Qualifiers
The Form I's were marked with all the individual and summary
data qualifiers and provided to software contractor for entry into a
database. A summary count of some selected flags used to qualify
the data are listed in Tables 5a-b, Summary of Data Validation
Qualifiers from Form I Database. The counts in both tables are by
analyte and not by sample. The counts for the individual qualifiers
should not total the counts of the summary qualifiers because more
than one individual qualifier may apply to the same data point in
the database.
The data presented in Table 5a gives the number and percentage
of each analyte by laboratory that were assigned each of the G, U,
and B summary qualifiers. The range of "G" analyses between
laboratories, 54.9 to 92.2 percent, is significant. Both
laboratories 3 and 8 had many samples which exceeded holding times
due to reextraction and reanalysis (see Table 5b). This is the
primary cause of the wide range in the percentage of "G" data. The
total "G" data, 72.4 %, should be considered an achievement for a
method of this complexity and the level of review to which the data
was subjected.
The data presented in Table 5b contains the distribution of a
selection of the individual data qualifiers by laboratory. The
flags chosen are some of those which apply to the field samples and
have an impact on the data. The definition of the flags are given
above. The table shows that the laboratories had different QA
defects in their data which indicates that the problems are not
method deficiencies.
8.3 Blind QC Sample Results
The BQC sample results, monitored during the study using the
RTQC software, were also evaluated during the Data Validation
process. The BQC sample Form I and RTQC results were compared to
the raw data of the BQC sample.
The results of the BQC samples were evaluated using the
criteria outlined in the QAPP, Exhibit A2-24, Section 3.13.2. If
two or more LCICs were out of the acceptance criteria, the BQC
sample was unacceptable. The BQC sample was considered acceptable
if fewer than two analytes were out of criteria, unless the same
analyte was out of criteria for three out of five consecutive BQC
samples. If this was the case, the three BQC samples with the
particular LCIC out of limits were unacceptable and the samples
associated with the unacceptable BQC samples were required to be
reextracted. The BQC samples were compared to the appropriate low
or high concentration acceptance windows as explained in the PE
Sample section of this report and as provided on Tables la-k.
-------
The results of the BQC samples were routinely uploaded by the
laboratories to the Bulletin Board at the beginning of the study;
however, towards the end of the study, a number of laboratories did
not send the reisults for evaluation or did not check the Bulletin
Board for the results of the evaluation by EMSL-LV. As a result,
during the validation process, a few samples were determined to be
unacceptable, which invalidated all the samples associated with
those BQC samples.
A review of the BQC sample raw data revealed that in most cases
where a "ND" (nondetect) was reported on the Form I, the analyte
either did not meet ion ratio, scan range or RRT criteria, but the
quantitation from the primary ion was within the acceptance
criteria. The BQC sample data entered into the Form I database was
condensed into summary tables by matrix, spike level, and
laboratory. When an analyte was reported as "ND" on the Form I
because it did not meet all of the identification criteria, the
concentration for the summary tables was calculated using the
primary ion. If the Form I indicated the need to use peak height or
secondary ion quantitation, the concentration reported was
calculated accordingly. A number of other anomolies with the data
were observed and are indicated as appropriate in Table A, Summary
of Flags Used to Qualify the BQC Sample Data.
Tables 1-3 present the data from the BQC samples which were
analyzed with each batch of samples. Tables la-k summarize the
percent recovery and percent relative standard deviation (% RSD) of
the individual spiking levels by laboratory and matrix. These
tables demonstrate that within the concentration ranges used, there
was no apparent concentration related recovery bias in the method.
The 0.2 ug/kg spike level (spike level E) has high recoveries for
1,2-dichlorobenzene and 1,2,4-trichlorobenzene due to traces of
these compounds in the method blanks. Tables 2a-b present the mean
percent recovery and mean % RSD averaging all spike levels by
laboratory and matrix. These tables display any interlaboratory
bias for each LCIC. Tables 3a-b are summaries of the overall mean
percent recovery and mean % RSD for the study for each LCIC by
matrix. These tables demonstrate the difference in analytical
behavior among the LCIC.
Two matrices were used for the BQC samples as described in the
Sample Preparation Section. The results of the two matrices are
reported separately to illustrate the performance of the method on
different matrices. In general, the recoveries and the % RSDs of
the LCIC from sand samples was lower than from the soil samples,
except for laboratory 7. The lower % RSDs for the sand samples was
anticipated whereas the lower recoveries were not. Neither of these
differences has an impact on the utility of the BQC samples as a QA
measure for determining that the laboratory's analytical system was
in control. The use of BQC samples was a valuable QA measure since
25
-------
many of the out of criteria samples exhibited deficiencies in other
QA criteria, such as surrogate recoveries. In general, the
laboratories performed well on the BQC samples; approximately 90
percent of the first analyses were acceptable, not including the
results from laboratory 4.
8.4 Data Validation Comments
The EMSL-LV estimated that the validation of the data for the
LCHSSA would take ten weeks if the data were received in three equal
portions, on schedule. The data packages were not received as
scheduled and the validation required approximately twelve weeks,
from mid-November, 1987 to mid-February, 1988. A number of other
problems occurred during the validation process which are discussed
below.
The Data Validation software (DVS), developed by software
contractor, was used as an organization and documentation tool
during the validation process. Due to schedule constraints, there
was inadequate time to thoroughly debug the software and properly
train the necessary personnel in its use. As a result, a number of
problems existed in the software that were identified when it was
already in use; the problems were corrected soon after they were
identified. The reports and checklists generated by the software
were used, but the data evaluation was not done at the keyboard as
anticipated. A new version of the software was provided to EMSL-LV
after data validation was complete. The data evaluation and
comments were entered into the computer retroactively so that the
information would be available in the LCHSSA database for future
reference.
The LabData Software provided to the laboratories generated the
Data Validation diskettes when the final data package forms were
produced. Many of the laboratories did not follow the directions
provided with LabData and in Project memos which caused problems
with the Data Validation software. For example, duplicate Lab Login
ID's were left in the databases. For a 16 hour analytical run, the
PC2 of the first shift was supposed to be named "PCMID"; this
convention was not followed. Unacceptable analyses were not removed
from the final database. All these anomolies caused the DVS not to
function properly.
The conventions for the laboratories to qualify their data were
not sufficiently clear in the SOW. For example, every laboratory
had low levels (less than 0.5 ppb and 0.6 ppb for 1,2-
dichlorbenzene) of some of the LCIC in the method/holding blanks.
Laboratory 1 qualified all samples associated with a blank that had
LCICs less than the 0.5/0.6 ppb with a "B" qualifier. No other
26
-------
laboratory did this. It is not clear from reading the SOW whether
the analytes should or should not be qualified with the "B". This
inconsistency still exists in the Form I database. In addition,
the laboratories did not use the "E" qualifier as defined in the
QAPP, p. A4-13, Section 2.4.A.
There is a field on the Form I for sample screening
information; the options to complete the field are MS or ECD. If
the sample was screened, that field must be completed. Only one
laboratory that participated in the study actually completed the
field. The only way for the data evaluator to determine if a sample
was screened was to search through the instrument run logs and the
screening raw data. There were too many instances where the
information was not provided on the Form I for the EMSL-LV to
request resubmissions to correct this deficiency.
The date of receipt, date of extraction, percent moisture, and
weight extracted needed for the Data Validation software were
provided from the same database as for the Sample Extraction Summary
Report, Form IX. The information entered into the database for the
Form IX was not reviewed thoroughly by some of the laboratories and,
as a result, the information on the forms and DVS did not always
match the information on the handwritten logbooks. This caused
problems during the validation process. In addition, a number of
laboratories had difficulty generating the Matrix Spike Form (Form
III) but the required resubmissions corrected the problems.
8.5 Laboratory Specific Comments
The following is a detailed discussion of the data from the
individual laboratories, including problems encountered and comments
on the overall data quality.
Laboratory 1 produced data of excellent quality. The data
packages were received on schedule. Very few resubmissions were
needed as a result of the thorough review the laboratory performed
on their own data. This laboratory was asked to participate in the
holding time study and to analyze the samples extracted by
laboratory 4. The Count of Individual Qualifiers from the Form I
Database, Table 5b, indicates a significant number of "El"
qualifiers for this laboratory. The explanation for this is
twofold. One, a few sand BQC samples that were extracted in batches
with Holding Time samples only, failed criteria and could not be
reextracted. This also accounts for the high % RSDs for the sand
samples, Table 2b. Two, the counts for laboratory 1 include the
counts for laboratory 4, which had trouble with the BQC samples.
Despite this aspect of the data, the overall work performed by
laboratory 1 was outstanding.
27
-------
Laboratory 2 had a number of specific problems. The
sensitivity of their instrument was very low, particularly for the
BHCs. As a result, many manual integrations were necessary.
Numerous resubmissions were requested for reevaluation of BHC
integrations in the method/holding blanks. In addition, the data
system did not always choose the proper peak for pyrene-dlO at m/z
213, which was not caught in-house. The SICP and quantitation
reports were resubmitted when this problem occured. Approximately
half of the Form Ill's were generated incorrectly due to a problem
with LabData and needed resubmission. Towards the end of the study,
laboratory 2 did not retrieve the results of the BQC sample
evaluations from the Bulletin Board. A number of the BQC samples
were deemed unacceptable during RTQC and data validation, which
invalidated data from numerous samples. This problem is reflected
in the number of "E7" flags for laboratory 2 in Table 5b. The
requested resubmissions were not received in a timely manner which
delayed validation of the data.
Laboratory 3 had difficulty meeting the BQC sample acceptance
criteria at the beginning of the study and three of the early
extracted batches required reextraction. As a result, the first
data package was very small and many of the samples that required
reextraction exceeded holding times. This is reflected in the
number of samples for this laboratory with "H" flags. This
laboratory did not thoroughly review their data in-house, and
numerous resubmissions were needed. At the beginning of the study,
peaks that met criteria were not reported on the Form I's; the
laboratory staff was informed of this problem and corrected the
problem for the later data packages. The data packages were put
together carelessly and often had items missing. The items listed
indicate a problem with organization and not technical quality. The
overall quality of the data was acceptable and all the requested
resubmissions were received promptly.
Laboratory 4 extracted their samples but were unable to perform
the analysis. The decision was made to send the extracts to
laboratory 1 to be analyzed. Many of the BQC samples did not meet
acceptance criteria so the remaining samples from those extraction
batches were never analyzed. The cause of the problems is unclear.
Laboratory 5 did not complete the qualification process.
Laboratory 6 did not submit the data packages on schedule.
Upon review of the first data package, it was discovered that the
proper analytical sequence was not followed for some shifts; the BQC
samples associated with the field samples were not analyzed in the
28
-------
same shifts. The laboratory was required to reanalyze the
approxiately 50 affected samples. The reanalyzed samples exceeded
the analysis holding times which accounts for the number of samples
with "H" qualifiers. The SICPs for 2-chloronaphthalene in the first
data package were not displayed on the appropriate scale and were
corrected through resubmissions. Some of the requested
resubmissions were not submitted in a timely manner which delayed
validation of the data. The majority of problems listed indicate
organizational problems and not technical problems. Overall, the
data was of acceptable quality.
Laboratory 7 had a problem with their data system and LabData
Software. As a result, the data packages were received late and in
one shipment, not three. The data system was set up so that many
manual integrations were necessary, which increased the time for
data validation. The data package was missing a few quantitation
reports and was not put together properly, but very few
resubraissions were needed. This laboratory had a problem with
internal standards out high, even after reinjection, as is reflected
in the number of "12" and "14" flags. The resubmissions that were
requested were received promptly. Overall, the data was of
acceptable quality as reflected in the data qualifier flags.
Laboratory 8 submitted the data packages late. A number of
unnecessary reextractions were performed because a surrogate or an
internal standard was within, but close to the QA criteria limit.
This caused confusion during data validation. For example, the BQC
sample was deemed acceptable by EMSL-LV but the data was not
submitted because the surrogate recovery was low, yet acceptable.
To the data evaluator, it appeared that the associated BQC sample
was not analyzed with the samples. Resubmissions for this type of
data were requested. Resubmissions were necessary for a number of
Form Ill's and Form I's where hand-written corrections were made.
The number of "H" qualifiers indicates that many reanalyzed samples
did not meet the holding time criteria. The requested resubmissions
were not submitted in a timely manner, however, most of the
deficiencies were corrected. Overall, the data was of acceptable
quality.
8.6 QA Criteria Summary
The following is a summary of the general performance of the QA
criteria.
- Initial Calibration: All the initial calibrations reported met
the QA criteria.
29
-------
- Performance Check: One PC2 failed percent valley criteria and the
remainder were acceptable.
- Continuing Calibration: All the continuing calibrations reported
met QA criteria.
- EPA Check Standard: All reported EPA check standard analyses met
the QA criteria.
- Internal Standard Area Difference: This criteria was not met
numerous times as indicated by the number of "I" qualifiers. The
failure to meet the criteria resulted in many reanalyses.
- Internal Standard Retention Time Differences: This criteria was
generally met.
- Surrogate Recovery: The number of reported surrogates recoveries
outside the QA criteria was low; however, a significant number of
samples were reextracted when the surrogate recovery did not meet
criteria.
- Matrix Spike/Matrix Spike Duplicate Samples (advisory): The
advisory recovery and RPD criteria were not consistently met. The
problems were not specific to any particular LCIC and time
constraints have not permitted a detailed examination of this data
by the EMSL-LV at this time.
- Method/Holding Blanks: The method/holding blanks generally met
the QA criteria after the detergent problem was solved. See
Section 4.4 for a detailed discussion of the blanks.
9.0 SUMMARY
The results of the data validation indicated that the QA
measures incorporated into the QAPP were effective in the production
of data of high and known quality. The combination of project
provided reagents and glassware, analyst training, rigorous QA
requirements and monitoring of the laboratories during the study was
used to reduce interlaboratory variation and to ensure the quality
of the data.
The method modifications introduced to lower the detection limit
and reduce interferences succeeded in lowering the effective
detection limit from approximately one ppb to approximately 0.3 ppb.
This resulted in many more samples with detected LCIC than in the
Pilot Study.
The training of the extraction analysts in the use of the method
by the chemists who developed it was instrumental in reducing the
time required for the laboratories to become competent in using the
new method. This, in combination with the analysis of PE samples,
resulted in the laboratories having fewer startup problems than
anticipated.
30
-------
The RTQC system performed adequately. The schedule demands of
the project required the system to be used before the software was
completely ready. This resulted in some delays in uploading data
for review. The system was of great value in the transmisson of the
Blind Quality Control sample results to the EMSL-LV for scoring.
The method/holding blank interpretation rules were successful in
permitting the use of large amounts of data which would have been
classified as having blank interference if only the usual blank
definition was used.
The data validation proceeded nearly as planned; some schedule
slippage was experienced due to delays in receipt of materials from
the laboratories. The detailed examination of the data showed that
when data defects were present the quality assurance measures
incorporated in the protocol indicated a problem. This indicates
that no additional quality assurance measures were needed.
The data produced by the laboratories is of known quality with a
large proportion of it suitable for statistical use in the LCHSSA.
The laboratories in general performed very well on a tight schedule
with a heavy sample load. The detailed comparison of the data
produced to the project Data Quality Objectives is presented
elsewhere in the LCHSSA project report.
31
-------
SUMMARY TABLE FOR LOVE CANAL SOIL BQC SAMPLES BY SPIKE LEVEL
SPIKE LEVEL A
TABLE la
SPIKE X RECOVERY
ANALYTE CONC. ACCEPTANCE
(ug/kg) WINDOW
1.2-D1CHLOR08ENZENE 0.60 56.1. - 170.5
1.2.4-TRJCHLOROBENZENE 0.1.5 53.9 - 162.2
1,2,3,4-TETRACHLOROBENZENE 0.50 50.7 - 107.7
AI.PHA-BHC 0.70 1.0. 0 - 125. 4
DELTA-BHC 0.60 28.6 - 184.5
BETA-BHC 0.75 11.2 - 168.9
GAMMA-BHC 0.65 J5.1 - 120.0
i ,4-DiBROMOBENZENE - 51.0 - 104.0
2,4,6-TRIBROMOBIPHENYL - 56.0 - 126.0
1,2.4,5-TETRABROMOBENZENE - 50.0 - 115.0
IAB 1
N MEAN XRSD
5 109.7 4.)
5 87.6 0.6
i 65.6 2.7
5 69 . 7 4.6
5 54. 0 5.9
S 62.7 5.2
5 67.ii 6.4
5 71.2 2.1.
5 78.1 8.3
5 64.9 6.0
MB 2
N MEAN XRSn
6 89.2 1.3
6 8S.9 1.0
6 70.3 (..0
6 60.2 9.9
6 50.8 6.5
6 56.7 13.8
6 57.9 8.6
6 85.1. |3.6
6 97.3 15.1
6 89.0 17.5
LAB 3
N MEAN XRSD
3 108.9 4.4
1 Ml-,! 3,9
3 93.3 2.7
3 11.9 8.4
3 77.2 8.5
3 lll.l 42.1
3 8G.5 16.0
3 74.2 0.9
3 97.0 12.4
3 86.2 7.2
UB 4
N MKAN XKSD
2 88.3 0.0
2 S!.! 0.»!
2 52.0 1.4
2 69 . 3 7.9
2 41.7 10.2
2 54.0 12.8
2 53.8 5.1
2 71.8 1.1
2 77.8 12.9
IAB 6
N MEAN XRSD
3 126.1 15.2
3 113.3 2.1
3 92.0 2.0
3 77.6 4.6
3 78.9 7.8
3 80.4 25.0
} 79.0 6.2
3 82.0 7.4
3 90.6 15.0
LAB 7
N MKAN XRSD
3 92.2 2.6
3 82. 2 I.!
3 64.0 0.8
3 57.1 3.2
3 60.6 3.2
3 81.8 9.0
3 63.6 5.8
3 73.0 13.9
3 92.8 36.6
LAB 8
N MEAN XRSD
2 130.8 1.6
2 ICi.i C.S
2 83.0 3.8
2 69.3 3.6
2 71.7 7.1
2 79.3 2.0
2 75.4 4.9
2 87.8 6.7
2 103.5 0.7
2 88.2 1.1
TABLE Ib
SPIKE LEVEL B
SPIKE X RECOVERY
ANALYTE CONC. ACCEPTANCE
(ug/kg) WINDOW
1,2-DICHLOROBENZENE 1.20 42.3 - 112.2
1,2,4-TRICHl.ORnBENZENE 0.90 45.6 - 109.3
1,2,3,4-TETRACHLOROBENZENE 1.00 44.3 - 93.1
2-CHLORONAPHTHALENE 0.80 30.6 - 74.8
ALPHA-BHC 1.40 37.1 - 116.2
DELTA-BHC 1.20 17.8 - 148.2
BETA-BHC 1.50 44.6 - 128.0
3AMMA-BHC 1.30 42.4 - 122.6
1,4-DIBROHOBENZENE - 51.0 - 104.0
2,4,6-TRIBROMOBIPHENYL - 56.0 - 126.0
1,2,4,5-TETRABROHOBENZENE - 50.0 - 115.0
LAB 1
N MEAN XRSD
8 72.7 14.0
8 68.3 6.5
8 6O.5 10.6
8 49.2 8.9
8 57.9 27.9
8 46.6 20.5
8 50.7 22.6
8 53.3 17.0
8 71.2 3.9
8 75.0 8.8
8 68.1 10.8
LAB 2
N MEAN XKSD
16 77.8 17.1
16 74.3 10.9
16 61.1 12.0
16 47.1 11.8
16 61.7 33.4
16 51.6 40.0
16 57.2 50.4
16 60.3 30.2
16 8H.I 7.0
16 104.2 11.6
16 92.4 9.2
LAB 3
N MEAN XRSD
8 79.1 24.4
8 92.4 9.2
8 81.5 8.8
8 46.1 9.0
8 74.5 15.1
8 86.5 61.8
8 73.9 29.7
8 78.6 21.0
8 68.3 7.2
8 96.0 8.7
8 137.6 93.0
I.AB 4
N MKAN XKSD
1 54.2
1 56.7
1 52.0
1 53.7
1 47.9
1 30.8
1 42.7
1 41.5
I 86.5
1 92.0
1 72.7
LAB 6
N MKAN XRSD
7 88.5 14.1
7 86.7 7.4
7 81.3 9.9
7 57.1 7.9
7 76.6 21.4
7 75.0 21.8
7 89.4 123.0
7 77.1 20.8
7 74.5 9.0
7 92.2 14.8
7 79.1 6.0
LAB 7
N MEAN XRSD
4 70.0 11.3
4 70.3 13.6
4 67.5 13.6
4 47.5 14.2
4 64.5 29.2
4 55.4 28.8
4 70.2 40.4
4 66.9 32.8
4 89.2 6.9
4 111.3 14.7
4 95.9 7.0
LAB 8
N MEAN XRSD
4 86.5 41.9
4 75.8 19.5
4 67.0 23.0
4 38.4 30.3
4 73.8 51.1
4 67.9 26.1
. 4 76.8 58.9
4 73.7 33.6
4 79.7 5.3
4 89.0 21.)
4 90.2 19.5
32
-------
SUMMARY TABLE FOR LOVE CANAL SOIL BQC SAMPLES BY SPIKE LEVEL
SPIKE LEVEL C
TABLE U
SPIKE * RECOVERY
ANALYTE OONC. ACCEPTANCE
(ug/kg) WINDOW
1,2-DICHLOROBENZEHE 1.80 1.2.3 - 112.2
1,2,4-TRICHLOROBENZENE 1.35 1.5.6 - 109.3
1,2,3,4-TETRACHLOROBENZENE 1.50 44.3 - 93.3
2-CHLORONAPHTHALENE 1.20 30.6 - 71.. 8
AI.PHA-BHC 2.10 37.1 - 116.2
DELTA-BHC 1.80 17.8 - 11.8.2
BETA-BHC 2.25 1.1.. 6 - 128.0
GAMMA-BHC 1.95 1.2.6 - 122.6
1.1.-DIBROHOBENZENE - 51.0 - 101.. 0
2,4,6-TRIBROMOBIPHENYL - 56.0 - 126.0
1,2,4,5-TETRABROMOBENZENE - 50.0 - 115.0
LAB 1
N MEAN XRSD
-
2 71.1 3.6
2 68.1 5.6
2 61.. 3 1.6
2 48.3 7.0
2 63.1 40.0
2 61.1. 51.8
2 59.3 2.7
2 63.3 45.7
2 73.7 2.8
2 82.9 10.9
2 68.9 8.6
LAB 2
N MEAN XRSD
5 67.1 10.3
5 63.4 8.3
5 65.9 34.2
5 47.2 14.2
5 58.4 28.0
5 45.8 46.0
5 59.6 88.0
5 58.6 33.4
5 87.3 12.2
5 90.3 5.7
5 84 . 3 8.8
LAB 3
N MEAN XRSD
2 77.8 9.8
2 95.6 6.0
2 88.0 12.1
2 34.2 0.0
2 68.3 45.6
2 59.4 21.4
2 69 . 3 73.4
2 68.7 68.2
2 73.3 4.5
2 91.8 9.5
2 104.0 11.3
IAB 4
N MKAN XRSD
1 53.3
1 60.7
1 50.0
1 46.7
1 49.5
1 42.2
1 45.8
1 47.2
1 '2.6
1 75.8
1 78.7
IAB 6
N KEAH %RSD
2 53.6 7.1
2 64.4 8.9
2 67.7 14.1
2 46.3 12.8
2 72.9 8.2
2 61.1 141.6
2 74.7 8.5
2 75.9 32.7
2 67.7 3.0
2 85.0 8.2
2 77.9 23.4
LAB 7
N MEAN XRSD
3 67.6 15.1
3 72.6 22.6
3 66.7 27.4
3 35.8 15.3
3 65.1 96.5
3 61.3 61.6
3 69.6 84.0
3 65.0 38.6
3 88.1 15.3
3 102.0 10.5
3 76.9 17.4
LAB 8
N MEAN XRSD
2 78.3 6.5
2 77.4 1.2
2 64.7 6.6
2 24.6 107.0
2 85.0 54.2
2 81.4 120.4
2 80.7 151.9
2 84.4 89.9
2 83.0 15.6
2 97.5 3.5
2 94.4 23.5
TABLE Id
SPIKE LEVEL D
SPIKE X RECOVERY
ANALYTE CONC. ACCEPTANCE
(ug/kg) WINDOW
1,2-DICHLOROBENZENE 2.40 42.3 - 112.2
1,2,4-TRICKLOROBENZEHE 1.80 45.6 - 109.3
1,2,3,4-TETRACHLOROBENZENE 2.00 44.3 - 93.3
2'CHLORONAPHTHALENE 1.60 30.6 - 74.8
ALPHA-BHC 2.80 37.1 - 116.2
DELTA-BHC 2.40 17.8 - 148.2
BETA-BHC 3.00 44.6 - 128.0
GAMMA-BHC 2.60 42.4 - 122.6
1,4-DIBROMOBENZENE - 51.0 - 104.0
2,4,6-TRIBROMOBIPHENYL - 56.0 - 126.0
1,2,4,5-TETRABROMOBENZENE - 50.0 - 115.0
IAB 1
N MEAN XRSD
5 67.8 87.4
5 68.7 51.5
5 64.1 68.2
5 52.2 41.2
5 61.4 92.8
5 56.7 42.7
5 61.6 25.9
5 60.1 85.9
5 71.7 8.2
5 70.4 12.2
5 73.4 9.0
IAB 2
N MEAN XRSD
7 68.0 70.7
7 70.4 56.8
7 62.6 63.8
7 49.6 21.4
7 59.8 93.8
7 46.3 96.7
7 49.0 174.0
/ 56.9 102.3
/ 81.7 9.9
7 89.1 13.3
7 69.6 6.9
LAB 3
N MEAN XRSD
6 60.3 80.6
6 73.3 56.9
6 70.3 46.3
6 37.4 110.3
6 75.5 140.6
6 54.4 134.2
6 71.0 158.3
6 77.3 154.1
6 62.1 6.0
6 91.6 17.5
6 92.6 14.5
LAB 4
N MEAN XRSD
-
-
-
-
-
-
LAB 6
N MEAN XRSD
6 67.6 104.4
6 79.3 30.4
6 80.3 41.9
6 42.6 106.4
6 85.4 144.0
6 86.0 135.3
6 102.0 424.7
6 94.8 114.5
6 80.4 6.2
6 94.8 11.7
6 99.8 43.6
IAB 7
N MEAN XRSD
3 66.5 83.5
3 67.6 12.6
3 75.8 66.0
3 60.8 50.5
3 65.2 249.2
3 62.9 150.1
3 72.7 224. q
3 64.9 182.8
3 83.7 9.4
3 103.3 40.6
3 87.3 20.0
LAB 8
N MEAN XRSD
4 76.1 92.6
4 77.9 23.2
4 67.3 39.3
4 45.9 34.9
4 83.7 118.6
4 69.8 90.7
4 73.2 135.0
4 83.8 82.2
4 84.7 7.1
4 1O0.4 3.5
4 91.5 12.4
33
-------
SUMMARY TABLE FOR LOVE CANAL SOIL BQC SAMPLES BY SPIKE I.EVEL
SPIKE LEVEL E
TABI£ le
SPIKE X RECOVERY
ANALYTE CONC. ACCEPTANCE
(ng/kg) WINDOW
1,2-DICHLOROBENZENE 0.20 56.1. - 170.5
1,2,4-TRICHLOROBENZENE 0.20 S3. 9 - 162.2
i,2,3,4-tETKACHLOROBENZENE 0.20 50.7 - 107.7
2-CHLORONAPHTHALENE 0.20 31.7 - 98.5
ALPHA-BHC 0.20 40.0 - 125.1.
DELTA- BHC 0.20 28.6 - 18I>.5
BETA-BHC 0.20 11.2 - 168.9
GAMMA-BHC 0.20 35.1 - 120.0
1,4-DIBROMOBENZENE - 51.0 - 104.0
2,4,6-TRIBROHOBlPHENYL - 56.0 - 126.0
1,2,4,5-TETRABRONOBENZENE - 50.0 - 115.0
1
]
1
1
LAB 1
1 MEAN XRSD
200.0 0.6
106.7 0.5
76.7 O.li
78.3 0.8
61.7 0.2
1.5.0 0.8
36.7 3.5
61.7 1.0
78.0 4.6
79.8 7.7
74.0 2.5
I>B 2
N KEAN XRSD
2 210.0 0.5
2 107.5 0.1
2 72.5 0.6
2 55.0 1.5
2 52.5 0.3
2 1.2.5 1.7
2 67.5 0.6
2 67.5 0.6
2 88.7 7.1
2 82.3 6.2
2 96.5 6.B 3
N HKAN XRSD
-
-
-
-
-
-
-
.
-
I.AB 1,
N HKAN USD
.
-
-
-
-
-
-
-
-
-
-
IAB 6
N MEAN XRSD
1 165.0
1 110.0
1 105.0
1 65.0
1 60.0
1 90.0
1 55.0
1 70.0
1 77.0
1 81.6
1 73.0
LAB 7
N MEAN XRSD
2 l'.0.0 0.6
2 107.5 0.4
2 87.5 0.2
2 62.5 0.7
2 52.5 0.3
2 37.5 0.4
2 21.2.5 2.6
2 80.0 0.0
2 98.0 0.1.
2 117.0 12.7
2 97.9 11.5
LAB 8
N MEAN XRSD
1 380.0
1 145.0
1 85.0
1 35.0
1 70.0
1 90.0
1 95.0
1 60.0
1 77.4
1 84.8
1 73.1
TABLE If
SPIKE LEVEL F
SPIKE X RECOVERY
ANALYTE CONC. ACCEPTANCE
(»g/kg) WINDOW
1,2-DICHLOROBENZENE 0.50 56.4 - 170.5
1,2,4-TRICHLOROBENZENE 0.50 53.9 - 162.2
1,2,3,4-TETRACKLOROBENZENE 0.50 50.7 - 107.7
2-CHLORONAPHTHALENE 0.50 31.7 - 98.5
ALPHA-BHC 0.50 40.0 - 125.4
DELTA-BHC 0.50 28.6 - 184.5
BETA-BHC 0.50 11.2 - 168.9
GAHNA-BHC 0.50 35.1 - 120.0
1,4-DIBROMOBENZENE - 51.0 - 104.0
2,4,6-TRIBROMOBIPHENYL - 56.0 - 126.0
1,2,4,5-TETRABROMOBENZENE - 50.0 - 115.0
LAB 1
N MEAN XRSD
3 108.0 0.0
3 69.3 1.1
3 57.3 2.Z
3 58.7 2.5
3 55.3 0.5
3 51.3 0.6
3 58.7 3.4
3 55.3 1.4
3 75.7 10.3
3 75.3 13.3
3 71.0 7.2
LAB 2
N MEAN XRSD
2 116.0 0.6
2 89.0 5.2
2 75.0 4.2
2 41.0 4.3
2 80.0 2.7
2 64.0 3.3
2 84.0 4.2
2 84.0 5.9
2 100.8 3.2
2 122.5 0.7
2 105.5 0.7
LAB 3
N MEAN XRSD
1 108.0
1 106.0
1 88.0
1 56.0
1 76.0
1 84.0
1 60.0
1 82.0
1 68.0
1 98 . 1
1 94 . 3
I>B 4
N MEAN XRSD
1 116.0
1 76.0
1 52.0
1 58.0
1 56.0
1 30.0
1 30.0
1 40.0
1 81. /
1 76.3
1 79.1
LAB 6
N MEAN XRSD
2 95.0 1.9
2 109.0 2.3
2 85.0 0.4
2 52.0 0.0
2 74.0 0.0
2 75.0 0.5
2 71.0 1.5
2 87.0 0.4
2 73.2 0.9
2 77.6 1.2
2 76.9 4.5
LAB 7
N MEAN XRSD
2 110.0 9.0
2 80.0 5.3
2 66.0 1.1
2 68.0 4.2
2 55.0 1.9
2 49.0 3.6
2 7O.O 2.0
2 58.0 1.2 '
2 86.3 2.1
2 110.0 15.5
2 93.9 4.0
LAB 8
N MEAN XRSD
5 152.0 2.6
5 75.6 3.1
5 58.8 5.1
5 54.8 7.7
5 52.8 7.4
5 51.6 4.2
5 69.2 5.8
5 60.0 9.9
5 84.0 12.1
5 86.6 15.5
5 82.4 13.5
-------
SUMMARY TABU: FOR LOVE CANAL SOIL BQC SAMPLES BY SPIKE LEVEL
SPIKE I.EVFX G
TABLE Ig
SPIKE X RECOVERY
ANALYTE CONC. ACCEPTANCE
(ug/kg) WINDOW
1,2-DICHI/OROBENZENE 2.OO b2.3 - 112.2
l,2,Ii-TRICHLOROBENZENE 2.00 1.5.6 - 109.3
l,2,3,li-TETRACHLOROBENZErIE 2.00 M..3 - 93.1
2-CHLORONAPHTflALENE 2.00 30.6 - 7 (..8
ALPHA-BHC 2.00 37.1 - 116.2
DELTA- BHC 2.00 17.8 - KB. 2
BETA-BHC 2.00 fcl>.6 - 128.0
GAftlA-BHC 2.00 1.2.1. - 122.6
1 .li-DlBROHOBENZENE - 51.0 - lOfc.O
2,4,6-TRIBROMOBIPHENYL - 56.0 - 126.0
1,2,11,5-TETRABROMOBENZENE - 50.0 - 115.0
LAB 1
N HEAN XRSD
<• 1.7.0 158.9
fc 52.9 39.2
l> 57.9 13.3
it >il. 6 1.0.5
* 59.3 fcO.2
* 51.5 18.2
k 57.8 1.0.6
It 59. b I.4.5
I. 73.9 2.8
<> 82. fc 2.9
* 71.9 4.8
IAB 2
N HEAN XRSD
.
-
-
.
.
.
.
-
-
LAB 3
N HEAN XRSD
2 1.3.5 231.. 1
2 61.3 6.9
2 67.0 109.8
2 38.3 166.1.
2 80.0 152.0
2 76.0 152.6
2 82.8 165.8
2 87.8 162.8
2 70.7 9.3
2 86.8 2.6
2 89.7 18.9
LAB t.
N HEAN XRSD
1 57.5
1 51.0
1 1.3.5
1 38.5
1 42.0
1 27.0
1 32.5
1 37.5
1 75.2
1 76.9
1 74.1
LAB 6
N MEAN XRSD
1 62.5
1 78.5
1 11,. 5
1 1.9.5
1 80.0
1 80.0
1 75.0
1 89.0
1 79.5
1 90.2
1 95.5
LAB 7
N HEAN XRSD
1 1.9.0
1 50.0
1 53.5
1 37.0
1 51.5
1 51.5
1 61.0
1 56.5
1 92.7
1 109.0
1 88.3
LAB 8
N MEAN XRSD
2 82.5 17.1
2 63.0 18.0
2 61.. 5 8.8
2 30.8 It. 6
2 76.5 18.5
2 60.8 86.1
2 83.8 15.2
2 82.8 18.8
2 75.6 1..7
2 80.1. 12.7
2 79.7 2.8
35
-------
SIM4AKY TABLE FOR IjOVE CANAL SAND BQC SAMPLES BY SPIKE LEVEL *
SPIKE LEVEL A
TABLE Ih
SPIKE \ RECOVERY
ANALYTE CONC. ACCEPTANCE
(ug/kg) WINDOW
1,2-DlCHLOROBENZENE 0.60 56. 4 - 170. S
!,J,!i-TR!CKLOROBEfiZEKE O.'-sS 53.9 - 162.2
1,2,3,4-TETRACHLOROBENZENE O.SO SO. 7 - 107.7
2-CHLORONAPHTHALENE 0.1.0 31.7 - 98. S
ALPHA-BHC 0.70 40.0 - 125.4
nn.TA-BHC 0.60 28.6 - 184.5
BETA-BHC 0.7S 11.2 - 168.9
GAHHA-BHC 0.65 35.1 - 120. 0
1,4-DIBROMOBENZENE - 51.0 - 104.0
2,14,6-TRIBROHOBIPHENYL - 56.0 - 126.0
1,2,4,5-TETRABROMOBENZENE - 50.0 - 115.0
LAB 1
N MEAN XRSD
2 83.3 33.9
2 Sii-t 26,0
2 44.0 12.9
2 65.0 16.3
2 56.4 1.8.3
2 53.3 1.8.6
2 60.7 d5.1
2 55. <4 51.1
2 80.<. 7.3
2 83.2 30.3
2 76. l> 33. 4
LAB 2
N MEAN XRSD
2 76.7 3.1
? 36-J 4.1
2 24.0 0.0
2 31.3 5.7
2 23.6 4.3
2 25.8 t>.6
2 26.0 3.6
2 30.0 3.6
2 95.1 1.0
2 116.5 1.8
2 90.5 8.7
I.AB 3
N MEAN XRSD
2 81.7 31.7
2 75.6 16.6
2 57.0 27.3
2 61.2 26.0
2 52.1 5.8
2 43.3 5.1.
2 56.7 8.3
2 53.8 0.0
2 85.7 3.4
2 101.5 0.7
2 92.5 3.5
UB 6
N MEAN XRSD
2 97.5 18.1
2 87.8 23.3
2 79.0 16.1
2 83.8 23.2
2 80.7 11.3
2 84. 2 4.2
2 84.0 6.7
2 80.8 1.3
2 85.3 11.2
2 79.5 19.3
2 78.1 6.5
LAB 7
N MEAN XRSD
1 90.0
1 84.1.
1 88.0
1 87.5
1 97.1
1 71.7
1 100.0
1 96.9
1 79.7
1 105.0
1 104.0
LAB 8
N MEAN XRSD
2 145.8 13.7
2 92.2 22.2
2 78.0 43.5
2 85.0 41.6
2 70.0 31.7
2 73.3 45.0
2 90.7 31.2
2 74.6 39.4
2 100.5 0.7
2 96.3 17.3
2 99.8 3.2
TABLE 11
SPIKE LEVEL B
SPIKE X RECOVERY
ANALYTE CONC. ACCEPTANCE
(ug/kg) WINDOW
1.2-DICHUIROBENZENE 1.20 42.3-112.2
1,2,4-TDICHLOROBENZENE 0.90 45.6 - 109.3
2-CHLORONAPHTHALENE 0.80 30.6 - 74.8
ALPHA-BHC 1.40 37.1 - 116.2
DELTA-BHC 1.20 17.8 - 148.2
BETA-BHC 1.50 44.6 - 128.0
GAMMA-BHC 1.30 42.4 - 122.6
1,4-DIBROHOBENZENE - 51.0 - 104.0
2,4,6-TRIBROMOBIPHENYL - 56.0 - 126.0
1,2,4,5-TETRABROMOBENZENE - 50.0 - 115.0
LAB 1
N MEAN XRSD
2 63.3 0.0
2 49.4 1.6
2 46.3 3.8
2 47.5 3.2
2 44.2 10.7
2 51.0 2.8
2 48.1 5.7
2 76.6 0.0
2 75.4 2.1
2 73.0 5.2
LAB 2
N MEAN XRSD
3 57.5 6.3
3 b5.6 21.1
3 35.4 17.4
3 38 . 8 17.1
3 28.1 14.0
3 38.9 22.3
3 43.8 16.7
3 R6.4 2.6
3 103.2 15.3
1 92.0 4.3
LAB 3
N MEAN XRSD
3 68.6 5.6
3 '8.9 6.5
3ff. •« SI
DO t J j . j
3 59.6 8.5
3 72.4 11.1
3 71.1 15.3
3 87.1 12.5
3 85.9 13.9
3 78.3 11.6
3 103.7 13.1
3 96.1 1.6
LAB 6
N MEAN XRSD
1 81.7
1 76.7
• if, n
1 IO .U
1 73.7
1 66.4
1 67.5
1 71.3
1 69.2
1 75.5
1 74.5
1 66.6
LAB 7
N MEAN XRSD
-
-
-
-
-
LAB 8
N MEAN XRSD
1 110.8
1 80.0
1 84 0
1 80.0
1 85.7
1 55.8
1 . 85.3
1 83.1
1 103.0
1 118.0
1 94.1
* LAB 4 did not analyze sand BQC samples.
36
-------
SUMMARY TABU FOR LOVE CANAL SAND BQC SAMPLES BY SPIKE LEVEL
SPIKE LEVEL C
TABLE Ij
SPIKE * RECOVERY
ANALYTE CONC. ACCEPTANCE
(ug/kg) WINDOW
_
1,2-DICHLOROBENZENE 1.80 1.2. } - 112.2
1,2,I.-TRICHLOROBENZENE 1.35 45.6 - 109.3
1,2,3,4-TETRACHLOROBENZENE 1.50 44.3 - 93.3
2-CHLORONAPHTHALENE 1.20 30.6 - 74.8
ALPHA-BHC 2.10 37.1 - 116.2
DELTA-BHC 1.80 17.8 - 148.2
BETA-BHC 2.25 44.6 - 128. 0
GAMMA-BHC 1.95 1.2. 4 - 122.6
1,4-DIBROMOBENZENE - 51.0 - 104.0
2,4,6-TRIBRONOBIPHENYL - 56.0 - 126.0
l,2, 10.7
' 15.2
LAB 7
N MEAN XRSD
1 98.3
1 84.4
1 90.0
1 77.5
1 86.2
1 77.2
1 96.4
1 93.3
1 94.4
1 119.0
1 109.0
LAB 8
N MEAN XRSD
2 87.8 1.8
2 71.1 10.3
2 66.7 1.4
2 62.1 0.9
2 62.4 3.2
2 35.0 2.2
2 62.2 1.0
2 67.7 11.8
2 103.4 9.1
2 90.4 24.4
2 91.5 5.6
TABLE Ik
SPIKE LEVEL D
SPIKE X RECOVERY
ANALYTE CONC. ACCEPTANCE
(ug/kg) WINDOW
1,2-DICHLOROBENZENE 2.40 42.3 - 112.2
1,2,4-TRICHLOROBENZENE 1.80 45.6 - 109.3
1,2,3,4-TETRACXLOROBENZENE 2.00 44.3 - 93.3
2-CHLORONAPHTHALENE 1.60 30.6 - 74.8
AI.PHA-BHC 2.80 37.1 - 116.2
DELTA-BIIC 2.40 17.8 - 148.2
BETA-BHC 3.00 44.6 - 128.0
GAMMA-BHC 2.60 42.4 - 122.6
1,4-DIBROMOBENZENE - 51.0 - 104.0
2,4,6-TRIBROHOBIPHENYL - 56.0 - 126.0
1,2,4,5-TETRABROMOBENZENE - 50.0 - 115.0
LAB 1
N MEAN XRSD
1 24.2
1 26.7
1 35.0
1 41.9
1 41.4
1 38.3
1 47.3
1 44.2
1 75.2
1 70.7
1 63.5
LAB 2
N HFAN XRSP
-
-
-
-
-
-
-
.
I>B 3
N MEAN XKSD
2 74.6 0.8
2 75.3 1.6
2 75.0 0.9
2 59.1 0.7
2 82.3 2.1
2 60.6 2.4
2 86.7 8.2
2 88.3 3.4
2 80.1) 0.5
2 91.1 2.5
2 91.5 0.5
LAB 6
N MEAN XRSD
1 46.7
1 53.3
1 50.5
1 46.9
1 48.6
1 47.5
1 51.7
1 53.8
1 84.5
1 92.8
1 74.7
LAB 7
N MEAN XRSD
1 93.3
1 80.0
1 83.0
1 76.3
1 73.6
1 74.6
1 93.0
1 99.2
1 82.9
1 92.1
1 83.1
LAB 8
N MEAN XRSD
1 64 . 2
1 53.3
1 53.5
1 46.9
1 48.9
1 35.4
1-49.7
1 49.2
1 100.0
1 114.0
1 95.5
* LAB 4 did not analyze sand BQC samples.
37
-------
IAVE CANAL SUMHARY SOIL BQC STATISTICS TABLE BY LABORATORY
TABLE 2a
ANALYTE
1 ,2-DICHLOROBENZENE
I ,3 ,i-TRICKLOROBEKZENE
1 , 2 , 3 ,l> -TETRACHLOROBENZENE
2 -CHLORONAPHTHALENE
ALPHA-BHC
DELTA-BHC
BETA-BHC
GAMHA-BHC
1,4-D1BRQMOBENZENE
2,4,6-TRIBROHOBIPHENYL
1 , 2 ,li , 5-TETRABROHOBENZENE
LAB 1
HEAN MEAN
N X REC. * RSD
30 90.8 1.0.5
30 73.5 16.2
30 63.2 16.8
30 54.0 15.6
30 61.1 31.8
30 51. 5 19.6
30 55. !• 17.5
30 59.3 29.1
30 72.9 4.9
30 76.8 9.0
30 69.9 7.6
I.AB 2
HEAN HEAN
N \ REC. X RSD
38 85.3 21.8
*O If- E • f. C-
JtJ 1 V , J M\l . V
38 64. 8 22.2
38 1.9.7 11.3
38 61.2 36.7
38 1.9.9 42.0
38 57.9 67.3
38 60.7 37.7
38 .87.1 9.1
38 98.3 10.8
38 87.5 9.5
IAB 3
HEAN HEAN
N X REC. X RSD
22 76.0 56.2
11 rtn n i •» f.
22 79.6 28.6
22 43.7 50.9
22 74.4 66.0
22 72.9 79.7
22 77.9 85.3
22 78.6 76.3
22 68.1 5.9
22 93.8 11.3
22 108.9 43.5
MB 4
HEAN HEAN
N % REC. X RSD
6 76.3 0.0
6 57.8 O.I*
6 50.3 1.4
6 51.6 1.5
6 55.7 7.9
6 35.6 10.2
6 43.2 12.8
6 45.7 5.1
6 /6.6 1.1
6 79.4 12.9
6 74.7 1.4
MB 6
HEAN HEAN
N X REC. X RSD
22 87.6 39.4
22 39. 0 13.1
22 82.3 17.8
22 53.1 36.6
22 78.0 52.2
22 78.2 63.6
22 86.4 175.2
22 83.2 45.9
22 76.7 6.5
22 90.1 11.9
22 85.1 18.8
LAB 7
HEAN HEAN
N X REC. X RSD
18 83.8 21.7
IS 76.3 10.3
IB 69.4 19.9
18 51.0 15.7
18 60.4 68.7
18 55.6 45.2
18 91.0 66.1
18 65.6 48.0
18 86.2 8.7
18 105.7 22.3 '
18 88.0 14.1
LAB 8
HEAN HEAN
N X REC. X RSD
20 118.7 31.7
2G 81.1 11.9
20 67.0 16.5
20 44.9 27.7
20 71.3 45.7
20 66.3 48.2
20 76.4 60.1
20 73.7 38.9
20 82.4 8.6
20 91.9 11.1
20 86.8 13.2
LOVE CANAL SUHHARY SAND BQC STATISTICS TABLE BY LABORATORY
ANALYTE
1 ,2-DICHLOROBENZENE
1 ,2 ,4-TRICHLOROBENZENE
1,2, 3,(. -TETRACHLOROBENZENE
2 - CHLORONAPHTHALENE
ALPHA-BHC
DELTA-BHC
BETA-BHC
GAHHA-BHC
1,4-DIBROHOBENZENE
2 ,4 ,6-TRIBROHOBIPHENYL
1,2,4 ,5-TETRABROHOBENZENE
LAB 1
MEAN HEAN
N X REC. X RSD
7 61.4 22.0
7 47.8 26.3
7 44.0 25.9
7 54.1 27.7
7 52.1 32.1
7 49.1 39.0
7 56.4 33.1
7 51.8 33.0
7 80.1 2.9
7 81.5 13.8
7 75.9 14.9
LAB 2
HEAN HEAN
N X REC. X RSD
8 62.2 9.3
8 46.0 13.2
8 36.1 12.4
8 34.1 12.1
8 34.7 8.3
8 29.2 8.4
8 41.3 11.2
8 41.0 10. 8
8 91. O 2.3
8 103.2 9.1
8 89.8 5.9
TABLE 2b
LAB 3
HEAN HEAN
N X REC. X RSD
10 68.4 17.8
10 69.2 17.8
10 61.9 19.2
10 55.8 19.0
10 63.7 14.1
10 56.3 13.8
10 71.6 15.5
10 70.6 12.7
10 80.6 4.8
10 97.6 8.4
10 92.6 2.2
IAB 6
HEAN HEAN
N X REC. X RSD
7 78.8 13.8
7 75.8 16.2
7 72.1 14.3
7 71.9 18.9
7 69 . 6 11.3
7 70.1 6.6
7 72.9 10.1
7 71.6 6.2
1 80.2 11.5
7 81.5 14.1
7 78.8 11.7
LAB 7
HEAN HEAN
N X REC. X RSD
3 93.9
3 83.0
87.0
80.4
85.6
74.5
96.5
3 96.5
3 85.7
3 105.4
3 98.7
LAB 8
MEAN HEAN
N X REC. X RSD
6 107.0 5.2
6 76.7 10.8
6 71.1 15.0
6 70.2 14.2
6 66.6 11.7
6 51.) 15.7
6 73.5 ' 10.7
6 69.5 17.0
6 101.8 3.3
6 100.9 13.9
6 95.3 2.9
* LAB 4 did not analyze sand BQC samples.
38
-------
LOVE CANAL SUMMARY SOIL BQC STATISTICS TABLE
TABLE 3a
ANALYTE
1 , 2-DICHLOROBENZENE
1,2, A -TRICHLOROBENZENE
1,2,3, A -TETRACHLOROBENZENE
2 - CHLORONAPHTHALENE
ALPHA-BHC
DELTA-BHC
BETA-BHC
GAMMA -BHC
1,4-DIBROMOBENZENE
2 , 4 , 6-TRIBROMOBIPHENYL
1,2,4, 5 -TETRABROMOBENZENE
i
N
156
156
156
156
156
156
156
156
156
156
156
ALL LABS
MEAN
% REC.
89.1
79.9
69.3
49.8
66.4
59.6
69.9
67.8
79.1
91.7
86.3
MEAN
% RSD
33.1
14.8
19.6
23.5
45.8
46.0
72.3
42.7
7.0
12.1
15.9
LOVE CANAL SUMMARY SAND BQC STATISTICS TABLE
TABLE 3b
ANALYTE
1, 2-DICHLOROBENZENE
1 , 2 , 4-TRICHLOROBENZENE
1,2,3 , 4-TETRACHLOROBENZENE
2 - CHLORONAPHTHALENE
ALPHA-BHC
DELTA-BHC
BETA-BHC
GAMMA -BHC
1,4-DIBROMOBENZENE
2 , 4 , 6-TRIBROMOBIPHENYL
1,2,4, 5 -TETRABROMOBENZENE
N
41
41
41
41
41
41
41
41
41
41
41
ALL LABS
MEAN
% REC.
75.3
64.3
58.7
57.9
59.1
52.8
65.4
63.5
86.0
94.2
87.7
MEAN
% RSD
13.1
16.0
16.5
17.2
14.5
15.7
16.0
15.2
4.5
11.0
6.4
39
-------
COUNT OF DATA VALIDATION SUMMARY QUALIFIERS
FROM FIELD SAMPLES IN FORM I DATABASE
TABLE 5a
FLAG
G
U
B
X
TOTAL
LAB 1
1351
45
212
0
1608
LAB 2
991
110
139
0
1240
LAB 3
708
471
53
0
1232
PERCENTAGE OF G,
FLAG
G
U
B
LAB 1
84.0
2.8
13.2
LAB 2
79.9
8.9
11.2
LAB 3
57.5
38.2
4.3
LAB 6
852
409
11
0
1272
U, B
LAB 6
67.0
32.1
0.9
LAB 7
922
57
21
0
1000
FLAGS
LAB 7
92.2
5.7
2.1
LAB 8
689
435
132
0
1256
LAB 8
54.9
34.6
10.5
TOTAL
5513
1527
568
0
7608
TOTAL
72.4
20.1
7.5
NOTE: The numbers represent analytes, not samples.
The counts for Lab 1 include the qualifiers for Lab 4.
41
-------
COUNT OF DATA VALIDATION INDIVIDUAL QUALIFIERS
FROM FIELD SAMPLES IN FORM I DATABASE
TABLE 5b
FLAG(S)
El
E3+E5+E8
E7
12+14
13+15
H1+H2+H3+H4
K2
Ql
Q6+Q7
Q8
QA
Zl
Z2
Z7
Z9
ZB
ZC
ZD
ZE
ZH+ZI
ZJ
TOTAL
LAB 1
87
0
0
0
0
8
0
0
0
0
1
80
56
56
136
0
0
0
0
0
0
424
LAB 2
36
0
66
6
0
256
0
4
1
0
0
16
0
0
120
0
0
0
0
48
0
553
LAB 3
25
6
5
4
0
784
0
3
7
8
0
16
56
48
8
0
0
24
0
0
8
1002
LAB 6
0
44
0
0
7
520
0
2
7
0
0
0
0
0
0
0
0
0
0
0
0
580
LAB 7
30
8
0
88
0
8
0
0
7
0
0
8
0
0
0
0
0
0
0
0
0
149
LAB 8
43
20
14
3
9
656
6
2
2
16
0
40
40
56
0
48
40
0
104
0
0
1099
TOTAL
221
78
85
101
16
2232
6
11
24
24
1
160
152
160
264
48
40
24
104
48
8
3807
NOTE: The numbers represent analytes, not samples.
The counts for Lab 1 include the qualifiers for Lab 4.
42
-------
PE SET 2
SAMPLE PERCENT RECOVERY USED IN THE
THE GENERATION OF THE LOW I.EVEI. PREDICTION INTERVALS
TABLE 6
PE SET 1
SAMPLE ID
LAB LOGIN ID
1,2-DICHLOROBENZENE
1 ,2 ,4-TRICHLOROBENZENE
1 ,2 ,3,4-TETRACHLOROBENZENE
2 -CHLORCKAPHTHALEKE
ALPHA-BHC
DELTA -BHC
BETA-BHC
GAMHA-BHC
1O04
B74644S
230
126
100
8<<
in
122
134
136
lOOIi Mint 6004 7001) 7004 8004 8004
B74645S 0911B7C06 091187C15 096/4R10 09674R11 0811704A 0811704C
201.
94
78
38
91.
88
104
108
101.
112
88
86
0
128
0
0
118
112
96
88
106
172
0
0
70
70
6
-------
SAMPLE PERCENT RECOVERY USED IN THE
GENERATION OF THE HIGH 1.EVEL PREDICTION INTERVALS
TABU 7
PE SET 1
SAMPLE ID 1001 1001 6001 6001 7OOI 7O01 8O01 8001 1005 1O05 6O05 6005 7005 7O05 8005 8005
LAB LOGIN ID B74635S B74636S O91187C08 091187C16 096740T1 096740T2 0811701C 0811701B B74647S 871,61.85 091187C11 0911B7C12 09671iR13 09674R14 0811705B 0811705C
1 , 2 -DICHLOROBENZENE
1,2,4-TRICHLOROBENZENE
1,2,3,4-TETRACHLOROBENZENE
2 - CHLORONAPHTHALENE
ALPHA-BHC
DELTA- BHC
BETA-BHC
GAHMA-BHC
97
76
66
SI
67
71
80
79
99
80
66
49
Bfc
77
92
90
71.
80
82
61
98
121>
101
106
62
69
67
54
90
113
98
88
51.
61
54
1.0
51
65
0
1.7
61
67
56
41
62
77
83
76
59
39
felt
38
1.7
57
53
56
62
60
71
(.2
69
103
62
75
113
87
71
52
73
1.8
99
89
113
89
77
63
88
60
116
110
90
91
80
69
99
93
131
110
74
84
76
61
91
101.
113
98
6(>
70
66
56
63
lib
76
81
57
57
1|7
50
1.0
27
0
61
76
62
72
1.8
59
58
74
78
69
1.1.
41
35
44
56
64
59
PE SET 2
SAMPLE ID
LAB LOGIN ID
1 ,2-DICHLOROBENZENt
1 ,2 ,4-TRICHLOROBENZENE
1,2,3,4-TETRACHLOROBENZENE
2 -CHLORONAPHTHALENE
ALPHA-BHC
DELTA-BHC
BETA-BHC
GAMMA- BHC
1019
75553
76
83
69
60
84
112
83
83
1002
75695
61
68
58
48
58
55
65
63
2O05
D0780
74
89
78
48
73
66
59
65
2OO9
D0781
74
84
77
54
64
66
61
58
3002
37559R
58
70
64
58
85
61
0
78
3017 4005 4017 6008 6009 7O02 7O12 8001 8008
37559R 17091439 17091451 10O387C04 100387C05 09806002 09806012 918801A 918802A
73
84
76
62
96
69
0
89
94
101
73
48
67
70
95
65
93
97
70
43
75
81
101
72
83
93
89
72
109
113
0
117
86
95
88
78
107
115
0
115
77
77
76
65
75
93
90
70
56
92
61
53
82
99
99
85
9J
87
77
39
101
153
88
108
95
87
72
43
98
159
88
103
NOTE: ALL REPORTED ZERO VM.UtS WERE NOT USED IN THE GENERATION OF THE CONFIDENCE INTERVALS
REVISED 11/24/87
-------
THE EMSL-LV
DATA VALIDATION
STANDARD OPERATING PROCEDURE
FOR THE LOVE CANAL
EMERGENCY DECLARATION AREA
HABITABILITY STUDY
SOIL ASSESSMENT -- INDICATOR CHEMICALS
by
D.L. Bogen
Environmental Programs
Lockheed Engineering and Management Services Company
Las Vegas, Nevada 89119
Contract No. 68-03-3249
Work Assignment Manager
Dr. J.D. Petty
Quality Assurance Division
Environmental Monitoring Systems Laboratory
Las Vegas, Nevada 89193
IJMVIRONMENTAL MONITORING SYSTEMS LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
LAS VEGAS, NEVADA 89193
-------
SOP TABLE OF CONTENTS
PAGE
Login SOP - 1
Case Narrative SOP - 1
QC Summary SOP - 1
Floppy Disk. , SOP - 1
Data Validation Software (DVS) SOP - 2
Organization of Data Package SOP - 2
Resubmission Requests SOP - 3
Initial Calibration SOP - 3
Shift Results SOP - 4
Sample Results SOP - 9
Screening Data SOP - 10
Silica Gel Verification Data SOP - 11
Deliverables SOP - 11
Form I Review SOP - 11
SOP - i
-------
SOP ILLUSTRATIONS
PAGE
Illus. 1 Example of Login Book Page SOP - 12
Illus. 2 Summary Concentration Table SOP - 13
Illus. 3 Error Listing of Samples/Analyses By
Lab Sample ID SOP - 14
Illus. 4 Error Listing of Samples/Analyses By
Error Type SOP - 15
Illus. 5 Audit Trail for Lab Database SOP - 16
Illus. 6 Initial Calibration Checklist SOP - 17
Illus. 7 Shift Results Checklist SOP - 18
Illus. 8 Sample Checklist SOP - 21
Illus. 9 Record of Telephone Conversation Form SOP - 24
Illus. 10 Summary of QA Criteria SOP - 25
Illus. 11 Initial Calibration Checklist Incomplete. . . . SOP - 27
Illus. 12 Shift Results Checklist Incomplete SOP - 28
Illus. 13 Sample Checklist Incomplete SOP - 29
Illus. 14 Example of Reviewed Form I SOP - 30
SOP - ii
-------
LCHSSA Data Validation Standard Operating Procedure
The EMSL-LV has been tasked to qualify all the data generated by
the seven laboratories participating in the Love Canal Emergency
Declaration Area Habitability Study, Soil Assessment (LCHSSA). A
database will be generated from the validated and qualified Form I's.
This SOP is designed to assist the data evaluator in organizing the
validation process and generating the appropriate deliverables for
this enormous task. It is important that the data evaluator be
completely familiar with the SOW, in particular, the method, QA
criteria, and deliverable requirements. The text of this SOP is
followed by illustrations that are intended to assist in the
validation, such as Data Validation Checklists, Summary of the QA
Criteria, and Form I with qualifiers. Remember to be patient and
thorough.
1. LOGIN - All data packages received must be logged into the Data
Package Login Book - include lab name, date of receipt, floppy
disks received, size of package, pages, courier, and the initials
of the person doing the login. (See illustration 1)
2. CASE NARRATIVE - Read the Case Narrative carefully and note
anything unusual and sample specific information. For example,
note any instrumentation problems, contamination problems,
decisions made by Project Management Staff, samples broken or lost
from sample receipt through analysis, and samples that required
reextraction or reanalysis and why.
QC SUMMARY - Spend a short time reviewing the QC Summary data to
get a feel for the data quality, problems encountered such as
surrogates or internal standards out of criteria, frequency of
initial calibrations, and number of reextractions and reanalyses.
Also, note carefully if the forms are properly completed.
4. FLOPPY DISK - Merge the floppy diskettes received with the data
package into the Data Validation software. See the Data
Validation System User's Guide prepared by the software contractor
for instructions on how to perform the merge. Login the merge of
the diskettes into the Love Canal Login Book. Print out the
Summary Concentration Report by Sample ID, Error Listing of
Samples/Analyses, and Audit Trail. Examine the information on all
SOP - 1
-------
of the reports, note any unusual incidences, compare the .error
messages to the comments in the Case Narrative, and compare some
of the concentration values to the values on the Form I. Refer
back to these summary tables throughout the data validation
process. See illustrations 2-5 for examples of the reports.
DATA VALIDATION SOFTWARE (DVS) - The DVS was developed with the
concept of a shift as the basis. It is important to understand
this idea as the entire validation process will be organized by
shift. An analytical shift in DVS is the same as outlined in the
SOW, Exhibit B, Section 3.2.2, except, for software purposes, a 16
hour analytical run is composed of two shifts. The first data to
review is from an initial calibration. The initial calibration is
associated with many shifts as well as being analyzed within a
shift; therefore, it is treated uniquely. Print the Validation
Review for the first initial calibration (Initial Calibration
Checklist). Then print the Validation Review for the first shift
file name (Shift Results Checklist). Finally, print the Sample
Checklists for all samples analyzed within the first shift. Print
the Validation Reviews for the remaining initial calibrations,
shifts and samples as above. See the Data Validation System
User's Guide prepared by the software contractor for details on
how to use the Data Validation software and print the reports.
See illustrations 6-8 for a copy of the Initial Calibration,
Shift Results and Sample Checklists generated by the DVS.
6. ORGANIZATION OF DATA PACKAGE - Arrange and label the data package
by shifts.
a) For a shift with an initial calibration, include the following
items in the order specified: Form V (Performance Check
Solution Summary), Form VII (Continuing Calibration Check),
Form Via (Initial Calibration Data), Form II (Surrogate
Percent Recovery Summary), and Form X's (Internal Standard
Response and Retention Time). For the raw data and remaining
forms, put the data in order of date and time of analysis and
put the sample specific forms with the corresponding raw data.
The remaining data must include the raw data for IC1 - IC5,
Form VIb (EPA Check Standard Summary) and the raw data for the
EPA Check Standard, Form IV (Blank Summary) and the Form I and
raw data for the blank if one was analyzed in the shift, Form
III (Matrix Spike/Matrix Spike Duplicate Recovery Summary) and
the Form I's and raw data for the MS and/or MSD samples if
they were analyzed in the shift and all other Form I's and
sample raw data analyzed in the shift. The end of the shift
is marked by the raw data for the PC2 analysis.
SOP - 2
-------
b) For a shift without an initial calibration, include the
following items in the order specified: Form V (Performance
Check Solution Summary), Form VII (Continuing Calibration
Check), Form II (Surrogate Percent Recovery Summary), and Form
X's (Internal Standard Response and Retention Time). For the
raw data and remaining forms, put the data in order of date
and time of analysis and put the sample specific forms with
the corresponding raw data. The remaining data must include
Form VIb (EPA Check Standard Summary) and the raw data for the
EPA Check Standard if one was analyzed in the shift, Form IV
(Blank Summary) and the Form I and the raw data for the blank
if one was analyzed in the shift, Form III (Matrix
Spike/Matrix Spike Duplicate Recovery Summary) and the Form
I's and the raw data for the MS and/or MSD samples if they
were analyzed in the shift and all other Form I's and the
sample raw data analyzed in the shift. The end of the shift
is marked by the raw data for the PC2 analysis.
When a shift is complete, label it with the shift file name.
Check each Form I to make sure that all the samples have the same
shift file name. Also, check for missing forms or raw data by
comparing the sample raw data to Forms II and X's. Arrange the
shifts in order of date and time of analysis and place in a
labeled banker box in that order. Include the lab name, data
package number and LCHSSA on the label of the box.
RESUBMISSION REQUESTS - If any form or raw data is missing, you
may call the laboratory and request the necessary resubmissions.
When requesting resubmissions, provide the laboratory with the
shift file name and all other pertinent information. Record the
phone call on a Record of Telephone Conversation Form and place in
the Love Canal Phone Log. If the list is very long, FAX the list
to the laboratory. Please make all requests to the Project
Manager of that laboratory. See illustration 9 for a Record of
Telephone Conversation Form.
It may be necessary during the data validation process to use the
Summary of QA Criteria, illustration 10, and the List of Data
Qualifiers.
7. INITIAL CALIBRATION - The initial calibration documents the
linearity of the instrument within the calibration range of the
method, establishes the relative response factors on which the
continuing calibration acceptance criteria are based, and verifies
the performance of the column before samples are analyzed. It is
important to check the initial calibration thoroughly as it
impacts all samples analyzed until the next initial calibration.
Check the following items from the initial calibration:
SOP - 3
-------
The DVS has areas that need to be completed by the data evaluator.
The options for the answers are: Y = yes, acceptable data; P =
problem encountered with the data; X = question is not applicable.
See the Data Validation Systems User's Guide.
a) Compare the information on the Form VI to the raw data. Do the
dates, times, instrument ID, and file names match? Spot check
the relative response factor (RRF), mean RRF, and percent
relative standard deviation (%RSD) calculations.
b) Does the Error Listings Report indicate any analyses out of
criteria? Compare the information from this report and check
the Form VI to see if the %RSD criteria are met. If the
numbers are close to the QC limits, check the calculations
carefully.
c) Check the quantitation reports for evidence of editing - manual
quantitation.
d) Examine the SICPs for interference, peak shape and
reasonableness. Note any problems on the DVS Initial
Calibration Checklist.
e) Are resubmissions needed? For example, resubmissions may be
necessary if the SICPs are not displayed on the appropriate
scale or the Form VI is not completed correctly.
f) If the IC2 is used for the PCI of the shift, the date and time
of analysis on Form VI for the IC2 must match that on Form V
for PCI and the RRF data on Form VI for IC2 must match that on
Form VII.
8. SHIFT RESULTS - A shift consists of a PCI, QC samples, field
samples, and a PC2. The quality of the field sample data within a
shift is impacted by the PCI, the QC sample results, and the PC2.
Therefore, the DVS Shift Results Checklist asks which QC data is
associated with the field samples. It is important to evaluate
the information in the Shift Results Checklist prior to evaluating
the field samples in a shift. See illustration 7 for the Shift
Results Checklist.
a) Deliverables - If any forms or raw data are missing from the
shift, list them on the DVS checklist in the space provided.
Complete the date resolved when the deliverables are received.
b) Performance Check 1 (PCI): The performance check solution is
analyzed to demonstrate column performance - resolution and
sensitivity - and must be analyzed at the proper frequency.
SOP - 4
-------
- Complete the date and time of analysis of the PCI. Is the
PCI analyzed at the proper time interval?
- Is this PCI a PC2 midpoint from the previous shift? If yes,
what is the shift file name from the previous shift? It is
important to complete this question for checking if the
appropriate QC samples were analyzed at the proper frequency.
See SOW, Exhibit B, Sec 3.2 for the requirements of a 16 hour
analytical run.
- Compare the PCI raw data to the information on the Form V.
If there are any discrepancies and can be corrected by
resubmissions, request them. If they are not correctable,
explain the problem on the checklist.
- Check the raw data to see if the PCI meets the percent
valley, signal-to-noise (S/N) ratio, and ion ratio criteria
and that the numbers on the Form V match the raw data.
- Check the chromatography - look at peak shape, interference,
and reasonableness.
- Check the quantitation report for evidence of editing and
manual quantitation. Was the manual quantitation necessary
due to interference or to meet criteria?
- Make any comments on the checklist, including qualifying
flags, that will need to be put on all or some of the
associated samples.
c) Continuing Calibration (CC): The CC solution is the same
solution as the PCI, but is examined for different reasons.
The CC analysis demonstrates that the RRFs have not changed
more than specified in the SOW as compared to the RRFs from the
initial calibration. Remember, the raw data for the PCI and CC
are the same.
- Complete the date and time of analysis of the CC, which must
be the same as that for the PCI.
- Check to see if the continuing calibration is compared to the
correct initial calibration.
- Compare the CC raw data to the information on the Form VII.
If there are any discrepancies and can be corrected by
resubmissions, request them. If they are not correctable,
explain the problem on the checklist.
- Check the raw data to see if the CC meets the percent
difference criteria and that the numbers on the Form VII
match the raw data.
- Check the internal standard areas and retention times. Are
the proper internal standard areas and retention times
correct on the Forms X-I - X-V? Be careful because the Mass
Spectrometer may use 100th of minutes and not seconds, 60th
of minutes.
- Make any comments on the checklist, including qualifying
flags, that will need to be put on all or some of the
associated samples.
SOP - 5
-------
d) Performance Check 2 (PC2): The PC2 is analyzed to demonstrate
that the column was within control throughout the time that
field samples were analyzed.
- Complete the date and time of analysis of the PC2. Is the
PC2 analyzed at the proper frequency?
- Compare the PC2 raw data to the information on the Form V.
If there are any discrepancies that can be corrected by
resubmissions, request them. If they are not correctable,
explain the problem on the checklist.
- Check the raw data to see if the PC2 meets the percent
valley, signal-to-noise (S/N) ratio, and ion ratio criteria
and that the numbers on the Form V match the raw data.
- Check the chromatography - look at peak shape, interference,
and reasonableness.
- Check the quantitation report for evidence of editing and
manual quantitation. Was the manual quantitation necessary
due to interference or to meet criteria?
- Make any comments on the checklist, including qualifying
flags, that will need to be put on all or some of the
associated samples. Explain the reason for the flag(s).
e) Blind QC Sample (BQC): The BQC sample must be analyzed with
the samples from the associated extraction batch and at least
one BQC sample must be analyzed in each shift (except if there
is a 16 hour analytical run, or an EMPC was analyzed in the
shift with rerun samples). See the SOW, Exhibit B, Sees 3.2
and 3.13. For a 16 hour analytical run, a BQC sample might
only be analyzed in one of the shifts composing the 16 hour
run.
- Complete the DVS checklist with the sample ID of the BQC
sample analyzed in the shift. If more than one BQC sample
was analyzed in a shift, make a comment on the checklist. If
a BQC sample is not analyzed in the shift, put not applicable
on the checklist.
- It is important to check if the correct BQC sample was
analyzed with the samples. To check this, examine the
extraction logbook for the date the BQC sample was extracted.
Look at the Form II and determine if the samples analyzed in
the shift were extracted with the BQC sample. If they were
not, use the appropriate flag from the Qualifier List.
- Calculate the recovery of each analyte in the BQC sample
using the sample key and determine if the recoveries are
within the appropriate high or low acceptance limits.
Compare the concentrations and recoveries to the data
received through the RTQC system. If there are
discrepancies, check the raw data to determine which value is
correct and make a comment on the checklist.
SOP - 6
-------
Examine the raw data and quantitation report carefully for
peak shape and interference.
Make any comments on the checklist, including qualifying
flags, that will need to be put on all or some of the
associated samples. Explain the reason for the flag(s).
f) EPA Check Standard (EMPC): The EMPC is analyzed with every
initial calibration and at least every two weeks. It can also
be analyzed in place of a BQC sample when samples from
different extraction batches are reinjected. The EMPC acts as
an independent verification of the standard solutions used for
the project. The following items must be checked for the EMPC:
- Complete the DVS checklist with sample ID of the EMPC
analyzed in the shift. If an EMPC is not analyzed in the
shift, put not applicable on the checklist.
- Calculate the recovery of each analyte in the EMPC and
determine if the recoveries are within the control limits.
Compare the concentrations and recoveries to the data
reported on the Form VIb. If there are discrepancies, check
the raw data to determine which value is correct and make a
comment on the checklist.
- Examine the raw data and quantitation report carefully for
peak shape and interference.
- Make any comments on the checklist, including qualifying
flags, that will need to be put on all or some of the
associated samples.
g) Method/Holding Blank (MB): The MB is stored with the field
samples and accompanies them through analysis. It is used to
evaluate possible laboratory contamination. The MB must be
analyzed with the samples from the associated extraction batch
and at least one MB must be analyzed in each shift (except if
there is a 16 hour analytical run, or if an instrument blank is
analyzed with rerun samples).
- Complete the DVS checklist with the sample ID of the MB
analyzed in the shift. If more than one MB was analyzed in a
shift, make a comment on the checklist. If a MB is not
analyzed in the shift, put not applicable on the checklist.
- It is important to check if the correct MB was analyzed with
the samples. To check this, examine the extraction logbook
for the date the MB was extracted. Look at the Form II and
determine if the samples analyzed in the shift were extracted
with the BQC sample. If they were not, use the appropriate
flag from the Qualifier List.
- Calculate the concentration of each ion of each analyte in
the MB and determine if the concentrations are within the
SOP - 7
-------
control limits. Compare the concentrations to the data on
Form IV. If there are discrepancies, check the raw data to
determine which value is correct and make a comment on the
checklist.
- Examine the raw data and quantitation report carefully for
peak shape and interference. Also look for peaks that are
within the RRT windows of the LCICs and are greater than 0.5
ppb equivalent concentration (0.6 ppb for 1,2-
dichlorobenzene). Check to see if the most interfering peak
in each RRT window is reported on the Form IV. Try to keep
in mind the contamination observed in the blank as you review
the data from the associated field samples. Qualify the
samples accordingly using the List of Qualifiers. Make any
comments on the checklist, including qualifying flags, that
' will need to be put on all or some of the associated samples.
h) Matrix Spike/Matrix Spike Duplicate Sample (MS/MSD): The
MS/MSD samples are extracted and analyzed at least every twenty
samples. These samples are used to demonstrate the accuracy
and precision of the analytical method.
- Complete the DVS checklist with the sample ID of the MS/MSD
analyzed in the shift. If more than one MS/MSD were analyzed
in a shift, make a comment on the checklist. If a MS/MSD was
not analyzed in the shift, put not applicable on the
checklist.
- Calculate the recovery of each analyte in the MS/MSD and
determine if the recoveries are within the limits. Compare
the concentrations and recoveries to the data reported on the
Form III. If there are discrepancies, check the raw data to
determine which value is correct and make a comment on the
checklist.
- Examine the raw data and quantitation report carefully for
peak shape and interference
- Make any comments on the checklist, including qualifying
flags, that will need to be put on all or some of the
associated samples. Explain the reason for the flag(s).
At this point, the Shift Results Checklist is complete. Review
your comments for the shift prior to starting the individual
Sample Results Checklists and keep a list of the qualifiers that
will need to go on the associated samples.
9. SAMPLE RESULTS - The thorough evaluation of the individual sample
data is extremely important. A database will be generated from
the validated and qualified Form I's. Before beginning the
individual Sample Checklists in DVS, pull all the Form I's for the
SOP - 8
-------
shift you are going to evaluate from the QC Summary package and
organize them in the order of date and time of analysis. On the
lower left corner of each Form I, write the correct extraction
date. Take the extraction date from the Form II; however, if the
extraction date does not appear to be correct, check the
extraction logbook. As you evaluate the sample data, you will
need the Form I's available. Evaluate the following items for
each sample within a shift.
a) Verify that the laboratory name, lab analysis ID, sample ID,
shift file name, instrument, and calibration date on the DVS
checklist are correct.
b) Compare the sample tracking information on the DVS checklist
with the information on the Form I. If the information is not
identical in both locations, a resubmission of either the Form
I or the DVS floppy diskette for that shift is necessary. If
you are unsure of what is needed, contact either the task
leader or the software contractor. A discrepancy in the login
date can be verified against the chain-of-custody forms, the
percent moisture, extraction analyst, and weight of sample
against the extraction log, and date of analysis against the
instrument run log. If the concentration/dilution factor is
not 1.0, verify the dilution against the screening data and the
instrument run log.
c) The Sample Checklist - Error Listing summarizes the QA
deficiencies detected by the DVS for the sample. The
deficiencies listed will include all items exceeding the QA
criteria listed in the Summary of QA Criteria at the end of
this SOP. Verify all of the items listed. If the raw data
indicates that the comment is incorrect, contact the software
contractor.
d) Check the surrogate recoveries against the raw data and
quantitation report. Spot check the calculations and compare
to the results on Form II. If the recoveries are not within
criteria, make a comment on the checklist and flag the data
accordingly. Read the reextraction and reanalysis requirements
in Exhibit C of the SOW. If a surrogate is out of criteria,
the sample needs reextraction. If after reextraction, the
criteria still are not met, both results get entered into the
Form I Database. Keep track of this type of occurance so that
only the appropriate Form I's get into the final database.
e) Check the internal standard retention time and area difference
criteria against the raw data. The data on Form X's must be
consistent with the raw data. If there is a discrepancy, make
SOP - 9
-------
a comment and determine if a resubmission is necessary. Are
all QA criteria met? If not, make a comment on the checklist
and qualify the data accordingly. Again, as with surrogates,
internal standards have rigid reanalysis requirements. See the
SOW, Exhibit C. Care must be taken to evaluate which Form I's
get entered into the database.
f) Review the raw data and quantitation report to see if the LCICs
reported meet all of the identification criteria (scan range,
ion ratio, all ions present). Also determine if all low level
peaks, greater than 0.2 ppb equivalent concentration, are
reported. Did the laboratory use the appropriate flags on the
Form I?
g) Spot check the calculations to determine if the appropriate
RRF, sample weight, percent moisture were used for
quantitation. If manual quantitation was used, was it
necessary?
h) It is imperative that all the information on the Form I
correspond with the raw data. Check everything carefully!!
After all the data has been reviewed, print out the DVS Checklist
Incomplete Reports for Initial Calibration, Shift Results, and
Samples (see illustrations 11-13). If any checklists are
incomplete, complete the record. If there is no raw data for the
missing sample(s) or there is any other discrepancy, try to
resolve it. Request any necessary resubmissions or qualify the
data accordingly.
10. SCREENING DATA - The screening data must be reviewed to determine
if the laboratory diluted the sample to get the analyte at the
highest concentration in the upper half of the calibration range.
Also, if a sample was screened, this should be indicated on the
Form I for that sample.
11. SILICA GEL VERIFICATION DATA - Compare the raw data for the
silica gel verification with the results reported on the Form
XI. If a batch of silica did not meet the QC criteria check to
see if it was used for the purification of the sample extracts.
12. DELIVERABLES - The SOW requires that the data package include a
long list of items that need to be checked to see if they are
present and reviewed only as needed. These items are listed in
the SOW, Exhibit D, Table D-3. Please check to see if everything
on the list was received. If not, request the appropriate
resubmissions.
SOP - 10
-------
13. FORM I REVIEW - When the data evaluator completes the data
validation and has put all the qualifiers on the Form I, put the
Forms in order of date and time. Also put the Data Validation
Checklists in sequential order. This package is then given to
the final data reviewer who will qualify each analyte on every
Form I with a G, U, B, or X (see illustration 14). See the Data
Qualifier List. The data reviewer will return the package with
questions and comments. Review these questions and comments and
make any revisions to the Form I qualifiers and DVS checklists.
You are now finished!!!!
SOP - 11
-------
tfl
H-
O
t— '
C/3
0
.
1
1— •
SJ
FLOPPY DISC INFORMATION
8 FLOPPY DISKETTES
5 FLOPPY DISKETTES
6 FLOPPY DISKETTES
E)
NUMBER
1
2
3
(AMPLE OF I
DATE
12/16/87
12/17/87
12/17/87
.OGIN BOOK
LAB
LAB 1
LAB 4
LAB 2
PAGE
PAGES
5561
3844
4487
BOXES
1 LARGE
1 SMALL
1 LARGE
1 LARGE
COURIER
FED EX
FED EX
FED EX
INITIALS
CMS
DLB
CMS
-------
Date 03/1B/8B
Tiie: 06:11:49
DATA VALIDATION SYSTEM
Suiiarv Concentration Table
Pane
Lab:
Concentrations bv LCIC
Suale No.
csxsessssa s
HS0171
HS0182
HS0191
HS0202
HS0215
HS0230
HS0241
HS0244
H50245
HS0246
HS0247
HS025B
HS0267
HS0275
HS02B4
HS02B9
HS0312
HS0315
HS0320
HS0331
HS0335
HS0337
HS0342
HS0350
HS0359
HS0365
HS03B4
HS03B5
HS0399
HS0404
H50432
HS0445
HS0446
HS0452
HS0467
HS0467
HS0473
H50477
HS0460
HS0490
HS0502
HS0503
HS0507
1
DCB
.SSSSSXSXSt Cl
0.539
1.01B
0.716
0.489
0.616
0.526
0.779
0.762
0.549
1.515
0.876
0.621
1.060
0.817
0.566
0.737
0.752
0.625
0.741
0.848
0.808
1.142
0.728
0.910
1.013
0.780
0.991
0.872
0.711
0.648
0.749
0.534
1.093
0.756
0.555
1.698
0.8BB
1.222
1.031
0.731
0.914
2.217
0.515
2
TCB
SSCSSSSSSS fSS
0.201
0.299
0.650
0.1B8
0.713
0.526
1.052
0.853
0.425
0.369
0.382
0.280
6.932
0.663
0.186
0.652
0.574
0.366
0.861
0.242
2.110
0.405
0.697
0.365
0.315
0.225
0.444
0.300
0.230
1.142
0.244
0.395
4.138
0.529
3.351
7.646
0.616
3.015
0.355
0.263
0.356
13.841
0.409
3
KB
&XXXZSSS
0.127
0.140
0.566
0.115
0.900
0.447
1.433
1.119
0.792
0.221
0.262
0.325
10.910
0.457
0.075
0.641
0.424
0.336
0.554
0.055
2.024
0.395
0.600
0.216
0.080
0.137
0.324
0.214
0.171
1.012
0.112
0.351
5.560
0.335
4.717
6.487
0.710
0.939
0.216
0.312
0.208
18.636
0.429
4
CN
ssssssssss: s
0.106
0.093
0.099
0.107
ND
0.137
0.097
ND
0.140
0.192
0.115
0.126
0.179
0.153
0.141
0.109
0.100
0.145
0.119
0.100
0.090
0.102
0.071
0.142
0.212
0.071
0.105
0.084
0.108
0.101
0.103
0.092
0.069
0.132
0.093
0.236
0.118
0.095
0.116
ND
0.120
0.189
0.103
5
A-BHC
ssssssssss s
0.325
0.096
0.165
0.083
0.261
0.094
0.326
0.725
0.092
0.117
0.151
0.114
7.192
0.145
ND
0.382
0.329
0.056
1.700
ND
2.317
0.131
0.270
0.080
ND
0.057
0.144
0.156
ND
0.570
ND
0.082
1.156
0.109
5.261
36.224
0.145
0.242
0.100
0.090
0.067
153.527
0.126
6
D-BHC
ssscsszrz:
ND
ND
0.082
ND
0.105
0.182
0.336
0.116
0.106
0.071
ND
0.055
2.089
ND
ND
0.180
0.189
ND
1.546
0.067
0.283
0.063
0.192
ND
0.152
0.656
0.161
0.092
0.173
0.279
0.296
0.060
0.469
0.086
0.281
0.740
0.099
0.141
0.468
ND
ND
5.B24
0.067
7
B-BHC
rsssrszssrs
0.153
0.067
0.195
0.051
0.346
0.162
0.257
1.417
2.B31
0.199
1.500
0.080
13.250
0.197
1.853
0.201
2.763
2.066
0.483
ND
2.462
0.254
0.768
0.062
4.994
0.230
0.219
0.198
1.858
0.911
0.166
0.113
2.085
0.119
2.343
14.691
0.664
O.S09
0.163
0.130
1.241
101.722
0.085
8
B-BHC
fSSSSSSSSSS
NO
ND
0.370
0.328
0.066
ND
0.204
0.167
0.172
0.557
0.092
ND
1.866
ND
ND
ND
0.153
0.056
0.711
ND
0.2::
0.125
0.050
ND
0.071
ND
0.095
ND
m
0.090
0.086
ND
0.212
ND
0.476
1.700
NO
0.084
0.067
ND
ND
20.993
0.055
Anahsis ID
rsarsesissssssss
8710222-04
8711011-05
8711126-01T
6710164-05
8711066-08
8710177-07
8711125-02
87U036-07A5
8711014-03
8711035-08
8711012-02
8711012-06A5
B710222-OBT
B71 1068-01
8711014-05
6710152-05
8710140-03
6711014-02
8711012-09AS
8710140-OB
8710143-06
8711036-02
8710219-03
6710290-04
6710140-01
B710290-06A5
B7I021B-02AS
8711011-07
B710140-02
B711035-03S
8711124-05
8710154-09
87101 43-02S
B710152-04T
8711125-06
8711125-06A
8710290-05
8710164-04
8711011-06
8710143-06
8711035-07
B711014-06AS
B711124-01A
Illustration 2
SOP - 13
-------
Date: 03/04/BB
Tilt: 09:1.1:1.3
Lib Analysis ID
&£££££££££££££££"
LC09974R1I
LC09974R12
LC09976004
LC09976009
LC09976010
LC09976R01
LC09976R05
LC09976T01
LC10005008
LC10005ROB
DATA VALIDATION SYSTEM Page 3
Error Listing of Saiples/Analyses Lab:
By Lab Saiple ID
Upper Loier Parameter
Saiple ID Paraietr Nate Criteria Criteria Value
fi£S££SS::£££t £££S££££££S£££S=££S£££SS£S£S££££££££££££££££S ££££££££££££ SS££fi£££££££ ££££££££££££
RHSHSA742 HS/HSD X REC FDR A-BHC OUTSIDE fit LIMIT 139.000 6S.OOO 6*..433
HS/HSD X REC FOR B-BHC OUTSIDE GC LIHIT 106.000 64.000 63.101
US/USD I REC FOR D-BHC OUTSIDE QC LIHIT 13B.OOO 64.000 60.10'.
RHSDHSA742 I RECOVERY FOR 1,4-DBB OUTSIDE 6A/OC LIMITS 104.000 51.000 46.068
AREA RESPONSE FOR DB-NAPH OUT QC LIHIT 50.000 -60.000 57.000
HS/HSD 1 REC FOR 1,2,3,4-QCB OUT QC LIHIT 113.000 63.000 59.050
HS/HSD I REC FOR 2-CNP OUTSIDE QC LIHIT 114.000 43.000 S7.977
HS/HSD X REC FOR A-BHC OUTSIDE QC LIHIT 139.000 65.000 62.939
HS/HSD t REC FOR B-BHC OUTSIDE QC LIHIT 106.000 64.000 47.006
R5/HSD X REC FOR D-BHC OUTSIDE QC LIHIT 138.000 64.000 56.389
HS/RSD I REC FOR 6-BHC OUTSIDE QC LIHIT 119.000 66.000 65.494
HSD RPD FOR 1,2-DCB OUTSIDE QC LIHIT 30.000 0.000 36.766
HSD RPD FOR 2-CNP OUTSIDE DC LIHIT 30.000 0.000 30.854
HS0710 I RECOVERY FOR E,4,6-TBBP OUTSIDE QC LIHITS 126.000 56.000 130.503
BLH110S X RECOVERY FDR 2,4,6-TBBP OUTSIDE QC LIHITS 126.000 56.000 137.555
QCEH7007 X RECOVERY FOR 2.4,6-TBBP OUTSIDE QC LIHITS 126.000 56.00C 136.4C9
HS0540 AREA RESPONSE FOR D10-ACENAPH OUT QC LIHIT 50.000 -60.000 56.000
RHS0147 AREA RESPONSE FOR D10-ACENAPH OUT QC LIHIT 50.000 -60.000 64.000
AREA RESPONSE FOR DB-NAPH OUT QC LIMIT 50.000 -60.000 60.000
RHS0540 AREA RESPONSE FOR D10-ACENAPH OUT GC LIHIT 50.000 -60.000 62.000
HS0555 AREA RESPONSE FOR D10-ACENAPH OUT QC LIHIT 50.000 -60.000 54.000
RHS0555 AREA RESPONSE FDR DIO-ACENftPK OUT QC LIHIT 50.000 -60.000 96.000
Illustration 3
SOP - 14
-------
Cite: 03/04/88
Tiu: 09:37:27
Parameter Hue
I RECOVERY FOR 1,2,4,5-QBB OUTSIDE DC LIHIT
I RECOVERY FDR 1,4-DBB OUTSIDE QA/8C LIHITS
J RECOVERY FOR 2.4,6-TBBP OUTSIDE BC LIHIT5
AREA RESPONSE FOR D10-ACENAPH OUT BC UNIT
DATA VftLIDAl
Error Listing of
By Errc
Analysis ID
BLR1 13087
1NSTBLK1025
LC09949R09
LC10020R07
LC101Z2R03
BLF.113087
INSTBLK1025
LC09949RI1
LC09949R14
LC09957R04
LC09957R04
LC09974002
LC09974011
LC09974012
LC09974R02
LC09974R11
LC09974R12
LC1002003
BLR1 130B7
INSTBLKI025
LC0996600C
LC09973002
LC09973002
LC09973003
LC09973003
LC09973004
LCC9973004
LC0997300B
LC0997300B
LC09973010
LC099730IO
LC09976004
LC09976009
LC09976010
LC1002011
LCI 0020 12
LC10071010
LC10110010
LC09974002
LC09974003
LC09974011
LC09974012
LC0997*R03
LC0997W07
riON SYSTEM
Stiples/Anilyses
ir Type
Suple ID
BLR1130
BLR1025
HS1371
KS080B
HS0773
BLR1130
BLR102S
BLHR1026
BCEHR70Ki
HS0471
HS047I
HS07I.2
NSHS0742
nSDHS07
-------
Date: 03/18/86
Tin: 07:38:13
Initial ShiU
Calibration ID Results ID Analysis ID
SSSSSCSCS5SES5 CSXSXZ5ESS XSS55CS3SES
IHP5970 102287
QS17102S.
8709189-01
8710140-01
8710140-02
8710140-03
B710140-OB
B710140-11A
BSB7102:.
BSB71029.
B7091B9-02
8710143-03
8710143-05
6710143-06
8710143-09
B710152-01
8710152-01-H5
B710152-01-NSD
8710152-02
8710152-06
B710152-07A
B710152-06A
B5871030.
8709189-03
8710154-04
8710154-05
8710134-09
8710154-12
8710164-02E
B710164-02C
8710164-03
DS871101.
8709189-12
8710164-04
8710164-06
8710164-07
8710164-06
8710164-10
8710164-11B
B710I77-01A
8710177-01B
8710177-01C
B710177-03
8710177-05
DATA VALIDATION SYSTEM
Audit Trail for
Saiole ID
BCER8002
HS035B
HS0399
HS0312
HS0331
BLH1016B7A
BCEH8003
HS0725
HS0603
HS0490
BLH101687C
H50B1B
HSHS081B
H5DHS0818
HS0151
HS074B
HS0707
H51364
BCEHB004
HS0080
HS0726
HS0445
BLH1019B7C
HSHS0659
HSDHS0659
HS0511
BCEM8016
HS0477
HS06B2
HS0732
HS1323
HSOS27
BLH1020A
HS0660
K5KS0660
HSDHS0660
HS0858
KS0092
Lab Database
Loo-in
Date
/ /
/ /
09/18/87
10/16/87
10/16/87
10/16/87
10/16/87
10/16/87
/ /
/ /
09/1B/87
10/23/87
10/16/67
10/16/87
10/16/87
10/19/87
10/19/87
10/19/87
10/19/87
10/23/87
10/19/67
10/19/87
/ /
10/31/87
10/19/87
10/19/87
10/19/87
10/19/87
10/20/87
10/20/87
10/20/87
/ /
10/21/87
10/20/87
10/20/87
10/20/87
10/20/87
10/20/67
10/20/67
10/21/87
10/21/67
10/21/67
10/21/87
10/21/87
Final Packaoe
Date
03/16/88
03/16/8B
03/16/88
03/16/88
03/16/88
03/16/88
03/16/68
03/16/86
03/16/BB
03/16/8B
03/16/88
03/16/86
03/16/86
03/16/8B
03/16/86
03/16/86
03/16/88
03/16/8B
03/16/68
03/16/88
03/16/68
03/16/86
03/16/86
03/16/68
03/16/88
03/16/66
03/16/88
03/16/86
03/16/BB
03/16/66
03/16/66
03/16/86
03/16/88
03/16/86
03/16/66
03/16/68
03/16/86
03/16/68
03/16/66
03/16/88
03/16/68
03/16/88
03/16/88
03/16/88
Paoe 1
Lab:
Final Packaoe Checklist
Tiie Coioleted
16:34
18:11
18:11
18:11
18:11
18:11
18:11
18:11
16:11
18:11
18:11
18:11
18:11
18:11
18:11
18:11
16:11
18:11
18:11
16:11
18:11
18:11
18:11
18:11
18:11
16:11
16:11
18:11
18:11
18:11
18:11
18:11
18:11
18:11
18:11
18:11
18:11
16:11
18:11
18:11
18:11
18:11
18:11
18:11
Illustration 5
SOP - 16
-------
03/21/BB
-AB:
LOVE CANAL SOFTWARE SUPPORT
Initial Calibration Checklist
Paoe 1
INSTRUMENT: IHP5970
CALIBRATION DATE: 10/22/87
1. Initial Calibration
Percent RSD exceotions checked.
Check quantitation reports for evidence o-f editing.
Examine chromatograms -for interference, peak shape, and reasonableness.
hi seel 1aneous.
Comments:
Initial Calibration Checklist - Error Listino
Parameter Name
Parameter Upper Lower
Value Criteria Criteria
NO ERROR RECORDS FOUND
Illustration 6
SOP - 17
-------
03/21/SB LOVE CANAL SOFTWARE SUPPORT Page 1
Shift Results Checklist
INSTRUMENT: IHP5970
LAB: SHIFT: QS171025 CALIBRATION DATE: 10/22/87
1. Deliverables Check
Are all deliverables present?
If not. enter missing deliverables below:
Missina Date Resolved
Performance Check 1 (PCI) Date: Time:
Is this a PC2 midooint for a 16 hour analytical run? vY/N)
If yes. olease enter first-half shift result file name:
Compare percent vallev SICPS to see if agrees with reoorted values.
Compare reported signal to noise ratios with raw data.
Check ion ratxo criteria
Check chromatograohy.
Check ouantitation reoorts for evidence of editing.
Mi seel 1aneous.
Comments:
Continuing Calibration (CO Date: Time:
Internal standard areas.
Check retention time differences.
Mi seel 1aneous.
Comments:
4. Performance Check 2 (PC2) Date: Time:
Compare percent valley SICPS with reoorted values.
Compare reported signal to noise ratios with raw data.
Check ion ratio criteria
Check chromatography.
Check quantitation reports for evidence of editing.
Mi seel1aneous.
Comments:
Illustration 7 SOP - 18
-------
03/21/BB LOVE CANAL SOFTWARE SUPPORT Page 2
Shift Results Checklist
INSTRUMENT: IHP5970
LAB: SHIFT: 05171025 CALIBRATION DATE: 10/22/87
Blind QC Samples Sample ID:
Recoveries within acceptance criteria.
Check chromatography.
Check quantitation reports.
Miscellaneous.
Comments:
6. EPA Check Standard.(EMPC) Samole ID:
Recoveries within acceptance windows and exceptions.
Check chromatography.
Check ouantitation reoorts.
Mi seellaneous.
Comments:
Method/Holding Blanks Sample ID:
Check: each SICP for interferences. Make sure SICP is displayed on a
scale to show contamination at the appropriate level.
Check that data is -flagged i-f blank is contaminated and. appropriate
action is taken.
Compare report form with chromatogram.
Mi seel 1aneous.
Comments:
8. MS/MSD Samele ID
Examine raw data for problems and discrepancies.
Outside criteria - see narrative for further discussion.
Examine SICPE and quantitation reoorts.
Mi seel 1aneous.
Comments:
Illustration 7 cont. SOP - 19
-------
03/21/88
LAB:
LOVE CANAL SOFTWARE SUPPORT
Shi-ft Results Checklist - Error Listing
Paae 3
SHIFT: QS171025
INSTRUMENT: IHP5970
CALIBRATION DATE: 10/22/87
Parameter Name
Parameter
Value
Upper
Criteria
Lower
Criteri a
NO ERROR RECORDS FOUND
Illustration 7 cont.
SOP - 20
-------
03/21/B8 LOVE CANAL SOFTWARE SUPPORT Page 1
Analysis Information
LAB: LAB ANALYSIS ID: B709189-01 SAMPLE: QCEMB002
SHIFT: OS171025. INSTRUMENT: IHP5970 CALIBRATION DATE: 10/22/87
1. Sample Tracking Dates:
Logged In: 09/18/87 Extraction: 10/21/87 Analysis: 10/25/87
Final Package: 03/16/88 Final Package Time: 18:11
2. Samole Data:
Weight o-f Aliquot Extracted grams: 20.00
Percent Moisture: 0.0
Concentration Dilution Factor: 1.000
3. Other Samnle Information:
Sample ID o-f blank created -for this sample: 8710140-11
Extraction Analyst:
Sulfur cleanuo performed:
Process -for screening high concentrations:
Illustration 8 SOP - 21
-------
03/21/SB LOVE CANAL SOFTWARE SUPPORT Page 2
Sample Checklist
LAB: LAB ANALYSIS ID: B709189-01 SAMPLE: QCEM8002
SHIFT: QS171025. INSTRUMENT: IHP5970 CALIBRATION DATE: 10/22/87
1. Surrogates
Check raw data to see :-f surrogates spiked into all samples.
Recoveries within criteria.
Mi seellaneous.
Comments:
2. Internal Standards
RT Criteria. Area Criteria. Miscellaneous.
Comments:
3. Identification
All ions maximize simultaneously.
Appropriate -flags used.
All peaks reported that meet identification criteria.
Low level peaks examined.
Miscellaneous.
Comments:
4. Quantitation
Aopropriate RRF's used ii nonstandard ouantitation used.
Check integration parameters i-f manual quantitation used.
Mi seellaneous.
Comments:
General
Review case narrative and address all problems.
Examine SICPS.
Examine Quantitation reports.
Miscellaneous.
Comments:
Illustration 8 cont. SOP - 22
-------
03/21/88 LOVE CANAL SOFTWARE SUPPORT Page 3
Sample Checklist - Data Qualifiers
LAB: LAB ANALYSIS ID: B7091B9-01 SAMPLE: QCEMB002
SHIFT: 05171025 INSTRUMENT: IHP5970 CALIBRATION DATE: 10/22/87
1
SAMPLE SPECIFIC FLAGS
ANALYTE SPECIFIC FLABS:
1.2-Di chlorobenrene
1.2.4-Trichlorobenzene
1.2.3.4-Tetrachlorobensene
2-Chrolonaphthalene
AlDha-BHC
Delta-BHC
Beta-BHC
Gamma-BHC
Sample Checklist — Error Listing
Parameter Name
Parameter
Value
Upper
Cri teria
Lower
Cr i teri a
NO ERROR RECORDS FOUND
Illustration 8 cont.
SOP - 23
-------
LLMECCJ KtTCCJWD OF 1ELLPHOWE COWVfR5«T ION
Date Region No.
Recorded by Facility
Talked with of
Nature of call Incoming Outgoing
Route to the following:
Information only Action
LOVE CANAL PHONZ LOG
Main subiect of the call
Items discussed
Illustration 9 SOP - 24
-------
SUMMARY OF QA CRITERIA
INITIAL CALIBRATION DATA (FORM Via):
Maximum %RSD = 30%, Delta-BHC = 35%
EPA CHECK STANDARD SUMMARY (FORM VIb):
QC Limits 80 - 120% recovery, Delta-BHC 70 -120% recovery
CONTINUING CALIBRATION CHECK (FORM VII):
Maximum %D = 25%, Delta-BHC = 30%
PERFORMANCE CHECK SOLUTION SUMMARY (FORM V):
PCI
Beta-BHC and Gamma-BHC baseline resolved: Y
% Valley (Chloronaphthalenes): < 30%
S/N Ratio (Beta-BHC m/z 217): > 2.5
S/N Ratio (1,2,4,5-Tetrachlorobenzene m/z 234): .... > 2.5
PC2
Beta-BHC and Gamma-BHC % valley: < 15%
% Valley (Chloronaphthalenes): < 50%
S/N Ratio (Beta-BHC m/z 217): > 2.5
S/N Ratio (1,2,4,5-Tetrachlorobenzene m/z 234): > 2.5
RATIOS
148/146 1,2-Dichlorobenzene: 0.52 - 0.78
182/180 1,2,4-Trichlorobenzene: 0.78 - 1.17
164/162 2-Chloronaphthalene: 0.26 - 0.39
216/214 1,2,3,4-Tetrachlorobenzene: 1.04 - 1.56
219/217 Alpha-BHC: 1.04 - 1.56
219/217 Beta-BHC: 1.04 - 1.56
219/217 Gamma-BHC: 1.04 - 1.56
219/217 Delta-BHC: 1.04-1.56
213/212 DIO-Pyrene: 0.14-0.23
Illustration 10
SOP - 25
-------
SOIL SURROGATE PERCENT RECOVERY (FORM II):
1,4-Dibromobenzene: 51 - 104
1,2,4,5-Tet.rabromobenzene: 50 - 115
2,4,6-Tribromobiphenyl: 56 - 126
MATRIX SPIKE/MATRIX SPIKE DUPLICATE RECOVERY SUMMARY (FORM III);
Recovery
RPD Limit
Compound Limit (Advisory)
1,2-Dichlorobenzene 30 39 - 105
1,2,4-Trichlorobenzene 30 42 - 106
2-Chloronaphthalene 30 43 - 114
1,2,3,4-Tetrachlorobenzene 30 62 - 113
Alpha-BHC 30 65 - 139
Beta-BHC 30 64 - 106
Gamma-BHC 30 66-119
Delta-BHC 30 64 - 138
BLANK SUMMARY (FORM IV):
QC Limit < 0.50 ppb, < 0.60ppb for 1,2-Dichlorobenzene
INTERNAL STANDARD RESPONSE AND RETENTION TIME VERIFICATION
DATA SHEET (FORM X-I TO X-V):
D4-1,4-Dichlorobenzene, D8-Naphthalene, DIO-Acenaphthene:
RT Limit: +/- 10 seconds
Area Response: -60% to +50%
DIO-Phenanthrene, DIO-Pyrene:
RT Limit: +/- 10 seconds
Area Response: -60% to +100%
Illustration 10 cont. SOP - 26
-------
Date: 03/IB/SB DATA VALIDATION SYSTEM • Page 1
Time: 07:47:02 Initial Calibration Checklist Incomplete Lab:
Package Date Initial Calibration ID
============ ======================
03/16/88 IHP5970 1022B7
03/16/88 IHP5970 112087
03/16/B8 IHP5970 120387
Illustration 11 SOP - 27
-------
Date: 03/1B/BB
Time: 07:48:17
DATA VALIDATION SYSTEM
Shi-ft Results Checklist Incomplete
Page
Lab:
Package Date
S===:===S=EC=
03/16/BB
03/16/88
03/16/88
03/16/BB
03/16/88
03/16/8B.
03/16/88
03/16/88
03/16/BB
03/16/BB
03/16/88
03/16/88
03/16/BB
03/16/BB
03/16/BB
03/16/BB
03/16/88
03/16/BB
03/16/88
03/16/88
03/16/88
03/16/BB
03/16/88
03/16/88
03/16/88
03/16/BB
03/16/8B
03/16/88
03/16/BB
03/16/88
03/16/88
03/16/88
03/16/88
Initial Calibration ID
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHPS970
IHP5970
IHP5970
IHP5970
1HP5970
IHP5970
IHP5970
IHP5970
IHP5970
1HP5970
IHP5970
1HP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
1HP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
1HP5970
IHP5970
1 HP5970
1022B7
1203B7
1022B7
102287
102287
1022B7
102287
1 022B7
1022B7
102287
1 02287
112087
1120B7
1120B7
112087
112087
112087
11 2087
1203B7
120387
120387
120387
120387
1203B7
120387
120387
120387
1203B7
1203B7
1203B7
120387
1203B7
1 02287
Shi-ft ID
SSSSSBCS
DS171025.
QS171203.
DS871022.
QSB71029.
DS871030.
DS871101.
QSB71104.
QSB71111.
QSB71114.
QSB71116.
QSB7111B.
DS871119.
DSB71120.
QSB71125.
CSB71129.
QSB71130.
DSB71201.
DSB71202.
DSB71203.
OSB71209.
QSB712I6.
DS871217.
QSB71218.
QSB71220.
QSB71221.
05871222.
DSB71223.
DS87122B.
OSB71229.
DSB71230.
DSB71231.
QS8B0107.
QS971118.
Illustration 12
SOP - 28
-------
Date:
Time:
03/18/88
07:55:11
DATA VALIDATION SYSTEM
Sample Checklist Incomplete
-Page 1
Lab:
Package Date
Initial
Calibration ID
Shi-ft ID
Lab Analysis ID
03/16/88
03/16/88
03/16/88
03/16/88
03/16/88
03/16/88
03/16/88
03/16/88
03/16/88
03/16/88
03/16/88
03/16/88
03/16/83
03/16/88
03/1 6/BB
03/16/88
03/1 6/BB
03/16/88
03/16/88
03/16/88
03/16/88
03/16/88
03/16/88
03/16/88
03/16/88
03/16/88
03/16/8B
03/16/88
03/16/88
03/16/88
03/16/88
03/16/88
03/16/88
03/16/88
03/16/88
03/16/88
03/16/88
03/16/88
03/16/88
03/1 6/BB
03/16/88
03/16/88
03/16/BB
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
1HP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
IHP5970
102287
102287
102287
1203B7
102287
102287
102287
102287
1 1 2087
112087
102287
102287
1022B7
102287
120387
120387
120387
120387
1 02287
120387
1022B7
120387
1203B7
1203B7
102287
102287
1022B7
102287
102287
1022B7
102287
102287
1022B7
1O22B7
102287
1022B7
120387
120387
120387
102287
102287
1O22B7
1022B7
OS171025.
QSB71029.
QSB71030.
QSB71223.
QS871111.
05871114.
DSB7111B.
QS871116.
DS871120.
QS871125.
DS871101.
QS171025.
QS171025.
DS171025.
QS871222.
OSB71203.
DSB71223.
QSB71203.
DS171025.
QSB71216.
QS171025.
QS871216.
DSB71222.
QSB71223.
QSB71029.
QSB71111.
QSB71029.
QSB71029.
QSB71111.
QSB71111.
QSB71029.
QSB71111.
QS871029.
QSB71029.
QSB71029.
DSB71029.
OSB71230.
QSB71230.
QSB71231.
QSB71111.
QSB71029.
DSB71029.
QSB71029.
8709189-
8709189-
8709189-
8709189-
8709189-
8709189-
8709189-
B7091B9-
8709189-
87091B9-
87091B9-
8710140-
B71014O-
8710140-
B710140-
8710140-
8710140-
B710140-
8710140-
871014O-
8710140-
8710140-
B710140-
8710143-
8710143-
8710143-
8710143-
8710143-
8710143-
8710143-
8710143-
8710143-
8710152-
S710152-
B710152-
B710152-
8710152-
8710152-
8710152-
8710152-
8710152-
8710152-
871O152-
01
•02
03
•03R
05
•06
07
•08
•09
•11
•12
•01
•02
•03
•O4A5
•05AS
•065
•07AS
•08
•09R
•11A
•11RE
•12B
•025
•03
-04
•05
•O6
•07AS
-08
•09
-10BH
•01
-01-MS
•01-MSD
-02
•030R
-03SRE
•O4T
-05
-06
-07A
•08A
Illustration 13
SOP - 29
-------
12/23/87 17:20
FINAL
Version: 2.30
LCIC ANALYSIS DATA FORM SAMPLE ID: HS0718
SEfll -VOLATILE LCIC
LABORATORY : ZA&
Lab Project No: 44358
Lab Login Id: 589945
Screened:
Analyst: TERMINAL
Semi-Volatile SIM: 12
Data Released/Authorised
28
PARAMETER (Semi-Volatile
1 ,2-Dicniorooenzene
// LL
1 ,2 ,4-Tr icnlorooenzene
G
Date/Tine
Analyzed
/17/87 18:33
b y :/c a^eu2.tb-L~<
> CAS NUMBER
000095-50-1
000120-82-!
1 ,2 ,3 ,4-TetrachloroDen:ene 000634-66-2
G
2-Chioronaonthaiene
G
rUofia-BHC
G,
Delta-SHC
Beta-BHC
G
Gamne-BHC
G
000091-58-7
000319-64-6
000319-86-8
000319-85-7
000058-89-9
Internal OC Report No: -
Percent Moisture: 19.E
Sulfur Cleanup:
Uelgnt Extracted: 20
Cone/Oil Shift Results File Nane/
Factor Data File Nane
1 QS1217Z7.id£>
>D1930::G6
fVu^aC-iu^
GC/MS SIM
Cone. Data ID Criteria
/ L A 1
fier All Scan Ion RRT
Ions Range Ratio
.39
.41
.39
2.03
ND +
NO • *
.12
ND +
COMMENTS:
XT
Forn I
Illustration 14
SOP -30
-------
Appendix I
Supporting Information on Results of
Sampling and Analytical Activities
-------
Appendix I
SUPPORTING INFORMATION ON RESULTS OF SAMPLING AND
ANALYTICAL ACTIVITIES
Figures 1-1 through 1-11 presented in this appendix show the
locations of all samples collected for the LCIC study.
Figures 1-1 through 1-7 show the location of the samples
collected in the seven EDA sampling areas, and Figures 1-8
through 1-11 show the locations of the samples collected in
Che.ektowaga, Tonawanda, Census Tract 221, and Census
Tract 225, respectively.
Table 1-1 shows the analytical results of the LCICs analyzed
for each of the sample locations shown on Figures 1-1
through 1-11. Table 1-1 indicates the following:
o Seven samples were not collected.
o Eight samples could not be extruded.
o Ninety-eight samples could not be analyzed.
o Seventy-six contingency samples were collected.
o Some LCICs did not pass QC criteria.
1-1
-------
Figure 1-1
SOIL ASSESSMENT - INDICATOR CHEMICALS
SAMPLING AREA 1 OF THE EDA
LOCATION OF ACTUAL SAMPLING POINTS
SCALE: V-3001
LEGEND
-- NEIGHBORHOOD BOUNDARY
— - FENCE LINE AROUND LOVE CANAL REMEDIATION SITE
263 SITE NO.
SAMPLING POINT
SOURCE: EDA BOUNDARIES TAKEN FROM NEW YORK
STATE PROPERTY TAX LAW ARTICLE 17, SECTION 1702
-------
1-
Uj
LU
cc
rr>
*- j
1
O
O
. J_
WHEATF1ELD AVE
•S3-
/
/ 51 •
.901 .87
.52.54
53 55
.56
57 ^
k ,.902
158
.83
59 82
'
,
*
50
61
62
53
.64R
*65
,977
.66
.67
77
*
.76
•
.70
^^_
^^
\
8f
~'
=5
K-
Lu
Lu
or
t--.
co
X>3
30
h—
CO
O
'9
75
'4
-^.
.86
.90
84
.9
92
•
.93
>81.94
a
•
96
*
.76
2
.97.98
^904
\
\~~
Lu
9S
100
•
.101
102 104
t •
.103
.105
^ ' UNPAVED RD
l f
CO
Q
•c.
OJ
o
^ ^
i
1.106
.107
122
*
/108
/a-it 108 123
13JD • •
907
;.iio
,111
.905
CENTER RL
jn,
k11J .114
.139.138
V-115
(~~
X
LU
CO
Q
Ct
CO
o
I-
LU
cc
«
I 140 -J
3
""\"^v^
"^\
Q
0
^^
116-^
.118 .119
7 121
120 7 *
.124 .125. 126
.127.128
1
117
[
.129 .
.130 ^133
jsi ;32 j
.135.136. 134R J37 J
J14,R
142 143 i-144
146
\ 14 / 14fi ^
M^5 * * *
^~^.
1
-149
Figure 1-2
SOIL ASSESSMENT - INDICATOR CHEMICALS
SAMPLING AREA 2 OF THE EDA
LOCATION OF ACTUAL SAMPLING POINTS
SCALE: 1"=300'
LEGEND
"- - NEIGHBORHOOD BOUNDARY
— - FENCE LINE AROUND LOVE CANAL REMEDIATION SITE
263 SITE NO.
• SAMPLING POINT
SOURCE: EDA BOUNDARIES TAKEN FROM NEW YORK
STATE PROPERTY TAX LAW ARTICLE 17, SECTION 1702
-------
~~^^ BOULB/ARD^^
-— _ pisr-
!L~
• Uj ,152
if
JCT) 154
jl—
!§>
V..—-.J
1
1
1
1
1
1
1
1
1
i
j
•
i
*
1
i
i
•
i
i
i
i
•
i
i
_ L
LU
^_
Cfl
• "
,0
-u
i
—
.153
.155
156,909
157
^955
.191.
.158 J59
.160
.161.162 .
.163 *
.164 2«>.
.165
166
• *
167R
.168
177
178
182
186
187
19:
I94
I96
197
199
204
>05
r —
200:-^
• ILU
.169
.966
o:
£. \\J |
on^,
r
h
Figure 1-3
SOIL ASSESSMENT - INDICAT
c. i , r^-
loo
i171
.172
WHEATF
Q
lf~-
216
.174
.-175
(.176
.179
*
.181
.183
.184
.185
188.189
.1!
.193
.195
19
.201
202
ISO
JJ
ct
Q
Z
o
i
•
.206 .207
209
*
4
2
.213 .2
215
.216*
.217
.219.220
12
14
'ELD AVE 1
222R/223
//910
(/.224 I
-S '22B \
^ T •228 J
.229 ^227 T
230 231 981
* • •
232 221
.233
>34
.235
.236 .237
•243,240 ,238
{ 242
\ { *
239 244
•
.245
.246
.247 250
X.249 ( 251
•252,253 -912
/ 254
>55
OR ^HFMI^AI S II
SAMPLING AREA 3 OF THE EDA Wti
LOCATION OF ACTUAL SAMPLING POINTS H
SCALE: V-300' «— v '/SM
LEGEND \ ^
-- NEIGHBORHOOD BOUNDARY • ^^^ LI
— - FENCE LINE AROUND LOVE CANAL REMEDIATION SITE ^^ 1
253 SITE NO. ^V^>^ 1
• SAMPLING POINT ^^^**""'1^
SOURCE: EDA BOUNDARIES TAKEN FROM NEW YORK
STATE PROPERTY TAX LAW ARTICLE 17, SECTION 1702
-------
_/f
1
Y-
— M- — _!_•_•• • JJ M. « •! ™ ^
\
MOSCHEL COURT
~ 1
L
• 1 i
'-U
Lu
n-
u^
oo
£
|
CO
o>
273~*
.256
257
• •
258x .280
913^ .26'
916—1 ,284
285^ 289-,
.259
.339 .293^
.296295
.297
.296
3QO/29C
260 303^91
• •
304
274
278
^.275
it!
y>915
»-^/a
LU
LU
CC
h~
I
O)
282.2*3
.286 .287
.288
.290 J291
.294
301
302
•
305
*
V V
K,
Lu
LU
Q:
co
3;
o
o
r
r ^^
f^
'307
.308.
.322
309
327
•
/332
.310/ .331
.311 330 '
.334
$* (
/315
.314/ .340
\ 316
1 •
\
t^
STREE'
336
O
^ -'
.321
.323
.324
.325
.326
.328
[329
.333
.335
.337
.336
V_BLACK CREEK DHIVt^/
.261 ^270 }
.272.271
.269
.252 .26,
P.5i _ 266.267 _2.K .
.317 .9^320
\ ^31q
X318
•
^ ~ 1
^Ifff^-rT BOULEVARD
COL
Figure I-A
SOIL ASSESSMENT - INDICATOR CHEMICALS
SAMPLING AREA 4 OF THE EDA
LOCATION OF ACTUAL SAMPLING POINTS
SCALE: 1"=300'
LEGEND
-- NEIGHBORHOOD BOUNDARY
— - FENCE LINE AROUND LOVE CANAL REMEDIATION SITE
263 SITE NO.
« SAMPLING POINT
SOURCE: EDA BOUNDARIES TAKEN FROM NEW YORK
STATE PROPERTY TAX LAW ARTICLE 17. SECTION 1702
-------
443
• ,930
f42/ 444R
i. GREENWALD
,931
*-440
.439R
438
. ^/l>-i-
X928Vi^
[22 M19-
Flgure 1-5
SOIL ASSESSMENT - INDICATOR CHEMICALS
SAMPLING AREA 5 OF THE EDA
LOCATION OF ACTUAL SAMPLING POINTS
SCALE: 1"'300'
LEGEND
NEIGHBORHOOD BOUNDARY
— . FENCE LINE AROUND LOVE CANAL REMEDIATION S!TE
263 SITE NO.
SAMPLING POINT
SOURCE: EDA BOUNDARIES TAKEN FROM NEW YORK
STATE PROPERTY TAX LAV*' ARTICLE 17, SECTION 1702
-------
J
ID
^
i1-
jco
I
;g
i^
03
.519.520
.522 ^23
524525
,526 527
530s
.531
*(
535 .
.532
533
535 .537 534
;>
.538.539
in
495
•
495 .499
540 541
•• •
.543
.547
542
,500 .-".•— - -
^939 .503 .506 J
509/.
549 550
• •
.551
552
.558^561
.560 / .562
.976 .555 .566.
.558
566
•
974565 567 _568
563
S61|
Uj
Uj
LO
Figure 1-6
SOIL ASSESSMENT - INDICATOR CHEMICALS
SAMPLING AREA 6 OF THE EDA
LOCATION OF ACTUAL SAMPLING POINTS
SCALE: 1"=3CC'
LEGEND
-- NEIGHBORHOOD BOUNDARY
— . FENCE LINE AROUND LOVE CANAL REMEDIATION SITE
253 SITE NO.
SAMPLING POINT
SOURCE: EDA BOUNDARIES TAKEN FROM NEW YORK
STATE PROPERTY TAX LAV,' ARTICLE 17, SECTION 1702
-------
u
-448
934"
4S7-
4714
472
474
475-
476-
478
.459
•„ &b
-468
607 609 x975°°
foa' I £i;
.614
613 616 616
i -615' '
/620 £23
521 622 '625
.624 .627
13
Figure 1-7
SOIL ASSESSMENT - INDICATOR CHEMICALS
SAMPLING AREA 7 OF THE EDA
LOCATION OF ACTUAL SAMPLING POINTS
SCALE: V-500'
LEGEND
NEIGHBORHOOD BOUNDARY
— - FENCE LINE AROUND LOVE CANAL REMEDIATION SITE
263 SITE NO.
SAMPLING POINT
SOURCE: EDA BOUNDARIES TAKEN FROM NEW YORK
STATE PROPERTY TAX LAW ARTICLE 17, SECTION 1702
/631 629 630 /
I ^32 JS37/.639
-------
Z//////X//////////////TZZZ
-HARLEM RD.
Figure 1-8
SOIL ASSESSMENT—INDICATOR
CHEMICALS
CHEEKTOWAGA COMPARISON AREA
LOCATION OF ACTUAL
SAMPLING POINTS
SCALE: 1"=650'
SOURCE: MAP DRAWN FROM NEW YORK
STATE PROPERTY TAX LAW ARTICLE 17. SECTION 1702
LEGEND
— - SAMPLING BOUNDARY
• SAMPLING POINT
764 SITE NO.
UNACCEPTABLE SAMPLE AREA
-------
Figure 1-9
SOIL ASSESSMENT—INDICATOR CHEMICALS
TONAWANDA COMPARISON AREA (census tract no. 80.03)
LOCATION OF ACTUAL SAMPLING POINTS
SCALE: 1"=650'
LEGEND
... SAMPLING BOUNDARY
• SAMPLING POINT
812 SITE NO.
UNACCEPTABLE SAMPLE AREA
SOURCE: MAP DRAWN FROM NEW YORK
STATE PROPERTY TAX LAW ARTICLE 17, SECTION 1702
-------
/
>
X
J
IJ
670
671
fc
I
^
672,
T
673"'
_^«^
575
676
679 .
V
/697R
/ .698
<&94R
' £ 35
"0*
f G/flA'iC AVEfjJt
.$85
.685
667
688
.689 .6*
.701
702
7:;
.695 .9 ;-
LINOBCHe* AWWL'i
I,
£
E
~— .
680-'
681-'
743
683Rs
6
_:~~^L
£
1
;4
1
691
5<:
^'\
7(i'
: .-)
£
.6%
'06R^
.7»a
.710
.711
.718
r-^53 ~j
/12R
>
i
1
x?13
744^ 7
715
£ .716
*;
-<
t
./17
719] /
• ; /
720' /
72 ^
.722
723-^
*
.725
.726
1 L.
.731
732^
733^
734
.735
736-,
73"
738
739
4
S
J .727
v>
728
,72a 73C
fc
ts
£
.740
74n
74
*
Tp
J
N
Figure 1-10
SOIL ASSESSMENT—INDICATOR CHEMICALS
CENSUS TRACT 221 COMPARISON AREA
LOCATION OF ACTUAL SAMPLING POINTS
SCALE: 1"«400'
LEGEND
_ _ CENSUS TRACT 221 BOUNDARY
SAMPLING POINT
674 j:TE NO.
SOURCE: BOUNDARIES TAKEN FROM NEW YORK
STATE PROPERTY TAX LAW ARTICLE 17, SECTION 1702
-------
879
O876"
944
PCWEfl TOAD
.883
888^
ess-'
^881
* .882
x
t
1
*
\
.BE
8
.M
£
\
886 ^948
89 lv
892 -.
.893
.890
Figure 1-11
SOIL ASSESSMENT—INDICATOR CHEMICALS
CENSUS TRACT 225 COMPARISON AREA
LOCATION OF ACTUAL SAMPLING POINTS
SCALE: T-400'
LEGEND
882
CENSUS TRACT 225 BOUNDARY
SAMPLING POINT
SITE NO.
SOURCE: BOUNDARIES TAKEN FROM NEW YORK
STATE PROPERTY TAX LAW ARTICLE 17. SECTION 1702
-------
SAMPLE
AREA
1-10
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
SITE
NUMBER
1
2
3
4
5
6
6
7
8
8
9
10
1 1
12
13
14
14
15
16
16
17
18
18
19
20
21
22
23
24
25
26
26
27
27
28
29
30
31
SAS
TABLE 1-1
SOIL SAMPLING ANALYSIS RESULTS BV SAMPLING AREA AND SITE
RESULTS GIVEN IN PARTS PER BILLION (PPB)
SPLIT DCB TCB TeCB CNP A-BHC D-BHC B-BHC
G-BHC
LAB
ID
X
X
X
X
X
X
X
X
X
X
X
X
NA
0.39
0.55
NA
0.84
0.81
0.50
1 .51
NA
1 . 18
1 .39
0.75
0.81
1 .50
0.98
1 .08
1.17
1 .20
0.55
0.62
0.40
0.88
0.97
0.94
QCF
QCF
0.52
0.82
0.88
0.56
0.95
1 .06
0.75
0.83
1 .64
1 .33
0.76
0.20
NA
0.80
10.02
NA
8.40
2.11
3. 10
7.21
NA
5.01
15.97
10.91
8.86
13.53
2.69
10.29
12.36
18.51
5.28
6.98
1 .32
6.97
11 .25
8.65
QCF
QCF
4.76
5.62
4.88
1.12
5.41
6.93
6.82
7.56
18.06
1 1 .35
4.76
0.15
NA
1 .31
27 .36
NA
10.87
2.02
3.67
6.70
NA
6.34
17.68
14.68
18.08
32.35
3.79
13. 10
14.75
21 .76
13.30
1 1 .34
3.53
9.12
15.60
1 1 .48
QCF
QCF
7.80
6.46
8.32
1.41
7.76
10.91
1 1 .44
12.00
27.04
QCF
9.04
0.11
NA
ND
0.09
NA
0.14
ND
ND
ND
NA
ND
0.17
ND
O.OS
0.21
ND
ND
ND
0.07
ND
ND
ND
0.09
ND
0. 10
ND
ND
0.06
ND
ND
0.07
0.17
ND
0.07
ND
ND
ND
ND
ND
NA
0.66
24.65
NA
3.73
2.32
2.39
5.02
NA
6.27
15.82
19.61
14.52
26.07
3.31
16.05
25.92
69.70
10.88
13.73
1 .78
10.68
18. 29
5.72
QCF
QCF
4.98
6.10
6. 18
1 . 16
4.81
7. 19
7.88
7.88
20.50
14. 49
6.19
ND
NA
ND
1 .82
NA
0.52
ND
0.36
ND
NA
1 .09
2.04
1 .93
1 .50
38.83
ND
3.39
ND
9.83
2.04
3.33
0.28
1 .20
2.32
0.79
QCF
QCF
ND
ND
0.57
0.30
1.21
ND
1 .39
1 .43
3.46
1 .77
0.70
ND
NA
2.40
QCF
NA
2.96
2.46
2.74
4.64
NA
6. 19
22.60
24.39
47.88
26.55
5.38
19.51
42.81
72.16
24. 17
24. 14
1 .89
10.87
22.29
7. 14
QCF
QCF
4.57
7.65
1 1 .77
2.92
7 .63
13.25
1 1 .58
11.72
55.51
26.63
10.73
ND
NA
ND
4.69
NA
ND
ND
0.35
0.79
NA
1 . 10
3. 24
3. 20
1.11
1 1 .43
0.66
3.38
ND
12.20
2.66
2.97
0.40
1 .87
3.05
1 .09
QCF
QCF
0.94
1 .42
1 .28
0.36
1 .30
1 .87
ND
1 .80
4. 24
2.66
1.14
ND
4
3
2
4
8
7
6
7
4
7
3
2
1
8
7
6
3
4
2
6
2
1
B
8
7
6
4
3
2
1
1
7
6
6
8
3
2
3
LEGEND:
QCF
MAC
ND
NA
NT
SAMPLE ANALYSED BUT FAILED QC CRITERIA
NOT ANALYSED - CONTINGENCY SAMPLE
SAMPLE ANALYSED.ANALYTE NOT DETECTED
SAMPLE NOT ANALYSED
SAMPLE NOT TAKEN/COLLECTED
-------
SAMPLE
AREA
1-10
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
SITE
NUMBER
32
33
34
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
895
896
897
898
899
900
962
51
52
53
54
55
55
56
57
58
59
60
SAS
TABLE 1-1
SOIL SAMPLING ANALYSIS RESULTS BY SAMPLING AREA AND SITE
RESULTS GIVEN IN PARTS PER BILLION (PPB)
SPLIT DCB TCB TeCB CNP A-BHC D-BHC B-BHC
G-BHC
LAB
ID
X
X
X
X
X
1 .50
1 .01
QCF
1 .22
2.47
5.65
2.68
0.92
4.22
2. 22
3.56
QCF
NT
NT
NT
2.46
1 .23
QCF
1 .79
QCF
NAC
MAC
NAC
NAC
NAC
NAC
NAC
NA
0.35
0.64
0.46
NA
0.29
0.61
0.52
NA
0.25
NA
15.73
6.26
QCF
10.99
24.02
45. 1 1
24.82
6.68
41 .74
13.84
31 .68
QCF
NT
NT
NT
12.13
8.69
QCF
10.85
5.65
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NA
1 .90
4.79
1 .05
NA
0.46
1 .82
1 .85
NA
0.49
NA
30.80
10.47
QCF
18.61
50.20
QCF
44.23
14.31
50.77
18.84
67.20
QCF
NT
NT
NT
23.09
12.76
QCF
QCF
8.32
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NA
3.16
6.97
1 .70
NA
0.70
3.18
2.49
NA
0.53
NA
ND
0.14
ND
ND
0. 17
ND
0.14
0. 15
ND
ND
ND
QCF
NT
NT
NT
ND
ND
QCF
ND
0.06
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NA
ND
ND
0.09
NA
0.03
ND
0.04
NA
ND
NA
18.06
3.71
QCF
8.25
15.36
34.68
13.79
18.44
35.09
QCF
18.12
QCF
NT
NT
NT
7.51
3.78
QCF
13.21
6. 10
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NA
0.51
3.92
0.28
NA
0.31
0.60
0.78
NA
0.21
NA
1 .79
ND
QCF
1.13
1 .65
ND
1.71
2.08
3.74
5.82
ND
ND
NT
NT
NT
1 .80
0.81
QCF
ND
ND
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NA
0.26
1 .47
0.16
NA
0.16
0.53
ND
NA
ND
NA
28.07
3.64
QCF
9.01
18.28
44.89
17.51
45. 14
38.99
101 .72
28. 12
QCF
NT
NT
NT
7.07
3.42
QCF
8.05
6.12
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NA
ND
2.20
0.18
NA
ND
0.27
0.13
NA
0.21
NA
2.92
ND
QCF
1 .58
3.09
9.53
2.43
2.95
5.88
20.99
4 . 74
QCF
NT
NT
NT
1 .55
0.80
QCF
2.93
0.92
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NA
0.20
1 .07
0.09
NA
0.13
0.20
0.15
NA
ND
NA
8
8
7
6
4
3
2
1
8
7
6
4
7
6
4
3
2
4
3
2
1
4
8
7
6
4
3
4
LEGEND:
QCF
NAC
ND
NA
NT
SAMPLE ANALYSED BUT FAILED QC CRITERIA
NOT ANALYSED - CONTINGENCY SAMPLE
SAMPLE ANALYSED.ANALYTE NOT DETECTED
SAMPLE NOT ANALYSED
SAMPLE NOT TAKEN/COLLECTED
-------
SAMPLE
AREA
1-10
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
SITE
NUMBER
60
61
62
63
63
65
65
66
67
68
69
70
70
71
72
72
73
74
75
76
76
77
78
79
80
81
82
83
84
84
85
86
87
88
89
90
90
91
SAS
TABLE 1-1
SOIL SAMPLING ANALYSIS RESULTS BY SAMPLING AREA AND SITE
RESULTS GIVEN IN PARTS PER BILLION (PPB)
SPLIT DCB TCB TeCB CNP A-BHC D-BHC B-BHC
G-BHC
LAB
ID
X
X
X
X
X
X
X
X
X
0.21
0. 17
0.20
QCF
QCF
0.75
0.37
0.61
0.44
1 .26
0.80
QCF
QCF
NA
QCF
QCF
1 . 23
0.41
0.57
NA
QCF
4 . 84
NA
0.47
0.68
1 .73
0. 19
QCF
QCF
0.35
0. 23
0.30
0. 28
QCF
0.33
0.40
QCF
0.50
0.19
0.35
0.82
QCF
QCF
1 .66
1 .42
3.81
1 .78
7.14
5.09
QCF
QCF
NA
QCF
QCF
11.18
1.51
3.37
NA
QCF
52.60
NA
1 .39
1 . 25
1 .23
0.65
QCF
QCF
0.75
0.32
0.83
0.57
0.17
0.69
0.92
QCF
1 .59
0. 23
0.67
1 .06
QCF
QCF
2.46
QCF
5.49
2. 29
1 1 .09
8.57
QCF
QCF
NA
QCF
QCF
10.71
2.31
3.63
NA
QCF
66.66
NA
1 .60
0.97
1 .37
0.96
QCF
QCF
0.94
0.26
0.68
0.84
0.27
1 .39
0.60
QCF
1 .53
ND
ND
0.06
QCF
QCF
0.12
0.09
ND
ND
0. 10
0. 17
ND
ND
NA
QCF
QCF
ND
ND
0.14
NA
QCF
0.24
NA
ND
QCF
0.15
0.06
QCF
QCF
ND
0.04
0.04
ND
0.05
0.08
0. 16
ND
ND
0.10
ND
0.18
QCF
QCF
0.61
0.77
1 .40
0.67
5.54
1 .53
QCF
QCF
NA
QCF
QCF
4. 20
0.54
1 .50
NA
QCF
7.17
NA
0.25
0. 28
0. 24
0. 19
ND
QCF
0.36
0.21
0.31
ND
ND
0.35
0.87
QCF
0.28
ND
ND
0.10
QCF
NO
r\ ic
0.22
ND
ND
1 .74
0.47
ND
QCF
NA
QCF
QCF
ND
ND
0.25
NA
ND
2.33
NA
ND
0.27
0.15
0.11
ND
QCF
0.20
ND
ND
ND
ND
0.25
0.68
QCF
ND
ND
ND
ND
ND
QCF
0.68
0.56
0.53
ND
4.26
1 .59
ND
QCF
NA
QCF
QCF
3.74
ND
1 .44
NA
QCF
6.77
NA
ND
0. 39
ND
ND
ND
ND
0.20
0.09
0.45
ND
ND
0.56
0.40
QCF
ND
ND
ND
ND
ND
ND
0.17
0.16
ND
ND
1 .20
ND
QCF
QCF
NA
QCF
QCF
ND
0.26
0.33
NA
QCF
1 .25
NA
ND
4.80
0.05
ND
ND
ND
0. 10
0.05
0.16
ND
ND
0.17
0.32
QCF
ND
2
8
8
7
7
i
4
3
2
1
B
7
7
6
4
4
3
2
1
4
7
6
4
3
2
1
8
7
7
6
4
3
2
1
8
1
7
6
LEGEND:
QCF
NAC
ND
NA
NT
SAMPLE ANALYSED BUT FAILED QC CRITERIA
NOT ANALYSED - CONTINGENCY SAMPLE
SAMPLE ANALYSED,ANALYTE NOT DETECTED
SAMPLE NOT ANALYSED
SAMPLE NOT TAKEN/COLLECTED
-------
SAMPLE
AREA
1-10
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
SITE
NUMBER
92
93
94
95
96
97
98
99
99
100
101
102
103
103
104
105
106
107
108
109
1 10
1 1 1
1 1 2
1 13
1 14
1 15
1 16
1 1 7
1 17
1 18
1 19
1 19
120
121
122
123
124
125
SAS
TABLE 1-1
SOIL SAMPLING ANALYSIS RESULTS BY SAMPLING AREA AND SITE
RESULTS GIVEN IN PARTS PER BILLION (PPB)
SPLIT DCB TCB TeCB CNP A-BHC D-BHC B-BHC
G-BHC
LAB
ID
X
X
X
X
X
X
X
X
0.31
0.22
0. 23
QCF
0.88
0.47
0.39
0.60
0.29
0.32
NA
0.88
0. 28
0. 25
ND
0.13
0.67
0.34
NA
QCF
NA
0.49
0.20
0.93
3.43
NA
QCF
NA
0.40
0.42
0.25
0.23
QCF
0.39
NA
0. 21
0.30
QCF
0.36
0.87
0.32
0.30
0.38
2.38
1 .66
0.45
0.49
1 .46
NA
1 .83
0.47
0.74
0.43
0.22
0.39
1 .55
NA
QCF
NA
1 .20
0.21
1 .60
5.13
NA
QCF
NA
0.52
0.56
0.41
0.35
QCF
1 .95
NA
0.40
0.75
QCF
0.48
1.11
0.31
0.47
0.26
2.77
2.46
0.54
0.67
1 .93
NA
2.19
0.60
0.84
0.77
0.31
0.48
2.11
NA
QCF
NA
1 .50
0.23
1 .48
8.36
NA
QCF
NA
0.54
0.38
0.62
0.66
QCF
2.00
NA
0.37
0.72
QCF
0.08
0.04
ND
0.06
ND
0.07
0.10
QCF
ND
ND
NA
ND
ND
ND
0.07
0.05
ND
ND
NA
QCF
NA
0.08
0.04
ND
ND
NA
ND
NA
ND
0.07
ND
0.07
QCF
0.05
NA
0.06
ND
ND
ND
0.48
ND
0.11
ND
0.66
1 .38
ND
ND
0.35
NA
0.53
ND
ND
0. 19
0.06
0.16
0.51
NA
QCF
NA
0.40
0.08
1 .47
5.92
NA
ND
NA
ND
0.11
0.18
0.16
QCF
0.56
NA
ND
0.28
ND
ND
ND
ND
0.03
ND
0.25
0.60
0.14
ND
0.11
NA
0.37
ND
ND
0.11
ND
0.05
0.42
NA
ND
NA
ND
ND
ND
0.83
NA
ND
NA
ND
ND
ND
ND
ND
0.20
NA
ND
ND
ND
0. 10
0.43
ND
ND
ND
0.49
1 .43
ND
ND
0. 16
NA
0.45
0.07
ND
0.19
ND
0.17
0.61
NA
QCF
NA
0.43
ND
ND
1 .65
NA
ND
NA
ND
0.14
ND
0. 19
ND
0.39
NA
ND
ND
ND
ND
0.27
ND
0.03
ND
0.16
0.40
ND
ND
ND
NA
0.49
ND
ND
0.08
ND
ND
ND
NA
ND
NA
ND
ND
1 .63
2.19
NA
ND
NA
ND
ND
ND
ND
ND
0.11
NA
ND
ND
ND
4
3
2
1
7
6
4
7
2
6
4
6
2
3
1
8
7
6
4
7
4
1
8
7
6
4
3
4
2
1
6
8
7
6
4
8
2
3
LEGEND:
QCF
NAC
ND
NA
NT
SAMPLE ANALYSED BUT FAILED QC CRITERIA
NOT ANALYSED - CONTINGENCY SAMPLE
SAMPLE ANALYSED,ANALYTE NOT DETECTED
SAMPLE NOT ANALYSED
SAMPLE NOT TAKEN/COLLECTED
-------
SAS
TABLE 1-1
SOIL SAMPLING ANALYSIS RESULTS BV SAMPLING AREA AND SITE
RESULTS GIVEN IN PARTS PER BILLION (PPB)
SPLIT OCB TCB TeCB CNP A-BHC D-BHC
B-BHC
G-BHC
LAB
ID
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2 .
2
3
3
1 26
127
1 28
129
1 29
130
131
132
133
135
36
37
37
38
39
140
142
143
144
145
146
147
148
149
150
901
902
903
904
905
906
907
908
64R
1 4 1R
134R
151
152
QCF
0. 29
NA
X QCF
X 0. 32
n or*
\j . *. ct
QCF
0. 30
1.01
NA
19.80
X 0.59
X 0. 26
QCF
0. 29
QCF
NA
0.42
0.66
QCF
0.76
0.54
NA
0.23
0.21
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
0.96
NAC
NAC
0.31
0. 17
QCF
0.81
NA
0.61
1 .25
0.64
0.91
1 .36
0.84
NA
167.33
0.60
0.81
0.64
1 .79
QCF
NA
1 .05
0.61
1 .05
0.85
2.75
NA
1 . 19
0.67
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
7.96
NAC
NAC
0.46
0.08
QCF
1.11
NA
0.81
1 .72
0.91
1 .23
1 .57
0.97
NA
44.83
0.65
1 .09
0.75
2.06
QCF
NA
1 .78
1 .05
1 .06
1.12
4.12
NA
1 .55
1 .21
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
9.85
NAC
NAC
0.32
ND
QCF
0.04
NA
0.05
0.06
0. 04
0.09
0.01
ND
NA
ND
0.06
0.06
0.13
0.07
QCF
NA
0.08
0.08
0. 13
ND
0.08
NA
0.04
0.06
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
0.07
NAC
NAC
0.07
0.06
QCF
0.34
NA
0.34
0.86
0.45
0.40
0.82
1 .30
NA
100.31
0.48
0.83
0. 19
0.50
QCF
NA
0.73
1 .25
0.86
0.73
2.54
NA
1 .48
0.72
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
1 .50
NAC
NAC
ND
0.01
ND
0.07
NA
ND
0.31
ND
0.26
0.16
0.55
NA
ND
0.06
ND
ND
ND
ND
NA
ND
ND
0.15
ND
0.86
NA
0.43
0.09
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
0.66
NAC
NAC
ND
ND
QCF
ND
NA
0.45
1.12
0 51
0.39
1 .47
ND
NA
29.97
QCF
0.63
QCF
0.48
ND
NA
0.76
QCF
1 .82
1 .42
2.78
NA
1 .49
ND
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
1.13
NAC
NAC
ND
ND
ND
0.09
NA
0.13
0.36
0. 15
0.15
0.14
0.51
NA
6.67
ND
0.14
ND
ND
ND
NA
0.14
0.25
0. 26
ND
0.93
NA
0.37
0.22
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
ND
NAC
NAC
ND
ND
7
6
4
1
6
2
1
8
7
4
3
2
8
1
8
7
4
3
2
1
7
6
4
3
8
6
6
3
2
LEGEND:
QCF
NAC
ND
NA
NT
SAMPLE ANALYSED BUT FAILED QC CRITERIA
NOT ANALYSED - CONTINGENCY SAMPLE
SAMPLE ANALYSED.ANALYTE NOT DETECTED
SAMPLE NOT ANALYSED
SAMPLE NOT TAKEN/COLLECTED
-------
SAMPLE
AREA
1-10
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
SITE
NUMBER
153
154
155
156
156
157
158
159
160
160
161
162
163
163
164
165
166
168
168
169
170
171
172
73
74
75
76
77
78
79
180
181
181
182
183
184
185
186
SAS
TABLE 1-1
SOIL SAMPLING ANALYSIS RESULTS BY SAMPLING AREA AND SITE
RESULTS GIVEN IN PARTS PER BILLION (PPB)
SPLIT DCB TCB TeCB CNP A-BHC D-BHC B-BHC
G-BHC
LAB
ID
X
X
X
X
X
X
X
X
X
X
0.48
0.12
0.75
NA
1 .37
NA
0.43
ND
0.50
0. 28
0. 24
1 . 19
0. 24
0.30
NA
0.43
0. 22
QCF
0.21
1 .42
0.29
NA
0. 27
0. 27
0.78
1 .05
0.26
QCF
0.46
0.42
0.31
0.34
0. 27
0.66
0.73
OCF
0.38
QCF
1.71
0.21
0.57
NA
0.82
NA
1 .49
1 .03
1 .85
1 .20
0.72
1 .57
0.38
0.61
NA
1 .68
0.56
1 . 25
0.89
4.85
0.96
NA
1.41
0.29
4.25
4.92
1 . 14
QCF
1 .33
0.57
0.77
0.30
0.43
0.65
6.70
QCF
1 .54
QCF
2.79
0. 19
0.42
NA
0.65
NA
0.57
0.81
1 .89
1 .29
1.21
2.49
0.58
0.92
NA
1 .93
1.14
2.63
1 .93
9.55
3.47
NA
1 .79
0.33
3.44
2.71
0.91
QCF
1 .29
0.44
0.76
0.27
0.38
0.62
4.48
OCF
2.27
QCF
0.08
ND
0. 10
NA
0.04
NA
ND
ND
0.09
0.06
0.07
ND
0.04
ND
NA
0.05
ND
0.06
0.04
ND
ND
NA
ND
ND
0.09
ND
ND
QCF
ND
0.09
0. 10
0. 14
0.08
0.10
ND
QCF
ND
QCF
0.60
0.11
0.33
NA
0.22
NA
ND
ND
ND
0.16
0. 19
0.61
0. 15
0.36
NA
ND
0.27
0.33
0.28
4.06
ND
NA
0.32
ND
0.94
5.06
1 .31
ND
0.30
0. 14
0.27
0.08
ND
ND
3.28
QCF
0.46
QCF
ND
ND
0. 19
NA
ND
NA
NO
ND
ND
ND
0.17
ND
ND
ND
NA
ND
ND
ND
0. 29
79.99
ND
NA
0.12
ND
0. 19
ND
1.81
ND
ND
ND
0.12
0.05
ND
ND
0.50
ND
ND
ND
0.46
ND
ND
NA
ND
NA
ND
ND
ND
ND
0.37
ND
NO
0.14
NA
ND
ND
0.30
0.71
27.91
0.45
NA
0.35
ND
1 .27
ND
0.62
ND
0.15
ND
0.26
ND
ND
ND
2.60
QCF
ND
ND
ND
0.06
0. 15
NA
ND
NA
ND
ND
0.32
ND
0.22
ND
ND
ND
NA
ND
ND
0.13
ND
ND
0.3B
NA
ND
ND
0. 34
12.68
1 .93
ND
ND
ND
0.26
0.02
ND
ND
0.97
QCF
ND
ND
1
8
7
4
6
4
3
3
1
2
8
7
6
3
4
3
2
1
8
7
6
4
3
2
8
7
6
4
3
2
1
1
8
7
6
4
3
2
LEGEND:
QCF
NAC
ND
NA
NT
SAMPLE ANALYSED BUT FAILED QC CRITERIA
NOT ANALYSED - CONTINGENCY SAMPLE
SAMPLE ANALYSED,ANALYTE NOT DETECTED
SAMPLE NOT ANALYSED
SAMPLE NOT TAKEN/COLLECTED
-------
SAMPLE
AREA
1-10
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
SITE
NUMBER
187
188
189
190
191
192
192
193
194
195
196
197
198
198
199
200
201
201
202
203
204
204
205
206
207
208
209
210
21 1
212
213
214
215
216
217
218
219
220
SAS
TABLE 1-1
SOIL SAMPLING ANALYSIS RESULTS BY SAMPLING AREA AND SITE
RESULTS GIVEN IN PARTS PER BILLION (PPB)
SPLIT DC8 TCB TeCB CNP A-BHC D-BHC B-BHC
G-BHC
LAB
ID
X
X
X
X
0.43
0.32
0.98
0.34
NA
2.52
2.37
QCF
0.39
1.11
0.39
NA
QCF
0.32
0.53
QCF
QCF
0.73
QCF
NT
QCF
QCF
0.52
QCF
QCF
0.62
QCF
0.48
QCF
NA
QCF
QCF
0.78
0.37
QCF
0.42
QCF
QCF
0.57
0.82
1 .79
0.92
NA
3.87
3.63
0.48
0.73
2.44
1 .68
NA
0. 29
0.62
1 .08
0.37
0.36
0.35
7.65
NT
QCF
QCF
1 .06
0.32
0.55
1 .30
QCF
3.45
QCF
NA
0.90
0.71
1 .05
1 .00
QCF
1 .84
0. 13
0.36
0.55
1 .06
2.16
1 .36
NA
7 ,49
2 .48
0.78
0. 74
2.36
1 .88
NA
0 .30
0.52
1 .40
0.41
0.43
0.47
6.49
NT
QCF
QCF
0.92
0. 26
0.37
1 .09
QCF
4.50
QCF
NA
0.69
0.87
1 .43
1 .59
QCF
2. 75
0. 10
0.37
0.09
0.10
ND
0.06
NA
ND
ND
0.09
0.12
ND
ND
NA
0.06
0.06
ND
0. 10
0.06
QCF
QCF
NT
QCF
ND
QCF
ND
0.10
QCF
ND
ND
QCF
NA
ND
0. 10
QCF
ND
QCF
ND
ND
0.10
0.20
0.26
ND
0. 27
NA
1 2.84
10.86
0.17
0.14
0.55
0.50
NA
0.12
0. 20
0.44
0.09
ND
0.13
QCF
NT
ND
ND
ND
ND
0.14
1 .85
ND
0.78
QCF
NA
1 .66
0. 18
0.33
0.38
ND
0.65
0.06
0.08
0.10
0.10
ND
ND
NA
ND
ND
ND
ND
0.42
0.27
NA
ND
ND
ND
ND
ND
ND
QCF
NT
QCF
QCF
ND
ND
ND
0.26
ND
0.75
QCF
NA
ND
0.11
ND
0.49
ND
0.45
ND
ND
0.14
ND
24.69
0. 24
NA
QCF
13.68
ND
ND
ND
0.33
NA
ND
ND
0.47
ND
ND
0.22
14.69
NT
QCF
QCF
0.13
ND
0.07
1 .87
ND
0.91
QCF
NA
3.53
0.75
0. 26
1 . 22
QCF
0.66
0.06
0.12
0.06
ND
2.39
0.15
NA
ND
0.47
ND
0.05
0. 20
ND
NA
ND
ND
0.20
0.01
ND
ND
1 .70
NT
QCF
QCF
ND
ND
0.03
0.34
ND
ND
QCF
NA
0.50
0.06
ND
0.17
ND
0.21
ND
ND
1
8
7
6
4
3
3
2
1
7
6
4
2
3
2
1
2
8
7
4
4
3
2
1
8
7
6
4
3
2
1
7
6
4
3
2
1
LEGEND:
QCF
NAC
ND
NA
NT
SAMPLE ANALYSED BUT FAILED QC CRITERIA
NOT ANALYSED - CONTINGENCY SAMPLE
SAMPLE ANALYSED,ANALVTE NOT DETECTED
SAMPLE NOT ANALYSED
SAMPLE NOT TAKEN/COLLECTED
-------
SAMPLE
AREA
1-10
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
SITE
NUMBER
221
221
223
224
225
225
226
227
227
228
228
229
230
231
231
232
233
234
235
236
237
238
239
240
240
241
242
243
244
245
246
246
247
248
249
249
250
251
SAS
TABLE 1-1
SOIL SAMPLING ANALYSIS RESULTS BY SAMPLING AREA AND SITE
RESULTS GIVEN IN PARTS PER BILLION (PPB)
SPLIT DCB TCB TeCB CNP A-BHC D-BHC B-BHC
G-BHC
LAB
ID
X
X
X
X
X
X
X
X
X
X
X
X
X
X
0.70
0.25
NA
0.64
0.39
0.49
0.75
0.29
QCF
0.66
0.73
0.38
NA
0.66
0.30
0.29
0.50
0.47
QCF
0.21
NA
QCF
0.31
NA
0.40
0. 18
0.73
0.47
0.35
0.65
0.44
0.42
0.42
QCF
0.31
0. 16
NA
0.32
0.97
1 .28
NA
1 .22
1.13
3.62
6.25
0.65
QCF
0.83
0.91
1 .46
NA
2.32
0.75
0.72
0.71
2.22
QCF
0.35
NA
QCF
0.92
NA
0.89
0.38
2.60
1 .36
0.54
1 .70
0.57
1 .22
0.24
QCF
0.23
0.29
NA
0.29
1 .48
1 .77
NA
1.89
1 .98
7.12
5.39
1 .00
QCF
1 .03
1 . 10
1 .47
NA
3.04
1 .43
0.66
0.65
1 .89
QCF
0.36
NA
QCF
0.95
NA
1 .09
0.55
2.24
1 .72
1 .22
1 .08
0.55
1.16
0.34
QCF
0.27
0.49
NA
0. 28
NO
0.05
NA
ND
0.09
ND
0.05
0.12
ND
ND
0.10
ND
NA
ND
ND
ND
0.12
ND
ND
ND
NA
ND
ND
NA
0.05
0.06
ND
0.04
0. 15
ND
ND
ND
0. 10
QCF
0.11
ND
NA
ND
0.27
0.24
NA
ND
0.87
499.09
16.05
1.10
QCF
1.13
1 .38
0.90
NA
0.24
ND
0.34
0.64
0.93
QCF
0.08
NA
QCF
0.22
NA
0. 29
0.16
0.17
0.30
0.12
ND
ND
0. 10
0.08
QCF
ND
ND
NA
ND
0.27
ND
NA
ND
ND
ND
2.11
ND
ND
ND
0.21
0. 22
NA
ND
ND
ND
ND
ND
ND
ND
NA
ND
ND
NA
ND
ND
ND
ND
ND
ND
ND
ND
0.13
QCF
ND
ND
NA
ND
ND
ND
NA
ND
QCF
216. 20
51.17
1 .73
QCF
1 .30
1 .59
0.85
NA
ND
ND
0.34
0.61
0.88
ND
ND
NA
QCF
ND
NA
0. 18
ND
ND
ND
ND
0.52
ND
ND
0.15
ND
ND
ND
NA
ND
0. 14
ND
NA
ND
0. 26
6.71
7 . 73
0.31
QCF
ND
0.52
0. 24
NA
ND
0.05
ND
0.17
0.26
QCF
0. 15
NA
ND
ND
NA
ND
ND
ND
ND
ND
ND
ND
ND
0.02
ND
ND
ND
NA
ND
7
8
4
3
2
8
6
8
3
7
7
6
4
3
3
2
1
8
7
6
4
3
2
4
1
8
8
6
8
3
2
3
1
7
2
6
4
3
LEGEND:
QCF
NAC
ND
NA
NT
SAMPLE ANALYSED BUT FAILED QC CRITERIA
NOT ANALYSED - CONTINGENCY SAMPLE
SAMPLE ANALYSED.ANALYTE NOT DETECTED
SAMPLE NOT ANALYSED
SAMPLE NOT TAKEN/COLLECTED
-------
SAS
TABLE 1-1
SOIL SAMPLING ANALYSIS RESULTS BY SAMPLING AREA AND SITE
RESULTS GIVEN IN PARTS PER BILLION (PPB)
SPLIT DCB TCB TeCB CNP A-BHC D-BHC B-BHC
G-BHC
LAB
ID
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
252
252
253
254
255
255
909
910
91 1
912
955
956
167R
981
222R
256
257
258
259
260
261
262
263
264
265
266
267
268
269
269
270
271
271
272
273
273
274
275
X NA
X QCF
0.46
0.33
X 0.55
X QCF
NAC
NAC
NAC
NAC
NAC
NAC
0. 36
NAC
0.40
0.71
2.23
NA
0.30
0.25
0. 29
0.14
0.54
1 .09
NA
0.45
0.36
0.52
0.16
0.17
ND
X NA
X 0.91
0.39
X 0.31
X 0. 27
0.28
QCF
NA
QCF
0.31
0.75
0.82
QCF
NAC
NAC
NAC
NAC
NAC
NAC
0. 24
NAC
1.12
0.23
0.22
NA
0.53
0.40
0.32
0.54
0.33
3.67
NA
0.96
1.13
1 .38
0.75
0.75
1 1 .92
NA
1 .76
1 .75
0.41
0.44
0.42
QCF
NA
QCF
0.45
1 .20
1 .32
QCF
NAC
NAC
NAC
NAC
NAC
NAC
0.37
NAC
1 .07
0.17
0. 19
NA
0.42
0.35
0.34
0.47
0.33
3.80
NA
0.54
1.16
1.51
QCF
QCF
15.51
NA
1 .46
2.01
0.63
0.35
0.28
QCF
NA
QCF
0.09
0.12
0.12
QCF
NAC
NAC
NAC
NAC
NAC
NAC
ND
NAC
0.05
0.11
0.05
NA
ND
0.06
0.05
ND
0.11
0.09
NA
0.05
0.07
0.13
QCF
QCF
ND
NA
ND
ND
0.01
0.06
0.09
QCF
NA
ND
ND
0.87
0.20
QCF
NAC
NAC
NAC
NAC
NAC
NAC
0.08
NAC
0.37
ND
0.04
NA
0.05
ND
ND
0.15
ND
2.36
NA
ND
0.81
1.41
0.60
0.61
152.53
NA
9.94
0.74
0.08
0.07
0.20
QCF
NA
ND
0.05
0.16
ND
ND
NAC
NAC
NAC
NAC
NAC
NAC
ND
NAC
0. 16
ND
ND
NA
ND
ND
ND
ND
0.08
0.98
NA
ND
0.29
0.40
0.31
0.33
ND
NA
ND
ND
ND
ND
ND
ND
NA
ND
ND
0.99
0.09
ND
NAC
NAC
NAC
NAC
NAC
NAC
ND
NAC
0. 27
ND
ND
NA
ND
ND
ND
ND
0.12
2.89
NA
ND
0.36
ND
0.42
0.41
1729. 10
NA
82.47
1.17
ND
ND
ND
QCF
NA
ND
ND
0.18
ND
ND
NAC
NAC
NAC
NAC
NAC
NAC
ND
NAC
0.12
ND
ND
NA
ND
ND
ND
0.14
ND
ND
NA
ND
ND
ND
0. 20
0. 18
47.05
NA
0.65
0.15
ND
ND
ND
ND
4
2
1
8
1
7
1
6
7
6
4
3
2
1
8
7
6
4
3
2
1
8
8
7
4
6
6
6
3
2
7
LEGEND:
QCF
NAC
ND
NA
NT
SAMPLE ANALYSED BUT FAILED QC CRITERIA
NOT ANALYSED - CONTINGENCY SAMPLE
SAMPLE ANALYSED,ANALYTE NOT DETECTED
SAMPLE NOT ANALYSED
SAMPLE NOT TAKEN/COLLECTED
-------
SAMPLE
AREA
1-10
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
SITE
NUMBER
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
296
297
298
298
299
300
301
302
303
304
305
307
308
309
310
31 1
312
SAS
TABLE 1-1
SOIL SAMPLING ANALYSIS RESULTS BY SAMPLING AREA AND SITE
RESULTS GIVEN IN PARTS PER BILLION (PPB)
SPLIT DCB TCB TeCB CNP A-BHC D-BHC B-BHC
10
G-BHC
LAB
ID
0.13
0.56
0.35
NA
0.29
0.31
0.37
0.51
0.17
NA
QCF
0.43
0.44
0.31
0.49
0.36
NA
0.25
0.24
0.36
X 0.37
X NO
0.85
0.32
0.30
NA
0.48
QCF
ND
0.53
0.53
0.21
QCF
0.40
0.56
0.66
0. 28
NA
0. 22
0.69
0.36
NA
0.37
0.47
0.14
0.11
0. 22
NA
QCF
0.30
0.21
0.32
0. 19
0.37
NA
0.71
0.28
0.38
0.48
0.26
4.07
0.53
0.47
NA
1 .98
QCF
ND
0.53
2.40
1 .36
0.41
0. 15
0.28
0.26
1 .09
NA
0.25
0.26
0.34
NA
0.32
0.58
0.10
0.04
0.16
NA
QCF
0.20
0.16
0.31
0.11
0.38
NA
1 .61
0.23
0.35
0.31
0.25
5.44
0.44
0.45
NA
1 .32
QCF
8.46
0.45
2.35
1 .02
0.39
0.13
0.25
0.21
0.64
NA
0.04
0.09
0.02
NA
ND
ND
ND
0. 10
0.03
NA
QCF
0.08
0.11
0. 10
QCF
ND
NA
ND
0.04
0.09
0.07
0.04
ND
0.04
0.05
NA
ND
ND
ND
ND
ND
ND
0.03
0.13
QCF
0.09
ND
NA
ND
0.54
ND
NA
0.07
ND
ND
ND
ND
NA
QCF
ND
0.03
0. 27
0.08
ND
NA
0.08
1 .96
0.06
0.12
ND
4.04
0. 15
0.16
NA
1.51
QCF
129.90
0.09
0.86
16.29
ND
ND
0.08
0.05
0.27
NA
ND
ND
ND
NA
ND
ND
ND
ND
ND
NA
QCF
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
ND
0.14
ND
NA
ND
QCF
ND
ND
1 .93
3.04
ND
ND
ND
ND
ND
NA
ND
0.41
ND
NA
ND
0. 20
ND
ND
ND
NA
ND
ND
ND
ND
0.05
ND
NA
ND
1.81
ND
0.77
ND
2.99
ND
0.14
NA
1 .36
QCF
4108.31
ND
1 .26
241 .25
0.12
ND
ND
ND
0. 18
NA
ND
ND
ND
NA
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
ND
NA
ND
ND
0.04
0.04
ND
0.77
ND
ND
NA
0.11
ND
85.60
ND
ND
5.76
ND
0.03
ND
ND
0.72
NA
8
7
6
4
3
2
1
7
6
4
3
2
1
8
7
6
4
3
2
1
1
8
7
6
6
4
3
7
1
7
6
4
2
1
8
7
6
4
LEGEND:
QCF
NAC
ND
NA
NT
SAMPLE ANALYSED BUT FAILED QC CRITERIA
NOT ANALYSED - CONTINGENCY SAMPLE
SAMPLE ANALYSED,ANALYTE NOT DETECTED
SAMPLE NOT ANALYSED
SAMPLE NOT TAKEN/COLLECTED
-------
SAMPLE
AREA
1-10
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
SITE
NUMBER
313
314
315
316
317
318
319
320
321
322
322
323
324
325
326
327
328
329
330
331
332
333
334
335
335
336
337
338
339
340
913
914
915
916
917
918
965
966
SAS
TABLE 1-1
SOIL SAMPLING ANALYSIS RESULTS BY SAMPLING AREA AND SITE
RESULTS GIVEN IN PARTS PER BILLION (PPB)
SPLIT DCB TCB TeCB CNP A-BHC D-BHC B-BHC
NO
QCF
2.03
41.17
QCF
0.47
NA
ND
ND
ND
QCF
QCF
0.07
NA
ND
QCF
0. 10
ND
ND
0.64
NA
0.22
ND
ND
ND
ND
ND
0.16
NA
ND
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
1 1
0.30
QCF
0.47
0.50
QCF
0.38
NA
0.50
QCF
QCF
QCF
QCF
0.22
NA
0.39
QCF
0.39
0.42
QCF
0.29
NA
0.29
QCF
QCF
QCF
0.31
QCF
0.22
NA
0.36
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
LEGEND:
OCF
MAC
ND
NA
NT
0.49
QCF
1 .86
0.94
QCF
1 . 16
NA
12.08
QCF
QCF
QCF
QCF
0.39
NA
0.71
QCF
0.48
0.18
QCF
0.62
NA
0.43
QCF
QCF
QCF
0.41
QCF
0.34
NA
0.91
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
= SAMPLE
0.43
QCF
1 . 15
0.47
QCF
0.89
NA
182.41
QCF
QCF
QCF
QCF
0.31
NA
0.52
QCF
0.48
ND
QCF
0.98
NA
0.32
QCF
QCF
QCF
0.48
QCF
0.38
NA
0.80
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
ANALYSED
= NOT ANALYSED -
= SAMPLE
= SAMPLE
= SAMPLE
ANALYSED
ND
QCF
0.11
ND
ND
0.05
NA
ND
QCF
QCF
QCF
QCF
ND
NA
0.07
ND
ND
ND
QCF
0.05
NA
0.07
ND
QCF
QCF
QCF
QCF
0.04
NA
ND
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
BUT FAILED
CONTINGENCY
0.13
QCF
1 . 10
10.47
QCF
0.49
NA
0.77
QCF
QCF
QCF
QCF
ND
NA
0. 18
QCF
0.27
0.12
QCF
0.94
NA
0.37
QCF
QCF
QCF
0.12
ND
0. 14
NA
0.28
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NO
ND
0.26
ND
QCF
0.12
NA
ND
ND
ND
NO
ND
0.10
NA
ND
ND
ND
ND
ND
0.12
NA
ND
ND
QCF
ND
ND
ND
ND
NA
ND
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
QC CRITERIA
SAMPLE
.ANALYTE NOT DETECTED
NOT ANALYSED
NOT TAKEN/COLLECTED
G-BHC
ND
ND
16.33
0.82
ND
0.16
NA
ND
QCF
QCF
QCF
ND
0.81
NA
0.20
QCF
0.12
ND
ND
0.12
NA
ND
ND
QCF
QCF
ND
ND
ND
NA
0.09
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
LAB
ID
3
2
1
8
7
6
4
3
2
1
1
7
6
4
3
2
1
8
7
6
4
3
2
1
1
8
7
6
4
3
-------
SAMPLE
AREA
1-10
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
SITE
NUMBER
306
341
342
342
343
344
345
346
347
348
349
349
350
351
352
353
354
354
355
356
357
358
359
360
361
363
364
365
366
367
367
368
369
369
370
371
372
373
SAS
TABLE 1-1
SOIL SAMPLING ANALYSIS RESULTS BY SAMPLING AREA AND SITE
RESULTS GIVEN IN PARTS PER BILLION (PPB)
SPLIT DCB TCB TeCB CNP A-BHC D-BHC B-BHC
12
G-BHC
LAB
ID
X
X
X
X
X
X
0.32
0.36
QCF
QCF
0.48
0.42
0.25
NA
0.39
0.52
QCF
QCF
0.47
0.30
NA
QCF
0.68
0.36
0.44
0.33
0.93
0.31
NA
1 . 19
0.58
1 .22
0.41
0.35
0.59
0.47
0.86
1 .04
0. 18
0. 14
QCF
0.63
0.39
0.58
0.26
0.20
QCF
QCF
0.21
1 .07
1.13
NA
1.12
2.02
QCF
QCF
0.20
0.84
NA
QCF
0.29
0.35
0.29
0.44
0.32
0.96
NA
2.97
1 .33
3.02
0.58
1 .07
1 .44
1.41
3.02
3.44
0.09
0. 10
QCF
0.60
0.46
3.22
0. 18
0.17
QCF
QCF
0.10
1 .05
0.92
NA
1 .52
1 .73
QCF
QCF
0. 19
0.28
NA
QCF
0.18
0.22
0.22
0.34
0.23
3.29
NA
2.61
2.35
0.94
0.62
1 . 19
1 . 10
0.93
2.13
2.46
QCF
0.12
QCF
0.47
QCF
3.06
ND
0. 10
ND
QCF
QCF
0.07
0.05
NA
ND
0. 10
QCF
QCF
0.10
0.09
NA
ND
0.08
ND
0.08
0.11
0.11
ND
NA
ND
ND
0.10
0.04
0.11
ND
ND
0. 10
0. 10
0.06
0.04
QCF
0.09
0.09
NO
ND
0.05
ND
ND
0.05
ND
ND
NA
0.65
2.36
QCF
QCF
ND
0.27
NA
ND
0.04
0.04
ND
ND
ND
ND
NA
5.39
0. 18
0. 24
ND
1 .00
ND
0.40
0.99
1 .55
ND
ND
QCF
0.55
ND
3.32
ND
ND
ND
ND
ND
ND
0.06
NA
ND
0. 13
ND
ND
ND
0.03
NA
ND
ND
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
0.14
0.06
ND
ND
0.70
ND
ND
ND
0.25
ND
ND
0.03
ND
ND
ND
0.07
ND
ND
NA
0.26
1 . 20
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
ND
ND
NA
24.72
ND
0.31
ND
0.42
ND
ND
0.50
0.27
ND
ND
ND
0.39
0. 16
1 .27
ND
ND
ND
ND
ND
ND
ND
NA
0.33
0.21
QCF
ND
ND
ND
NA
ND
ND
ND
ND
ND
ND
ND
NA
2.21
ND
ND
ND
0.22
ND
ND
ND
1.41
ND
ND
QCF
ND
ND
1 .59
3
1
8
8
7
6
8
4
2
1
8
8
7
8
4
3
7
2
1
8
7
6
4
3
2
7
6
4
3
2
6
1
4
8
7
7
4
3
LEGEND:
QCF
NAC
ND
NA
NT
SAMPLE ANALYSED BUT FAILED QC CRITERIA
NOT ANALYSED - CONTINGENCY SAMPLE
SAMPLE ANALYSED,ANALVTE NOT DETECTED
SAMPLE NOT ANALYSED
SAMPLE NOT TAKEN/COLLECTED
-------
SAMPLE
AREA
1-10
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
SITE
NUMBER
374
375
376
377
378
379
360
381
382
383
383
384
385
386
387
388
389
390
391
392
392
393
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
SAS
TABLE 1-1
SOIL SAMPLING ANALYSIS RESULTS BY SAMPLING AREA AND SITE
RESULTS GIVEN IN PARTS PER BILLION (PPB)
SPLIT DCB TCB TeCB CNP A-BHC D-BHC B-BHC
13
G-BHC
LAB
ID
X
X
X
X
0.33
0.25
0.26
0.34
0.91
0.26
NA
QCF
0.27
0. 24
0.34
0.14
1 .09
3. 19
NA
0.39
0.30
0.86
0.57
0.25
QCF
1 .20
1 . 19
NA
0.29
0.66
0.87
1 .73
0.31
NA
NA
QCF
0.57
0.66
QCF
0.27
NA
0.30
0.43
0. 19
0.36
0.31
0.45
0.26
NA
QCF
0.34
0.36
0.44
0.27
4.14
35.65
NA
0.59
0.39
2.28
0.61
0.14
QCF
9.96
10.93
NA
0.68
0.98
3.48
11.14
0.54
NA
NA
QCF
2.07
0.39
QCF
0.91
NA
0.50
0.43
0.17
0.48
0.33
0.25
0. 17
NA
QCF
0.33
0.26
0. 26
0. 29
5.56
51 .97
NA
0.38
0.21
1 .39
0.43
0.09
QCF
168.64
QCF
NA
0.48
0.85
6.61
18.40
0.61
NA
NA
QCF
0.60
0.20
QCF
0.75
NA
0.28
ND
ND
ND
0.11
0.16
0.02
NA
ND
0.05
0.06
0.05
0.04
ND
ND
NA
0.07
0.04
0.07
ND
0.05
QCF
ND
0.03
NA
0.07
ND
0.12
ND
ND
NA
NA
ND
0.13
QCF
QCF
ND
NA
ND
ND
ND
0.06
ND
0.09
ND
NA
QCF
0.17
0.06
ND
ND
1 . 16
83.51
NA
0.04
ND
0.60
ND
ND
QCF
57.60
QCF
NA
0.31
0.20
4.33
10.83
0. 13
NA
NA
QCF
0.20
0.11
QCF
0. 16
NA
0.04
ND
ND
ND
0.05
ND
ND
NA
ND
0.06
ND
ND
ND
ND
ND
NA
ND
ND
0. 19
0.09
ND
ND
9.96
10.11
NA
ND
ND
0.84
4.64
ND
NA
NA
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
ND
ND
NA
ND
ND
0.07
ND
ND
2.08
663.49
NA
ND
0. 19
0.77
ND
0.10
QCF
89.45
QCF
NA
ND
ND
2.80
13.27
ND
NA
NA
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
ND
ND
NA
ND
0.06
ND
ND
ND
ND
13.30
NA
ND
ND
0. 18
ND
ND
ND
26.33
QCF
NA
ND
0.06
4.68
14.02
0.05
NA
NA
ND
ND
ND
ND
0.04
NA
ND
2
2
1
8
7
6
4
3
6
1
2
8
7
6
4
3
2
1
8
1
7
6
6
4
3
2
1
7
6
4
3
2
3
8
7
6
4
3
LEGEND:
QCF
NAC
ND
NA
NT
SAMPLE ANALYSED BUT FAILED QC CRITERIA
NOT ANALYSED - CONTINGENCY SAMPLE
SAMPLE ANALYSED.ANALYTE NOT DETECTED
SAMPLE NOT ANALYSED
SAMPLE NOT TAKEN/COLLECTED
-------
SAS 14
TABLE 1-1
SOIL SAMPLING ANALYSIS RESULTS BV SAMPLING AREA AND SITE
RESULTS GIVEN IN PARTS PER BILLION (PPB)
SAMPLE SITE SPLIT DCB TCB TeCB CNP A-BHC D-BHC B-BHC G-BHC LAB
AREA NUMBER ID
1-10
5 409
5 410
5 411
5 412
5 413
5 414
5 415
5 416
5 417
5 417
5 418
5 419
5 420
5 421
5 422
5 423
5 424
5 425
5 426
5 427
5 428
5 429
5 430
5 431
5 432
5 433
5 434
5 435
5 436
5 436
5 437
5 438
5 440
5 441
5 442
5 443
5 445
5 919
LEGEND:
QCF = SAMPLE ANALYSED BUT FAILED QC CRITERIA
NAC = NOT ANALYSED - CONTINGENCY SAMPLE
NO = SAMPLE ANALYSED,ANALYTE NOT DETECTED
NA = SAMPLE NOT ANALYSED
NT = SAMPLE NOT TAKEN/COLLECTED
X
X
X
X
0.31
0.39
0.74
0.29
NA
QCF
0. 18
0.36
0.42
0.36
0.71
0. 24
NA
NA
0. 17
0.50
0.40
0.58
0. 25
NA
0.50
0.45
0.36
0.54
QCF
0.49
0.21
0.28
0.34
QCF
0.36
0.87
0. 16
0. 26
0.32
0.37
QCF
NAC
0.17
0.63
0.65
0.41
NA
QCF
0.31
0.08
1 .75
0.22
0.27
0.42
NA
NA
0.22
0.67
0.82
0.44
0.47
NA
1 .91
1 . 16
0.40
1 .83
QCF
2.48
0.31
0.76
0.27
QCF
0.34
0.30
0. 15
0.21
0.34
0.29
QCF
NAC
0.14
0.54
0.64
0.40
NA
QCF
0.22
ND
2.32
0.46
0. 20
0. 37
NA
NA
0.12
0.39
0.51
0.25
0.49
NA
QCF
1 . 22
0.33
2.59
QCF
1 .33
0. 22
1 . 10
0. 25
QCF
0.31
0.21
0.09
0. 16
0.24
0.25
QCF
NAC
ND
0.06
QCF
ND
NA
ND
0.03
0.07
0.03
0.12
0.08
0.04
NA
NA
0.05
ND
0. 15
0.08
0.06
NA
ND
ND
0.08
0.12
ND
ND
0.07
ND
0.06
ND
0.08
0.08
0.06
ND
0.07
0.08
QCF
NAC
ND
0.16
0.38
0.28
NA
QCF
ND
ND
1 .66
ND
ND
0. 20
NA
NA
ND
0.74
0.58
0.16
0.21
NA
0.51
0.32
0.17
0.94
QCF
0.94
ND
0.14
ND
QCF
0.14
0.16
ND
ND
ND
0.03
QCF
NAC
ND
ND
ND
ND
NA
ND
ND
ND
0. 34
ND
ND
ND
NA
NA
ND
0.11
0.13
ND
ND
NA
ND
ND
ND
0.17
ND
0. 23
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
NAC
ND
ND
ND
ND
NA
QCF
ND
ND
2.49
ND
ND
ND
NA
NA
ND
0.64
0.48
ND
ND
NA
ND
ND
ND
1.11
ND
0.68
ND
ND
ND
ND
ND
ND
ND
ND
ND
0.01
QCF
NAC
ND
0.06
ND
0.20
NA
ND
ND
ND
0.49
ND
ND
ND
NA
NA
ND
ND
ND
ND
0.22
NA
ND
ND
0.04
0.32
QCF
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
NAC
2
1
7
6
4
3
2
1
6
8
7
6
4
4
2
1
8
7
6
4
3
2
1
8
7
6
4
3
2
3
1
7
4
3
2
1
7
-------
SAS
TABLE 1-1
SOIL SAMPLING ANALYSIS RESULTS BV SAMPLING AREA AND SITE
RESULTS GIVEN IN PARTS PER BILLION (PPB)
SPLIT DCS TCB TeCB CNP A-BHC D-BHC B-BHC
G-BHC
LAB
ID
15
920
921
922
923
924
925
920
927
928
929
930
931
964
362R
984
439R
444R
481
482
483
484
485
486
487
488
489
490
491
493
494
495
496
497
498
499
500
500
501
NAC
NAC
NAC
NAC
NAC
NAC
N Av-
NAC
NAC
NAC
NAC
NAC
NAC
0.40
NAC
NAC
0.30
QCF
0.36
0.27
0.62
0. 26
NA
QCF
QCF
0.38
0.30
0.68
0.21
0.23
QCF
0.56
0.22
0.65
0.31
X 0.52
X 0.25
0.43
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
0.86
NAC
NAC
0.57
QCF
0.21
0.49
0.33
0.55
NA
QCF
QCF
0.52
0. 19
0.38
0.13
0.19
QCF
0.34
0. 15
0.51
0.30
0.28
0. 18
0.87
NAC
NAC
NAC
NAC
NAC
NAC
NA u
NAC
NAC
NAC
NAC
NAC
NAC
0.82
NAC
NAC
0.51
QCF
0.13
0.39
0. 29
0 .44
NA
QCF
QCF
0.38
0.25
0.22
0.04
0.11
QCF
0.32
0.11
0.46
0.32
0.30
0.14
0.73
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
0.07
NAC
NAC
ND
NO
0.07
0.06
0.09
ND
NA
ND
ND
ND
0.11
0.05
0.08
ND
ND
0. 10
0.05
ND
ND
0.05
0.11
ND
NAC
NAC
NAC
NAC
NAC
NAC
MM V*
NAC
NAC
NAC
NAC
NAC
NAC
0.21
NAC
NAC
0.19
QCF
ND
0.18
0.23
0.18
NA
ND
QCF
0.16
0.07
ND
ND
ND
ND
0.08
ND
0.16
ND
ND
ND
0.15
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
ND
NAC
NAC
ND
ND
ND
ND
ND
0. i 1
NA
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
0. 15
NAC
NAC
0.16
ND
ND
ND
ND
0.11
NA
ND
ND
0.14
ND
ND
ND
ND
ND
0.11
ND
0.36
ND
ND
ND
ND
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
0.06
NAC
NAC
ND
ND
ND
ND
ND
0.08
NA
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
1
6
2
1
8
7
6
4
3
2
1
8
7
4
3
2
1
8
7
6
1
4
3
LEGEND:
QCF
NAC
ND
NA
NT
SAMPLE ANALYSED BUT FAILED QC CRITERIA
NOT ANALYSED - CONTINGENCY SAMPLE
SAMPLE ANALYSED,ANALVTE NOT DETECTED
SAMPLE NOT ANALYSED
SAMPLE NOT TAKEN/COLLECTED
-------
SAMPLE
AREA
1-10
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
SITE
NUMBER
502
503
504
505
506
507
508
509
510
51 1
512
513
514
515
516
517
518
519
520
521
522
523
524
525
525
526
527
528
529
530
530
531
532
533
534
534
534
535
SAS
TABLE 1-1
SOIL SAMPLING ANALYSIS RESULTS BY SAMPLING AREA AND SITE
RESULTS GIVEN IN PARTS PER BILLION (PPB)
SPLIT OCB TCB TeCB CNP A-BHC D-BHC B-BHC
16
G-BHC
LAB
ID
X
X
X
X
X
X
X
0.29
0.48
0.78
0. 27
0.26
QCF
0.31
0.53
0.33
1 .08
0.17
NA
0.20
0.30
0.28
0.35
0.35
0.76
0.28
NA
QCF
0.23
0.44
0. 29
0.16
0.97
0.36
NA
0.24
0.26
0.38
0.36
0.24
1 .03
QCF
0.32
0.17
QCF
0.26
0.30
1 .20
0.91
0. 19
QCF
0.40
0.28
0.23
0.79
0.17
NA
0.11
0.79
0.71
0.52
0.57
0.53
0.37
NA
QCF
0.39
0.26
0. 13
0.17
0.37
0.46
NA
0.37
0. 13
0.42
0.29
0.28
0.36
QCF
0.12
0.33
QCF
0.16
0.61
0.83
0.99
0. 15
QCF
0.34
0. 27
0.23
0.58
0.17
NA
0.08
0.41
0.66
0.47
0.61
0.33
0.24
NA
QCF
0.28
0.22
0.05
0.10
0.12
0.39
NA
0.24
ND
0.26
0.24
0.27
0.22
QCF
QCF
0.31
QCF
ND
0.09
0.11
0.04
0.10
ND
ND
0.10
0.09
ND
ND
NA
0.03
ND
0.08
ND
ND
0. 13
ND
NA
QCF
0.05
O.OB
0.05
0.03
0. 18
ND
NA
0.02
ND
ND
ND
0.05
0.12
QCF
QCF
ND
QCF
ND
0.43
0.36
0.23
ND
ND
ND
0.06
ND
0.33
ND
NA
ND
0.49
0. 29
0.07
0.19
0.11
0.07
NA
QCF
ND
ND
ND
0.03
ND
0.12
NA
ND
ND
0.03
ND
ND
ND
ND
0.02
0.05
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
3.44
ND
NA
ND
ND
ND
ND
ND
ND
ND
NA
ND
ND
ND
0.06
ND
ND
ND
NA
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
0.21
ND
ND
0.10
ND
1 .04
ND
NA
ND
0.59
ND
ND
ND
ND
ND
NA
QCF
ND
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
0.89
ND
ND
ND
ND
ND
ND
ND
ND
ND
NA
ND
0.43
ND
ND
ND
0.04
ND
NA
ND
ND
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
2
1
7
6
4
3
2
1
a
7
6
4
3
2
2
1
8
7
6
4
3
2
1
1
8
7
6
4
3
2
3
1
8
7
2
2
6
4
LEGEND:
QCF
NAC
ND
NA
NT
SAMPLE ANALYSED BUT FAILED QC CRITERIA
NOT ANALYSED - CONTINGENCY SAMPLE
SAMPLE ANALYSED,ANALYTE NOT DETECTED
SAMPLE NOT ANALYSED
SAMPLE NOT TAKEN/COLLECTED
-------
SAS 17
TABLE 1-1
SOIL SAMPLING ANALYSIS RESULTS BV SAMPLING AREA AND SITE
RESULTS GIVEN IN PARTS PER BILLION (PPB)
SAMPLE SITE SPLIT DCB TCB TeCB CNP A-BHC D-BHC B-BHC G-BHC LAB
AREA NUMBER ID
1-10
6 536 0.16 0.04 NO ND ND NO ND ND 3
6 537 0.29 0.45 0.35 ND ND ND ND ND 2
6 538 0.25 0.14 0.08 0.10 0.03 ND ND ND 1
6 539 0.62 0.37 0.34 0.15 ND ND ND ND 7
6 540 0.21 0.47 0.51 0.04 0.19 ND ND ND 6
6 541 QCF QCF QCF QCF NO ND ND ND 4
6 542 0.36 0.88 0.70 ND 0.09 ND ND ND 3
6 543 0.53 1.49 1.18 0.14 1.22 0.25 0.85 0.30 2
6 544 0.64 3.51 2.90 0.15 1.99 0.97 3.15 0.47 1
6 545 0.38 0.89 1.08 0.10 0.48 ND ND ND 8
6 546 0.87 0.48 0.32 0.18 ND ND ND ND 7
6 547 0.22 0.71 0.68 ND 0.38 ND ND ND 6
6 548 QCF QCF QCF QCF QCF ND ND ND 4
6 549 0.52 0.86 0.68 0.10 0.56 ND ND ND 3
6 550 0.55 0.26 ND 0.09 0.20 ND ND ND 2
6 551 0.50 0.85 0.72 0.11 0.52 ND 0.34 0.22 1
6 552 0.42 0.51 0.53 0.15 0.23 ND ND 0.13 8
6 553 1.51 0.37 0.22 0.19 ND ND ND 0.56 7
6 554 0.22 0.59 0.58 0.04 0.24 ND ND 0.08 6
6 555 NA NA NA NA NA NA NA NA 4
6 556 0.39 0.49 0.29 ND 0.06 0.19 ND ND 3
6 557 QCF QCF QCF QCF ND ND ND ND 2
6 558 0.42 1.00 1.22 0.10 0.53 0.09 0.26 0.05 1
6 559 0.91 0.36 0.21 0.12 ND ND ND ND 7
6 560 0.24 0.34 0.25 0.05 0.13 ND ND ND 6
6 561 NA NA NA NA NA NA NA NA 4
6 562 0.45 1.09 0.90 0.09 0.28 ND ND ND 3
6 563 QCF 0.44 0.35 0.08 0.28 ND 0.31 ND 2
6 564 0.39 0.47 0.50 0.08 0.28 ND 0.20 0.07 1
6 565 0.26 0.79 1.27 0.07 0.13 0.03 ND ND 8
6 566 0.69 0.33 0.28 ND ND ND ND ND 7
6 567 0.24 0.32 0.1B ND ND ND ND ND 6
6 568 NA NA NA NA NA NA NA NA 4
6 569 0.45 1.22 1.00 0.09 0.45 ND ND ND 3
6 570 QCF 0.39 0.29 ND 0.22 ND ND ND 2
6 571 QCF 0.22 0.19 0.11 0.06 ND ND ND 1
6 572 0.28 0.35 0.38 0.08 0.14 ND ND ND 8
6 573 1.14 0.40 0.39 0.10 ND ND ND ND 7
-------
SAS
TABLE 1-1
SOIL SAMPLING ANALYSIS RESULTS BY SAMPLING AREA AND SITE
RESULTS GIVEN IN PARTS PER BILLION (-PPB)
SPLIT DCB TCB TeCB CNP A-BHC D-BHC B-BHC
G-BHC
18
LAB
ID
574
576
577
578
579
580
581
582
582
583
584
585
586
587
588
937
938
939
940
941
942
974
976
492R
446
447
448
449
450
451
452
453
454
455
456
457
457
458
0.27
0.31
0.64
QCF
0.69
0.25
NA
0.37
QCF
0.64
0.42
0.32
0.65
0.85
NA
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
0. 29
NA
0.41
0. 34
0.43
0.34
0.94
0.38
NA
0.32
0. 31
0.41
X 0.62
X 0.20
0.76
0.32
0.30
0. 19
0.21
0.30
0.39
NA
0.46
QCF
0.16
0.56
0.36
0.41
0. 15
NA
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
0.78
NA
0.45
0.59
1.31
QCF
0.63
0.56
NA
0.48
0.33
0.95
0.25
0.39
0.64
0.35
0.25
0.12
0.16
0.14
0. 44
NA
0.20
QCF
0.10
0.54
0.27
0.25
0.11
NA
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
0.90
NA
0.30
1 .24
1 . 25
0.29
0.64
1.11
NA
0.36
0.28
0.79
0.26
0.38
QCF
ND
0.07
0.05
0.08
0.08
ND
NA
ND
ND
0.08
0. 10
0.11
ND
0.03
NA
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
ND
NA
ND
ND
0.08
0.10
0.13
ND
NA
ND
ND
0.06
0.08
0.04
ND
0.05
0.04
ND
0.11
ND
0.18
NA
ND
QCF
0.09
0. 18
0.17
0.18
ND
NA
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
0.49
NA
0.17
0.14
0.26
0.13
0.25
0. 15
NA
0.13
0.14
0. 15
0.05
0. 10
0. 16
ND
ND
ND
ND
ND
0.11
NA
ND
ND
ND
ND
ND
ND
ND
NA
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
0.24
NA
ND
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
ND
ND
ND
ND
QCF
ND
0.27
NA
ND
ND
ND
ND
ND
0.22
ND
NA
NAC
NAC
NAC
NAC
NAC
NAC
NAC
NAC
0.47
NA
ND
ND
0.38
0.18
0. 19
ND
NA
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
0.28
NA
ND
QCF
ND
0.11
ND
ND
ND
NA
NAC
NAC
NAC
•NAC
NAC
NAC
NAC
NAC
0.20
NA
ND
ND
0.12
ND
ND
ND
NA
ND
ND
0.05
ND
ND
ND
6
3
2
1
7
6
4
3
3
2
1
8
7
6
4
6
4
3
2
1
8
7
6
4
3
2
1
1
8
7
LEGEND:
QCF
NAC
ND
NA
NT
SAMPLE ANALYSED BUT FAILED QC CRITERIA
NOT ANALYSED - CONTINGENCY SAMPLE
SAMPLE ANALYSED,ANALYTE NOT DETECTED
SAMPLE NOT ANALYSED
SAMPLE NOT TAKEN/COLLECTED
-------
SAMPLE
AREA
1-10
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
SITE
NUMBER
459
460
461
462
462
462
463
464
465
466
467
467
468
468
469
470
47 1
472
472
473
473
474
475
476
476
477
478
479
480
575
589
590
591
592
592
592
593
594
SAS
TABLE 1-1
SOIL SAMPLING ANALYSIS RESULTS BY SAMPLING AREA AND SITE
RESULTS GIVEN IN PARTS PER BILLION (PPB)
SPLIT DCB TCB TeCB CNP A-BHC D-BHC B-BHC
19
G-BHC
LAB
ID
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
0.45
NA
QCF
QCF
0.42
0.54
0.48
0.86
0.56
0.43
NA
0.41
0.31
QCF
0.38
0.73
0.72
0.45
0.38
NA
0.37
0.23
0.39
0.63
QCF
0.46
0.69
0.33
NA
NA
0.79
0.31
0.36
QCF
0. 19
0.19
ND
NA
1 . 15
NA
QCF
QCF
1.10
1 .36
0.59
4.49
0.59
0.86
NA
0. 17
0.80
QCF
0.30
3.52
0.93
0.65
0.51
NA
0.80
0.21
0.27
0.44
QCF
0.70
0.58
0.89
NA
NA
3.77
0. 18
0.26
QCF
0.34
0.34
0.22
NA
1 . 10
NA
QCF
QCF
1 .21
1 .22
0.60
3.06
0.67
0.99
NA
0.07
0.84
QCF
0. 10
2.63
0.98
0.53
0.44
NA
0.72
0.11
0.26
0.42
QCF
0.76
0.49
1 .02
NA
NA
3.91
0.12
0. 19
QCF
0.46
QCF
0.25
NA
0.02
NA
ND
QCF
0.05
ND
ND
0.12
ND
0.06
NA
0.05
0.08
ND
ND
ND
0.09
ND
ND
NA
0.04
ND
ND
0.11
ND
0.09
0.13
ND
NA
NA
0.08
0.09
0.03
QCF
ND
QCF
0.02
NA
ND
NA
QCF
QCF
0.34
0.43
ND
0.49
0.14
0.21
NA
ND
ND
QCF
0.05
0.89
0. 19
ND
0.10
NA
ND
ND
ND
0.14
ND
0.21
0.25
0.21
NA
NA
2.42
0.02
0. 10
ND
ND
0.17
ND
NA
ND
NA
ND
ND
ND
ND
ND
ND
ND
ND
NA
ND
0.15
QCF
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
ND
ND
0.12
NA
NA
ND
ND
ND
ND
ND
0.05
ND
NA
ND
NA
ND
ND
ND
0. 26
ND
ND
0.32
ND
NA
ND
0.59
ND
ND
2.01
0.76
ND
ND
NA
ND
ND
ND
0.35
ND
ND
0. 16
0.14
NA
NA
ND
ND
ND
ND
ND
ND
ND
NA
ND
NA
QCF
ND
ND
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
ND
ND
0.07
NA
NA
0.61
NO
ND
ND
ND
ND
ND
NA
6
4
3
2
6
6
1
a
7
6
4
2
8
3
2
1
7
2
6
4
6
3
2
1
3
8
7
6
4
4
2
1
8
7
8
8
6
4
LEGEND:
QCF
NAC
ND
NA
NT
SAMPLE ANALYSED BUT FAILED QC CRITERIA
NOT ANALYSED - CONTINGENCY SAMPLE
SAMPLE ANALYSED,ANALYTE NOT DETECTED
SAMPLE NOT ANALYSED
SAMPLE NOT TAKEN/COLLECTED
-------
SAMPLE
AREA
1-10
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
SITE
NUMBER
595
596
597
598
599
600
601
602
603
604
604
605
606
607
608
609
610
61 1
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
SAS
TABLE 1-1
SOIL SAMPLING ANALYSIS RESULTS BY SAMPLING AREA AND SITE
RESULTS GIVEN IN PARTS PER BILLION (PPB)
SPLIT DCB TCB TeCB CNP A-BHC D-BHC B-BHC
20
G-BHC
LAB
ID
0.36
0.47
0.43
0.33
QCF
0. 16
NA
0.35
0.35
X 0.41
X 0.26
0.29
QCF
0.36
NA
0.32
0.31
0.28
0.79
QCF
NA
0 .44
QCF
QCF
0.33
0.86
0.31
NA
QCF
0.44
QCF
0.46
0.86
0.37
NA
0.34
QCF
0.47
1 . 19
0.77
0.50
0. 24
QCF
0. 17
NA
0.56
0.90
0.24
0.34
0. 25
QCF
0.95
NA
0.76
0.57
0.42
0.67
QCF
NA
1.15
QCF
0.31
0.34
0.47
0.78
NA
QCF
0.37
0.65
1.14
1 .43
1 .33
NA
1 .02
QCF
0.81
0.87
0.64
0. 27
0.35
QCF
0.15
NA
0.56
0.83
0.20
0.32
0.20
QCF
1.11
NA
0.63
0.54
0.31
0.47
QCF
NA
0.86
QCF
0.36
0.37
0.30
1 .05
NA
QCF
0.43
0.62
1 . 18
0.78
1 .65
NA
1.21
QCF
0.71
NO
0.06
0.09
0.10
QCF
0.04
NA
ND
NO
0. 10
0.08
0.07
ND
0.05
NA
ND
0.06
0.07
0.16
ND
NA
0.10
QCF
0.08
0.12
0. IB
0.04
NA
ND
QCF
0. 12
0.11
0.14
ND
NA
0.06
QCF .
0.13
ND
0.20
0.08
ND
ND
ND
NA
0.14
0.51
0.07
ND
0.12
QCF
0.45
NA
0. 29
0.23
0. 12
ND
QCF
NA
ND
QCF
0.17
0. 13
ND
0.18
NA
QCF
0.36
0. 15
0.33
0.56
0.52
NA
0.23
QCF
0.46
ND
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
NA
ND
QCF
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
0. 15
NA
ND
NO
0. 12
ND
ND
ND
ND
ND
ND
NA
ND
0.38
ND
0.29
ND
ND
0.43
NA
0.31
0.25
ND
ND
ND
NA
ND
QCF
QCF
ND
ND
ND
NA
QCF
0.60
QCF
ND
1.11
QCF
NA
0.51
ND
0.29
0.13
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
QCF
NA
ND
QCF
ND
ND
ND
0.08
NA
QCF
ND
0.04
ND
ND
0.26
NA
ND
ND
0.12
3
2
1
8
7
6
4
3
2
1
3
8
7
6
4
3
2
1
7
6
4
3
2
1
8
7
6
4
3
2
1
8
7
6
4
3
2
1
LEGEND:
QCF
NAC
ND
NA
NT
SAMPLE ANALYSED BUT FAILED QC CRITERIA
NOT ANALYSED - CONTINGENCY SAMPLE
SAMPLE ANALYSED,ANALVTE NOT DETECTED
SAMPLE NOT ANALYSED
SAMPLE NOT TAKEN/COLLECTED
-------
SAMPLE
AREA
1-10
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
SITE
NUMBER
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
SAS
TABLE 1-1
SOIL SAMPLING ANALYSIS RESULTS BV SAMPLING AREA AND SITE
RESULTS GIVEN IN PARTS PER BILLION (PPB)
SPLIT DCB TCB TeCB CNP A-BHC D-BHC B-BHC
0.35
ND
NA
0.21
QCF
NO
ND
0.10
QCF
NA
1 .92
ND
1 .24
ND
0.40
0.05
ND
0.28
ND
0.07
ND
ND
ND
ND
ND
0.08
ND
QCF
0. 16
ND
0.84
0.95
1 .74
0.43
ND
0.27
QCF
ND
21
0.62
0. 24
NA
0.54
QCF
0.47
0.23
0.60
0.31
NA
0.50
0.59
0.91
0.20
0.91
0.29
0.21
0.27
0.53
0.34
0.68
0.35
0.31
0. 23
0.56
0.37
0.73
QCF
0.30
0.21
0.70
0.65
QCF
0.69
0.52
0.40
QCF
0.46
LEGEND:
QCF =
NAC =
ND =
NA =
NT =
0.71
0.33
NA
1 .79
QCF
0 75
0.23
0.33
0.64
NA
1 .92
1.15
4.04
0.21
0.64
0.39
0. 18
1.12
0.42
0.52
0.37
1 .20
0.36
0.51
0.40
0.29
0.30
QCF
0.52
0.15
2.30
0.99
1 .73
0.54
0.41
1 .06
QCF
0.95
SAMPLE
0.90
0.38
NA
1.31
QCF
0 54
0.17
0.29
0.44
NA
1 .65
0.57
1.61
0.22
0.51
0.51
0.17
0.65
0. 27
0.37
0.30
1.17
0.43
0.60
0.22
0.31
0.31
QCF
0.53
0.12
2.64
0.61
2.86
1 .03
0.43
1 .20
QCF
1 .73
ANALYSED
NOT ANALYSED -
SAMPLE
SAMPLE
SAMPLE
ANALYSED
ND
ND
NA
0.10
QCF
0.15
0.05
0.04
0.02
NA
ND
0.13
0.17
0.05
QCF
0.03
ND
0.04
0.10
0.11
ND
0.05
0.09
ND
0.12
0.10
QCF
ND
0.05
0.06
NO
ND
0.13
QCF
0.10
0.06
QCF
QCF
0.26
0.11
NA
0.28
QCF
0.21
0.05
ND
0.27
NA
ND
ND
1 .69
ND
0.23
0.06
ND
0.22
0.09
0.11
ND
0.30
ND
0.23
0.12
0.09
0.05
ND
0. 16
ND
1 .00
0.81
1 .00
0.36
ND
0.26
ND
0.47
ND
ND
NA
ND
ND
ND
ND
ND
0.11
NA
1 .25
ND
0. 23
ND
ND
ND
ND
ND
ND
ND
ND
0.34
ND
ND
ND
ND
ND
ND
0.05
ND
0.98
ND
0.76
0. 18
ND
0.22
ND
ND
BUT FAILED QC CRITERIA
CONTINGENCY SAMPLE
, ANALYTE
NOT DETECTED
NOT ANALYSED
NOT TAKEN/COLLECTED
G-BHC
ND
ND
NA
ND
QCF
0. 04
0.12
ND
0.07
NA
ND
ND
0.40
ND
ND
ND
ND
0.09
ND
0.05
ND
0.22
ND
ND
ND
ND
ND
ND
ND
ND
0.36
0.35
0.34
0.16
ND
0.09
ND
ND
LAB
ID
7
6
4
3
2
1
8
7
6
4
3
2
1
8
7
6
4
3
2
1
7
6
4
3
2
1
8
7
6
4
3
2
1
8
7
6
4
3
-------
SAMPLE
AREA
1-10
7
7
7
7
7
7
7
8
8
8
B
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
a
8
8
SITE
NUMBER
932
933
934
935
936
973
975
745
746
747
748
749
750
750
751
752
753
754
754
755
756
757
758
759
760
761
762
763
764
765
766
766
767
768
768
769
770
770
SAS
TABLE 1-1
SOIL SAMPLING ANALYSIS RESULTS BY SAMPLING AREA AND SITE
RESULTS GIVEN IN PARTS PER BILLION (PPB)
SPLIT DCB TCB TeCB CNP A-BHC D-BHC B-BHC
22
G-BHC
LAB
ID
X
X
X
X
X
X
X
X
X
X
NAC
NAC
NAC
NAC
NAC
NAC
NAC
0.26
0.37
0.20
1.01
0.52
0.29
0.85
0.30
NT
0.33
0.38
0.11
0.85
0.44
NA
QCF
0.92
0.27
0.12
0.85
0.35
NA
QCF
0.29
0.35
0.42
0.60
QCF
0.55
NA
0.27
NAC
NAC
NAC
NAC
NAC
NAC
NAC
0. 15
0.13
0.07
0.31
0. 13
0.09
0.24
0.20
NT
0.14
0.09
0.09
0. 14
0. 12
NA
QCF
0.25
0.09
0.07
0. 15
0.25
NA
QCF
0. 15
0.18
0. 1 1
0. 12
QCF
0. 15
NA
0.40
NAC
NAC
NAC
NAC
NAC
NAC
NAC
0.07
0.05
ND
0.08
0.03
0.02
0.06
ND
NT
0.04
0.02
0.04
ND
0.03
NA
QCF
0.05
ND
ND
0.03
0.18
NA
QCF
0.03
ND
0.03
0.03
QCF
0.04
NA
0.05
NAC
NAC
NAC
NAC
NAC
NAC
NAC
0.06
0.08
0.07
0.21
0.06
0.04
0. 10
ND
NT
ND
0.06
ND
ND
0.05
NA
ND
0.06
0.07
0.03
0.06
0.02
NA
ND
0.07
ND
0.09
0.07
ND
0.04
NA
ND
NAC
NAC
NAC
NAC
NAC
NAC
NAC
ND
ND
ND
ND
ND
ND
ND
ND
NT
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
ND
NA
ND
ND
ND
0.01
ND
ND
0.09
NA
ND
NAC
NAC
NAC
NAC
NAC
NAC
NAC
ND
ND
ND
ND
ND
ND
ND
ND
NT
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
ND
NA
QCF
ND
ND
ND
ND
ND
ND
NA
ND
NAC
NAC
NAC
NAC
NAC
NAC
NAC
ND
ND
ND
ND
ND
ND
ND
ND
NT
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
ND
ND
NA
ND
NAC
NAC
NAC
NAC
NAC
NAC
NAC
ND
QCF
ND
ND
ND
ND
ND
ND
NT
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
QCF
ND
NA
ND
2
1
8
7
6
1
7
3
6
1
8
7
6
4
3
2
1
8
7
6
4
7
1
2
1
7
3
6
4
2
LEGEND:
QCF
NAC
ND
NA
NT
SAMPLE ANALYSED BUT FAILED QC CRITERIA
NOT ANALYSED - CONTINGENCY SAMPLE
SAMPLE ANALYSED.ANALYTE NOT DETECTED
SAMPLE NOT ANALYSED
SAMPLE NOT TAKEN/COLLECTED
-------
SAMPLE
AREA
1-10
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
SITE
NUMBER
771
772
773
774
774
775
776
777
777
778
779
780
780
781
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
800
801
802
803
SAS
TABLE 1-1
SOIL SAMPLING ANALYSIS RESULTS BY SAMPLING AREA AND SITE
RESULTS GIVEN IN PARTS PER BILLION (PPB)
SPLIT DCB TCB TeCB CNP A-BHC 0-BHC B-BHC
23
G-BHC
LAB
ID
X
X
X
X
X
X
X
X
X
X
0.18
0.30
0.42
0. 23
0.11
0.72
0.21
NA
0.33
0. 18
0.20
0.38
0.21
0.22
0.32
0.95
0.16
NA
0.32
0.21
0.35
1 .02
0.17
NA
1 .38
0.24
0.36
0.22
0.75
0.44
QCF
0.18
0.53
0.37
0.46
0.25
0.56
0.38
0. 15
0. 19
0.13
0.08
0.08
0.16
0.08
NA
0.08
0.11
0.17
0. 10
0.10
0. 15
0.15
0.20
0.08
NA
ND
0.11
0.14
0.30
0.09
NA
0.27
0. 16
0. 10
0.09
0.24
0. 16
QCF
0.11
0.29
0.12
0.17
0. 13
0.11
0.92
0.04
0.08
0.05
0.03
ND
0.04
0.03
NA
0.07
ND
0.07
0.02
0.02
0.07
0.05
0.10
0.05
NA
0.14
0.07
0.07
0.14
0.08
NA
0.10
0.11
0.07
0.06
0.11
0.07
ND
0.05
0.11
0.05
0.05
0.05
ND
0.84
ND
ND
0.09
ND
ND
ND
ND
NA
0.07
ND
0.03
0.08
ND
ND
0.09
0.13
0.01
NA
ND
ND
0.08
ND
0.04
NA
ND
ND
0.12
ND
ND
ND
ND
0.03
ND
0.14
0.09
0. 10
ND
ND
ND
ND
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
0.17
ND
ND
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
NA
0.12
ND
ND
5.36
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
0.36
ND
ND
0.04
ND
ND
ND
ND
NA
2.54
ND
ND
ND
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
3
2
1
2
8
7
6
4
1
3
6
1
3
6
8
7
6
4
3
2
1
7
6
4
3
2
1
8
7
6
4
3
2
1
2
8
7
6
LEGEND:
QCF
NAC
ND
NA
NT
SAMPLE ANALYSED BUT FAILED QC CRITERIA
NOT ANALYSED - CONTINGENCY SAMPLE
SAMPLE ANALYSED,ANALYTE NOT DETECTED
SAMPLE NOT ANALYSED
SAMPLE NOT TAKEN/COLLECTED
-------
SAMPLE
AREA
1-10
B
8
8
8
8
8
8
8
8
8
8
8
8
8
8
B
8
8
8
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
SITE
NUMBER
804
805
806
807
808
809
810
81 1
812
813
814
815
816
817
818
819
949
950
951
670
671
672
673
674
675
675
675
678
679
680
681
682
684
685
686
687
688
689
SAS
TABLE 1-1
SOIL SAMPLING ANALYSIS RESULTS BY SAMPLING AREA AND SITE
RESULTS GIVEN IN PARTS PER BILLION (PPB)
SPLIT DCB TCB TeCB CNP A-BHC D-BHC B-BHC
24
G-BHC
LAB
ID
QCF
0.99
QCF
QCF
0.74
0.36
QCF
0.31
QCF
QCF
0.38
0.67
0. 23
NA
0.33
QCF
NAC
MAC
NAC
0.53
0.32
0.58
0.33
NA
X 0. 34
X 0.33
X QCF
0. 23
QCF
0.32
NA
1 . 15
0.71
0.41
0.73
0.41
0.29
0.40
QCF
0.61
0.12
0.14
0.20
0.11
QCF
0.08
0.11
0.12
0.09
0.16
0.12
NA
0.28
QCF
NAC
NAC
NAC
0.70
0.37
0.64
0.87
NA
0.45
QCF
QCF
0.42
QCF
0.98
NA
3. 12
0.97
0.65
0.70
0.99
0.44
0.57
QCF
0.27
0.04
0.05
0.07
ND
QCF
ND
0.07
0.04
ND
0.06
ND
NA
0. 13
ND
NAC
NAC
NAC
0.69
0.28
0.42
0.81
NA
QCF
QCF
QCF
0.35
QCF
0.74
NA
1 .33
0.89
0.76
0.60
1.11
0.57
0.54
QCF
ND
0.05
ND
QCF
ND
QCF
ND
0.04
0.10
QCF
0.08
ND
NA
ND
ND
NAC
NAC
NAC
0.09
0.12
0.07
0.06
NA
QCF
QCF
ND
0.06
ND
0.03
NA
ND
0. 10
0.08
0.07
0.05
0.11
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
NA
ND
ND
NAC
NAC
NAC
0.14
ND
0.23
0.18
NA
0.08
0. 10
QCF
ND
QCF
0. 28
NA
0.78
0.61
ND
0. 27
0.44
1 . 19
0.30
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
NA
ND
ND
NAC
NAC
NAC
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
0.94
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
NA
ND
ND
NAC
NAC
NAC
0. 13
ND
ND
ND
NA
ND
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
0.79
0.40
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
NA
ND
ND
NAC
NAC
NAC
ND
ND
0. 20
ND
NA
ND
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
0.37
ND
4
3
2
1
7
6
4
3
2
1
8
7
6
4
3
2
1
8
7
6
4
8
8
3
8
7
6
4
3
1
8
7
6
4
3
LEGEND:
QCF
NAC
ND
NA
NT
SAMPLE ANALYSED BUT FAILED QC CRITERIA
NOT ANALYSED - CONTINGENCY SAMPLE
SAMPLE ANALYSED,ANALYTE NOT DETECTED
SAMPLE NOT ANALYSED
SAMPLE NOT TAKEN/COLLECTED
-------
SAMPLE
AREA
1-10
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
SITE
NUMBER
690
690
691
692
695
696
698
698
699
700
700
701
702
703
704
705
706
707
709
709
710
7 1 1
71 1
713
713
714
714
715
716
717
718
719
720
721
722
723
724
725
SAS
TABLE 1-1
SOIL SAMPLING ANALYSIS RESULTS BY SAMPLING AREA AND SITE
RESULTS GIVEN IN PARTS PER BILLION (PPB)
SPLIT DCS TCB TeCB CNP A-BHC D-BHC B-BHC
25
G-BHC
LAB
ID
X
X
X
X
X
X
X
X
X
X
X
X
X
X
0.64
0.36
0.59
QCF
NA
0.32
0.29
QCF
QCF
0.3B
0.29
NA
0.38
0.40
0. 19
0. 19
1 .09
0.30
QCF
0.50
0.34
NA
QCF
0.37
0.57
NA
QCF
0.40
0.37
QCF
0.43
0.88
0. 19
NA
0.33
0.33
0.39
0.39
0.56
1 .03
0.58
QCF
NA
0.81
0.48
QCF
QCF
0.53
0.62
NA
1 .25
0.83
0.46
0. 25
1 .40
0.60
0.72
1 .03
0.74
NA
1 .03
1 .06
1.81
NA
0.37
0.91
0.64
0.53
0.76
0.38
0.38
NA
0.39
0.63
0.48
0.46
1 .25
0.99
0.43
QCF
NA
0.56
0.50
QCF
QCF
0.32
0.46
NA
0.32
0.56
0.38
0.28
0.79
0.51
0.69
0.56
0.57
NA
0.64
0.91
1 .37
NA
0.24
0.84
0.72
0.50
0.62
0.22
0.39
NA
0.34
0.48
0.27
0.29
0.09
NO
0.10
QCF
NA
ND
ND
ND
QCF
0.09
ND
NA
O.OG
ND
ND
0.09
0.08
ND
0.06
ND
ND
NA
0.06
ND
ND
NA
0.06
0.09
ND
0.11
0. 17
0.16
ND
NA
ND
0.05
0.08
0.11
0.81
0.30
0.21
QCF
NA
0.35
0.11
QCF
ND
ND
ND
NA
0. 29
ND
0. 10
0.06
0.38
0.33
0.23
0.27
0.21
NA
0.25
0.60
0.45
NA
0.09
0.08
ND
0. 19
0. 18
ND
0. 18
NA
0.06
ND
0.10
0.08
ND
ND
ND
ND
NA
ND
0.04
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
ND
ND
ND
ND
NA
ND
ND
0.25
NA
ND
ND
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
ND
0. 15
ND
NA
ND
0.81
QCF
ND
ND
0.11
NA
ND
ND
ND
ND
0. 16
0.56
0.83
1 .03
ND
NA
ND
0.38
ND
NA
ND
ND
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
0.73
ND
ND
ND
NA
ND
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
ND
0.08
ND
ND
NA
0.06
0.14
0.09
NA
ND
ND
ND
0.04
0.48
ND
ND
NA
ND
ND
ND
ND
1
6
1
7
4
2
8
3
7
1
6
4
3
2
8
8
7
6
1
3
2
4
1
6
3
4
1
3
2
1
8
7
6
4
3
3
1
8
LEGEND:
QCF
NAC
ND
NA
NT
SAMPLE ANALYSED BUT FAILED QC CRITERIA
NOT ANALYSED - CONTINGENCY SAMPLE
SAMPLE ANALYSED,ANALYTE NOT DETECTED
SAMPLE NOT ANALYSED
SAMPLE NOT TAKEN/COLLECTED
-------
SAMPLE
AREA
1-10
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
10
10
10
10
. 10
10
10
10
SITE
NUMBER
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
952
953
954
694R
71 2R
683R
708R
697R
693R
820
821
822
822
823
824
825
826
SAS
TABLE 1-1
SOIL SAMPLING ANALYSIS RESULTS BV SAMPLING AREA AND SITE
RESULTS GIVEN IN PARTS PER BILLION (PPB)
SPLIT DCB TCB TeCB CNP A-BHC D-BHC B-BHC
26
G-BHC
LAB
ID
0.70
0.22
NA
0.39
NA
1 .03
0.48
QCF
NA
0.40
QCF
NA
0.34
0.82
QCF
NA
0. 19
0.54
QCF
NAC
NAC
NAC
QCF
QCF
0.47
NAC
0.45
0.56
NT
NT
0.28
0.36
X 0.66
X 0.54
0.49
0.73
QCF
NA
0.56
0.35
NA
0.69
NA
1 .03
0.37
QCF
NA
1 .04
QCF
NA
0.32
0.68
QCF
NA
0.52
0.70
0.68
NAC
NAC
NAC
QCF
QCF
0.39
NAC
2.91
1.17
NT
NT
0.23
0.36
1 .72
1.41
0.37
0.26
QCF
NA
0.42
0.33
NA
0.50
NA
0.84
0. 23
QCF
NA
0.67
OCF
NA
0.26
0.46
OCF
NA
0.45
0.52
0.47
NAC
NAC
NAC
QCF
QCF
0. 24
NAC
1 .20
1.14
NT
NT
0. 15
0.30
1 .39
1 .20
0.37
0.31
QCF
NA
0. 13
0.05
NA
0.09
NA
0.09
0.05
NO
NA
NO
QCF
NA
0. 22
0.15
ND
NA
ND
0.11
0.12
NAC
NAC
NAC
QCF
QCF
0.09
NAC
ND
0.05
NT
NT
0.08
ND
0.12
0.06
0. 19
NO
ND
NA
ND
0.20
NA
0.33
NA
0.13
0.06
QCF
NA
0.17
ND
NA
0.09
ND
QCF
NA
ND
0.25
0. 18
NAC
NAC
NAC
ND
QCF
0.20
NAC
0.29
0.33
NT
NT
ND
0.12
1 . 10
0.79
ND
ND
QCF
NA
ND
ND
NA
ND
NA
0.08
ND
ND
NA
ND
ND
NA
ND
ND
ND
NA
ND
ND
ND
NAC
NAC
NAC
ND
ND
ND
NAC
ND
0.30
NT
NT
ND
ND
0.27
ND
ND
ND
ND
NA
ND
ND
NA
ND
NA
QCF
ND
QCF
NA
0.08
ND
NA
ND
ND
ND
NA
ND
ND
0.07
NAC
NAC
NAC
ND
ND
ND
NAC
0.37
0.44
NT
NT
ND
ND
0.54
0.52
ND
0.13
QCF
NA
ND
ND
NA
ND
NA
ND
ND
ND
NA
ND
ND
NA
0.50
ND
ND
NA
ND
ND
0.30
NAC
NAC
NAC
ND
ND
ND
NAC
80.81
ND
NT
NT
ND
ND
0.23
0.12
ND
ND
QCF
NA
7
6
4
3
2
1
7
6
4
3
2
1
8
7
6
4
3
2
1
4
7
2
4
1
6
3
2
1
1
8
7
6
4
LEGEND:
QCF
NAC
ND
NA
NT
SAMPLE ANALYSED BUT FAILED QC CRITERIA
NOT ANALYSED - CONTINGENCY SAMPLE
SAMPLE ANALYSED,ANALYTE NOT DETECTED
SAMPLE NOT ANALYSED
SAMPLE NOT TAKEN/COLLECTED
-------
SAS 27
TABLE 1-1
SOIL SAMPLING ANALYSIS RESULTS BV SAMPLING AREA AND SITE
RESULTS GIVEN IN PARTS PER BILLION (PPB)
SAMPLE SITE SPLIT DCB TCB TeCB CNP A-BHC D-BHC B-BHC G-BHC LAB
AREA NUMBER ID
1-10
10 827
10 827
10 828
10 829
10 830
10 831
10 832
10 834
10 835
10 836
10 837
10 838
10 839
10 840
10 841
10 841
10 842
10 843
10 844
10 845
10 846
10 847
10 848
10 849
10 850
10 850
10 851
10 851
10 852
10 852
10 853
10 854
10 854
10 855
. 10 856
10 857
10 858
10 859
LEGEND:
QCF = SAMPLE ANALYSED BUT FAILED QC CRITERIA
NAC = NOT ANALYSED - CONTINGENCY SAMPLE
NO = SAMPLE ANALYSED,ANALYTE NOT DETECTED
NA = SAMPLE NOT ANALYSED
NT = SAMPLE NOT TAKEN/COLLECTED
X
X
X
X
X
X
X
X
X
X
0.43
0.69
0.48
0. 27
0.20
0.57
0.36
QCF
0.30
0.84
0.43
0.69
0. 18
NA
0.43
QCF
0.23
0.42
0.72
0.45
NA
QCF
0.43
0.49
NA
0.26
0.35
0.53
NA
0.35
NA
NA
0.23
0.34
0.83
0.27
0.54
0.38
1 .59
2. 23
1 .77
0.33
0.41
0. 19
0.83
QCF
1 .34
3.93
1.12
0.37
0.25
NA
2.89
QCF
0. 24
0.17
0.56
0.88
NA
QCF
0.73
0.58
NA
0.12
0.54
0.40
NA
0.62
NA
NA
0. 28
0.57
2.89
1 .00
0.20
0.86
2.41
2.68
1 .07
0. 24
0.38
0.08
0.82
QCF
1 .67
4.07
21 .02
0.61
0.21
NA
2.47
QCF
0.21
0.12
0.71
0.92
NA
QCF
0.54
0.37
NA
0.23
0.44
0.35
NA
0.52
NA
NA
0.27
0.57
1 .72
0.92
0. 13
0.64
ND
ND
0.10
0.07
0.06
0.14
0.07
ND
0.08
0.11
0.10
0.04
ND
NA
ND
ND
0.09
0.06
ND
0.09
NA
ND
0.07
0.12
NA
ND
0.08
0.09
NA
ND
NA
NA
0.04
0.06
0.08
0. 10
0.11
0.04
0.51
0.62
0.33
0.06
ND
ND
0. 19
QCF
0.79
3.14
2.49
ND
0.06
NA
10.88
QCF
ND
0.08
ND
0. 15
NA
QCF
2.28
0. 15
NA
0.05
0.08
ND
NA
0. 1 1
NA
NA
0.03
0.67
0.61
0.85
0.33
0.83
ND
ND
ND
ND
ND
ND
ND
QCF
ND
0.43
0.49
ND
ND
NA
ND
QCF
ND
ND
ND
ND
NA
ND
ND
ND
NA
ND
ND
ND
NA
ND
NA
NA
ND
ND
0. 17
ND
ND
0.07
ND
0.67
ND
ND
0.29
ND
0.70
QCF
0.69
0.82
1 .75
ND
ND
NA
5.16
QCF
ND
0.07
0.40
ND
NA
ND
2.32
0.12
NA
ND
0.06
ND
NA
ND
NA
NA
ND
0.29
0.43
46.95
ND
1 .06
ND
ND
ND
ND
ND
ND
ND
QCF
0.20
3.23
1 .55
ND
0.04
NA
0.23
QCF
ND
ND
0.12
ND
NA
QCF
ND
ND
NA
ND
0.04
ND
NA
ND
NA
NA
ND
ND
0.27
ND
ND
0.13
6
3
2
1
8
7
6
3
2
1
8
7
6
4
3
3
2
1
7
6
4
3
2
1
4
8
1
7
4
6
4
3
2
2
1
8
7
6
-------
SAMPLE
AREA
1-10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
SITE
NUMBER
860
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
880
881
882
883
884
885
886
887
888
889
890
891
892
892
893
894
943
944
SAS
TABLE 1-1
SOIL SAMPLING ANALYSIS RESULTS BV SAMPLING AREA AND SITE
RESULTS GIVEN IN PARTS PER BILLION (PPB)
SPLIT DCB TCB TeCB CNP A-BHC D-BHC B-BHC
28
G-BHC
LAB
ID
0.20
0.25
0.96
QCF
0.56
QCF
0.30
0.37
0.35
1 .40
0.55
0. 29
QCF
0.34
1 .07
0.23
0.32
0.65
0.22
X NA
X 0.63
0.28
0.63
QCF
0.84
0. 27
NA
0.51
0.66
QCF
0. 23
0.77
1 . 13
0.99
NA
0.54
NAC
NAC
0.40
0.25
3.23
QCF
0.91
QCF
0.69
0.33
0.59
33.07
0.43
2.10
QCF
0.76
2.37
0.17
1 . 20
0.65
0.70
NA
0.39
0.44
0.38
0.20
0.35
0. 21
NA
1 . 13
0.13
0. 16
2. 20
0.92
QCF
24. 1 1
NA
2.15
NAC
NAC
QCF
0. 19
12.51
QCF
0.97
QCF
1 .37
0.39
0.70
64. 25
0.79
2.10
QCF
0.79
2.31
0.12
2.81
0.33
0.83
NA
0.40
0.56
0.39
0.21
0.25
0.14
NA
1 .23
0.06
0. 12
0.96
0.39
6.61
6.30
NA
0.83
NAC
NAC
0.07
ND
0.09
QCF
0.06
QCF
ND
0.11
0.12
0.32
0.14
ND
QCF
0.07
0.09
0.08
0.07
0.12
ND
NA
0.07
0.05
0.06
0.10
ND
ND
NA
0.08
0.08
0.11
0.05
0.11
ND
ND
NA
ND
NAC
NAC
ND
0.21
2.92
QCF
0.41
ND
ND
0.16
0.62
33.97
0.09
0.52
ND
ND
1 .07
0.02
0.39
0.14
0.26
NA
0.35
0.13
0.10
ND
ND
ND
NA
0.03
ND
0.03
ND
ND
0.26
ND
NA
ND
NAC
NAC
ND
ND
ND
ND
ND
QCF
ND
ND
0.27
5.40
ND
0.51
ND
ND
ND
ND
0. 15
ND
ND
NA
ND
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
ND
ND
NA
ND
NAC
NAC
0.38
ND
ND
QCF
0.33
QCF
0.20
ND
0.14
50.57
ND
0.70
ND
ND
0.99
0.01
ND
ND
ND
NA
QCF
ND
ND
ND
ND
ND
NA
ND
ND
ND
ND
ND
0. 19
ND
NA
ND
NAC
NAC
0.03
ND
1 .77
ND
ND
QCF
ND
ND
0.58
15.83
ND
0.54
ND
ND
0.2B
ND
0.37
ND
0.14
NA
ND
ND
0.04
ND
ND
ND
NA
ND
ND
ND
ND
ND
ND
ND
NA
ND
NAC
NAC
4
2
1
7
6
4
3
2
1
8
7
6
4
3
2
1
8
7
6
4
2
3
2
1
7
6
4
3
2
1
8
7
6
6
4
3
LEGEND:
QCF
NAC
ND
NA
NT
SAMPLE ANALYSED BUT FAILED QC CRITERIA
NOT ANALYSED - CONTINGENCY SAMPLE
SAMPLE ANALYSED,ANALYTE NOT DETECTED
SAMPLE NOT ANALYSED
SAMPLE NOT TAKEN/COLLECTED
-------
SAS 29
TABLE 1-1
SOIL SAMPLING ANALYSIS RESULTS BY SAMPLING AREA AND SITE
RESULTS GIVEN IN PARTS PER BILLION (PPB)
SAMPLE SITE SPLIT DCB TCB TeCB CNP A-BHC D-BHC B-BHC G-BHC LAB
AREA NUMBER ID
1-10
10 945 NAC NAC NAC NAC NAC NAC NAC NAC
10 946 NAC NAC NAC NAC NAC NAC NAC NAC
10 947 NAC NAC NAC NAC NAC NAC NAC NAC
10 948 NAC NAC NAC NAC NAC NAC NAC NAC
10 861R 0.41 0.95 0.78 ND NO ND ND ND 3
10 971 NAC NAC NAC NAC NAC NAC NAC NAC
LEGEND:
QCF = SAMPLE ANALYSED BUT FAILED QC CRITERIA
NAC = NOT ANALYSED - CONTINGENCY SAMPLE
ND = SAMPLE ANALYSED,ANALYTE NOT DETECTED
NA = SAMPLE NOT ANALYSED
NT = SAMPLE NOT TAKEN/COLLECTED
-------
TABLE 1-1
SOIL SAMPLING ANALYSIS RESULTS BY SAMPLING AREA AND SITE
LCIC LEGEND:
ABBREVIATION
DCS
TCB
TeCB
CNP
A-BHC
D-BHC
B-BHC
G-BHC
COMPOUND NAME
1 ,2-DICHLOROBENZENE
1,2,4-TRICHLOROBENZENE
1 ,2,3,4-TETRACHLOROBENZENE
2-CHLORONAPHTHALENE
ALPHA-BHC
DELTA-BHC
BETA-BHC
GAMMA-BHC
-------
Appendix J
Summary of QA/QC Results
-------
Appendix J
SUMMARY OF QA/QC RESULTS
INTRODUCTION
The soil assessment for indicator chemicals was designed to
achieve the QA/QC objectives of precision, accuracy, rep-
resentativeness, comparability, completeness, and detection
limit, as specified in the sample collection and analysis
QAPPs (CH2M HILL, 1987a and 1987b).
The following QA/QC measures were developed to assist in
monitoring and achieving these QA/QC objectives:
o Holding times (comparability)
o Instrument calibration (accuracy, precision,
sensitivity)
o EPA check standard analyses (accuracy)
o Surrogate recoveries (accuracy and precision)
o Matrix spike recoveries (accuracy)
o Matrix spike duplicate recoveries (precision)
o Field and laboratory blanks (comparability)
o Internal standard area and retention time differ-
ences (comparability)
o Blind QC sample recoveries (accuracy and precision)
The QC data generated by each participating laboratory for
each sample were reviewed to provide a useful indication of
data quality. The findings of this QA/QC data review are
summarized in Table J-l and are discussed in the paragraphs
below.
HOLDING TIMES
All samples were initially extracted within 30 days of re-
ceipt by the laboratory and analyzed within 50 days of re-
ceipt, meeting the QAPP requirement. However, many samples
had to be reanalyzed because they did not meet stringent QC
criteria and thus ultimately did not meet the holding time
criteria.
J-l
-------
Table J-l
QC DATA SUMMARY
QA Measure
MHB
Laboratory
1
2
3
6
7
8
Data Meeting QC Criteria
Goal (%)
90
Achieved
(Y or N)
Y
Y
Y
Y
Y
Y
Surrogate Spike
Recovery
1
2
3
6
7
8
90
Y
Y
Y
Y
Y
Y
MS Recovery
(Advisory Only)
1
2
3
6
7
8
90
N
N
N
N
N
N
MSD Precision
(Advisory Only)
1
2
3
6
7
8
90
Y
Y
N
Y
Y
N
BQC Recovery
1
2
3
6
7
8
90
Y
Y
Y
Y
Y
Y
EPA Check
Standard
Recovery
1
2
3
6
7
8
100
Y
Y
Y
Y
Y
Y
J-2
-------
Table J-l
(continued)
QA Measure
Internal Stand-
ard Area and Rt
Variation
Performance
Check Standard
(PCI and PC2)
Initial Calibra-
tion Standard
Continuing
Calibration
Standard
Holding Time
(for original
extraction and
analysis only)
Laboratory
1
2
3
6
7
8
1
2
3
6
7
8
1
2
3
6
7
8
1
2
3
6
7
8
1
2
3
6
7
8
Data Meeting QC Criteria
Goal (%)
90
100
100
100
100
Achieved
(Y or N)
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
J-3
-------
INSTRUMENT CALIBRATIONS
All laboratories met the QC criteria for initial calibrations,
performance check standard calibrations, and continuing cali-
brations. In general, this indicates a low probability of
large instrument errors in the LCIC concentration estimates
and increases the confidence of the reported concentrations.
EPA CHECK STANDARD ANALYSES
The recoveries of all reported EPA check standard analyses
were within the required 80 to 120 percent of the theoret-
ical concentration. Although the EPA check standard mea-
surements do not include the variability associated with
sampling, transportation, storage, preparation, and extrac-
tion of samples, the data indicate the calibration method
had good accuracy and precision.
SURROGATE RECOVERIES
More than 95 percent of surrogate analyses had recoveries
within the control limits as specified in the QAPP. A num-
ber of the recoveries exceeded the control limits because of
dilution analysis of the sample extract. Surrogate recov-
eries for diluted extracts were not required to meet the
control limits specified in the QAPP. Also, a number of
outliers were due to not meeting the advisory upper control
limit for 2,4,6-tribromobiphenyl. Having high percentages
of reported surrogate recoveries within the QC criteria gen-
erally increases the confidence in both the methodology em-
ployed and in the reported concentrations.
MATRIX SPIKE/MATRIX SPIKE DUPLICATE ANALYSES
The matrix spike/matrix spike duplicate (MS/MSD) samples
were analyzed to assess the accuracy and precision of the
analytical method. The percent recovery and relative per-
cent difference (RPD) control limits of the MS/MSD analysis
are advisory only. This is because the results of MS/MSD
analysis are dependent upon the matrix of individual samples
and are not under the direct control of the laboratory.
This follows the example of the EPA CLP where MS/MSD control
limits are also advisory.
The percent recovery and RPD results from the MS/MSD analy-
sis were evaluated to determine if they achieved the goal of
meeting QC criteria for 90 percent of the analysis. As
shown in Table J-l, none of the six laboratories used for
the soil assessment achieved the recovery goal. This is
probably because the stringent control limits specified in
J-4
-------
the QAPP were developed based on method validation results
from only the two experienced laboratories used in the
method development work. The time constraints prevented a
second multi-laboratory MS/MSD study to revise the control
limits. The samples were not reanalyzed because the control
limits were advisory only. Four of the six laboratories did
meet the 90 percent goal for MS/MSD precision. This demon-
strated, in general, the reproducibility of quantification
for most of the laboratories.
FIELD AND LABORATORY BLANK ANALYSES
The laboratory method/holding blank results met the 90 per-
cent completeness goal as specified in the QAPP. The lab-
oratory method/holding blank was required to contain no more
than 0.5 yg/kg of LCICs (0.6 yg/kg for 1,2-dichlorobenzene)
or LCIC interferences. Except 1,2-dichlorobenzene, all the
method/holding blank results showed very low levels of blank
contamination. For 1,2-dichlorobenzene, all the laboratories
had about a 0.2-ppb level of contamination, with the excep-
tion of Laboratory 8, which had about 0.5 ppb. Field han-
dling blanks were also analyzed by the laboratories. The
field handling blank results generally showed slightly higher
LCIC concentrations than the laboratory method/holding blanks,
This was not unexpected because the field handling blanks
underwent all the same steps of sample collection, extrusion,
mixing, shipping, and analysis as a normal sample. Another
contributing factor may be the fact that a native soil matrix
is used for the field handling blank while a sand matrix is
used for the laboratory method/holding blanks.
INTERNAL STANDARD AREA AND RETENTION TIME DIFFERENCES
The 90 percent completeness goal for internal standard area
and retention time differences was met by all the labora-
tories. The retention time difference criteria were always
met, but the internal standard area difference criteria were
not always met. The laboratories reinjected the samples
when necessary, as required by the QAPP. The internal stan-
dard area response and retention time difference observed
over the course of the study did not vary significantly,
thus increasing the confidence in the reported values.
BLIND QC SAMPLE ANALYSES
The blind QC sample results were generally within the con-
trol limits specified by EMSL-LV. The 90 percent data com-
pleteness goal was met by all the laboratories, except
Laboratory 4, as noted in Section 6.1. In general, the lab-
oratories performed well on the blind QC samples and no
J-5
-------
apparent concentration-related recovery bias was observed
within the concentration ranges used in the blind QC samples
(discussed in Appendix H).
REFERENCES
CH2M HILL. 1987a. Love Canal Habitability Study—Soil Sam-
ple Collection and Preparation Quality Assurance Project
Plan (Final Revised Version).
CH2M HILL. 1987b. Love Canal Habitability Study—Soil Sam-
ple Laboratory Quality Assurance Project Plan.
8855B/056
J-6
-------
APPENDIX K
Supplementary Results for the
Statistical Comparisons
-------
Appendix K
SUPPLEMENTARY RESULTS FOR THE STATISTICAL COMPARISONS
This appendix supplements the results presented in
Section 6.4. Included are the following:
o Two-sided p-values for the results presented in
Tables 6-8 and 6-9 (Tables K-l and K-2).
o Spearman's (rank) correlations between LCICs for
each of the seven EDA sampling areas and the three
comparison areas (Tables K-3a to K-3j).
o Results of the application of the univariate and
multivariate blocked Wilcoxon test to all data
that are either "good" or "uncertain" (Tables K-4
to K-8).
o Results of the application of the univariate and
multivariate blocked Wilcoxon test to all "good"
data with concentrations below 1 ppb treated as
nondetectable (Tables K-9 to K-12).
o Results of the retrospective power simulations for
the univariate and multivariate tests.
The two-sided p-values presented in Table K-l correspond to
the results of the univariate comparisons between EDA sam-
pling areas and comparison areas reported in Table 6-8 of
Section 6.4. Similarly, the p-values presented in Table K-2
correspond to the results of the multivariate comparisons
reported on Table 6-9 of Section 6.4.
The Spearman's rank correlations (Conover, 1980) presented
for each area in Tables K-3a to K-3j were computed by calcu-
lating the Spearman's correlation within each laboratory and
then averaging these within-laboratory Spearman's correla-
tions over all laboratories. Nondetectable values were
treated as tied in the usual way. The sample size (N)
reported for each area is the number of observations over
all laboratories.
The Spearman's correlations should be considered as explora-
tory statistics; they were not used directly in the compari-
son method described in Section 5.5. Generally, the results
show that: (1) LCICs in the chlorobenzene group (di-, tri-,
and tetrachlorobenzene) are strongly correlated,
(2) chloronaphthalene is relatively uncorrelated with the
remaining seven LCICs, (3) alpha-BHC is relatively strongly
correlated with all the remaining LCICs, and (4) there is
modest correlation among the BHC group (alpha-, delta-, beta-,
and gamma-BHC). Approximately 95 percent confidence bounds
for a two-sided test against a null hypothesis of zero
K-l
-------
correlation are +2/N-2, so that correlations with absolute
value less than about 0.3 should not be considered significant,
Tables K-4, K-5, and K-6 are analogous to Tables 6-8, 6-9,
and 6-10 of Section 6.4, except that the analyses were per-
formed with "good" and "uncertain" data, rather than just
"good" data. In Tables K-4 and K-5, any changes in direc-
tion or magnitude of significance between the comparisons
based on the "good" data and those based on the "good" +
"uncertain" data are recorded with two sets of symbols: the
second set of symbols (in parentheses) denotes the original
result based on the "good" data alone. The two-sided p-values
for the univariate and multivariate comparisons with the
"good" + "uncertain" data are given in Tables K-7 and K-8,
respectively.
Table K-9 and K-10 are analogous to Tables 6-8 and 6-10,
except that an artificial detection limit of 1.0 ppb was
imposed to test the sensitivity of the results to relatively
"large" concentrations only. As for the previously discussed
analyses based on the "good" + "uncertain" data, any change
in direction or magnitude of significance between the com-
parisons based on the "good" data and those based on the
"good" data with an artificial detection limit of 1 ppb are
recorded with two sets of symbols. The two-sided p-values
for the univariate comparisons for these data are given in
Table K-ll. No multivariate analyses were performed for
this data set because the artificial 1 ppb detection limit
resulted in several cases where all observations were less
than the detection limit for some of the LCICs; this pre-
cluded computation of a non-singular variance-covariance
matrix based on ranks.
ABBREVIATIONS
The following chemical abbreviations are used in Appendix K;
DCB
TCB
QCB
CNP
A-BHC
B-BHC
D-BHC
G-BHC
1,2 Dichlorobenzene
1,2,4 Trichlorobenzene
1,2,3,4 Tetrachlorobenzene
2 Chloronaphthalene
Alpha BHC
Beta BHC
Delta BHC
Gamma BHC
REFERENCE
Conover, W.J. 1980. Practical Nonparametric Statistics
John Wiley & Sons, New York, New York.
8856/010
K-2
-------
Table K-1
TWO-SIDED p-VALUES FOR THE UNIVARIATE STATISTICAL COMPARISONS
WITH OBSERVATIONS CLASSIFIED AS GOOD
LCIC
DCB
TCB
OCB
CNP
A-BHC
D-BHC
B-BHC
G-BHC
comparison | •
Area 1
221
225
C&T
221
225
C&T
221
225
C&T
221
225
C&T
221
225
C&T
221
225
C&T
221
225
C&T
221
225
C&T
0.000
O.OOCI
O.OOCI
0.000
0.000
0.000
0.000
0.000
0.000
0.110
0.047
0.69:5
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
2
0.805
0.539
0.008
0.001
0.010
0.000
0.000
0.000
0.000
0.055
0.013
0.914
0.000
0.001
0.000
0.000
0.000
0.000
0.000
0.105
0.000
0.000
0.014
0.000
— CUM sampling
3 4
0.980
0.583
0.002
0.014
0.032
0.000
0.000
0.005
0.000
0.059
0.005
0.814
0.069
0.207
0.000
0.000
0.150
0.000
0.000
0.460
0.000
0.001
0.179
0.000
0.003
0.025
0.915
0.011
0.077
0.000
0.040
0.054
0.000
0.597
0.107
0.103
0.630
0.410
0.000
0.032
0.573
0.000
0.016
0.755
0.000
0.020
0.711
0.000
areas -•
5
0.617
0.779
0.058
0.268
0.828
0.000
0.446
0.250
0.000
0.311
0.018
0.605
0.270
0.942
0.000
0.015
0.560
0.000
0.453
0.118
0.000
0.278
0.461
0.000
6
0.009
0.312
0.146
0.000
0.005
0.000
0.000
0.000
0.000
0.383
0.322
0.193
0.028
0.192
0.000
0.563
0.372
0.008
0.247
0.001
0.004
0.921
0.035
0.002
7'
0.992
0.556
0.022
0.123
0.808
0.000
0.446
0.982
0.000
0.965
0.762
0.018
0.509
0.577
0.000
0.206
0.667
0.002
0.072
0.632
0.000
0.457
0.140
0.000
- comparison areas -
221 225 C&T
1.000
0.757
0.007
1.000
0.368
0.000
1.000
0.346
0.000
1.000
0.824
0.091
1.000
0.346
0.000
1.000
0.075
0.070
1.000
0.033
0.000
1.000
0.095
0.007
0.757
1.000
0.194
0.368
1.000
0.000
0.346
1.000
0.000
0.824
1.000
0.010
0.346
1.000
0.000
0.075
1.000
0.001
0.033
1.000
0.000
0.095
1.000
0.000
0.007
0.194
1.000
0.000
0.000
1.000
0.000
0.000
1.000
0.091
0.010
1.000
0.000
0.000
1.000
0.070
0.001
1.000
0.000
0.000
1.000
0.007
0.000
1.000
a C&T = Cheektowaga and Tonauanda
221 = Census Tract 221
225 = Census Tract 225
-------
Table K-2
TWO-SIDED p-VALUES FOR NONPARAMETRIC MULTIVARIATE STATISTICAL COMPARISONS
WITH OBSERVATIONS CLASSIFIED AS GOOD
Comparison | EDA Sampling Areas | |- Comparison Areas -
Area 1 2 3 4 5 6 7 221 225 C&T
221 0.000 0.000 0.000 0.000 0.037 0.001 0.004 1.000 0.017 0.000
225 0.000 0.000 0.022 0.095 0.023 0.019 0.778 0.017 1.000 0.000
C&T 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 1.000
-------
LCIC
Table K-3a
LCIC RANK CORRELATIONS FOR EDA SAMPLING AREA 1 a
N = 31 b
1.00
0.79
0.80
0.36
0.65
0.46
0.52
0.62
1.00
0.93
0.21
0.74
0.53
0.70
0.64
1.00
0.35
0.81
0.62
0.67
0.65
1.00
0.27
0.39
0.21
0.28
1.00
0.83
0.84
0.85
1.00
0.62
0.67
1.00
0.83
DCB
TCB
QCB
CNP
A-BHC
D-BHC
B-BHC
G-BHC 0.62 0.64 0.65 0.28 0.85 0.67 0.83 1.00
DCB TCB QCB CNP A-BHC D-BHC B-BHC G-BHC
Table K-3b
LCIC RANK CORRELATIONS FOR EDA SAMPLING AREA 2 a
N = 60 b
LCIC
DCB
TCB
OCB
CNP
A-BHC
D-BHC
B-BHC
G-BHC
1.00
0.77
0.72
0.29
0.60
0.42
0.30
0.47
DCB
1.00
0.94
0.33
0.77
0.45
0.50
0.51
TCB
1.00
0.33
0.80
0.48
0.59
0.58
OCB
1.00
0.34
0.29
0.42
0.29
CNP
1.00
0.55
0.56
0.76
A-BHC
1.00
0.47
0.60
D-BHC
1.00
0.41
B-BHC
1.00
G-BHC
a. Correlations were computed by calculating correlations within
laboratories and then averaging correlations across laboratories.
b. N denotes total sample size over all laboratories.
-------
LC1C
Table K-3c
LCIC RANK CORRELATIONS FOR EDA SAMPLING AREA 3 a
N = 62 b
1.00
0.68
0.59
-0.02
0.40
0.10
0.42
0.27
DCS
1.00
0.85
-0.27'
0.61
0.18
0.51
0.37'
TCB.
1.00
-0.30
0.59
0.15
0.43
0.30
OCB
1.00
-0.13
0.03
-0.15
0.01
CNP
1.00
0.27
0.67
0.34
A-BHC
1.00
0.32
0.11
D-BHC
1.00
0.41
B-BHI
DCB
TCB
QCB
CNP
A-BHC
D-BHC
B-BHC
G-BHC 0.27 0.37' 0.30 0.01 0.34 0.11 0.41 1.00
G-BHC
Table K-3d
LCIC RANK CORRELATIONS FOR EDA SAMPLING AREA 4 a
N = 45 b
LCIC
DCB
TCB
OCB
CNP
A-BHC
D-BHC
B-BHC
G-BHC
1.00
0.61
0.42
0.12
0.41
0.17
0.10
0.17
DCB
1.00
0.84
-0.17
0.52
0.33
0.31
0.34
TCB
1.00
-0.24
0.65
0.34
0.45
0.37
QCB
1.00
-0.08
0.25
-0.25
-0.34
CNP
1.00
0.13
0.63
0.34
A-BHC
1.00
0.14
-0.10
D-BHC
1.00
0.32
B-BHC
1.00
G-BHC
a. Correlations were computed by calculating correlations within
laboratories and then averaging correlations across laboratories.
b. N denotes total sample size over all laboratories.
-------
LCIC
LCIC
Table K-3e
LCIC RANK CORRELATIONS FOR EDA SAMPLING AREA 5 a
N = 68 b
1.00
0.73
0.79
0.09
0.59
0.49
0.63
0.56
1.00
0.92
-0.01
0.75
0.50
0.69
0.59
1.00
0.01
0.68
0.52
0.67
0.62
1.00
0.02
0.10
0.04
0.03
1.00
0.44
0.77
0.71
1.00
0.53
0.44
1.00
0.73
DCB
TCB
CO
CNP
A-BHC
D-BHC
B-BHC
G-BHC 0.56 0.59 0.62 0.03 0.71 0.44 0.73 1.00
DCB TCB QCB CNP A-BHC D-BHC B-BHC G-BHC
Table K-3f
LCIC RANK CORRELATIONS FOR EDA SAMPLING AREA 6 a
N = 76 b
DCB 1.00
TCB 0.58 1.00
QCB 0.53 0.92 1.00
CNP 0.33 0.31 0.33 1.00
A-BHC 0.36 0.70 0.72 0.26 1.00
D-BHC 0.23 0.34 0.36 0.04 0.32 1.00
B-BHC 0.33 0.45 0.42 -0.01 0.37 0.42 1.00
G-BHC 0.30 0.30 0.28 0.23 0.36 0.23 0.30 1.00
DCB TCB QCB CNP A-BHC D-BHC B-BHC G-BHC
a. Correlations were computed by calculating correlations within
laboratories and then averaging correlations across laboratories.
b. N denotes total sample size over all laboratories.
-------
LCIC
Table K-3g
LCIC RANK CORRELATIONS FOR EDA SAMPLING AREA 7 a
N = 65 b
1.00
0.68
0.59
0.49
0.34
0.18
0.18
0.18
DCB
1.00
0.93
0.35
0.57
0.24
0.38
0.34
TCB
1.00
0.28
0.58
0.22
0.43
0.29
QCB
1.00
0.17
0.11
0.05
0.11
CNP
1.00
0.11
0.45
0.25
A-BHC
1.00
0.17
0.21
D-BHC
1.00
0.10
B-BHI
DCB
TCB
QCB
CNP
A-BHC
D-BHC
B-BHC
G-BHC 0.18 0.34 0.29 0.11 0.25 0.21 0.10 1.00
G-BHC
Table K-3h
LCIC RANK CORRELATIONS FOR CENSUS TRACT 221 a
N = 43 b
LCIC
DCB
TCB
OCB
CNP
A-BHC
D-BHC
B-BHC
G-BHC
1.00
0.47
0.39
0.56
0.37
0.08
0.21
0.01
DCB
1.00
0.82
-0.11
0.58
0.14
0.33
0.28
TCB
1.00
-0.15
0.47
0.14
0.28
0.18
OCB
1.00
0.05
-0.00
-0.25
-0.12
CNP
1.00
0.10
0.29
0.29
A-BHC
1.00
0.24
-0.05
D-BHC
1.00
0.11
B-BHC
1.00
G-BHC
a. Correlations were computed by calculating correlations within
laboratories and then averaging correlations across laboratories.
b. N denotes total sample size over all laboratories.
-------
LCIC
LC1C
Table K-3i
LCIC RANK CORRELATIONS FOR CENSUS TRACT 225 a
N = 52 b
DCB
TCB
QCB
CNP
A-BHC
D-BHC
B-BHC
G-BHC 0.27 0.55 0.56 0.04 0.53 0.43 0.59 1.00
Table K-3j
LCIC RANK CORRELATIONS FOR CHEEKTOWAGA AND TONAUANDA
N = 46 b
1.00
0.60
0.60
0.13
0.34
0.15
0.37
0.27
DCB
1.00
0.91
0.06
0.53
0.27
0.47
0.55
TCB
1.00
0.16
0.52
0.27
0.49
0.56
QCB
1.00
0.30
0.16
0.07
0.04
CNP
1.00
0.37
0.60
0.53
A-BHC
1.00
0.35
0.43
D-BHC
1.00
0.59
B-BHI
1.00
0.52
0.21
0.29
0.16
0.00
0.06
0.09
DCB
1.00
0.66
0.08
0.08
0.00
0.03
0.07
TCB
1.00
0.16
0.02
0.00
0.01
0.02
QCB
1.00
0.01
0.00
-0.10
0.02
CNP
1.00
0.00
0.09
-0.03
A-BHC
1.00
0.00
0.00
D-BHC
1.00
-0.03
B-BHI
DCB
TCB
QCB
CNP
A-BHC
D-BHC
B-BHC
G-BHC 0.09 0.07 0.02 0.02 -0.03 0.00 -0.03 1.00
a. Correlations were computed by calculating correlations within
laboratories and then averaging correlations across laboratories.
b. N denotes total sample size over all laboratories.
-------
Table K-4
RESULTS OF NONPARAMETRIC UNIVARIATE STATISTICAL COMPARISONS
WITH OBSERVATIONS CLASSIFIED AS GOOD OR UNCERTAIN a,b,c,d
p— _
LOT
LCIC
DCB
TCB
QCB
CNP
A-BHC
D-BHC
B-BHC
G-BHC
TOTALS
A 1 1 Symbo I s
.
[tyDari son
Area
221
225
C&T
221
225
C&T
221
225
C&T
221
225
C&T
221
225
C&T
221
225
C&T
221
225
C&T
221
225
C&T
++
+
0
-
--
Median
(ppb)
0.40
0.42
0.36
0.64
0.60
O.K
0.55
0.64
0.05
0.06
0.07
0.03
0.18
0.11
ND
NI)
ND
ND
N!)
ND
ND
ND
ND
ND
1 2
1.01 0.41
++ 0
•f-f O
++ + (++)
8.67 0.88
++ ++
++ + (++)
++ ++
12.93 1.12
•f-f ++
++ ++
++ ++
ND 0.04
O 0
-
0 0
9.47 0.36
++ ++
•f-f ++
++ ++
1.20 ND
•f-f 4"f
•f-f ++
+f ++
11.68 0.19
•f-f ++
•f-f O
++ ++
1.73 0.05
•f-f ++
•f-f +
++ ++
21 15
0 3
2 5
1 1
0 0
24 24
CPiA Ca
tuA ba
3
0.42
o
o
++
0.89
++ (+)
+
+*
1.07
++
•f-f
++
ND
0
• ( -• )
o
0.22
+ (0)
o
•f-f
ND
++
o
++
0.07
++
o
++
ND
++
0
++
13
2
8
1
0
24
nip 1 1 ny Ar63S
4 5
0.38
--
o (-)
o
0.45
-
o
+*
0.42
o (-)
o
++
0.04
o
o
0
0.12
0
o
*+
ND
•f-f (+)
0
*+
ND
+
o
*+
ND
•f
0
*+
7
2
13
1
1
24
0.39
0
0
0
0.57
o
0
*+
0.46
0
o
++
0.06
o
-
o
0.14
o
o
++
ND
•f
0
++
ND
O
o
++
ND
O
o
++
6
1
16
1
0
24
6
0.36
• (••)
0
0
0.38
--
--
++•
0.32
--
--
++
0.05
o
o
o
0.07
-
o
++
ND
0
O
•f (++)
ND
O
--
++
ND
o
-- (-)
++
5
1
10
2
6
24
. . . 1
71
0.41
0
0
+
0.59
o
o
+*
0.55
o
o
++
0.05
o
o
+
0.15
o
o
++
ND
0
O
++
ND
O
o
++
ND
o
o
++
6
2
16
0
0
24
.
I • Compari son Areas • |
221 225 C&T
0.40
0
++
0.64
0
++
0.55
o
++
0.06
o
0
0.18
o
++
ND
- (o)
o
ND
-
++
ND
- (o)
+
5
1
7
3
0
16
0.42
0
0
0.60
0
++
0.64
0
*+
0.07
0
+
0.11
0
++
ND
+ (o)
•ff
ND
+
++
ND
+ (0)
++
6
4
6
0
0
16
0.36
--
0
0.14
--
--
0.05
--
0.03
o
-
ND
--
--
ND
0
--
ND
--
--
ND
- (--)
--
0
0
3
2
11
16
a. ++: EDA > Comparison area at 0.01 significance level
+: EDA > Comparison area at 0.05 significance level
o: No significant difference at 0.05 significance level
-: EDA < Comparison area at 0.05 significance level
--: EDA < Comparison area at 0.01 significance level
b. All test results reported are based on two-sided p-values.
c. The first entry in each column is median concentration. ND indicates non-detect.
d. Any change in results between GOOD and GOOD + UNCERTAIN is indicated by
two sets of symbols; the second set of symbols denotes the original
result for the GOOD observations.
e. C&T = Cheektowaga and Tonawanda
221 = Census Tract 221
225 = Census Tract 225
-------
Table K-5
RESULTS OF NONPARAMETR1C MULT WAR IATE STATISTICAL COMPARISONS
WITH OBSERVATIONS CLASSIFIED AS GOOD OR UNCERTAIN a,b,c
Comparison | EDA Sampling Areas | |- Comparison Areas •
Area 1 2 3 4 5 6 7 221 225 C&T
221 ++ ++ 4+ •• - (+) -- -- (++) ++ (+)
225 ++ *+ o (+) o - -- <-) o -- (-)
C&T ++ ** ++ ++ ++ ++ ++ ** ++•
a. ++: EDA > Comparison area at 0.01 significance level
+: EDA > Comparison area at 0.05 significance level
o: No significant difference at 0.05 significance level
-: EDA < Comparison area at 0.05 significance level
--: EDA < Comparison area at 0.01 significance level
b. The direction of the difference between the EDA and Comparison areas is based on the sign of the sum
of the elements of the S x 1 vector of rank-suns (See Appendix B).
c. Any change in results between GOOD and GOOD + UNCERTAIN is indicated by two sets of symbols;
the second set of symbols denotes the original result for the GOOD observations.
d. C&T = Cheektowaga and Tonawanda
221 = Census Tract 221
225 = Census Tract 225
-------
Table K-6
SUMMARY OF NONPARAMETRIC UN WAR IATE STATISTICAL COMPARISONS
WITH OBSERVATIONS CLASSIFIED AS GOOD OR UNCERTAIN a,b
Totals over all LCICs
Comparison
Area
CT221
CT225
C&T
++
+
0
1
7
0
1
0
0
7
0
0
1
0
7
0
1
0
0
2
6
0
2
0
0
3
2
2
1
0
6
1
1
0
0
— CUA sampling
3 4
5
1
2
0
0
1
1
5
1
0
7
0
1
0
0
1
2
3
1
1
0
0
8
0
0
6
0
2
0
0
Areas -•
5
0
1
7
0
0
0
0
7
1
0
6
0
2
0
0
6
0
0
4
2
2
0
0
4
0
4
5
1
2
0
0
7'
o •
0
8
0
0
0
0
8
0
0
6
2
0
0
0
|- uomparison Areas -|
221 225 C&T
0
0
5
3
0
5
1
2
0
0
0
3
5
0
0
6
1
1
0
0
0
0
2
1
5
0
0
1
1
6
a. ++: EDA > Comparison area at 0.01 significance level
+: EDA > Comparison area at 0.05 significance level
o: No significant difference at 0.05 significance level
-: EDA < Comparison area at 0.05 significance level
--: EDA < Comparison area at 0.01 significance level
b. All test results reported are based on two-sided p-values.
c. C&T = Cheektowaga and Tonawanda
221 = Census Tract 221
225 = Census Tract 225
-------
Table K-7
TWO-SIDED p-VALUES FOR THE UNIVARIATE STATISTICAL COMPARISONS
WITH OBSERVATIONS CLASSIFIED AS GOOD OR UNCERTAIN
LCIC
DCB
TCB
OCB
CNP
A-BHC
D-BHC
B-BHC
G-BHC
Area 1
221
225
C&T
221
225
C&T
221
225
C&T
221
225
C&T
221
225
C&T
221
225
C&T
221
225
C&T
221
225
C&T
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.086
0.044
0.609
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
2
0.802
0.301
0.013
0.003
0.015
0.000
0.000
0.000
0.000
0.051
0.023
0.877
0.000
0.001
0.000
0.000
0.001
0.000
0.000
0.071
0.000
0.000
0.045
0.000
— CUH sampling
3 4
0.992
0.201
0.000
0.006
0.029
0.000
0.000
0.007
0.000
0.083
0.016
0.695
0.029
0.277
0.000
0.000
0.191
0.000
0.000
0.588
0.000
0.000
0.349
0.000
0.004
0.210
0.833
0.019
0.192
0.000
0.081
0.105
0.000
0.501
0.139
0.125
0.638
0.391
0.000
0.006
0.428
0.000
0.010
0.961
0.000
0.015
0.759
0.000
«reas -•
5
0.389
0.468
0.098
0.263
,0.787
0.000
0.429
0.248
0.000
0.258
0.018
0.653
0.361
0.978
0.000
0.012
0.817
0.001
0.290
0.147
0.000
0.154
0.366
0.000
6
0.016
0.657
0.114
0.000
0.004
0.000
0.000
0.000
0.000
0.426
0.412
0.155
0.010
0.064
0.000
0.523
0.201
0.036
0.308
0.000
0.004
0.933
0.006
0.004
7'
0.546
0.318
0.025
0.123
0.910
0.000
0.496
0.832
0.000
0.927
0.745
0.033
0.748
0.713
0.000
0.144
0.567
0.007
0.076
0.347
0.000
0.270
0.058
0.000
|- comparison Areas -
221 225 C&T
1.000
0.359
0.002
1.000
0.330
0.000
1.000
0.333
0.000
1.000
0.910
0.095
1.000
0.393
0.000
1.000
0.034
0.291
1.000
0.014
0.000
1.000
0.021
0.015
0.359
1.000
0.318
0.330
1.000
0.000
0.333
1.000
0.000
0.910
1.000
0.014
0.393
1.000
0.000
0.034
1.000
0.004
0.014
1.000
0.000
0.021
1.000
0.000
0.002
0.318
1.000
0.000
0.000
1.000
0.000
0.000
1.000
0.095
0.014
1.000
0.000
0.000
1.000
0.291
0.004
1.000
0.000
0.000
1.000
0.015
0.000
1.000
a. C&T = Cheektowaga and Tonawanda
221 = Census Tract 221
225 = Census Tract 225
-------
Table K-8
TWO-SIDED p-VALUES FOR NONPARAMETRIC MULTIVARIATE STATISTICAL COMPARISONS
WITH OBSERVATIONS CLASSIFIED AS GOOD OR UNCERTAIN
Comparison | EDA Sampling Areas | |- Comparison Areas -|
Area 1 2 3 4 5 6 7 221 225 C&T
221 0.000 0.000 0.000 0.000 0.017 0.000 0.002 1.000 0.004 0.000
225 0.000 0.000 0.077 0.123 0.037 0.009 0.777 0.004 1.000 0.000
C&T 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 1.000
a. C&T = Cheektowaga and Tonauanda
221 = Census Tract 221
225 = Census Tract 225
-------
Table K-9
RESULTS OF NONPARAMETRIC UNIVARIATE STATISTICAL COMPARISONS
WITH OBSERVATIONS CLASSIFIED AS GOOD a,b,c,d
All observations <= 1 ppb treated as Non-detect
Median
LCIC
DCB
TCB
OCB
CNP
A-BHC
D-BHC
B-BHC
G-BHC
TOTALS
All
loiiiparisun
Area
221
225
C&T
221
225
C&T
221
225
C&T
221
225
C&T
221
225
C&T
221
225
C&T
221
225
C&T
221
225
C&T
++
•f
o
-
--
Symbols
uuni.
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
1 2
1.01 ND
++ o
++ o
++ 0 (++)
8.67 ND
++ ++
++ 0 (++)
++ ++
11.48 1.17
++ ++
++ ++
++ ++
ND ND
0 0
o (-) o (-)
o o
8.25 ND
++ ++
•ft 0 (++)
•M. ++
1.13 ND
++ o (++)
++ o C++)
++ o C++)
11.58 ND
++ ++
++ o
++ ++
1.73 ND
++ 0 (++)
++ 0 (+)
++ ++
21 10
0 0
3 14
0 0
0 0
24 24
- CUM sai
3
ND
o
o
0 <++)
ND
++ ( + )
0 ( + )
+*
1.06
++
+ C++)
++
ND
o
o (--)
o
ND
++ (0)
0
++
ND
0 C++)
0
0 (++)
ND
+ (+*)
0
+*
ND
0 (++)
0
+ (++)
7
3
14
0
0
24
inning HI
4
ND
o {--)
o (-)
o
ND
o (-)
o
**
ND
+ (-)
0
**
ND
0
o
o
ND
++ (0)
o
+*
ND
0 (+)
o
o C++)
ND
++ ( + )
0
++
ND
0 ( + )
o
0 (++)
6
1
17
0
0
24
reas
5
ND
0
o
o
ND
0
o
++
ND
++ (0)
O
•f-f
ND
0
o (-)
o
ND
+ (o)
0
*+
ND
0 ( + )
0
0 (++)
ND
+ (0)
0
+ (++)
ND
o
o
+ C++)
4
4
16
0
0
24
6
ND
o (--)
o
o
ND
--
--
0 (++)
ND
0 (--)
--
0 (++)
ND
o
0
o
ND
o (-)
- (o)
o C++)
ND
o
0
o C++)
ND
0
o (--)
0 (++)
ND
o
-
0 (++)
0
0
19
2
3
24
7'
ND
- (0)
0
0 ( + )
ND
o
0
**
ND
++ (0)
0
*+
ND
O
0
0 ( + )
ND
0
-- (o)
0 (++)
ND
o
o
0 (++)
ND
0
0
0 (++)
ND
0
- (o)
0 (++)
3
0
18
2
1
24
i- lompai
221
ND
o
0 (++)
ND
o
**
ND
" (0)
+ (++)
ND
o
0
ND
- (o)
o C++)
ND
0
0
ND
-
0 (++)
ND
0
0 (++)
1
1
11
2
1
16
rison AH
225
ND
o
0
ND
0
++
ND
++ (0)
+*
ND
o
o (•*•)
ND
+ (o)
+*
ND
o
0 (++)
ND
+
+ (++)
ND
0
+ C++)
4
4
8
0
0
16
;as -
C&T
ND
0 (
0
ND
--
--
ND
- (
--
ND
0
0 (
ND
0 (
--
ND
o
0 (
ND
0 (
- (
ND
0 (
- (
0
0
9
3
4
16
EDA > Comparison area at 0.01 significance level
EDA > Comparison area at 0.05 significance level
No significant difference at 0.05 significance level
-: EDA < Comparison area at 0.05 significance level
--: EDA < Comparison area at 0.01 significance level
b. All test results reported are based on two-sided p-values.
c. The first entries in each column are median concentrations. ND indicates non-detect.
d. Any changes in results between the original analysis and this analysis is indicated by two sets of
symbols; the second set of symbols denotes the original results for the GOOD observations.
e. C&T = Cheektowaga and Tonawanda
221 = Census Tract 221
225 = Census Tract 225
-------
Table K-10
SUMMARY OF NONPARAMETRIC UNIVARIATE STATISTICAL COMPARISONS
WITH OBSERVATIONS CLASSIFIED AS GOOD a,b
All observations <= 1 ppb treated as non-detect
Totals over all LCICs
Comparison
Area
CT221
CT225
C&T
'l
7
0
1
0
0
7
0
1
0
0
7
0
1
0
0
2
4
0
4
0
0
1
0
7
0
0
5
0
3
0
0
— EUH :
3
3
1
4
0
0
0
1
7
0
0
4
1
3
0
0
sampling
4
2
1
5
1
1
0
0
8
0
0
4
0
4
0
0
Areas -•
5
1
2
5
0
0
0
0
8
0
0
3
2
3
0
0
6
0
0
7
0
1
0
0
4
2
2
0
0
8
0
0
7
1
0
6
1
0
0
0
6
1
1
2
0
6
0
0
l- uonc
221
0
0
5
2
1
1
1
6
0
0
arisen /
225
1
2
5
0
0
3
2
3
0
0
\reas -
C&T
0
0
6
1
1
0
0
3
2
3
a. ++: EDA > Comparison area at 0.01 significance level
+: EDA > Comparison area at 0.05 significance level
o: No significant difference at 0.05 significance level
-: EDA < Comparison area at 0.05 significance level
--: EDA < Comparison area at 0.01 significance level
b. All test results reported are based on two-sided p-values.
c. C&T = Cheektowaga and Tonawanda
221 = Census Tract 221
225 = Census Tract 225
-------
Table K-11
TWO-SIDED p-VALUES FOR THE UNIVARIATE STATISTICAL COMPARISONS
WITH OBSERVATIONS CLASSIFIED AS GOOD
All observations <= 1 ppb treated as Non-detect
LCIC
DCB
TCB
OCB
CNP
A-BHC
D-BHC
B-BHC
G-BHC
comparison |
Area 1
221
225
C&T
221
225
C&T
221
225
C&T
221
225
C&T
221
225
C&T
221
225
C&T
221
225
C&T
221
225
C&T
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
1.000
1.000
1.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
2
0.201
0.112
0.126
0.000
0.056
0.000
0.000
0.002
0.000
1.000
1.000
1.000
0.001
0.219
0.000
0.162
0.454
0.113
0.001
0.165
0.001
0.086
0.551
0.009
•-- CUM sampling
3 4
0.500
0.186
0.128
0.001
0.352
0.000
0.000
0.043
0.000
1.000
1.000
1.000
0.007
0.756
0.002
0.141
0.487
0.078
0.013
0.512
0.004
0.220
0.925
0.017
0.435
0.971
0.678
0.585
0.260
0.000
0.022
0.408
0.000
1.000
1.000
1.000
0.001
0.421
0.001
0.205
0.755
0.335
0.004
0.182
0.001
0.206
0.862
0.063
Areas -•
5
0.456
0.190
0.189
0.080
0.969
0.000
0.005
0.778
0.000
1.000
1.000
1.000
0.041
0.740
0.005
0.255
0.728
0.165
0.028
0.805
0.024
0.097
0.627
0.016
6
0.534
0.697
0.809
0.009
0.000
0.063
0.759
0.000
0.063
1.000
1.000
1.000
0.949
0.011
0.251
0.508
0.814
0.386
0.861
0.056
0.792
0.182
0.013
1.000
. comparison areas -
7 221 225 C&T
0.019
0.095
0.052
0.486
0.241
0.000
0.006
0.405
0.000
1.000
1.000
1.000
0.963
0.010
0.258
0.398
0.786
0.423
0.269
0.421
0.251
0.169
0.010
1.000
1.000
0.495
0.693
1.000
0.121
0.000
1.000
0.003
0.030
1.000
1.000
1.000
1.000
0.016
1.000
1.000
0.289
1.000
1.000
0.042
1.000
1.000
0.203
0.317
0.495
1.000
0.629
0.121
1.000
0.000
0.003
1.000
0.000
1.000
1.000
1.000
0.016
1.000
0.005
0.289
1.000
0.317
0.042
1.000
0.050
0.203
1.000
0.050
0.693
0.629
1.000
0.000
0.000
1.000
0.030
0.000
1.000
1.000
1.000
1.000
1.000
0.005
1.000
1.000
0.317
1.000
1.000
0.050
1.000
0.317
0.050
1.000
a. C&T = Cheektowaga and Tonawanda
221 = Census Tract 221
225 = Census Tract 225
-------
Table K-12
TWO-SIDED p-VALUES FOR THE UNIVARIATE STATISTICAL COMPARISONS
WITH OBSERVATIONS CLASSIFIED AS GOOD
All observations <= 1 ppb treated as Non-detect
LCIC
DCB
TCB
QCB
CNP
A-BHC
D-BHC
B-BHC
G-BHC
comparison (-•
Area 1
221
225
C&T
221
225
C&T
221
225
C&T
221
225
C&T
221
225
C&T
221
225
C&T
221
225
C&T
221
225
C&T
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
1.000
1.000
1.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
2
0.201
0.112
0.126
0.000
0.056
0.000
0.000
0.002
0.000
1.000
1.000
1.000
0.001
0.219
0.000
0.162
0.454
0.113
0.001
0.165
0.001
0.086
0.551
0.009
— EU« sampling
3 4
0.500
0.186
0.128
0.001
0.352
0.000
0.000
0.043
0.000
1.000
1.000
1.000
0.007
0.756
0.002
0.141
0.487
0.078
0.013
0.512
0.004
0.220
0.925
0.017
0.435
0.971
0.678
0.585
0.26C
0.000
0.022
0.408
0.000
1.000
1.000
1.000
0.001
0.421
0.001
0.205
0.755
0.335
0.004
0.182
0.001
0.206
0.862
0.063
Areas • •
5
0.456
0.190
0.189
0.080
0.969
0.000
0.005
0.778
0.000
1.000
1.000
1.000
0.041
0.740
0.005
0.255
0.728
0.165
0.028
0.805
0.024
0.097
0.627
0.016
6
0.534
0.697
0.809
0.009
0.000
0.063
0.759
0.000
0.063
1.000
1.000
1.000
0.949
0.011
0.251
0.508
0.814
0.386
0.861
0.056
0.792
0.182
0.013
1.000
7'
0.019
0.095
0.052
0.486
0.241
0.000
0.006
0.405
0.000
1.000
1.000
1.000
0.963
0.010
0.258
0.398
0.786
0.423
0.269
0.421
0.251
0.169
0.010
1.000
|- comparison areas -
221 225 C&T
1.000
0.495
0.693
1.000
0.121
0.000
1.000
0.003
0.030
1.000
1.000
1.000
1.000
0.016
1.000
1.000
0.289
1.000
1.000
0.042
1.000
1.000
0.203
0.317
0.495
1.000
0.629
0.121
1.000
0.000
0.003
1.000
0.000
1.000
1.000
1.000
0.016
1.000
0.005
0.289
1.000
0.317
0.042
1.000
0.050
0.203
1.000
0.050
0.693
0.629
1.000
0.000
0.000
1.000
0.030
0.000
1.000
1.000
1.000
1.000
1.000
0.005
1.000
1.000
0.317
1.000
1.000
0.050
1.000
0.317
0.050
1.000
a. C&T = Cheektowaga and Tonawanda
221 = Census Tract 221
225 = Census Tract 225
-------
Table K-13
Simulated power (at alpha = 0.05) for univariate tests
LCIC
Delta 1 2345678
0.00 .050 .050 .050 .044 .048 .056 .058 .042
0.05 .122 .088 .086 .140 .062 .076 .074 .070
0.10 .244 .144 .130 .322 .096 .088 .104 .084
0.20 .616 .286 .254 .756 .136 .124 .152 .156
0.50 .988 .808 .756 1.000 .338 .354 .412 .452
1.00 1.000 .990 .982 1.000 .676 .740 .808 .832
2.00 1.000 1.000 1.000 1.000 .958 .984 .994 .998
5.00 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000
10.0 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000
a. Estimated parameters for power simulation from "good" soil
LCIC study data.
b. LCIC concentrations in EDA sampling areas are drawn from a
statistical population reliited to that of the comparison area
by the factor 1+delte; delta = 10.0 corresponds to an order of
magnitude shift. The goal for the soil LCIC study was to have
a power of 0.90 at a delta of 10.0.
Table K-14
Simulated power 'for multivariate test
ALPHA
.0100 .0250 .0500 .1000 .2000
Deltaa
.00
.05
.10
.20
.50
1.00
2.00
5.00
10.00
.000
.000
.000
.014
.256
.920
1.000
1.000
1.000
.006
.000
.004
.048
.402
.968
1.000
1.000
1.000
.022
.022
.034
.096
.528
.990
1.000
1.000
1.000
.066
.072
.102
.174
.656
.994
1.000
1.000
1.000
.176
.192
.216
.340
.808
.998
1.000
1.000
1.000
a. Estimated parameters for power simulation from "good" soil
LCIC study data.
b. LCIC concentrations in EDA sampling areas are drawn from a
statistical population related to that of the comparison area
by the factor 1+delta; delta = 10.0 corresponds to an order of
magnitude shift. The goal for the soil LCIC study was to have
a power of 0.90 at a delta of 10.0.
------- |