Hazard Ranking System Issue Analysis:
Consideration of Contaminant Concentration
MITRE
-------
Hazard Ranking System Issue Analysis:
Consideration of Contaminant Concentration
Thomas F. Wolfinger
June 1987
MTR-86W40
SPONSOR:
U.S. Environmental Protection Agency
CONTRACT NO.:
EPA-68-01-7054
The MITRE Corporation
Civil Systems Division
7525 Colshire Drive
McLean, Virginia 22102-3481
-------
Department Approval:
MITRE Project Approval:
.
JUk.
ii
-------
ABSTRACT
This report examines the issue of the extent to which
concentration data can play an increased role within the Hazard
Ranking System (HRS) for the evaluation of uncontrolled waste sites.
The report examines general limitations in concentration data,
limitations arising from programmatic constraints in the site
evaluation program, and limitations arising during the data
development process. The report concludes that there is a large
number of important limitations that restrict the utility of
concentration data in the HRS. The report also examines, to the
limited extent possible, the potential costs of developing
representative concentration data at uncontrolled waste sites.
Several options for increasing the role of concentration data in the
HRS are described. These options were designed within the context of
limitations in the availability and extent of concentration data.
Suggested Keywords: Superfund, Hazardous ranking, Hazardous waste,
Concentration, Hazardous constituents.
iii
-------
ACKNOWLEDGMENT
The author would like to acknowledge the contributions of
John A. Cioffi of the Chemical and Biological Systems Department of
the MITRE Corporation. Mr. Cioffi's preliminary work on the issue of
the use of concentration data in the Hazard Ranking System formed the
basis for many of the ideas, examination, and conclusions contained
in this report.
iv
-------
TABLE OF CONTENTS
Page
LIST OF ILLUSTRATIONS vii
LIST OF TABLES vlii
1.0 INTRODUCTION !
1.1 Background 1
1.2 Issue Description 3
1.3 Scope and Approach of Analysis 5
1.4 Organization of the Report 6
2.0 GENERAL LIMITATIONS IN CONCENTRATION DATA 9
3.0 LIMITATIONS ARISING FROM PROGRAM CONSTRAINTS 13
3.1 Funding Constraints 13
3.2 Time Constraints 14
3.3 Variation in Levels of Expertise 14
3.4 Quality Control and Quality Assurance 15
4.0 LIMITATIONS IN CONCENTRATION DATA QUALITY 17
4.1 Spatial Variation 18
4.1.1 Environmental Data 20
4.1.2 Wastes 45
4.2 Temporal Variation 48
4.2.1 Environmental Data 49
4.2.2 Wastes 56
4.3 Limitations in the Data Development Process 57
4.3.1 Sampling 58
4.3.2 Handling 75
4.3.3 Analysis 78
4.3.4 Interpretation 90
5.0 COSTS 93
5.1 Problems in Estimating Costs 93
5.2 Media 96
5.3 Contaminants 100
-------
TABLE OF CONTENTS (Concluded)
5.4 Number of Sampling Locations
5.5 Number of Samples Per Location HI
5.6 Unit Sampling, Handling, and Analysis Costs H4
5.7 Illustrative Example 117
6.0 SUMMARY OF PRECEDING DISCUSSIONS AND IMPLICATION H9
FOR THE HRS
7.0 OPTIONS FOR EMPLOYING CONCENTRATION DATA 123
7.1 Basic Philosophy 123
7.2 Benchmarks 126
7.3 Options 127
7.3.1 Options in the HRS Release Category 127
7.3.2 Options in the HRS Waste Characteristics 131
Category
7.3.3 Options in the Targets Category 142
8.0 OVERALL CONCLUSIONS AND RECOMMENDATIONS 147
9.0 REFERENCES AND BIBLIOGRAPHY 153
APPENDIX A - COMMENTS ON THE HRS RELATED TO HAZARDOUS 177
WASTE CONCENTRATIONS
APPENDIX B - DATA QUALITY 183
vi
-------
LIST OF ILLUSTRATIONS
Figure Number Page
1 DCE Profile at Site in February 1984 23
2 PCE Profile at Site in February 1984 24
3 Thomas Solvent Ground Water Total VOC 26
Concentration Profile, August 1984
4 Vertical Distribution of Representative 27
Organic Compounds From Upgradient and
Dowugradient Multilevel Monitoring System
5 Vertical Distribution of Representative 28
Inorganic Compounds From Upgradient and
Downgradient Multilevel Monitoring System
6 Vertical Nitrate and Dissolved Oxygen 32
Distribution in Shallow Sandy Aquifer (after
Hendry et al., 1983)
7 Range and Mean for Vinyl Chloride Monitoring 40
Near BKK Landfill, West Covina, California
8 Isoarea Map of Kriging Estimate of Lead 43
Concentrations (ug/g) in Soil
9 Fluctuations of Nitrate Concentrations in 51
Water from Individual Shallow Wells Over Time
10 Change in Shallow Plume Over Time, as Measured 53
by Em Conductivity
11 Water Quality Fluctuations with Time of Pumping 67
VI1
-------
LIST OF TABLES
Table Number
1 Horizontal Variation in Ground Water 30
Contamination Biscayne Aquifer/Dade County,
Florida
2 Vertical Variation in Ground Water 31
Contamination Biscayne Aquifer/Dade County,
Florida
3 Spatial/Temporal Variability in Atmospheric 34
Benzene Concentrations: Kin-Buc Landfill
(1976)
4 Spatial/Temporal Variability in Atmospheric 35
Carbon Tetrachloride Concentrations: Kin-Buc
Landfill (1976)
5 Spatial/Temporal Variability in Atmospheric 36
Chloroform Concentrations: Kin-Buc Landfill
(1976)
6 Ranges of Landfill Gas Concentrations at BKK 38
Landfill
7 Coefficients of Variations (%) for Soil 41
Parameters Reported in Mason (1983)
8 Examples of Spatial Variation in Surface Water 44
Contaminant Concentrations
9 Spatial Variability in Waste Concentration 47
10 Concentrations of Total Organic Compounds in 52
Water from Four Long Island, New York Wells
Over a One-Year Period (1977-1978)
11 Potential Sources of Bias in Sampling Ground 63
Water
12 Sampling Methods for Airborne Organics 71
13 Sampling Equipment for Particular Waste Types 76
viii
-------
LIST OF TABLES (Continued)
Table Number Pa8e
14 RCRA Appendix VIII Compounds for Which No 80
Suitable Analytical Parameter Methodology
or Standards Exist, or for Which a Surrogate
Has Been Suggested as an Analyte
15 Acceptable Ranges for Precision and Accuracy 81
in the EPA Contract Laboratory Program
16 CLP Performance Evaluation Sample Results, 83
1982
17 Analysis of CLP Performance Evaluation Sample 84
Data, 1982
18 Summary of Inter-Laboratory Comparison 85
19 Summary of Inter-Laboratory Comparison 86
20 Comparison of Gas Chromatography (GC) and Gas 88
Chromatography/Mass Spectroscopy (GC/MS)
Results
21 Magnitude of Laboratory Variability 89
22 Ranges of Selected Organic Contaminant 92
Concentrations in Urban Air
23 Summary of Sampling Costs 99
24 Recommended Media in Which to Sample RCRA 101
Appendix VIII Hazardous Constituents
25 Expected Confidence Intervals for a Parameter 113
Mean as a Function of Number of Samples
26 Example Representative Pre-Sampling Costs 116
Ten^Well Sampling Field
27 Hypothetical Examples of Release Category 130
Option
28 Illustrative Concentration Factor Table 133
ix
-------
LIST OF TABLES (Concluded)
Table Number Page
29 Hypothetical Examples of Concentration Factor 135
Option
30 Illustrative Concentration Factor 139
31 Illustrative Toxicity Factor Matrix 141
32 Illustrative Table for Evaluating Population 144
(Ground Water Pathway)
33 Illustrative Tables for Evaluating Population 145
(Air Pathway)
34 Guidance on Concentration Data Development: 149
Selections from the Literature
-------
1.0 INTRODUCTION
1.1 Background
The Comprehensive Environmental Response, Compensation, and
Liability Act of 1980 (CERCLA) (PL 96-510) requires the President to
identify national priorities for remedial action among releases or
threatened releases of hazardous substances. These releases are to
be identified based on criteria promulgated in the National
Contingency Plan (NCP). On July 16, 1982, EPA promulgated the
Hazard Ranking System (HRS) as Appendix A to the NCP (40 CFR 300;
47 FR 31180). The HRS comprises the criteria required under CERCLA
and is used by EPA to estimate the relative potential hazard posed
by releases or threatened releases of hazardous substances.
The HRS is a means for applying uniform technical judgment
regarding the potential hazards presented by a release relative to
other releases. The HRS is used in identifying releases as national
priorities for further investigation and possible remedial action by
assigning numerical values (according to prescribed guidelines) to
factors that characterize the potential of any given release to
cause harm. The values are manipulated mathematically to yield a
single score that is designed to indicate the potential hazard posed
by each release relative to other releases. This score is one of
the criteria used by EPA in determining whether the release should
be placed on the National Priorities List (NPL).
-------
During the original NCP rulemaklng process and the subsequent
application of the HRS to specific releases, a number of technical
Issues have been raised regarding the HRS. These Issues concern the
desire for modifications to the HRS to further improve its
capability to estimate the relative potential hazard of releases.
The issues include:
• Review of other existing ranking systems suitable for
ranking hazardous waste sites for the NPL.
• Feasibility of considering ground water flow direction and
distance, as well as defining "aquifer of concern," in
determining potentially affected targets.
• Development of a human food chain exposure evaluation
methodology.
• Development of a potential for air release factor category
in the HRS air pathway.
• Review of the adequacy of the target distance specified in
the air pathway.
• Feasibility of considering the accumulation of hazardous
substances in indoor environments.
• Feasibility of developing factors to account for
environmental attenuation of hazardous substances in ground
and surface water.
• Feasibility of developing a more discriminating toxicity
factor.
• Refinement of the definition of "significance" as it relates
to observed releases.
• Suitability of the current HRS default value for an unknown
waste quantity.
• Feasibility of determining and using hazardous substance
concentration data.
-------
• Feasibility of evaluating waste quantity on a hazardous
constituent basis.
• Review of the adequacy of the target distance specified in the
surface water pathway.
• Development of a sensitive environment evaluation methodology.
• Feasibility of revising the containment factors to increase
discrimination among facilities.
• Review of the potential for future changes in laboratory
detection limits to affect the types of sites considered for
the NPL.
Each technical issue is the subject of one or more separate but
related reports. These reports, although providing background,
analysis, conclusions, and recommendations regarding the technical
issue, will not directly affect the HRS. Rather, these reports will
be used by an EPA working group that will assess and integrate the
results and prepare recommendations to EPA management regarding future
changes to the HRS. Any changes will then be proposed in Federal
notice and comment rulemaking as formal changes to the NCP. The
following section describes the specific issue that is the subject of
this report.
1.2 Issue Description
It is intuitively and intellectually attractive that knowledge of
the actual concentrations resulting from releases from uncontrolled
waste sites should lead to better characterizations of the risks
associated with the sites. The extent of (actual or potential) damage
to human health or the environment caused by a release of hazardous
substances is a function of dose which is, in turn, directly related
3
-------
to the concentration of the substance found at the point of exposure
and the duration of time that the concentration is maintained.
Thus, risk from contaminant releases is a direct function of the
exposure concentrations.
Individuals and organizations commenting on the MRS and its
subsequent application to uncontrolled waste sites and other releases
have expressed several considerations. The commenters have indicated
that concentration data are needed to: 1) determine if a contaminant
has been released from the site, 2) estimate the quantity of
hazardous substances present at the site, and 3) determine the human
and environmental threat posed by the substances at or released from
the site. A more extensive review of the comments received by EPA
concerning the concentration issue can be found in Appendix A.
Currently, concentration data are used in the HRS to determine
whether a release of contaminants has occurred, termed an "observed
release". These data are also used to assist in determining the
substances present at the site.
The avoidance of increased reliance on concentration data to
assess site risks using the HRS is based on: 1) the current
role of the HRS as a screening tool, 2) shortcomings in the
representativeness of concentration data, 3) the quality of data
that can be collected, 4) the costs of data collection, and 5)
limitations in inferring exposure concentrations from environmental
and waste concentrations.
-------
1.3 Scope and Approach of Analysis
Several different types of concentration data can be developed
during preliminary assessments (PA) and site inspections (SI).
Generically, the types of data can be grouped by media as follows:
• Ground water
• Surface water
• Soil
• Air
• Wastes
For the purposes of this report, the first four media are considered
"environmental" media as are samples collected from those media.
For example, ground water is considered an environmental medium and
ground water samples are considered environmental samples. For
purposes of distinction, waste samples are considered separately. A
sixth medium of concern, sediments, is discussed, as applicable,
under the categories of soil or wastes.
Several factors determine the role that concentration data can
play in any regulatory or scientific program. The primary factor is
the objectives of the program. Of nearly equal importance, however,
are the inherent limitations in concentration data, the programmatic
constraints placed on the development of such data, and the quality
of the data that can be developed in support of the program given
programmatic and other constraints. The role of the data should
further the achievement of the objectives within the constraints
-------
posed by the quality and other limitations of the data. In the
context of the HRS, the objective of the program is to develop a
ranking of uncontrolled waste sites in a fashion that reflects the
relative, multi-media, risks of the sites as accurately as possible
within the resource and knowledge constraints of the program. The
resulting list of sites is used in evaluating sites for further,
more detailed, investigation. Given this objective, a thorough
understanding of the limitations in concentration data is needed to
define the appropriate role for concentration data in the HRS.
In order to determine the appropriate role for concentration
data, an examination of the limitations in concentration data
developed during activities such as preliminary assessments and site
inspections was undertaken. The purpose of this examination was to
delineate the limitations in such data that would restrict its use
in the HRS. Once the limitations were understood, options of
employing concentration data within the HRS were developed.
1.4 Organization of the Report
This report presents the results of an examination of the
nature of concentration data developed during PAs and Sis and the
results of the effort to develop options for using such data in the
HRS. Sections 2, 3, and 4 discuss the limitations associated with
the types of concentration data collected during site inspections as
they relate to the utility of such data in the HRS. A limited
discussion of the probable costs associated with developing
-------
representative concentration data is presented in Section 5.
Section 6 discusses the implications of data limitations for the
use of concentration data in the HRS while Section 7 presents
several options for using concentration data within the framework
of the HRS. These options were developed to make optional use of
the concentration data as warranted by limitations in the data.
Ultimately, the question of which uses of concentration data are
acceptable, given the limitations, must be decided by EPA. Section 8
presents the overall conclusion and recommendations of the issue
analysis. A bibliography is presented in Section 9. Appendix A
contains a discussion of the comments made by interested parties on
the issue of using concentration data in the HRS. Appendix B
contains a short discussion on the meaning of representativeness in
terms of concentration data.
-------
2.0 GENERAL LIMITATIONS IN CONCENTRATION DATA
A principal limitation with concentration data is that they
rarely represent the concentrations to which individuals will
actually be exposed. This limitation is important in assessing risk
since human health risk from environmental contamination is a
function of exposure concentrations; there are generally no simple
relationships between overall environmental or waste concentrations
and exposure concentrations. This limitation may be minimized when
sampling is performed at the point of exposure, although other
problems may arise when sampling "exposure" concentrations. For
example, samples drawn from drinking water taps are likely to yield
samples representative of exposure concentrations, although such
data may suffer from interferences from nonsite-related contaminants
(such as chloroform arising from water chlorination or lead from
lead solder in pipes). Moreover, the technology of sampling
exposure concentrations is, in some cases, not in an advanced stage
of development. For example, special equipment was developed to
assess personal airborne contaminant exposure during the Total
Exposure Assessment Methodology (TEAM) EPA project (Wallace, 1985).
A second important limitation in concentration data is that
concentration data are fundamentally retrospective in nature. At
best, the data represent contaminant concentrations at the sampling
locations at some point in the past. This point may be a few minutes
or several months in the past. As such, concentration data are
-------
never totally complete. It is only with the aid of models (Including
simple models that assume that conditions and concentrations will
not change over time) that one can infer future concentrations from
past concentrations. The importance of this limitation in terms of
the utility of concentration data in the HRS depends upon the
objectives for the use of the data in the HRS. If the purpose of
the HRS remains to assess relative risks associated with actual or
potential contaminant releases, then this limitation may be critical.
The above limitations are important; however, in perform-frig risk
assessments scientists have found it necessary to accommodate then.
A third important limitation associated with concentration data
arises from the statistical nature of concentration data. Even
given a high state of quality control for a particular data set, one
cannot be certain of the true concentration. Ihere is always some
probability that the true contaminant concentration is significantly
higher or lower than that Indicated by the data. In general, the
probability that the true concentration will be significantly higher
(or lower) than the estimated concentration will decrease as the
representativeness of the concentration data increases. Conversely,
uncertainty will increase as the degree of representativeness
decreases. This is particularly important if the data set
represents only a portion of the true spatial/temporal distribution
of concentrations, or if it represents only one of the many
spatial/temporal concentration distributions associated with the
10
-------
site due to problems of specificity, homogeneity, accuracy, and
completeness. This is discussed further in Section 4.
Further, the temporal limitation discussed above could act to
increase the uncertainty associated with inferences associated with
low levels of detected contamination. This concern is particularly
important when inferring that a site poses no risk to individuals
living near the site or when inferring that the site will not
release contaminants (e.g., when contaminants have simply not yet
been released from the site).
11
-------
3.0 LIMITATIONS ARISING FROM PROGRAM CONSTRAINTS
Several aspects of the current structure of the CERCLA
preliminary assessment and site inspection (PA/SI) program limit the
representativeness and utility of the concentration data developed
for use in evaluating sites. The most important of these aspects are:
• Funding constraints
• Time constraints
• Variation in levels of expertise
• Current quality assurance and quality control (QA/QC) program
Since the PA/SI program would likely be revised in concert with
revisions to the HRS, these aspects of the program may become less
limiting in the future.
3.1 Funding Constraints
Currently, a site inspection cost is about 35-50 thousand dollars
(Caldwell, 1986). Preliminary assessment costs are a relatively minor
component of overall PA/SI costs. Some site inspections cost less
than the average while others cost considerably more. Given the range
of costs associated with data development (addressed in Section 5),
it is certain that the availability of monetary resources will
restrict the extent of the collection of concentration data at most
sites. The effects of these constraints are seen in the current
program, in part in the emphasis on ground water and soil sampling*
and on the number of samples taken.
*This emphasis also derives from a perception that ground water
and soil contamination are the sources of the principal component
of site risks.
13
-------
3.2 Time Constraints
A similar constraint arises from the need to inspect sites as
quickly as possible to ensure that the risks from those sites
requiring cleanup are mitigated as soon as possible. This, generally
implicit, time constraint is reflected in the Congressional
discussions on requirements for sites to be placed on the NPL, for
remedial investigations to be performed, and for sites to be cleaned
up by specific dates. This concern is addressed in the Superfund
Amendments and Reauthorization Act of 1986 Conference Report. The
Joint Conference Committee states in this report that the requirement
that the HRS rank sites as accurately as possible does not require
that long-term monitoring be conducted during site inspections and
that the role of the HRS is to expeditiously identify sites for
response actions. This time constraint restricts the duration of
sampling and, together with other resource constraints, may restrict
the spatial, media, and contaminant completeness of the data
development effort.
3.3 Variation in Levels of Expertise
A third limitation arises from the wide mix of interested
parties involved in concentration data development. The PA/SI
program is designed to make use of a wide range of resources
available to the Federal government, State and local governments
and other interested parties. The use of these resources is
generally desirable for many reasons including cost and time savin
14
-------
However, three aspects of this practice are of concern if increased
reliance is to be placed on concentration data. First, are the
differences in the objectives for which the samples are collected.
For example, some samples are collected to ensure investigator safety,
utilizing measurements to support short-term acute effects risk
assessments. This objective frequently results in the use of
specialized sampling techniques that may not yield sufficiently
reliable data to support longer-term chronic risk determinations.
The second aspect of concern is the wide variation in the quality
of the data developed outside of EPA. This variation results in part
from variations in the level of technical expertise available to
interested parties that might develop concentration data. This
variation indicates the need for using trained professionals in site
inspections, as indicated by Glaccum, Benson, and Noel, 1982, to
ensure the representativeness of data collected. The final aspect is
the degree of control that EPA can exercise over concentration data
development activities in order to ensure that adequate QA/QC
practices are employed by all parties collecting data to support site
evaluations. Almost all of the Superfund program is implemented by
the States and the degree to which EPA can ensure QA/QC for site
inspections conducted by such other parties is problematic.
3.4 Quality Control and Quality Assurance
A final programmatic constraint relates to potential problems in
the field and laboratory quality control and quality assurance program
15
-------
that is applied to the concentration data, particularly to those
data developed outside of the EPA PA/SI program. The uncertainty of
data that have not been developed subject to a uniformly applied,
comprehensive field and laboratory QA/QC program must be considered
to be high. Increased reliance on concentration data in the HRS
would require accompanying changes in the QA/QC program, especially
in regard to data developed outside of EPA, to insure that the
concentration data were developed properly, in all cases, to the
extent possible given current capabilities.
The Superfund program is not alone in requiring a high state of
QA/QC for the collection of concentration data. Cooperative efforts
with other organizations can help promote consistency in data
collection. Other organizations emphasizing the need for
concentration data QA/QC in decision making include the American
Chemical Society (Keith, Lawrence H. et al., 1983), the U.S.
Geological Survey (Claassen, 1982) and parts of EPA outside of
the Superfund program (Scalf et al., 1981; Riggin, 1983; and Gosse
et al., 1986).
16
-------
4.0 LIMITATIONS IN CONCENTRATION DATA QUALITY
The purpose of this chapter is to present a general discussion
of the limitations, uncertainties, and confounding factors that
affect the quality of concentration data and its utility in the
HRS. Quality, as discussed in Appendix B, is determined by comparing
the representativeness of a data set with the requirements for
representativeness imposed by the uses to which the data will be
put. The determination of what constitutes an acceptable degree of
representativeness is an issue that must be resolved by the user (in
this case, EPA) based on the objectives of the program and the role
the data play in achieving those objectives.
It is apparent in examining the literature that a large number
of factors interact to affect concentration data. These factors
determine the degree to which the results of a particular sampling
program are representative. These factors generally fall into two
categories, those reflected in "natural" variation in concentrations
(i.e., arising from variation in the environmental and waste disposal
processes that influence concentrations) and those of human origin
that relate to errors and resulting variations in concentrations
that arise during data development. A lack of knowledge of these
factors and the degree to which they affect an individual data set
increases the uncertainty associated with that data set, limiting
its usefulness. These factors include, for example, contamination
of well samples by materials used in well construction, temporal
17
-------
variations in atmospheric characteristics inducing variations in
contaminant concentrations, safety considerations in sampling sealed
drums, spatial variations in soil contaminant concentrations caused
by differential deposition patterns, gaps in knowledge of how to
analyze nonnormal data, and gaps in technologies and procedural
standards for determining sample concentrations for many contaminants
of concern. A large amount of information has been published
concerning such confounding factors (e.g., Barcelona, 1983; Driscoll,
1986; Fetter, 1983; Geraghty and Miller, Inc. and American Ecology
Services, Inc., 1985; Gillham et al., 1983; Keith et al., 1983a and
I983b; Scalf et al., 1981; Sgambat and Stedinger, 1981; Weber and
Mims, 1981; and Zachowski and Borgianini, 1984). Procedures to
partially alleviate the effects of these factors have been developed
in many cases. A truly comprehensive discussion of these factors in
all five of the media of interest is beyond the scope of this paper.
This chapter deals with an overview of these factors and their
implications for limiting the use of concentration data in the MRS.
The factors play greater or lesser roles depending on the media to
be evaluated. The chapter is divided into three sections, the first
two address spatial and temporal variability, respectively. The
third section addresses limitations in the data development process.
4.1 Spatial Variation
Spatial variation is the first of the two natural sources of
variation in concentration data. Spatial variation is defined as
18
-------
the variation in a parameter of interest (e.g., contaminant
concentrations) seen at a given point in time, between different
locations on or near a site. It is important to note that both
horizontal and vertical spatial variation are important at many
sites. Spatial variation in contaminant concentrations is three
dimensional and is induced by several environmental factors and the
spatial variability of these factors. Factors of particular
importance include transport media characteristics (e.g., for ground
water, the direction of flow, stability, porosity, and soil clay
content), contaminant and waste characteristics (e.g., for air, the
physical state, vapor pressure, solubility, and molecular weight);
and site characteristics (such as location and containment of
deposited materials). These characteristics may act singly to
induce variations or may interact, such as in the case of
photodegradation.
Spatial variation is a critical factor in determining the
number, density, and location of sampling points. Generally, those
sites expected to have a high degree of spatial variability would
require a greater number of sampling locations to ensure a given
level of representativeness in the data. Conversely, spatially
homogeneous sites (those with a low degree of spatial variability)
would require fewer sampling locations. As might be expected, few
sites are spatially homogeneous in any of the five media of concern.
19
-------
A major problem in dealing with spatial variation is determining
the degree of spatial variability. It is necessary to know the state
of variability to determine how many samples must be collected to
obtain an adequate representation of the range and arithmetic mode of
site concentrations. A determination of the degree of homogeneity
itself may require extensive sampling.
4.1.1 Environmental Data
This section discusses examples from the literature illustrating
the extent and importance of spatial variation in the four
environmental media.
Ground Water. Several authors indicate that there is a fairly
high degree of spatial variability in ground water conditions and
contaminant concentrations (e.g., Berg, 1982; Driscoll, 1986; Gillham
et al., 1983; Gosse et al., 1986; McKown, Schalla, and English, 1984;
Nazar, Prieur, and Threlfall, 1984; Quinn, Wittmann, and Lee, 1985;
Sgambat and Stedinger, 1981; and U.S. Environmental Protection
Agency, 1977 and 1985b). Many indicate that spatial variability is
most evident in the vertical direction (Driscoll, 1986; Singh et al.,
1984; and U.S. Environmental Protection Agency, 1985b). As noted by
Sgambat and Stedinger (1981). ground water is not well mixed, in any
dimension.
Among the parameters that vary spatially and might affect ground
water contaminant concentrations are (McKown, Schalla, and English
1984):
20
-------
• Horizontal and vertical conductivity
• Coefficient of storage
• Porosity
• Specific yield and retention
Additional parameters include hydraulic gradient and rate of
transport, contaminant chemical properties, geochemical and
biological properties of the aquifer matrix, and discharge/recharge
parameters. Of particular importance are discontinuities in soil
strata, particularly with regard to fractures (Gillham et al., 1983
and Gosse et al., 1986). It is important to note that contaminants
do not necessarily move in the same direction and at the same speed
as the ground water (Aller et al., 1985).
A final, critical set of parameters that affect the spatial
distribution of contaminant concentrations are the location and
release characteristics of the waste site and any other possible
contamination sources (such as other, unrelated sites).
The extent of spatial variability in ground water contaminant
concentrations can be seen in studies by Sgambat and Stedinger
(1981); Quinn, Wittmann, and Lee (1985); Nazar, Prieur, and Threlfall
(1984); Singh et al. (1984); and Gillham et al. (1983).
For example, Sgambat and Stedinger (1981) reproduced a table in
their article showing the results of an analysis of variance (ANOVA)
conducted on observed nitrate concentrations in shallow wells located
on Long Island, New York. The ANOVA indicates that 70 percent of
21
-------
the total variation in nitrate concentrations could be accounted
for by well-to-well (or spatial) variation. About 27 percent could
be accounted for by within-well variation (temporal and other well-
specific variation) while only about 3 percent was considered to be
due to measurement error. The authors believe that the spatial
variation in concentrations was due to spatial variation in the
principal sources of nitrate contamination: cesspool leachate and
lawn fertilizers.
Quinn, Wittmann, and Lee (1985) analyzed the spatial
distribution of contamination at the Verona Well Field near Battle
Creek, Michigan as part of a CERCLA remedial investigation (RI).
Figures 1 and 2, illustrate the spatial distribution of 1,2-
dichloroethene (DCE) and tetrachloroethylene (or perchloroethylene,
PCE) in the area of the well field. The study identified three
probable sources of DCE and PCE: two Thomas Solvent facilities and
the railroad marshalling yard. DCE concentrations ranged from over
10,000 near one of the sources, declining to 50 ug/1 and below
within the well field. Environmental PCE concentrations arising
from the Thomas Solvent facilities (the southern PCE plume) have a
much smaller range, from 50 to 100 ug/1 near the source, and below
detection limits further out. Concentrations in the eastern PCE
plume range from about 100 ug/1 to below detection limits within
the boundaries of the well field. The authors stress that the two
wells immediately downgradient of the facility (source) showed no
22
-------
Thomas.
Solvent
Annex
Thomas Solvent
Raymond Road
0 300 600
Scale
In Feet
Note: Contour lines show 1,2 DCE concentration pg/l (Dashed where inferred)
Source: Quinn, Witman, and Lee, 1985.
FIGURE 1
DCE PROFILE AT SITE IN FEBRUARY 1984
23
-------
Thomas Solvent ,
Raymond Road lKf'\$
0 300 600
Scale
In Feet
Note: Contour lines show PCE concentrations in ?g/l (Dashed where inferred)
Source: Qulnn, Wltman, and Lee, 1985.
FIGURE 2
PCE PROFILE AT SITE IN FEBRUARY 1984
24
-------
contanilnation above the detection limit. Presumably, these two wells
missed the plume, indicating the importance of vertical variability.
The importance of hydraulic and chemical characteristics in
determining contaminant spatial distributions is also evident in
this study. The authors indicate that the probable causes for the
differences in spatial distributions of concentrations between the
DCE plume and the southern PCE plume are: 1) the higher retardation
factor for PCE and 2) the fact that DCE is a transformation product
of PCE. Wide variation in the concentrations in the eastern plume
is indicated in Figure 2, a map of the railroad yard PCE
concentrations. Contaminant concentrations ranged from 1,000 to
10 ug/1 within a distance of less than 400 feet.
A wide variation in ground water contaminant concentrations
is also indicated by the spatial distribution of soil gas VOC
concentrations, illustrated in Figure 3. Within a distance of
less than 200 feet, soil gas VOC concentrations range over 5 orders
of magnitude.
The potential for variability in the vertical dimension is
further evident from the concentration profiles presented in Nazar,
Prieur, and Threlfall (1984) (Figures 4 and 5). These samples were
taken in the area of an abandoned chemical plant that produced
specialty intermediates. There is a large degree of vertical
variation in the concentrations of both organic and inorganic
contaminants. For example, chlorobenzene ranges from about 0.01 mg/1
25
-------
\
7 /
RAYMOND RD
m
30 50
1—(—
Scale
In Feet
100
Note: Contour lines show total groundwater VOC concentrations in
Source: Quinn, Witman, and Lee, 1985.
FIGURE 3
THOMAS SOLVENT GROUND WATER TOTAL
VOC CONCENTRATION PROFILE, AUGUST 1984
26
-------
0.0 -
10.0 -
50.0 -1—
001
Water Table
jenac ^ oichloroethane
Chlorobenzene
Upgradient
I I I I I I I I I I I I I I I I I I I i I I I I I I I
0.10
1.00
Concentration (mg./L)
100.00
Water Table
40.0 -
50.0 -
Downgradlent
0.01
Source: Nazar, Prleur, and Threlfall, 1984.
100
Concentration (mg./L)
100.00
FIGURE 4
VERTICAL DISTRIBUTION OF REPRESENTATIVE ORGANIC
COMPOUNDS FROM UPGRADIENT AND DOWNGRADIENT
MULTILEVEL MONITORING SYSTEM
27
-------
30.0 -
50.0-
Upgradlent
I I I I I I I ll
X
So"
I I I I I I I I
100
Concentration (mgJL)
1000
10.000
0.0 -
10.0 -
0 10
Source. Nazar, Prleur, and Threltall, 1984
100
Concentration (mg/L)
10,000
FIGURE 5
VERTICAL DISTRIBUTION OF REPRESENTATIVE INORGANIC
COMPOUNDS FROM UPGRADIENT AND DOWNGRADIENT
MULTILEVEL MONITORING SYSTEM
28
-------
(at a depth of less than 10 feet) to about 10 mg/1 (at a depth of
between 200 and 300 feet). Iron concentrations also exhibit strong
variations ranging from about 20 mg/1 at less than 100 feet,
reaching a maximum of over 5,000 mg/1 at between 100 and 200 feet,
and returning to a concentration below 100 mg/1 at a depth of about
400 feet. Nazar, Prieur, and Threlfall believe that the overall
vertical distributions of the upgradient and downgradient
contaminant concentrations indicate the presence of multiple sources.
Sometimes when specific sources of contamination are unknown
and/or multiple sources are present, a large area, such as an aquifer
can be designated as a site. Such sites can have marked spatial
variation. Both horizontal and vertical variations in ground water
contaminant concentrations were found in a study of contamination in
the Biscayne aquifer in Florida (Singh et al., 1984). As indicated
in Tables 1 and 2, contaminant concentrations varied by up to two
orders of magnitude between locations within the aquifer and between
the three depths sampled. In another aquifer, significant vertical
variation in nitrate and dissolved oxygen concentrations was
observed by Gillham et al., 1983 (Figure 6).
Air. Overall, the atmosphere is the most variable medium of
concern associated with uncontrolled waste sites. Atmospheric
conditions and contaminant concentrations can change significantly
within the space of seconds and within distances measured in feet.
Contaminant concentrations can vary over orders of magnitude, in
29
-------
TABLE 1
HORIZONTAL VARIATION IN GROUND WATER CONTAMINATION
BISCAYNE AQUIFER/DADE COUNTY, FLORIDA
(ug/1)
Area
Hialeah
Upper Miami Springs
Lower Miami Springs
Airport (36th Street)
58th Street Landfill
Unsewered Industrial
Area
Mean
Total VOC
57
33
20
10
6.2
1.1
Mean
Chloride
23
17
8.7
3.5
0.31
0.25
Trans-1, 2-
dichloroethene
28
7.3
3.6
1.1
0.53
0.25
Source: Singh et al., (1984).
30
-------
TABLE 2
VERTICAL VARIATION IN GROUND WATER CONTAMINATION
BISCAYNE AQUIFER/BADE COUNTY, FLORIDA
(ug/1)
Parameter
Vinyl Chloride
Trans-1, 2-Dichloroethene
Total VOCs
Upper
Level
0.35
0.36
7.8
Middle
Level
12
6.7
22
Lower
Level
10
4.3
19
Source: Singh et al., (1984).
31
-------
0.0
CD
TD
c
2.0 -
4.0 -
1 6.0-
m
ZL.
Q.
8.0
10.0
N03 -(mgL-i)
10
20
1
DO (mgL1)
10 20
Water Table
• Location of Sampling points
• Measured Data
Note: The dashed-lines represent the weighted average concentrations
Source: Gillham et al., 1983.
FIGURE 6
VERTICAL NITRATE AND DISSOLVED OXYGEN DISTRIBUTION
IN SHALLOW SANDY AQUIFER (after Hendry et al., 1983)
32
-------
both time and space as the wind direction shifts and plumes meander.
Spatial variation in the atmosphere has been identified as important in
determining concentrations by several authors including Gosse et al. ,
1986; Riggin, 1983; and Zachowski and Borgianini, 1984. Spatial
variability in atmospheric contaminant concentrations arises from
three different sources: variation in the atmosphere itself (e.g.,
wind speed, wind direction, and stability), variation in locations
and release characteristics of contaminant sources, and effects of
terrain. Each of these sources of variation play an important role in
determining the spatial (and temporal) distribution of atmospheric
contaminant concentrations.
The extent of spatial variation in atmospheric contaminant
concentrations is illustrated by sampling data reported in Pellizzari,
1978 (Tables 3 through 5). These data are abstracted from an extensive
set of sample data based on several days of sampling conducted on and
near the Kin-Buc Landfill in New Jersey. Analyses were performed for
a large number of organic contaminants. The data presented in this
report are for benzene, carbon tetrachloride, and chloroform. Temporal
average (mean) benzene concentrations varied by more than a factor of
3 3
2 (13 ug/m to 30 ug/m ) between locations within a one-mile
radius of the site. Individual benzene concentrations varied by over
an order of magnitude off-site. The highest detected on-site
concentration (July 1, 191,000 ng/m ) was nearly 30 times the lowest
detected off-site concentration taken at the same time (6,875 ng/m ) .
33
-------
TABLE 3
SPATIAL/TEMPORAL VARIABILITY IN ATMOSPHERIC
BENZENE CONCENTRATIONS: KIN-BUG LANDFILL (1976)
(ng/m3)
Locations*
Date
6-29
6-30
7-1
Mean
RSD
Time**
1207
1607
1029
1457
1006
1425
Tower
Marina
20,343
7,718
NS
trace
15,969
10,156
13,457
42%
Meadow
Road
93,750
14,093
5,906
NS
NS
6,875
30,156
141%
East
NS
10,656
NS
NS
7,000
27,343
14,500
72%
On-Site
NS
NS
NS
NS
trace
191,000
NA
*Locations: • Tower Marina is approximately 1 mile from the site
at a direction of 255 degrees.
• Meadow Road is approximately 0.25 miles from the
site at a direction of 345 degrees.
• The East location is approximately 0.1 miles from
the site at a direction of 40 degrees.
**Approximate time sampling began at each location.
NA: Not applicable.
NS: Not sampled.
RSD: Relative standard deviation (or coefficient of variation); the
ratio of the sample standard deviation to the sample mean.
Source: Adapted from Pellizzari, 1978.
34
-------
TABLE 4
SPATIAL/TEMPORAL VARIABILITY IN ATMOSPHERIC
CARBON TETRACHLORIDE CONCENTRATIONS: KIN-BUG LANDFILL (1976)
(ng/nr)
Locations*
Date
6-29
6-30
7-1
Mean
RSD
Time**
1207
1607
1029
1457
1006
1425
Tower
Marina
ND
12,687
NS
1,127
ND
3,125
5,646
109%
Meadow
Road
ND
13,687
trace
NS
NS
625
7,156
129%
East
NS
7,250
NS
NS
ND
7,000
7,125
2%
On-Site
NS
NS
NS
NS
ND
10,600
NA
*Locations: • Tower Marina is approximately 1 mile from the site
at a direction of 255 degrees.
• Meadow Road is approximately 0.25 miles from the
site at a direction of 345 degrees.
• The East location is approximately 0.1 miles from
the site at a direction of 40 degrees.
**Approximate time sampling began at each location.
NA: Not applicable.
ND: Not detected.
NS: Not sampled.
RSD: Relative standard deviation (or coefficient of variation); the
ratio of the sample standard deviation to the sample mean.
Source: Adapted from Pellizzari, 1978.
35
-------
TABLE 5
SPATIAL/TEMPORAL VARIABILITY IN ATMOSPHERIC
CHLOROFORM CONCENTRATIONS: KIN-BOC LANDFILL (1976)
(ng/m3)
Locatioas*
Date
6-29
6-30
7-1
Mean
RSD
Time**
1207
1607
1029
1457
1006
1425
Tower
Marina
6,389
1,999
NS
186
17,222
944
5,348
132%
Meadow
Road
ND
trace
12,333
NS
NS
2,500
7,416
94%
East
NS
trace
NS
NS
8,334
260
4,297
133%
On-Site
NS
NS
NS
NS
19,444
27,200
23,322
24%
*Locations: • Tower Marina is approximately 1 mile from the site
at a direction of 255 degrees.
• Meadow Road is approximately 0.25 miles from the
site at a direction of 345 degrees.
• The East location is approximately 0.1 miles from
the site at a direction of 40 degrees.
**Approximate time sampling began at each location.
ND: Not detected.
NS: Not sampled.
RSD: Relative standard deviation (or coefficient of variation); the
ratio of the sample standard deviation to the sample mean.
Source: Adapted from Pellizzari, 1978.
36
-------
Similar variations can be seen in the carbon tetrachloride and
chloroform concentration data. Measured chloroform concentrations
varied 100-fold between the East location and the on-site location
on July 1.
The sources of the variation seen at Kin-Buc are undetermined.
The site is located near a chemical plant and the New Jersey
Turnpike. An analysis of the full data set indicates that both of
these sources contribute to organic contamination near the landfill.
The data also indicate that the landfill is emitting organic
contaminants (e.g., the 191 ug/1 measurement on-site). The
difficulty in determining the contribution of the landfill supports
the observation of Riggin (1983) that the determination of waste
site contributions to polluted atmospheres may require a spatial
sampling resolution of tens or hundreds of meters; finer than was
used in the Kin-Buc study.
The potential for spatial variation in atmospheric
concentrations near waste sites is also evidenced by the data on
landfill gas contaminant concentrations presented in California
Department of Health Services, California Air Resources Board, and
South Coast Air Quality Management District, 1983 (Table 6). As can
be seen in this table, the measured emission concentration varied
greatly at BKK Landfill in California, by as much as two, possibly
three, orders of magnitude. Furthermore, emissions were not always
detected. It is uncertain, however, whether these data represent
37
-------
TABLE 6
RANGES OF LANDFILL GAS CONCENTRATIONS AT BKK LANDFILL
Compound Range (ppa)
Vinyl Chloride 83 - 12,800
Vinylidene Chloride ND - 1,200
Acetylene Bichloride ND - 800
Trichloroethylene ND - 1,000
Tetrachloroethylene KD - 1,500
Ethylene Dichloride ND - 5,000
Benzene 10 - 2,000
Chlorobenzene ND - 500
ND: Not detected.
Source: Adapted from California Department of Health Services,
California Air Resources Board, and South Coast Air Quality
Management District, 1983.
38
-------
spatial or temporal variation, although it is likely that both are
involved since BKK has a large number of gas vents. Regardless of
the source of the variation, vinyl chloride concentrations around
BKK Landfill vary greatly from location to location (see Figure 7).
Soil. Soil contaminant concentrations and other soil
characteristics also exhibit a significant degree of spatial
variation (Bruehl, Chung, and Diesl, 1980; Flatman, 1984; Ford and
Turina, 1985; Gosse et al., 1986; Mason, 1983; and Schweitzer and
Black, 1985). Variation in natural parameters as well as variations
in waste deposition and migration patterns, induce variations in
soil contaminant concentrations.
At an investigation at an unnamed site, Bruehl, Chung, and Diesl
(1980) measured chromium concentrations in soils ranging from 20 to
280,000 mg/kg. This compares with an average site background
concentration of about 100 mg/kg. In contrast, these authors note
that naturally occurring chromium concentrations in soils in the
United States range from 5 to 3,000 mg/kg.
Mason (1983) summarizes the results of several studies of
variability in soil properties (see Table 7). Mason states that
coefficients of variation (the ratio of the sample standard deviation
to the sample mean) in natural soil properties reported in the
literature range from a low of 1 to 2 percent to as high as
850 percent. As indicated in the table, natural soil properties
and phosphorus properties of soils varied over a range from 6 to
39
-------
60
40
20
.a
Q.
0.
B 10
ra
0)
o
c
o
O
0 A
/-^. Mean for
^-' Laboratory 1
A Mean for
Laboratory 2
o
_L
_L
A B C D E p
Monitoring Stations
Source: California Department of Health Services, California Air Resources Board, and south Coast Air
Quality Management District, 1983.
FIGURE 7
RANGE AND MEAN FOR VINYL CHLORIDE MONITORING
NEAR BKK LANDFILL, WEST COVINA, CALIFORNIA
40
-------
TABLE 7
COEFFICIENTS OF VARIATIONS (%) FOR SOIL PARAMETERS
REPORTED IN MASON (1983)
Original Sources
Parameter
Range (%)
White and Hakonsen (1979)
Mathur and Sanderson (1978)
Plutonium
Natural soil
constituents
62 - 840
5.6 - 75.2
Harrison (1979)
Hind in,
May, and Duns tan (1966)
Phosphorus
properties
Insecticide
residue*
11 - 140
156
*Soil square 30 inches on a side.
Original Sources:
Harrison, A. F., "Variation of Four Phosphorus
Properties in Woodland Soils," Soil Biology and
Biochemistry, Vol. 11, 1979, pp. 393-403.
Hindin, E., D. S. May, and G. H. Dunstan,
"Distribution of Insecticide Sprayed by Airplane
on an Irrigated Corn Plot, Organic Pesticides in
the Environment, A. A. Rosen and H. F. Kraybill,
eds.,(Advances in Chemistry Series Number 60),
American Chemical Society, Washington, DC, 1966,
pp. 132-145.
Mathur, S. P- and R. B. Sanderson, "Relationships
Between Copper Contents, Rates of Soil Respiration
and Phosphatase Activities of Some Histosols in
an Area of Southwestern Quebec in the Summer and
Fall, Canadian Journal of Soil Science, Vol. 58,
No. 5, 19/», pp. 125-134.
White, G. C. and I. E. Hakonsen, "Statistical
Considerations and Survey of Plutonium
Concentration Variability in Some Terrestrial
Ecosystem Components," Journal of Environmental
Quality, Vol. 8, No. 2, 1979, pp. 176-182.
41
-------
140 percent. Variation in plutonium concentrations was significantly
higher; coefficients of variation ranged from over 60 percent to
840 percent. Mason also notes that insecticide residue concentrations
have been seen to vary significantly (156 percent coefficient of
variation) within a soil square 30 inches on each side.
Flatman (1984) and Schweitzer and Black (1985) present estimates
of the spatial distributions of lead contamination in soil at the
Dallas Lead Smelter site. These estimates, seen in Figure 8, were
created using the estimation method known as kriging. As can be seen
from the figure, soil lead concentrations vary considerably, ranging
from less than 100 ug/g to over 2,500 ug/g. It is interesting to note
both the non-uniform shape of the contamination isoareas and the
location of zones of less than 100 ug/g within higher contamination
zones.
Surface Water. Depending on the size and other characteristics
(such as degree of turbulence) of a water body as well as the
location and other characteristics of contaminant sources, surface
water contaminant concentrations may vary greatly or nearly
not-at-all. Well-known phenomena that lead to significant spatial
variation in surface water bodies include thermal stratification, in
lakes and impoundments, and for large water bodies, the formation of
contaminant plumes (Gosse et al., 1986).
Table 8 illustrates spatial variation in contaminant
concentrations in streams at two CERCLA NPL sites, Whitehouse Oil
42
-------
12500
10500
8500
S 6500
4500
2500
500
Site: DMC
| | <100
frX'l 100 < 250
K,'j| 250 < 500
P%;| 500 < 1000
\fffj\ 1000 < 2500
till 2500 < 5000
I
I
I
500
2500
4500
6500 8500
X (feet)
10500 12500
Source: Flatman, 1984.
FIGURES
ISOAREA MAP OF KRIGING ESTIMATE OF
LEAD CONCENTRATIONS (ug/g) IN SOIL
43
-------
TABLE 8
EXAMPLES OF SPATIAL VARIATION IN SURFACE WATER
CONTAMINANT CONCENTRATIONS
Location
Upstream
Downstream
Whitehouse Oil Pits (ug/1)
Distance (ft) Iron Magnesium
300
1,000
2,500
3,800
1,200
1,700
1,000
1,200
700
2,200
1,700
1,700
Zinc
20
80
30
70
Location
Upstream
Downstream
Fields Brook (ug/kg)
Distance (ft) Trichloroethene Tetrachloroethene
10
635
1,429
1,694
2,118
2,753
3,547
3,812
6,989
7,011
9,107
9,319
ND
7.5
74
64
41
61
25
NQ
ND
62-65
51
31
72
48
NQ
ND: Not detected.
NQ: Not quantifiable.
Note: Distance is measured upstream or downstream from the probable
point of entry of contaminants from the source to the surface
water body.
Source: Adapted from Burger and Kushner, 1986.
44
-------
Pits, and Fields Brook. At the Whitehouse Oil Pits site, a sizeable
relative increase in magnesium and zinc concentrations was detected
between the upstream and the downstream samples accompanied by a
smaller relative increase in iron concentrations, all within a
distance of 1,300 feet of the probable point of entry of contaminants
from the site. Both magnesium and iron concentrations remained
somewhat stable for almost 3,000 feet downstream while zinc
concentrations fluctuated, declining by over 50 percent and then
more than redoubling within the same distance.
At Fields Brook, concentrations of both trichloroethene and
tetrachloroethene increased significantly between upstream and
downstream sampling points. Concentrations of both contaminants
decline with distance, although both exhibit fluctuations at
different distances from the site. Although the reason for these
fluctuations is unknown, it is probable that they are caused by a
combination of measurement error and/or the presence of additional
contaminant sources.
4.1.2 Wastes
The concentration of hazardous substances in the wastes at a
site varies considerably at different locations at the site. Even
at a given location, waste concentrations may vary with depth. The
reasons for this variation are straightforward. Variation arises
from different patterns in the original deposition of wastes; from
physical, chemical, and biological processes that act on the wastes
45
-------
after deposition; and from differences in the migration patterns of
the waste constituents.
An example of the extent of on-site variation in waste
constituent concentrations is indicated in Table 9. These data were
abstracted from site inspection and remedial investigation reports
as part of the HRS issue analysis of waste constituent concentrations
(Wusterbarth, 1986). The values shown in the table represent the
sum of the concentrations of CERCLA contaminants found in the wastes
at various locations on each site. As can be seen from the data,
there is significant variation in total hazardous substance
concentrations within sites. Relative standard deviations range
from a low of 47 percent in the waste pile(s) at the Micronutrients
International site to a high of 265 percent at the drum portion of
the Morgantown Ordnance site.
The data also support the conclusion drawn by others (e.g., in
Ford and Turina, 1985) that concentrations of hazardous constituents
in drummed waste is highly variable between drums. The relative
standard deviations of the hazardous substance concentrations
measured in drums are consistently above 100 percent while the ranges
are very large (generally over two to about five orders of magnitude)
indicating a high degree of spatial variation in waste contaminant
concentrations.
Ford and Turina (1985) also indicate that waste contaminant
concentrations may vary with depth within a drum, as a result of
46
-------
TABLE 9
SPATIAL VARIABILITY IN WASTE CONCENTRATION
(SUM OF HAZARDOUS CONSTITUENT; ppm)
Site Name
Cosden Chemical
Morgantown Ordnance
Hunterstown Road
Cleve Reber
Old Mill
Cecil Lindsay
Mill Creek
Micronutrients Int.
Texarkana Wood
Preserving
Laskin Poplar
Laskin Poplar
Old Inger
Number of
Type Samples Mean
Drum
Drum
Drum
Drum
Drum
Drum
Drum
Pile
Pond
Sludge
Tank
Sludge
Tank
Liquid
Tank
5
8
28
7
11
3
4
10
5
4
32
9
143,629
110,243
50,843
11,158
136,764
352,391
30,537
201,194
84,605
30,298
3,750
486
Standard
Deviation
266,393
292,660
92,715
14,462
159,130
365,028
46,829
94,641
78,261
45,509
3,193
560
RSD
(%)
185
265
182
130
116
104
153
47
93
150
85
115
Min.
427
202
4
17
61
4,870
469
38,678
494
3,751
28
2
Max.
610,400
834,120
459,060
41,374
451,800
732,712
99,765
339,759
163,100
98,339
12,882
1,431
Source: Wusterbarth, 1986.
-------
stratification. The extent to which the data in Table 9 represent
horizontal versus vertical variation is unknown.
4.2 Temporal Variation
Temporal variation is the second of the two "natural" sources
of variation in concentration data (i.e., variation resulting from
environmental and waste disposal processes as opposed to human error).
Temporal variation is defined as the variation in a parameter of
interest at a given location at different points in time. The effect
of undetected temporal variation on data utility is similar to that of
spatial variation. Together, the two represent the "natural" variation
in estimated contaminant concentrations. Temporal variation is caused
by many of the same factors that induce spatial variations. Generally,
temporal variations in contaminant concentrations arise from temporal
variations in environmental factors such as flow rates and stability.
However, temporal variations in contaminant concentrations may arise
from the interaction of spatially varying and temporally varying
factors. For example, a flood may inundate a landfill located in the
flood plain releasing additional contaminants into a stream. This
phenomenon can be viewed as arising from the interaction of a
temporally varying factor (stream flow) and a spatially varying factor
(waste concentration). Similarly; spatial variations can arise from
the interaction of temporally and spatially varying factors such as
the interplay of atmospheric deposition and solar radiation in
determining soil contaminant concentrations of photodegradable
materials.
48
-------
Temporal variation is a critical factor in determining the
frequency and duration of sampling. As in the case of spatial
variation, those sites with a high degree of temporal variation
generally require a greater sampling frequency and longer sampling
duration to achieve a particular level of representativeness.
Conversely, sites that are temporally homogeneous generally require
lower sampling frequencies and durations. Few sites and media are
temporally homogeneous.
4.2.1 Environmental Data
This section discusses examples from the literature illustrating
the extent and importance of temporal variation in determining
contaminant concentrations in the five environmental media.
Ground Water. Natural temporal variation in ground water
contaminant concentrations at a specific location and the
characteristics that determine those concentrations is considered
to be small (Sgambat and Stedinger, 1981). These and other authors
indicate, however, that in areas affected by human activities,
temporal changes in ground water contaminant concentrations may be
relatively large, although the variation is usually evident only on
a seasonal or longer scale (Everett, 1981; Gillham et al., 1983;
McKown, Schalla, and English, 1984; Nacht, 1983; Possin, 1983;
Schuller, Gibb, and Griffin, 1981; and Sgambat and Stedinger, 1981).
As cited by these and other authors, causes of temporal variation in
ground water quality and other related parameters include:
49
-------
• Variations in contaminant source release characteristics.
• Changes in recharge patterns due to climatic variation.
• Variations in water use patterns.
Everett (1981) also indicates that seasonal variations are likely in
areas characterized by the presence of highly permeable soils and
geologic materials lying above the water table and in areas in which
wells tap the shallow portions of aquifers.
The extent of temporal variation of ground water contaminant
concentrations is illustrated by information contained in Sgambat
and Stedinger, 1981 (Figure 9 and Table 10) and Benson, Glaccum, and
Beam, 1981. The former figure shows nitrate concentrations in a
single well varied by a factor of three over the period of
approximately one year. The ranges and standard deviations of the
concentration of total organic compounds in the four wells listed in
Table 10, further illustrates the extent of temporal variation. In
two of the four wells, the standard deviation exceeded the mean
while in the third it was nearly 80 percent of the mean.
As an additional example, Benson, Glaccum, and Beam (1981)
illustrate the temporal variation in the extent of contamination in
a shallow plume arising from a spill as measured by Em conductivity
(Figure 10).
In contrast, in their investigation of contamination in the
Biscayne Aquifer in Florida, Singh et al. (1984) found little
variation in contaminant concentrations over a period of six months.
50
-------
12
Well 1
12
Well 2
O)
E
0
12
Well 3
III I I I 1 1 1 1 I I I I I I I I I I I I I
III I I I I I I I I I I I I I I I I I I I
0 I I I I I I I I I I 1 I I I I I I I I I I I I
16
12
Well 4
I I I I I I I I I I I I I I I I I I I I I I
JFMAMJJASOND
JFMAMJJASOND
Time, in Months
Source: Sgambat and Stedinger, 1981.
FIGURE 9
FLUCTUATIONS OF NITRATE CONCENTRATIONS
IN WATER FROM INDIVIDUAL SHALLOW WELLS OVER TIME
51
-------
TABLE 10
CONCENTRATIONS* OF TOTAL ORGANIC COMPOUNDS
IN WATER FROM FOUR LONG ISLAND, NEW YORK WELLS
OVER A ONE-YEAR PERIOD (1977-1978)
Well No.
N-8326
N-8327
N-5962
S-27259
Number of
Analyses
25
14
13
9
Mean
93
148
33
19
Standard
Deviation
110
115
17
30
Range
0
4
10
1
- 430
- 455
- 46
- 79
Concentration in ug/1.
Source: Sgambat and Stedinger, 1981.
52
-------
Point of Spill
Regional Ground-Water Flow
I Acre
A) I Day
B) 26 Days
Source: Benson, Glarcson, and Beam, 1981.
FIGURE 10
CHANGE IN SHALLOW PLUME OVER TIME,
AS MEASURED BY Em CONDUCTIVITY
53
-------
Air. As stated previously, the atmosphere is probably the most
variable medium of concern with regard to measurements of contaminant
concentrations and other related characteristics from uncontrolled
waste sites. Contaminant concentrations at a given location can vary
over orders of magnitude within a period of hours. Numerous authors
identify temporal variation as important in determining concentrations
(for example, Gosse et al. , 1986; Riggin, 1983; and Zachowski and
Borgianini, 1984). The same factors that determine spatial
variability, also determine temporal variability: variation in the
atmosphere itself (e.g., wind speed, wind direction, and stability),
variation in locations and release characteristics of contaminant
sources, and effects of terrain.
The extent of temporal variability in atmospheric concentrations
can be seen in the data presented earlier on Kin-Buc Landfill (as
reported in Pellizzari, 1978; see Tables 3 through 5). Concentrations
of benzene, carbon tetrachloride, and chloroform varied substantially
over a three day period, particularly at the Meadow Road location.
The temporal coefficients of variance (or relative standard
deviations) at this location were 141 percent, 129 percent, and
94 percent, respectively. Variation at the East location (0.18 km
from the landfill) was high for both benzene and chloroform
(72 percent and 133 percent, respectively), while variation in
measured carbon tetrachloride concentrations was small (about
2 percent). A particularly high degree of variation was seen at th
54
-------
on-site location. Within a matter of four hours, detected benzene
3
concentrations increased from a trace level to 191,000 ng/m , while
carbon tetrachloride concentrations increased from below the detection
3
limit to 10,600 ng/m . Chloroform concentrations were relatively
3 3
more stable, increasing from 19,444 ng/m to 27,200 ng/m . The
on-site benzene concentration variation clearly illustrates the
problems that temporal variation in concentrations can cause in
assessing site risks. A single measurement taken in the morning
would indicate that the site posed no risk from benzene emissions. A
measurement taken only four hours later would lead to the opposite
conclusion.
Surface Water. It is well known that surface water contaminant
concentrations and other characteristics exhibit a significant degree
of temporal variability, both short-term and longer-term (e.g.,
seasonal). The sources of variation include changes in the locations
and release characteristics of contaminant sources, natural variation
in background contaminant concentrations, and climatic variations in
rainfall and other factors (Gosse et al., 1986 and Sherwani and
Moreau, 1975). According to Gosse et al. (1986), sampling surface
water bodies should be sensitive to a number of temporally varying
factors including:
• Mixing
• Freeze over
• Dry/wet seasons
55
-------
• Flow and elevation
• Micro- and macrobiotic communities
• Runoff patterns
• Diurnal variations in chemistry
• Contaminant sources
• Plant growth patterns
Soil. The extent of temporal variation in soil contaminant
concentrations is problematic. Significant variation is possible due
to chemical and biological transformation and degradation. This
possibility is illustrated by the previously discussed Verona Well
Field ground water example in which a plume of DCE originated from a
source of PCE through chemical transformation (see Figures 1 and 2).
4.2.2 Wastes
Temporal variations in waste contaminant concentrations is also
problematic. Significant variations in waste concentrations over
time are possible due to differences in contaminant transformation,
degradation, and migration patterns. Especially important in this
context is the potential for waste concentrations to decline over
time as materials escape into the environment. A "site" may have
little or no hazardous constituents remaining in it (and thus have
low waste constituent concentrations) but may have created a
substantial plume of contaminated ground water. in such cases,
waste quantity as indicated by the quantity of hazardous constituents
remaining on a site (calculated from waste contaminant concentrations)
56
-------
may be a very poor indicator of the potential magnitude and duration
of releases from the site.
Additionally, contaminants not deposited at the site may be
created at the site through chemical interactions. The previously
discussed Verona Well Field ground water example illustrates this
point. No DCE was deposited at the site and yet DCE can now be found
at the site. This phenomena is also evidenced by the generation of
methane and vinyl chloride in municipal landfills (see, for example,
Lipsky and Jacot, 1985).
4.3 Limitations in the Data Development Process
In addition to the natural factors affecting data quality,
numerous factors of human origin also affect data quality in each of
the four semi-independent stages of the data development process:
• Sampling
• Handling
• Analysis
• Interpretation
The following discussion addresses each of these stages
separately. It is important to note that the problems identified
in data development vary greatly in importance from site to site.
Further, many of the problems are well known and their effects can
be minimized by the use of proper procedures, reinforced by a
comprehensive QA/QC program.
57
-------
4.3.1 Sampling
Several factors interact to adversely affect the quality of data
obtained from environmental and waste samples. So«e of these factors
are environmental in nature while others are related to sampling
techniques and technologies.
4.3.1.1 Environmental Factors. A comprehensive discussion of
the environmental factors affecting contaminant concentrations is
beyond the scope of this report. However, numerous authors (e.g.,
Barcelona et al., 1985; Berg, 1982; Driscoll, 1986; Ford and Turina,
1985; Ford, Turina, and Seeley, 1984; Geraghty and Miller, Inc. and
American Ecology Services, Inc., 1985; Gibb, Schuller, and Griffin,
1981; Gillham et al. , 1983; Gosse et al. , 1986; Kazaann, 1981; Mason,
1983; McKown, Schalla, and English, 1984; Nacht, 1983; Riggin, I983?
Scalf et al., 1981; Schuller, Gibb, and Griffin, 1981; Sisk, 1981;
Tirsch and Male, 1983; U.S. Environmental Protection Agency, 1977;
U.S. Environmental Protection Agency. 1985a and I985b; and Zachowski
and Borgianini, 1984) have emphasized the importance of adequately
characterizing the site-specific environment before a sampling plan
is developed and sampling activities are begun. The aspects of the
environment to be characterized consist of the major components of
the processes that determine concentrations (e.g., media flow rates,
directions, and diffusion characteristics). Knowledge of these
factors is also necessary to assess the representativeness of the
samples once they are taken.
58
-------
The types of background environmental information that should be
collected at a particular site, before sampling, depend on the
physical characteristics of the site, the types of wastes deposited
at the site, and the nature of the threat posed by the site (e.g.,
human health). Several authors have expressed opinions on the
minimum information that they feel is necessary to determine sampling
locations, frequencies, and methods for ground water and air
sampling. These authors indicate that such sampling parameters must
be developed on a site-specific basis. They believe that it is not
possible to develop values for such parameters generically. The
principal guidance documents on ground water sampling (Barcelona
et al., 1985; Claassen, 1982; Gibb, Schuller, and Griffin, 1981;
Gosse et al., 1986; Scalf et al., 1981; U.S. Environmental Protection
Agency, 1977; and U.S. Environmental Protection Agency, 1985a and
1985b, as well as other authors) list the following as important
ground water parameters that should be assessed before sampling:
• Ground water flow gradient
• Ground water flow rate
• Aquifer transmissivity and related parameters (e.g.,
conductivity, porosity, and permeability)
• Extent of fracturing
• Locations of recharge/discharge areas
The feasibility of measuring many of these parameters within the
confines of a PA/SI is open to question (Gerstein, 1986). The use
of other information (e.g., regional ground water flow parameters),
59
-------
while possible, may not provide sufficiently accurate, site-specific
values. This concern is reinforced by the known effects of drinking
water well capture zones on local ground water flow patterns (see,
for example, Quinn, Wittmann, and Lee, 1985 and Absalon and
Starr, 1980). The removal of water from aquifers by pumping
drinking water wells may affect the local hydraulic gradient. The
regional gradient may be determined largely by geologic conditions
and may be unaffected by pumping the wells. The size of the
affected area and the degree of change in the gradient will depend
on local conditions such as the rate of pumping, the number of
wells, and local hydraulic conditions.
A similar set of background information requirements to support
air sampling is indicated by authors such as Riggin (1983) and Gosse
et al. (1986). These authors list the following parameters that
should be assessed prior to sampling:
• Wind speed
• Wind direction
• Temperature
• Pressure
• Relative humidity
• Precipitation patterns
• Locations and characteristics of sources
Ideally, these parameters would be evaluated at the site, although
the data are more frequently available only from the nearest airport.
60
-------
The limitations, however, in using such "regional" information to
infer local conditions has been noted by authors such as laccarino
et al., 1984, particularly in areas of complex terrain.
Thus, it is clear from the literature that a fairly extensive
program of background data collection and sampling is needed prior
to contaminant sampling to ensure that the samples are as
representative as possible within the limitations (e.g., cost and
time constraints) placed on the sampling program. However, such
extensive background data collection is rarely performed in support
of current site inspections. At best, information from secondary
sources, often regional information, is used to infer local
conditions. The lack of actual site specific knowledge of these
important environmental parameters increases the uncertainty in the
data, limiting its usefulness. As a simple final example, without
site specific information on the wind direction during air sampling
(as is the case for some site investigations, one cannot determine
if a sample constitutes a background (i.e., upwind) sample or a site
(downwind) sample.
4.3.1.2 Sampling Techniques and Technologies. In addition to
the environmental factors, several factors of human origin can
adversely affect the quality of concentration data during the
sampling process. Two inter-related concerns arising from the need
to produce representative data with minimal errors, are of particular
importance during sampling. These concerns are: 1) that the
61
-------
samples collected represent to the greatest extent possible, in situ
concentrations at the instant of sampling and 2) that the samples
do not become contaminated during the sampling process. Several
additional concerns common to most media can be identified. These
include the potential for sample contamination from a wide variety
of sources, potential for interaction between the sample and the
environment after the sample has been taken, gaps in sampling
technology (both in terms of contaminants that can be sampled and in
terms of critical operating parameters such as minimum sampling
times), lack of standardization and validation of some existing
techniques, cross-contamination between samples taken with the same
sampler, safety considerations that might preclude the collection of
otherwise needed samples (e.g., in buried drums), and the potential
for needed, but unvalidated, modifications to sampling equipment to
meet site-specific conditions. The following discussion illustrates
these concerns in each of the five environmental media and in wastes.
Ground Water. The literature yielded more information on
problems associated with ground water sampling technologies and
techniques than for any of the other media, reflecting, as indicated
by several authors, the relative difficulties inherent in ground
water sampling in comparison with other media. Gillham et al. (1983)
presents the best overall discussion of the alteration in chemical
concentrations that can be induced in ground water samples. Table 11
lists the sources of sample bias identified in their analysis.
62
-------
TABLE 11
POTENTIAL SOURCES OF BIAS IN SAMPLING GROUND WATER
INORGANIC AND
RADIOACTIVE
PARAMETERS
DIRECT
INDIRECT
INORGANIC
Contamination Parameters
Electrical Cond.
PH
Redox Condition
Non Toxlo Conatltuenta
Chloride
SuIfate
Sodium
Ammonium
Calolum
Magnesium
Iron, Manganese
Toxlo Constituents
Nitrate
Fluoride
Araenlo
Selenium
Barium
Cadmium
Chromium
Lead
Silver
Mercury
RADIOACTIVE
Radium
Gross Alpha and Beta
temperature, suspended particles, precipitation, adsorption / exchange
O>2 degassing, precipitation of carbonates and oxides
oontaot with akin
precipitation of gypsum, reduction
adsorption / exchange, leaohlng from glass, oontact with akin
adsorption / exchange, volatilization at low pH
adsorption / exchange
adsorption / exchange, precipitation
denltrlfloatlon, nitrification
precipitation
adsorption / exchange / leaohlng from sampling equipment,
precipitation
adsorption / exchange, precipitation, degasalng
degassing, changes In Eh, pH
J 0, Invasion, temperature, mloroblal aotl
vlty
02 Invasion, C0? degassing
C0» degassing, 0- Invasion
' Cross-contamination Is a problem for all parameters, but la
Indicated here as a source of bias only If It Is especially
likely to occur.
Source: Gilham et al., 1983.
-------
TABLE 11 (Concluded)
BIOLOGICAL
AND ORGANIC
PARAMETERS
DIRECT
INDIRECT
BIOLOGICAL
Colifora Bacteria
ORGANIC
Drlnklng-Water SKla.
Endrln
Llndana
Methoxychlor
Touphene
2,»-D
2,*,1-1T Sllvex
Quality Paraaatera
Phenols
Contamination Parantera
Total Organic Carbon
Total Organic Halogai
Caaollne Coaponanta
B«nz«n«
Tolu«na
I»l«n«
Mathjrl t-butyl «th*r
croaa-contaalnatlon
aorptlon, orono-oontaalnallon ,
blodagradatlon during atoraga
aorptlon, oroaa-oontaalnatlon ,l«aohlng,
blodegradatlon during atoraga
aorptlon, oroaa-oontaailnatlon , leaching, volatlllr.atlon,
blodagradatlon during atoragA
dlffualon through plaatloo"] aorptlon, oroaa-conta>lnatlon
J laaohlng, volatllnation,
blodagradation during atoraga
0. Invasion, pH, precipitation of
other ooapounde
0 Invasion
• Croaa-oontaalnatlon la a problaai for all paraautara, but la
Indicated here aa a source of blae only IT It le (specially
likely to ooour.
Source: Gilham et al., 1983.
-------
Possibly the most frequently identified problem in ground water
sampling is the problem of sample contamination by stagnant casing
water (see, for example, Berg, 1982; Driscoll, 1986; Gibb, Schuller,
and Griffin, 1981; Gosse et al., 1986; McKown, Schalla, and
English, 1984; Nacht, 1983; Scalf et al., 1981; and Sisk, 1981).
Concentrations in the casing water are not representative of the
concentrations actually in the aquifer. Thus, this water must be
purged from the well before sampling. Generally, most authors
recommended purging between 4 and 10 well volumes from the well to
achieve a level of no more than 5 percent casing water. McKown,
Schalla, and English (1984), however, raise the possibility that
purging of five well volumes may not appreciably decrease the
probability of significant sample contamination. These authors
indicate that greater reliance should be placed on natural purging.
In either case, the resultant observed concentration may not reflect
risk with respect to the actual concentration of pollutants to which
the user is exposed (e.g., in water from a tap).
Additionally, numerous authors have identified problems with
ground water sampling techniques and technologies. Most authors,
such as Barcelona et al. (1985) and Claassen (1982) stress the
importance of not overly disturbing the geomedia during well
construction and sampling so as to ensure that natural conditions
are maintained and in-situ conditions are sampled.
65
-------
Both Driscoll (1986) and Claassen (1982) identified pump and
well residues as important sources of sample contamination.
Driscoll specifically identifies entrained sediments as a source of
contamination. Claassen (1982) identified the problem of
contamination of ground water by chemicals resulting from the
degradation of well materials in old wells. Ihis problem is of
particular interest since current PA/SIs rely heavily on existing
wells in order to reduce sampling costs.
Several other authors have identified well construction and
drilling equipment as potential sources of sample contamination (for
example, Absalon and Starr, 1980; Fetter, 1983; Geraghty and Miller,
Inc. and American Ecology Services, Inc., 1985; Gillham et al., 1983;
and Keith et al., 1983b).
Along similar lines, McKown, Schalla, and English (1984) note
the effect that well diameter has on estimates of transmissivity,
indicating that transmissivity estimates in two inch diameter wells
range over two orders of magnitude.
The duration of pumping has been specifically identified as an
important contributor to sample uncertainty by Driscoll (1986);
Geraghty and Miller, Inc. and American Ecology Services, Inc. (1985);
Nacht (1983); Keith et al. (1983b); and Sgambat and Stedinger (1981).
Figure 11 illustrates the changes that can occur in contaminant
concentration and related parameters during pumping. As indicated
in this figure, contaminant concentrations as well as other
66
-------
350
300
250
25
20
15
10
15
10
EC, ^mho/cm
Cl, mg/L
- N03, mg/L
26.0
25.5
Temp., °C
1
10
100
Time after Start of Pumping, Minutes
Source: Keith et a/., 1983b.
FIGURE 11
WATER QUALITY FLUCTUATIONS WITH TIME OF PUMPING
67
-------
indicators of the stability of the ground water environment may vary
considerably over time as a result of pumping. Claassen (1982) also
identified the importance of pump operation in adversely affecting
sample quality.
Time for withdrawal of samples from a well and the possibility
of changes in the water samples during withdrawal have been identified
as important sources of uncertainty (Driscoll, 1986; Gillham et al.,
1983; Geraghty and Miller, Inc. and American Ecology Services, Inc.,
1985; Ford and Turina, 1985; and Scalf et al., 1981). The concern is
that the water environment of the sample will change even during the
short time for withdrawal, altering the contaminant concentrations.
Particular phenomena identified include degassing and volatilization
of volatile compounds (Gillham et al., 1983), introduction of oxygen
and possible changes in the oxidizing/reducing nature of the sample
(Gillham et al., 1983), changes in the carbon dioxide concentration
in the water (Geraghty and Miller, Inc. and American Ecology Services,
Inc., 1985), changes in biological activity (Geraghty and Miller, Inc.
and American Ecology Services, Inc., 1985), and changes in sample pH
and temperature (Driscoll, 1986; Gillham et al., 1983; Ford and
Turina, 1985; and Scalf et al., 1981). As noted by Ford and Turina
(1985), the rates of many chemical reactions (including gas exchange,
microbial growth) double with each increase of 10 degrees Centigrade
in temperature. They also note that the presence of sunlight and
oxygen may produce near instantaneous changes in sample chemistry.
68
-------
Gillham et al. (1983) also identify the potential for sample
contamination arising from sampler and well materials (e.g.,
bentonites and plastics). Geraghty and Miller, Inc. and American
Ecology Services, Inc. (1985) also identify the possibility of
contamination by the sampling apparatus.
The problem of cross contamination arising from using the same
drilling or sampling equipment in two different locations was noted
by Absalon and Starr (1980); deVera et al. (1980); and Seanor and
Brannaka (1981).
Gibb, Schuller, and Griffin (1981) indicate that the type of
pump used can have an adverse effect on sample quality. The use of
air or nitrogen pumps appears to affect sample concentrations while
peristaltic pumps or bailing do not. This conclusion is supported by
Seanor and Brannaka (1981) who indicate that air squeeze, piston,
jet, and submersible pumps are those most likely to cause
contamination problems.
The extent of these problems was investigated under laboratory
conditions by Ho (1983). Ho investigated the effects that transport
line materials, pump rate, lift (travel distance in the sampler), and
initial contaminant concentration had on final sample concentration.
He determined that all were important factors in determining sample
concentration although to differing degrees for different
contaminants. Overall, he found that line material was the most
consistently important factor, pump rate was important for volatiles,
69
-------
and initial contaminant concentration was important for the more
volatile compounds. Lift was found to be important for compounds
_2 3
with Henry's constants greater than 10 atm-m /mol.
Air. The most common problems associated with techniques and
technologies for air sampling cited in the literature, disregarding
those directly related to the extreme spatial and temporal
variability of the atmosphere and its effects on contaminant
concentrations, are gaps in the availability of sampling equipment
(Ford, Turina, and Seeley, 1984; Gosse et al., 1986; Harrison,
undated; Riggin, 1983; and Zachowski and Borgianini, 1984).
This problem is primarily apparent in the lack of commonly
accepted methods for the sampling of many organic vapors, gases, and
particles. Ford, Turina, and Seeley (1984) note that currently
available portable equipment is unable to detect certain compounds of
possible interest in site investigations. This problem is primarily
related to limitations of sorbents such as Tenax and to limitations
in field gas chromatography. Gosse et al. (1986) present
a list of methods available for sampling organic vapors and gases
(Table 12). It is interesting to note the relatively small number of
compounds listed and the lack of methods for maleic anhydride, methyl
acetate, and phthalic anhydride in this short list (all CERCLA
hazardous substances). Riggin (1983) also notes that no standard
methods are available for many toxic air pollutants at low ambient
concentrations. This is addressed by Harrison (undated) who
70
-------
TABLE 12
SAMPLING METHODS FOR AIRBORNE ORGANICS
Hazardous Constituent
Tenax Charcoal Cryogenic Hl-Vol DNPH
Cartridge Cartridge Trapping PUF Impinger
Other
Acetaldehyde
Acrolein
Acrylonitrile
Ally! chloride
Benzene
Benzyl chloride
Carbon tetrachloride
Chlorobenzene
Chloroform
Chloroprene
Cresol
Cumene
1,4-DCB
1,2-DCB
Dichlorome thane
Dioxin
Epichlorohydrin
Ethylbenzene
Ethylene oxide
Formaldehyde
X
X
X X
X X
XXX
X X
XXX
X X
X X
X X
X
X X
X X
X X
X X
Aarvan, 1981
X
X X
X X
X
Source: Gosse et al., 1986.
-------
TABLE 12 (Concluded)
TABLE 8-13 (continued)
Hazardous Constituent
Tenax Charcoal Cryogenic Hl-Vol DNPH
Cartridge Cartridge Trapping PDF Impinger
Other
Hexachlorobutadlene
Hexachlorocyclopentadiene
Hydrogen cyanide
Maleic anhydride
Methyl acetate
N-Dlmethylnitrosamine
Naphthalene
Nitrobenzene X
Phenol X
Phosgene
Phthalic anhydride
PCBs
Propylene oxide
1,1,2,2-tetrachloroethane X
Tetrachloroethylene X
Toluene X
1,1,1-Trichloroethane X
Trichloroethylene X
Vinyl chloride
Vinylidene chloride
Xyleae X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
No method identified
No method identified
ASTM STP 721, 198 p. 80-91
NIOSH Method 3502
Ruggle, et al., 1979
No method identified
Source: Gosse et al., 1986.
-------
states that the analysis of low molecular weight hydrocarbons at low
concentrations requires a difficult chromatographic separation.
Although the exact reasons for the problems encountered are not
discernible, Weber and Mims (1981) rejected the walk-through survey
method for detecting organic compound leaks because of severe
problems with the reproducibility of organic compound measurements
using portable devices. To the extent that walk-through surveys of
chemical plants are similar to air contaminant surveys of waste sites
with portable equipment, this study raises questions about the use
of portable equipment surveys in site inspections. Ford, Turina,
and Seeley (1984) also note problems of insensitivity and slow
response time in organic samplers.
The problem of equipment gaps affects more than simply the
samplers. Zachowski and Borgianini (1984) found that no suitable
sampling pumps were available to support their investigations of
uncontrolled waste sites in New Jersey. The authors had to fabricate
their own pump in order to calculate daily exposures using a 24-hour
sampling period and relatively low flow velocities. This problem is
also noted by Ford, Turina, and Seeley (1984).
In comparison, the state of the art in particle sampling is
fairly well advanced (see, for example, Farthing, 1982; Harrison,
undated; and Stevens, 1984). The outstanding problems in these
areas are concerned with the details of sampler design and with the
extraction of organics from filter media.
73
-------
Overall, these problems indicate the possibility that, given the
variability of air concentrations and the wide range of contaminants
found at uncontrolled waste sites, concentrations of concern for some
contaminants may not be detected during site inspections.
Wastes. Four problems were identified with techniques and
technologies for sampling wastes. The primary problem associated
with waste sampling is maintaining sampling personnel safety (see,
for example, Wetzel, Wagner, and Tafuri, 1982). This problem also
occurs in sampling any environmental media. It is a more significant
problem in sampling waste, however. As noted by Ford, Turina, and
Seeley (1984), care must be taken to minimize disturbance to the
waste matrix during sampling. Some of the reasons for this are the
same as for ground water, as discussed above. Unlike ground water,
however, sufficient disturbance of the wrong wastes may cause them to
explode or otherwise react with serious consequences. For these
reasons, for example, it is considered unwise to sample the interiors
of landfills unless waste locations and waste characteristics are
known.
Associated with the need to minimize disturbance for safety
reasons and to ensure collection of a representative sample is the
need to avoid increasing the mobility of wastes that are not
currently mobile. For example, drilling into a landfill not only may
pose a safety risk, but it may also cause an increase in the rate of
release of gaseous contaminants from the landfill. Further, drilling
74
-------
through waatea into ground water may increase the rate of waste
migration to ground water.
Accessibility is a third problem in waste sampling which is also
common to sampling any environmental media. It may be extremely
difficult or hazardous to sample wastes in certain locations, e.g., from
the middle of a large surface impoundment or from buried containers.
This difficulty has led to the development of specific technologies
such as dippers with telescoping handles (Gosse et al., 1986).
Finally, according to Gosse et al. (1986), different sampling
techniques are needed to sample wastes at different sites depending on
the type of waste and waste management unit (see Table 13). Waste
samplers have been developed to address most types of wastes in most
types of waste management units. However, Gosse et al. note that
equipment for the following situations must be selected or designed on
a site-specific basis: sludges, sand or powdered granules from waste
piles, landfills, land treatment facilities, surface impoundments
(including ponds and lagoons), and pits as well as sand or powdered
granules from tanks and other containers.
Other Media. The problems associated with the techniques and
technologies of sampling soils, surface water, and sediments are
similar to the problems already discussed.
4.3.2 Handling
Generally, once a sample has been taken, it must be transported
to an analysis location. This location may be off-site (such as an
75
-------
TABLE 13
SAMPLING EQUIPMENT FOR PARTICULAR WASTE TYPES
Source Unit Type
Storage Waste piles. Ponds,
tanks landfills, lagoons.
Waste Type Drua or bins & Land treataent & pits
Free flowing COLIWASA* Weighted N/A Dipper
liquids and glass tube bottle
slurries
Sludges Trier Trier a «
Moist powders Trier Trier Trier, Trier,
cr granules scoop scoop
Irv powder j Thief Thief Thief Thief
or granules
Sand or Auger a a a
packed powders
and granulea
Large Large Large Large Large
grained Trier Trier Trier, Trier,
solids scoop scoop
Other Waste Saaple Sources
Sacks Open bed Closed Convery
and baga truck bed truck belt Pipe
N/A N/A COLIWASA* N/A Dipper
glass tube
N/A Trier Trier Shovel, Dipper
scoop
Trier Trier Trier Shovel, Dipper
acoop
Thief Thief Thief Shovel, Dipper
scoop
Auger Auger Auger Dipper dipper
Large Large Large Trier Dipper
Trier Trier Trier
•Adapted frosi U.S. EPA, 1982.
This situation can present significant logistical sampling problems) equlpMnt Bust be specifically selected or designed
based on sit* and wast* conditions.
N/A - Hot applicable.
Source: Gosse et al., 1986.
-------
EPA Contract Laboratory) or on-site (as in a mobile laboratory).
Alternately, there may be no appreciable handling component as in the
case of automatic sampling and analysis equipment such as mobile gas
chromatography/mass spectroscopy (GC/MS) units. In any case, the
problems encountered in handling samples, though important, are not
as numerous as those of sampling and analysis.
There is one pervasive problem associated with sample handling.
This is the need to minimize changes in the sample between the time
of sampling and the time of analysis. This problem is noted by
numerous authors including Berg, 1982; Geraghty and Miller, Inc. and
American Ecology Services, Inc., 1985; Gillham et al., 1983;
Kirchmer, 1983; Plumb, 1984; Riggin, 1983; and Schuller, Gibb, and
Griffin, 1981.
There are two semi-independent aspects to the problem. First,
the integrity of the sample must be maintained and sample
contamination avoided. This usually means that the sample must be
sealed and handled in a sealed container. This is required to
minimize the interaction between the sample and the atmosphere and to
prevent the accidental introduction of foreign substances into the
sample. Sealing prevents, for example, outgassing of volatiles from
liquid and sorbent samples and changes in sample chemistry that may
induce changes in contaminant concentrations. These problems are
similar to those discussed previously concerning the interaction of
samples with the non-in-situ environment.
77
-------
The second aspect of the problem is that of sample stability. A
minimum of "natural" changes in sample chemistry should be allowed to
occur between the time of sampling and time of analysis. Naturally
occurring reactions (those that were in progress in-situ) that would
otherwise affect contaminant concentrations should be halted. This
commonly requires that the sample be either frozen or preserved with a
preservative such as hydrochloric acid.
These two aspects of the handling problem also interact to create
the need to minimize the interaction between the sample and its
container. This problem is particularly wellknown in regard to liquid
samples where care must be taken to prevent loss of metals through ion
exchange with the container material (Fitzgerald, 1986).
An additional problem, noted by Geraghty and Miller, Inc. and
American Ecology Services, Inc. (1985), is the potential contamination
that may occur when an aliquot of a sample is withdrawn from a sample
for analysis.
4.3.3 Analysis
Once a sample has been taken and transported (if necessary) to a
laboratory, the process of analyzing that sample begins. As mentioned
above, the analysis may be performed off-site, on-site, or as an
integral part of sampling. Though it presents more problems than that
of handling, the analysis component of data development is considerably
more amenable to control than the sampling component. The analysis
component occurs generally in a laboratory where conditions are, in
principle, controllable.
78
-------
One of the most critical problems with sample analysis, one noted
by many authors (including Geraghty and Miller, Inc. and American
Ecology Services, Inc., 1985; Gurka et al., 1982; Keith et al., 1983a;
Kirchmer, 1983; Plumb, 1984; and Riggin, 1983), is the lack of
consistent, acceptable analytical methods for many contaminants of
concern at the concentrations encountered at waste sites. Table 14
presents a list of over 90 RCRA hazardous constituents for which there
are analytical problems (Gosse et al., 1986). These constituents
represent almost 25 percent of the RCRA hazardous constituents (all of
which are CERCLA hazardous substances). EPA has recently proposed
changes in required ground water analyses for RCRA facilities, reducing
the list of RCRA contaminants to be analyzed in part because of
analytical constraints (51 FR 26632, 24 July 1986). In addition,
Fitzgerald (1986) indicates that a significant number of CERCLA
contaminants cannot be analyzed by the EPA Contract Laboratories at
this time. Overall, these gaps in analytical capabilities lead to
possibly undetected gaps in the coverage of contaminant compounds at
sites.
In support of this contention, Keith et al. (1983a) note the
acceptable ranges for accuracy and precision employed in the EPA
Contract Laboratory Program (see Table 15). The acceptable accuracy
of 200 percent indicates that the estimated concentration of a
contaminant may be consistently twice as high as the true value and
still be acceptable.
79
-------
TABLE 14
RCRA APPENDIX VIII COMPOUNDS FOR WHICH NO SUITABLE
ANALYTICAL PARAMETER METHODOLOGY OR STANDARDS EXIST,
OR FOR WHICH A SURROGATE HAS BEEN SUGGESTED AS AN ANALYTE
l-acetyl-2-thiourea
acrylamide
aflatoxins
aldicarb
5-(aminomethyl)-3-isoxazolol
amitrole
auramine
azaserine
benz(c)acridine
benzenearsonic acid
benzo(j)fluoranthene
brotnoacetone
brucine
chlorambucil
chlornaphazine
citrus red no. 2
cycasin
2-cyclohexyl-A,6-dinitrophenol
cyclophospharaide
daunomycin
diallate
dibenz(a,h)acridine
dibenz(a,j)acridine
7H-dibenzo(c,g)carbazole
dichlorophenylarsine
1,2:3,4-diepoxybutane
diethylarsine
diethylestilbesterol
dihydrosafrole
3,4—dihydroxy- -(methylamino)
methyl benzyl alcohol
2,4-dithiobiuret
3,3-dimethyl-l-(methyl thio)-
2-butanone, 0-[(methyl amino)
carbonyl] oxime
endothal
ethyl carbamate
ethylenebis dithiocarbamic acid,
salts and esters
ethylenethiourea
2-fluoroacetamide
fluoroacetic acid, sodium salt
formic acid
hydrazine
hydroxydimethylarsine oxide
iron dextran
lasiocarpine
lead subacetate
maleic hydrazine
raelphalan
mercury fulminate
methanethiol
raethomyl
2-methylaziridine
methylthiouracil
mitomycin C
n-methyl-n'-nitri-n-nitrosoguanidine
n-napthyl-2-thiourea
N,N-diethyIhydrazine
n-nitrodiethanolamine
4-nitroquinoline-l-oxide
n-nitroso-n-ethylurea
n-nitroso-n-methylurea
n-nitroso-n-methylurethane
n—nitrosoarcosine
n-nitrosomethylvinylamine
n—nitrosonornicotine
n-phenylthiourea
n-popylamine
nickel carbonyl
nicotine and salts
nitrogen mustard-N-oxide and
hydrochloride salt
nitroglycerine
l-(o—chlorophenyl) thiorea
o,o,o-triethyl phosphorothionate
o,o-diethyl S-methyl ester of
phosphorodithioic acid
octamethylpyrophosphoramide
phenyl mercury acetate
phenylenediamine
phosphine
propylthiouracil
reserpine
saccharin and salts
selenourea
streptozotocin
strychnine and salts
sym-trinitrobenzene
tetraethyl lead
thiosemicarbazide
thiourea
thiuram
toluenediamine
0-toluidine hydrochloride
tris(1-aziridinyl) phosphine sulfide
trypan blue
uracil mustard
warfarin and salts
Source: Gosse et al., 1986
80
-------
TABLE 15
ACCEPTABLE RANGES FOR PRECISION AND
ACCURACY IN THE EPA CONTRACT LABORATORY PROGRAM
Class of Compound
Volatiles
Acids
Base/Neutrals
Precision (%)
15
40
50
Accuracy (%)
40 -
50 -
15 -
200
200
200
Source: Keith et al., 1983a.
81
-------
Gurka et al. (1982) summarize the results of an early (pre-1983)
performance evaluation performed on the CLP EPA Contract Laboratory
Program (CLP) (see Tables 16 and 17). For example, the authors note
that at that time several laboratories were unable to detect
benzidine and endosulfan II. The results for chloroform, a common
waste site contaminant, regarded from a low of 14 ug/1 to a high of
303 ug/1, in comparison with a true value of 120 ug/1. All but one
laboratory estimated the concentration as below 70 ug/1. Further,
for endrin there was about a five order of magnitude difference
between the highest and lowest measurement among the laboratories.
These data also indicate a wide variation between laboratories. A
similar evaluation is not available for the CLP program at the
present time.
Isaacson, Eckel, and Fisk (1985) list three factors most likely
to account for the low occurrence of many compounds found at waste
sites: non-optimum analytical methodologies, degradation of
compounds
prior to analysis, and incorporation into sample matrices. Low
occurrence compounds are those that were either: 1) undetected
in the CLP Analytical Data Base of the EPA National Enforcement
Investigation Center Data Base or 2) detected at less than one-tenth
the average frequency of compound occurrence in each data base.
The problem of inter-laboratory differences is also a common
problem outside of the CLP. As indicated in Tables 18 and 19,
analytical results can vary markedly between laboratories.
82
-------
TABLE 16
CLP PERFORMANCE EVALUATION SAMPLE RESULTS, 1982
Contractor L»boratorle«
Conpound
Chlorobenzene
1.1,2, 2,-Tetrachloro-
ethane
Hethylene chloride
1,1, 2-Trlchloro-
ethane
Chlorofom
1 , 4-Dlchlorobenzene
Napthalene
Acenapthalene
Isophorone
llexachlorocyc lo-
pentadlene
ftenzldlne
Dlbenz(a,h)anthracene
N-Nltroao-dlphenyl
anlne
2-Chloronapthalene
4-Nt tropheaol
PentacKlorophenol
Phenol
1-BHC
p.p'-DDD
Endoaulfan II
Endrln
1
46
NDb
20
10
14
160
100
40
100
380
ND
ISO
330
50
375
210
170
ND
140
ND
ND
95
59
99
150
69
ND
41
60
3200
280
ND
ND
1300
220
160
220
98
130
213
NO
180
57
110
31
160
42
100
200
100
590
710
42
120
830
170
279
190
82
120
150
300
220
32
63
ND
100
37
140
140
70
660
200
160
80
580
100
360
230
no
100
150
170
300
41
10
65
121
43
ND
133
ND
ND
1105
ND
29
926
130
246
429
80
148
ND
224
0.06
51
94
54
130
36
160
170
96
370
ND
10
62
520
130
ND
520
69
130
150
50
180
60
107
58
152
31
116
98
86
ND
747
ND
10
1016
132
443
285
248
ND
218
ND
222
64
18
88
148
61
214
223
109
ND
1818
ND
ND
894
I7S
ISO
304
II)
61
136
150
159
39
ND
35
148
303
ND
2710
1220
ND
4600
ND
120
13400
1800
2050
1840
1330
SI2
167
ND
212
10
29
76
16
120
39
140
130
75
290
530
ND
160
610
100
285
280
130
ND
200
ND
ND
II
38
63
69
100
37
96
131
70
222
ND
ND
95
493
113
168
238
6
140
130
160
160
12
42
90
55
112
40
285
285
169
ND
ND
ND
ND
1118
220
754
1080
160
99
197
197
217
A
40
68
68
85
50
130
150
84
280
360
20
50
510
120
180
360
120
160
240
20
240
EPA 1
B
39
73
64
115
51
121
129
83
280
441
ND
76
396
103
344
454
104
113
119
190
189
C
^J
99
17
94
40
120
130
100
200
440
33
190
10
130
190
360
75
130
290
ND
220
D
a
a
a
a
a
186
180
143
328
461
5
237
722
123
144
462
87
S
231
5
345
e
42
90
II
103
40
152
168
106
220
190
8
345
780
157
262
326
82
117
174
23
340
True
Value
80
100
160
150
120
260
200
95
300
500
400
175
600
150
400
350
200
150
200
175
250
* Laboratory ~n- did not analyze the aanple for volatile organic*.
b ND - Not detected.
Source: Gurka et al., 1982.
83
-------
TABLE 17
ANALYSIS OF CLP PERFORMANCE EVALUATION SAMPLE DATA, 1982
Laboratory No.
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Compound
Chlorobenzene
1,1,2.2-Tetra-
chloroethane
Methylene chloride
1.1,2-Trlchloro-
e thane
Chloroform.
1 ,4-Dlchlorobenzene
Napthalene
Acenapthalene
Hexachlorocyc lo-
pentadiene
Benzldlne
Dibenzanthracene
N-Nitroso-
dlphenylamine
2-Chloronapthalene
4-Nttrophenol
Pentachlorophenol
Phenol
beta-BHC
p,p'-DDD
Endosulfan II
Endrln
1
46
ND
20
10
41
160
100
40
380
ND
180
350
50
375
210
170
ND
140
ND
ND
3
57
110
31
160
42
180
200
110
710
42
120
830
170
270
190
82
120
150
300
200
6
51
94
54
130
36
160
170
58
ND
10
62
520
130
ND
520
69
130
150
50
180
9
39.5
ND
35.1
148
30.3
ND
2710
1220
4600
ND
120
13400
1800
2050
1840
1330
512
167
ND
212
True
Value
80
100
160
150
120
260
200
95
500
400
175
600
150
400
360
200
150
200
175
250
Mean3
49.5
75
54.2
116.0
44.3
165
160
96
617
78.3
109
851
143
294
368
122
132
186
179
219
Standard Performance Perfornance
Deviation Window Window
(S.D.) 0.5T 3T Mean +2 S.D.
17.4
29.6
26.3
42
10.7
53.4
69.5
35.8
450
70.9
74.4
309.0
48
169
237
48
33
50
84
87
40
50
80
75
60
130
100 -
47.5
250
200
87.5
300
75
200
175
100
75
100
87.5
125 -
240
300
480
450
360
780
600
285
1500
1200
525
1800
450
1200
1050
600
450
600
525
750
14.7
15.8
1.6
32.3
22.9
58.4
20.2
26.4
0
0
0
249
47
0
0
26
46
86
11
45
84.3
134.0
- 107.0
- 200.0
- 66.7
- 272.0
- 300.0
170.0
1517
- 220
258
1468
239
632
- 842
218
178
286
247
393
The Hated mean values for compounds No. 1 to 5 are baaed on 16 laboratories. The mean
values for compounds No. 6 to 13 are based on 16 laboratories and do not Include values
reported by laboratory No. 9. Mean values for compounds No. 14 to 20 are based on 17
laboratories and do not Include values reported by laboratory No. 9.
Source: Gurka et al., 1982.
84
-------
TABLE 18
SUMMARY OF INTER-LABORATORY COMPARISON
Compound
Trichloroethylene
Tetrachloroethylene
Carbon Tetrachloride
1,1,1 trichloroethane
1,2 dichloroethane
Vinyl Chloride
Benzene
1,4-dichlorobenzene (as VOC)
1,1 dichloroethylene
Results
(20 ug/1)
23 +
22 +
20 +
22 +
21 +
20 +
21 +
23 +
23 +
3.0
2.6
2.6
3.0
3.5
11.3
3.5
3.9
6.4
Variation (%)
13
12
13
14
17
57
17
17
28
Source: Adapted from Geraghty and Miller, Inc. and American Ecology
Services, Inc., 1985.
85
-------
TABLE 19
SUMMARY OF INTER-IABORATORY COMPARISON
Compound
2-Chlorophenol
1, 4-dlchlorobenzene
(as extractable)
Dimethyl Phthalate
Heptachlor
PCB-1260
Results
(200 ug/1)
156 + 47
144 + 56
13 + 5.6*
41 + 159
171 + 73
136 + 58
Variation (%)
30
39
43
388
43
43
*Results for 20 ug/1.
Source: Adapted from Geraghty and Miller, Inc. and American Ecology
Services, Inc., 1985.
86
-------
An additional problem is that different equipment may yield
different results when analyzing the same samples. Geraghty and
Miller, Inc. and American Ecology Services, Inc. (1985) present the
data listed in Table 20, comparing the results of an analysis of a
soil sample using gas chromatography (GC) and GC coupled with mass
spectroscopy (MS). The results show that there can be an order of
magnitude difference in the analysis of the same sample using
different equipment.
Overall laboratory variability in ground water contaminant
analysis is summarized in Table 21. As indicated in this table, good
precision and accuracy can be achieved in laboratories although the
laboratory variability for TCE (one of the most common waste site
contaminants) is large.
Riggin (1983) notes that one problem with many techniques used
to analyze air filter and other samples is that the techniques are
destructive in nature. This is important since it implies that
samples must be collected in sufficient quantity to allow the sample
to be split whenever more than one analysis is to be undertaken using
a destructive technique. The list of destructive techniques includes
gas chromatography/mass spectroscopy, the principal analytical method
used to analyze waste samples for organic contaminants.
Numerous other problems are noted by various authors. Kirchmer
(1983) notes that there are numerous sources of bias and other
problems in laboratory analyses. These include:
87
-------
TABLE 20
COMPARISON OF GAS CHROMATOGRAPHY (GC) AND
GAS CHROMATOGRAPHY/MASS SPECTROSCOPY (GC/MS) RESULTS
(SOIL SAMPLE - ug/kg)
Compound
1, 1-Dichloro ethane
t-1 , 2-Dichloroethane
1,1, 1-Trichloroethane
Tetrachloroethane
GC
32
64
240
27
GC/MS
310
1,900
1,200
140
Source: Adapted from Geraghty and Miller, Inc. and American Ecology
Services, Inc., 1985.
88
-------
TABLE 21
MAGNITUDE OF LABORATORY VARIABILITY
Component
Precision (%)
Accuracy (%)
Indicators
pH, Special Conductivity
TOG, TOX
TON, TOG, COD, IDS
Organics
Inorganics
Nonmetals
Metals
ICAP
Trichloroethylene
10
5-10
10-50
19-48
5-20
4-6
10-25
50 to 1,000
10
5-20
up to 50
up to 50
0-11
2-7
10
50 to 1,000
Source: Geraghty and Miller, Inc. and American Ecology Services,
Inc., 1985.
89
-------
• Sample instability
• Interferences
• Biased calibrations
• Biased blank correction
Plumb (1984) identifies the following as being of particular
importance; impurities in purge gases, potential for compound
outgassing, inter-sample contaminant carry-over, and other
interferences. McKown, Schalla, and English (1984) also discuss
the problem of interferences and contamination, noting that in one
occurrence, the benzene concentration in distilled water to be used
in field blanks was 33 ug/1. These authors also identify the problem
of possible absorption of contaminants during filtration as a problem
of concern.
4.3.4 Interpretation
As indicated by Schweitzer (1981), interpretation of hazardous
waste site concentration data is more of an art than a science. It
is a characteristic of such data that different persons can come to
different conclusions upon viewing the same data. This phenomena
arises for several reasons. First, the variability in the environment
and the potential for error discussed previously make interpretation
difficult. Second, it is frequently difficult to differentiate
background from site-related contaminant concentration, particularly
in a highly varying medium such as the atmosphere. As an example
consider the ranges of background contaminant concentrations in urban
90
-------
air reported in Riggin (1983) (Table 22). Detecting site emissions
in an urban environment in the presence of other sources is a complex
task. The applicability of advanced techniques, such as receptor
modeling (see, for example, Core, 1981), of such cases awaits
consideration. Similarly, the lack of standards for many contaminants
(as noted by Zachowskl and Borgianini (1984) and others) makes the
interpretation of the risk from contaminant concentrations difficult.
Third, at many sites data are lacking on the local direction of flow
of the media during and preceding times of sampling. Lacking such
data, the designation of a sample as "background" and another as a
"site sample" is subjective at best. Fourth, there are no uniformly
acceptable methods for analyzing concentration data. Many statistical
methods suffer from reliance on unrealistic assumptions, such as
normality, and are thus not generally applicable. Recently developed
methods, such as variography and kriging (e.g., Moore and Mclaughlin,
1980 and Flatman, 1984), have also been criticized (e.g., Journel,
1984).
Finally, it is unreasonable to expect that a single set of
techniques for interpreting concentration data can be developed.
There is too much variation between sites, too many gaps in
background information, and too many deficiencies in interpretive
techniques to allow development of a uniformly applicable collection
of such techniques.
91
-------
TABLE 22
RANGES OF SELECTED ORGANIC CONTAMINANT
CONCENTRATIONS IN URBAN AIR
(ppb)
Contaminant Concentration Range
Methylene Chloride 0.049 - 9.4
Chloroform 0.019 - 5.1
Carbon Tetrachloride 0.11 - 2.9
Trichloroethylene 0.005 - 2.5
Benzene 0.11 - 37.0
Formaldehyde 6.6 - 41.0
Source: Riggin, 1983.
92
-------
5.0 COSTS
This chapter presents a limited discussion of the costs
associated with developing concentration data at uncontrolled waste
sites. A discussion of the general problems with estimating data
development costs is presented followed by discussions of the
principal factors that determine site-specific costs. The chapter
concludes with an illustrative example of the magnitude of sampling
costs.
It would be desirable to estimate the resources, both in terms
of time and money, necessary to collect representative concentration
data, to the limits provided by current technology and knowledge,
within the context of a preliminary assessment (PA) and site
inspection (SI). Regrettably for reasons discussed below, this
cannot be done generically. This chapter presents what information
is available in order to provide an indication of a lower bound on
the costs of developing representative data. Based on this
information, a considerable increase in resources would be needed to
collect representative concentration data at all waste sites for all
media. Even a ten-fold increase in the current average per-site SI
expenditures may not be sufficient.
5.1 Problems in Estimating Costs
The first problem in estimating costs for developing
representative concentration data is the problem of specifying the
necessary degree of representativeness. As indicated in Appendix B,
93
-------
this is a significant problem since representativeness is poorly
defined and can only be assessed subjectively. Moreover, the
necessary degree of representativeness depends on the objectives of
the program the data are to support and the uses to which the data
will be put as part of the program. Since the exact objectives of
the HRS and the specific role that concentration data are to play
within the HRS are still being reviewed by EPA, no clear-cut,
unequivocal set of requirements can be stated.
For the purposes of this discussion, therefore, the following
subjective statement of the representativeness requirements for
concentration data will be used: the data will be of sufficient
quality to support inferences (either subjectively or through models)
of the degree of risk posed by current and near-term releases of
contaminants from the sites and resulting exposures. In order to
meet this requirement, three-dimensional profiles of contaminant
concentrations will be required in both environmental media (i.e.,
ground water, surface water, air, soils, and sediments) and wastes.
The sampling, handling, and analysis process must be performed
according to a comprehensive QA/QC procedure. All sample analyses
used in the HRS must achieve the standards for accuracy and
precision established for the EPA Contract Laboratory Program.
Finally, estimates must be developed for all contaminants of concern,
to the extent possible given limitations on the state of sampling
and analysis technology.
94
-------
Second, as indicated by numerous authors (including Barcelona
et al., 1985; Berg, 1982; Claassen, 1982; deVera et al., 1980; Ford,
Turina, and Seeley, 1984; Ford and Turina, 1985; Gibb, Schuller, and
Griffin, 1981; Gosse et al., 1986; Hanisch and McDevitt, 1984;
JRB Associates, 1985; iMason, 1983; Nacht, 1983; Plumb, 1984;
Popkin, 1983; Porcella, 1983; Riggin, 1983; Scalf et al., 1981;
Sisk, 1981; Turpin, 1983; and U.S. Environmental Protection Agency,
1977, 1985a, and 1985b) numerous site-specific factors play critical
roles in determining many of the components of a sampling plan and
thus would affect the costs associated with a sampling effort. A
partial list of these factors would include:
• Types of contaminants present at the site.
• Degree of heterogeneity in the environment surrounding the
site.
• Site characteristics, including location of wastes and
release characteristics.
• Size of the site and the surrounding potentially affected
area.
• Demographics of the surrounding population.
• Relative importance of various transport and exposure
pathways.
This problem would not be critical if, in general, there tended
to be similarities among uncontrolled waste sites, in terms of these
critical factors. Even after a cursory review of the sites on the
National Priorities List, one is struck by the wide variety of sites.
Sites range from traditional waste disposal sites (e.g., landfills,
95
-------
surface impoundments, piles, tanks, containers, and landfarms) to
contaminated sediments, well fields, roads and soil, to identify
just a few of the variations encountered during preliminary
assessments and site inspections. There have been almost 500
different contaminants identified as being present at the 888 sites
proposed or listed on the NPL through Update 5. Sites are located
in both urban and rural areas, in both fairly homogeneous and
heterogeneous areas. There have been fairly small sites (Kin-Buc
Landfill is about 20 acres) and very large sites (BKK Landfill is
nearly 300 acres and there are even larger sites measured in terms
of square miles). Nearly any setting in the United States may
contain a CERCLA waste site. In total, it is not possible to define
a reasonably small number of generic "sites" and to thus estimate
costs for different cases. However, certain components of total
costs, as well as other factors that affect costs, can be evaluated.
The following sections address some of these costs and factors with
regard to:
• Media
• Contaminants
• Sampling locations and numbers of samples
• Unit sampling and analysis costs
5.2 Media
The cost of any particular sampling effort depends on the
environmental media and waste to be addressed. Total costs for any
96
-------
particular site would not be simply equal to the sum of costs
incurred as a result of sampling each applicable media, due to
economies of scale in manpower, mobilization, and other cost
components. Soil and air sampling could be performed concurrently
rather than on separate trips, for example.
Based on a subjective assessment of the difficulties to be
encountered in developing representative data (as indicated in
Sections 2 through 4), as well as the limited data available in
reports cited above, an attempt has been made to provide an
approximate ranking of the media of concern in terms of the relative
costs to collect representative concentration data in each medium.
Ground water is probably the most expensive medium in which to
sample. Costs to install wells are high, and the background data
requirements necessary to ensure development of representative
samples are extensive. Unit installation costs are high, and it
would appear that in the best of situations a minimum of four wells
are needed to obtain valid concentration data (U.S. Environmental
Protection Agency, 1977 and 1985b). Further, several wells may have
to be installed at each location, each screened in a different part
of the aquifer(s) of concern to develop data in all three dimensions.
Also, at least three parameters that have been mentioned as
prerequisites for sampling (gradient, transmissivity, and
conductivity) are expensive to develop on a site-specific basis
(see Gerstein, 1986). It is problematic whether regional estimates
97
-------
are sufficiently accurate to be used instead. Overall, the
difficulties inherent in ground water sampling and the resulting
high unit costs indicate that it would be the most expensive medium
from which to develop representative concentration data.
Waste sampling is probably the second most expensive type of
sampling due to extreme spatial variability, safety considerations,
and potentially high unit costs of sampling. Air is probably less
expensive, even though it is the most variable medium. Air sampling
equipment is fairly well developed for both vapors and particulates.
Air sampling to ensure investigator safety is routinely performed
during all site investigations (U.S. Environmental Protection
Agency, undated) so unit costs should not be too extreme.
Finally, it is likely that soil and surface water are the least
costly media to sample. Despite its high degree of spatial
variability, soil sampling costs are probably relatively low. The
high spatial variability would indicate the need for many sampling
locations but this could be offset by low unit sampling costs and
the potential for compositing samples (see Mason, 1983). Similarly,
the high temporal variability and potential for spatial variability
in large water bodies would be offset by the low unit sampling costs.
Overall, handling and analysis costs should be roughly
comparable for each media.
These conclusions are nearly consistent with the ordering of
costs indicated by JRB Associates (1985) and presented in Table 23.
98
-------
TABLE 23
SUMMARY OF SAMPLING COSTS*
Nominal Complex
Hydrogeologic Investigation** Case Case
Develop plan, select contractors 4,000 8,000
Perform tasks (tasks undefined) 1,000+ 3,000+
Sample wells 5,600+ 10,000
Analyze samples variable variable
Report preparation 8,000 16,000
TOTAL (exclusive of analysis costs) 18,600+ 37,000+
Onsite waste sampling
Lagoons 1,540+ 12,850+
Drums and barrels 7,040+ 23,400+
Surface water, leachate and sediment,
sampling and analysis 6,100+ 9,800+
Air quality sampling and analysis 1,520+ 1,660+
*Exact coverage of costs is unknown.
**Note: For comparison, the reader should note the results of a
recent (January 1987) EPA Region 1 analysis of data gathering
costs which indicates that a current site investigation
requires 300-400 hours of technical level of effort hours
(including analytical) at a cost of $27,000 to $51,000.
Source: JRB Associates, 1985.
99
-------
The validity of the JRB Associates cost data, however, could not
be verified as the report provides little, consistent documentation
concerning their underlying assumptions or derivation.
5.3 Contaminants
One of the principal determinants of the cost to develop
representative concentration data at a site is the list of
contaminants whose presence is suspected at or near the site.
This list includes, for any given site, not only the contaminants
known to have been deposited on the site, but also any of their
transformation products that might pose a risk as well. Analysis of
additional contaminants may also be required to support advanced
interpretive techniques (e.g., receptor modeling). As discussed
in Section 2, in many cases special monitoring instruments are
required to measure the concentration levels of many of these
contaminants. Also, it is important to note in this context that
there are over 700 substances that are considered hazardous under
CERCLA. Gosse et al. (1986) present a partial list of these
contaminants and the principal media of concern for each (see
Table 24). The EPA Contract Laboratory Program analyzes for only
130 organic contaminants plus 27 metals and cyanide as part of their
routine analysis of water and soil samples. Special analytical
services to address additional contaminants (e.g., PCB or TCDD) are
also available on a special request basis.
100
-------
TABLE 24
RECOMMENDED MEDIA IN WHICH TO SAMPLE RCRA
APPENDIX VIII HAZARDOUS CONSTITUENTS
Ground
water
acenaphthalene
acenaphthene
acetaldehyde
acetonitrile
acetophenone
2-acetylaminof luorene
acetyl chloride
acrolein
acrylonitrile
aldrin
allyl alcohol
alpha-BHC
aluminum phosphide
4-aminobiphenyl
aniline
anthracene
antimony (total)
aramite
aroclor 1016
aroclor 1221
aroclor 1232
aroclor 1242
aroclor 1248
aroclor 1254
aroclor 1260
arsenic (total)
arsenic acid
arsenic pentoxide
arsenic trioxide
asbestos
barium (total)
barium cyanide
benz(a)anthracene
benzene
benzene, dichloromethyl
benzenethiol
benzidine
benzo(a)pyrene
benzo( b ) f luoranthene
benzo(ghi)perylene
benzo(k) fluoranthene
p-benzoquinone
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Surface
water
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Saturated
soil
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Unsaturated Subsurface
soil gas Air
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Source: Gosse et al., 1986.
101
-------
TABLE 24 (Continued)
Ground
water
benzot rich lor ide
benzylchloride
beryllium (total)
beta-BHC
bromodichlorome thane
b romome t hane
4-bromophenyl phenyl
ether
2-butanone peroxide
butyl benzyl phthalate
2-sec-butyl-4,6-
dinitrophenol
cadmium (total)
calcium chromate
calcium cyanide
carbon disulfide
carbon oxyfluoride
carbon tetrachloride
chloral
chlordane
chloroacetaldehyde
p-c hlo roan i line
chlorobenzene
chlorobenzilate
2-chloro-l , 3-butadiene
p-chloro-m-cresol
chlorodibromome thane
chloroethane
b is ( 2-c h loroethoxy )
methane
bis(2-chloroethyl) ether
2-chloroethyvinyl ether
l-chloro-2 , 3-epoxypropene
chloroform
bis(2-chloroisoprpyl )
ether
chlorome thane
bis(chloromethyl ) ether
chloromethyl methyl ether
2-chloronapthalene
2-chlorophenol
3-chloropropene
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Surface
water
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Saturated
soil
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Unsaturated
soil
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Subsurface
gas Air
X
X
X
X
X
X
X
X
X X
X
Source: Gosse et al., 1986.
102
-------
TABLE 24 (Continued)
Ground
water
3-chloropropionitri le
chromium (total)
chrysene
copper (total)
copper cyanide
crotonaldehyde
creosol
cumene
cyanide
cyanogen
cyanogen bromide
cyanogen chloride
ODD
DDE
DDT
delta-BHC
deno(l ,2, 3-cd)pyrene
di-n-butyl-phthalate
di-n-octyl phthalate
di-n-propylnitrosamine
dibenzo(a,e)pyrene
dibenzo(a.h) anthracene
dibenzo(a,h)pyrene
dibenzo(a, i)pyrene
1 , 2-dibromo-3-chloro-
propane
dibromome thane
1 , 2-dibromoethane
tris (2 , 3-dibromopropyl )
phosphate
m-dichlorobenzene
o-dichlorobenzene
p-dichlorobenzene
3, 3 '-dichlorobenzidine
trans-l,4-dichloro-
2-butene
dichlorodif luoro-
me thane
1, 1-dichloroethane
1 ,2-dichloroethane
trans-1 ,2-dichloroethene
1 , 1-dichloroethylene
1 , 2-dichloroethylene
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Surface
water
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Saturated
soil
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Unsaturated
soil
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Subsurface
gas Air
X
X
X
X
X
X
X
X
X X
X X
Source: Gosse et al., 1986.
103
-------
TABLE 24 (Continued)
Ground
water
die h lor ome thane
2 ,4-dichlorophenol
2 ,6-dichlorophenol
2 ,4-dichlorophenoxy-
acetic acid
1 , 2-dichloropropane
cis-1 ,3-dichloropropane
trans-l,3-dichloro-
propene
dieldrin
o,o-diethyl o-2-parazinyl
phosphor othionate
o,o-diethyl phosphoric
acid, o-p-nitrophenyl
ester
diethyl phthalate
diisopropylf luoro-
phosphate
3, 3'-dimethoxybenzidine
dimethoate
p-dimethylaminoazo-
benzene
7, 12-dimethylbenz(a)
anthracene
dimethyl benzene
3,3'-dimethylbenzidine
dimethylcarbatnoyl
chloride
1 , 1-diraethylhydrazine
1 , 2-dimethylhydrazine
, -dimethylphenethyl-
amine
2,4-dimethylphenol
dimethyl phthalate
dimethyl sulfate
m-dinit robenzene
4, 6-dinitro-o-cresol
2,4-dinitro phenol
2, 4-dinitro toluene
2,6-dinitrotoluene
1 ,4-dioxane
diphenylamine
1 , 2-diphenylhydraz ine
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Surface
water
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Saturated
soil
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Unsaturated Subsurface
soil gas Air
X XX
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Source: Gosse et al., 1986,
104
-------
TABLE 24 (Continued)
Ground
water
disulfoton
endosulfan I
endosulfan II
endrin
endrin aldehyde
ethyl benzene
ethyl cyanide
bis(2-ethylhexyl)
phthlate
ethyl methacrylate
ethyl methanesulfonate
ethylene oxide
ethyleneimine
famphur
f lourene
fluoride
fluorine
fluouranthene
formaldehyde
garama-BHC
g lye idyl aldehyde
heptachlor
heptachlor epoxide
hexachlorobenzene
hexachlorobutadiene
hexachlorocyclo-
pentadiene
hexachlorodibenzo-p-
dioxins
hexachlorodibenzofurans
hexachloroe thane
hexachlorophene
hexachloropropene
hexaethyltetraphosphate
hydrocyanic acid
hydroflouric acid
hydrogen sulfide
icocyanic acid,
methyl ester
iodome thane
isobutyl alcohol
isodrin
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Surface
water
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Saturated
soil
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Unsaturated Subsurface
soil gas Air
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Source: Gosse et al., 1986.
105
-------
TABLE 24 (Continued)
Ground
water
kepone
lead (total)
lead acetate
lead phosphate
maleic anhydride
malonitrile
mercury (total)
methacrylonitrile
methane
met hapyri line
methoxychlor
3-methycholanthrene
methyl acetate
2-methylactonitril
4)4'-methylene bis
(2-chloroaniline)
methyl ethyl ketone
methyl hydrazine
methyl methacrylate
methyl tnethanesulfonate
methyl parathion
mustard gas
napthalene
1 ,4-napthoquinone
1-napthylamine
2-napthylamine
nickel (total)
nickel cyanide
4-nitophenol
nitric oxide
p-nitroaniline
nitrobenzene
nitrogen dioxide
nitrogen mustard and
hydrochloride salt
n-nitrosodi-n-
butylamine
n-nitroso diethylatnine
n-nitrosodimethylamine
n-nitrosodiohenyl-
amine
n-nitroso raethy lethyl-
amine
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Surface
water
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Saturated
soil
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Unsaturated
soil
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Subsurface
gas Air
X
X
X
X
X
X
X
X
X
Source: Gosse et al., 1986.
106
-------
TABLE 24 (Continued)
Ground
water
n-nitrosomorpholine
n-nitrosopi peri dine
n-nitrosopyrrolidine
5-nitro-o-toluidine
ortho cresol
osmium (total)
osmium tetroxide
para cresol
paradelyhe
parathion
pentachlorobenzene
pent achlo rod ibenzo-p
dioxins
pentachlorodibenzo-
furans
pen tachloroe thane
pentachloronitrobenzene
pentachlorophenol
phenacetin
phenanthrene
phenol
phorate
phosgene
phthalic anhydride
2-picoline
potassium cyanide
potassium silver
cyanide
pronamide
1,3-propane sultone
propylene oxide
2-propyn-l-o
pyrene
pyridine
resorcinol
safrole
selenium (total)
selenium sulfide
seleniuos acid
silver (total)
silver cyanide
silvex
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Surface
water
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Saturated
soil
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Unsaturated Subsurface
soil gas Air
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Source: Gosse et al., 1986.
107
-------
TABLE 24 (Continued)
Ground Surface Saturated Unaaturated
water water soil soil
sodium cyanide
strontium sulfide
sulfide
2,4,5-T
1,2,4, 5-tet rach loro-
benzene
2,3,7, 8-tetrachloro-
dibenzo-p-dioxin
tetrachlorodibenzo-p-
dioxins
tetrachlorodibenzo-
furans
tetrachloroethene
1,1,1, 2-tetrachloro-
ethane
1,1,2, 2-tetrachloro-
ethane
2,3,4,6-tetrachloride
tetraethyldithio-
pyro phosphate
tetraethylpyrophosphate
t e t ran i t rome t hane
thallic oxide
thallium (I) acetate
thallium (I) carbonate
thallium (I) chloride
thallium (I) sulfate
thallium (total)
thallium selenite
t hioacetamide
toluene
toluene diisocyanate
toxaphene
t ribromome thane
1 ,2 ,4-t rich lor obenzene
t richloroethene
1,1, 1-t rich loroe thane
1,1,2-trichloroethane
trichlorome thane thiol
t rich lor omonof luoro-
methane
2,4, 5— t rich lor ophenol
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Subsurface
gas Air
X
X X
X
X
X X
X X
X X
X
Source: Gosse et al., 1986.
108
-------
TABLE 24 (Concluded)
Ground Surface Saturated Unsaturated Subsurface
water water soil soil gas Air
2,4,6-trichlorophenol
1,2, 3-t rich loropro pane
vanadic acid,
ammonium salt
vanadium (total)
vanadium pentoxide
vinyl chloride
xylenes
zinc (total)
zinc cyanide
zinc phosphide
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X X
X
Source: Personal correspondence with Robert April (EPA-OSW-LDB).
"Guidance on Issuing Permits to Facilities Required to
Analyze Ground Water for Appendix VIII Constituents."
Source: Gosse et al., 1986.
109
-------
5.4 Number of Sampling Locations
The required number of sampling locations (in three dimensions)
needed to achieve representative samples depends heavily on site and
waste specific conditions. The two most critical site specific
factors are the size of the potentially affected area and the size
of the zones of homogeneity at the site. A zone of homogeneity (in
three dimensions) is a volume of space in which it can be assumed
that the distribution of contaminant, in the medium in question is
homogeneous. Since each zone of homogeneity must be sampled in at
least one location, the number of sampling locations must generally
exceed the ratio of these two sizes. Moreover, multiple samples
should be taken within each zone of homogeneity to account for
sampling and other random errors. Statistical sampling theory
provides some approaches for estimating the number of samples needed
within each zone, depending on the desired level of confidence.
However, most of these approaches depend on assumptions that are, at
best, unlikely to be true about contaminant concentrations at
uncontrolled waste sites (e.g., normality of the contaminant
distribution). Further, they also require some knowledge about the
parameters of the distribution such as the coefficient of variation
which is almost always unknown. Overall, there is no objective set
of approaches that can be employed routinely to estimate the number
of sampling locations needed at a given waste site. Judgment is
required.
110
-------
As an example of the number of sampling locations that might
be required to sample ground water, the draft RCRA Ground-Watej:
Monitoring Technical Enforcement Guidance Document of August 1985
(U.S. Environmental Protection Agency, 1985b) indicates that a
minimum of four wells are required (one upgradient, three
downgradient) while a spacing of one well each 150 feet (or less) is
required whenever, for example, liquid wastes are present. This
density of wells is required to support a detection monitoring
system, i.e., one that would detect a release, but not necessarily
characterize the concentrations of that release from the site. In
general, more wells would be needed to assess the extent and
concentrations of a release. The installation of even four wells
would be expensive in terms of current site inspection costs at many
sites (see Gerstein, 1986).
As another example, Hanisch and McDevitt (1984) present a method
for estimating the number of sampling locations required in a zone
of homogeneity of area A (in square meters). In their equation, the
1/2
number of locations must exceed 6 + 0.1A . Thus, the number of
2
sampling points in a homogeneous zone of size 10,000 m in a
surface impoundment of area must exceed 16.
5.5 Number of Samples Per Location
There are three aspects to consider in determining the number
of samples to be taken at each sampling location: frequency,
duration, and replication. The first two are determined by the
111
-------
temporal variability of the contaminant concentrations in the
potentially affected area and the type of health (and other) effects
expected as a release of the release (e.g., acute or chronic). For
example, assessments of chronic effects risks might require long-term
average concentrations requiring possibly less frequent samples over
a long period of time, while assessments of acute effects risks may
require that a great many samples be taken in a short period of time.
In between, the assessment of risks of subchronic effects may require
fairly frequent sampling over a relatively long period of time.
Replication, the third aspect to the number of samples,
addresses the need take multiple samples to account for uncertainty
brought on by random sampling errors.
With the exception of replication, there are again no objective
approaches to assessing the frequency or duration of sampling.
However, representative sampling would, in principle, require a
fairly large number of samples at each location, if for no other
reason than to account for random errors.
Provost (1984) presents a table that is useful in estimating
the sampling frequency needed to achieve specified confidence
intervals about a mean as a function of a known coefficient of
variation and an assumption about the underlying distribution (see
Table 25). As indicated in this table, for example, 24 samples from
a normally distributed process with a coefficient of variation of
50 percent would yield a 95 percent confidence interval about the
112
-------
TABLE 25
EXPECTED CONFIDENCE INTERVALS FOR A PARAMETER MEAN AS A FUNCTION OF NUMBER OF SAMPLES
Expected Confidence Intervals for a Parameter Mean aa a Function of Number of Samples (Measurements)
Expected
Variability
Measurement
(Coefficient
of Variation
51
10Z
25Z
50Z
100Z
200Z
500Z
1000Z
10000Z
of
95Z Confidence Interval About the Mean Estimate ( percent )b
Distribution n - 4
)a (Model) (Quarterly)c
Normal Model
Lognormal Model
Normal Model
Lognormal Model
Normal Model
Lognormal Model
Normal Model
Lognormal Model
Normal Model
Lognormal Model
Lognormal Model
Lognormal Model
Lognormal Model
Lognormal Model
+8.0
+8.1, -7. 5
+15
+17, -15
+40
+48, -32
+80
+110, -53
+160
+280, -73
+650, -87
+1700, -94
+3000, -96
+13000. -99
n - 6 n - 12
(Bl-monthly)c (Monthly)0
+5.2
+5.4", -5.1
+11
+11, -9. 9
+26
+2 9, -23
+52
+62, -39
+110
+104, -58
+280, -74
+570-85
+900-90
+2300, -96
L _ _-
+3.2
+3. 2, -3.0
+6.4
+6. 5, -6.1
+16
+17, -14
+32
+35, -26
+64
+70, -41
+124.-55
+220, -68
+300, -7 5
+590, -85
n - 24 n - 52
(Seml-monthly)c (Weekly)c
+2.1
+2.1, -2.0
+4.2
+4. 3, -4.1
+11
+1 1 , -9 . 9
+21
+22, -18
+42
+42, -30
+71, -42
+120, -53
+150, -60
+260, -7 2
+1.4
+1.4
+2.8
+2. 8, -2. 7
+6.9
+7.1,6.6
+14
+14, -12
+28
+26, -21
+43, -30
+65, -40
+82, -4 5
+130. -57
n - 365
(Dally)c
+0.5
+0.5
+1.0
+1.0
+2.6
+2. 6, -2. 5
+5.2
+5.0, -4. 8
+11
+9.0, -8. 2
+14, -12
+20, -17
+24, -20
+37, -27
value for the coefficient la assumed In this table. If the coefficient of variation Is greaterthan 100Z,
the normal model Is not realistic.
bn - number of samples or data points.
cMonltorlng frequencies required for the specific value of n If the duration of study was one year.
Source: Provost, 1984.
-------
mean of about 21 percent. This means that in practice if semi-
monthly samples were taken at a sampling location that experienced
temporal variations of 50 percent (expressed as the coefficient of
variation), then one would be 95 percent confident that the true
average lay within 21 percent of the sample average.
5.6 Unit Sampling, Handling, and Analysis Costs
The development of generic cost estimates is also hampered by a
lack of data on unit costs for background information collection,
sampling, handling, and analysis in many of the media and under the
conditions applicable to uncontrolled waste sites. The compilation
of such a unit cost data set is beyond the scope of this report.
However, some unit costs estimates were identified as part of this
review. Additional unit cost information concerning ground water
sampling can be found in Gerstein, 1986.
Few cost data concerning background information collection
could be collected. The only indication of the magnitude of these
costs is the current cost of Preliminary Assessments. These
assessments currently cost approximately $1,200 per site. It
is likely that the cost of background information to support
representative data development (e.g., ground water flow direction)
would be substantially higher.
Some very limited information is available on sampling,
handling, and analysis costs. As discussed previously, Table 23
presents cost data derived from JRB Associates, 1985. The quality
114
-------
of these data is unknown, given the lack of adequate documentation.
Further, based on the limited documentation provided in that report,
some of the estimates (e.g., the estimates for air sampling) appear
to be unreasonably low when compared with the requirements of
representative data development. It should be stressed, however,
that the JRB Associates report is a draft report.
Table 26 presents an example of the costs to establish a ground
water sampling well field consisting of 10 wells drilled to a depth
of 100 feet each (adapted from Gerstein, 1986). A representative
mobilization cost (including plan development, contracting, and
other administrative costs) is estimated to be about £6,000. A
representative cost to install each well is estimated to be $4,000
(or about $40 per foot). Unit costs cited in the literature ranged
from $30 per foot to $100 per foot (JRB Associates, 1985 and U.S.
Environmental Protection Agency, 1985c). A representative cost to
determine hydraulic gradient by potentiometric mapping (a necessary
prerequisite to sampling) is estimated as $4,500. A representative
cost to determine hydraulic conductivity (a second necessary
prerequisite) by either a bailing test or a slug test is estimated
to be ijl,500 per well, or $15,000 overall. These costs do not
include costs for site security or disposal of drilling waste (a
potentially hazardous waste regulated under RCRA), nor do they
include any increased costs that might result from additional
concerns for worker safety over and above normal drilling risks.
115
-------
TABLE 26
EXAMPLE REPRESENTATIVE PRE-SAMPLING COSTS
TEN-WELL SAMPLING FIELD
Cost Component
Costs ($)*
Mobilization
Well installation (10, 4-inch diameter wells,
100 feet deep)
Potentiometric Mapping
Conductivity Tests (10 wells at $1500/well,
slug or bail test
Total
6,000**
40,000
4,500
15,000
65,500
*Costs do not include site security, disposal of drill spoils, or
additional costs associated with worker safety.
**Note: EPA Region 1 currently (January 1987) experiences a $55,000
mobilization fee for a ten-well drilling program, in
addition to costs of drilling and installation of wells.
Source: Gerstein, 1986.
116
-------
EPA Region 1 currently experiences a $55,000 mobilization fee for a
10-well drilling program, in addition to the cost of drilling and
installation of wells.
Consistent unit cost information for collecting ground water
samples is not available. JRB Associates (1985) listed costs
ranging from $300 to $400 per well. The average costs for RCRA
Interim Status facilities is estimated to be $200 per well (U.S.
Environmental Protection Agency, 1985c). The assumptions concerning
the number of samples taken and the cost component (e.g., manpower)
included in both these estimates are unknown.
In contrast, consistent information was available on the cost
of analyzing water and soil samples using the CLP (Kolb, 1986). The
current average analytical cost for the standard organic and
inorganic analyses are about $960 and $140, respectively. The
costs for special analytical services are variable.
Finally, little information was available on the cost of QA/QC
programs, although Schweitzer (1981) states that quality assurance
costs range from 0 to 20 percent of monitoring costs.
5.7 Illustrative Example
To put this information into perspective, consider as an example
the 10-well field discussed above. If the ground water were subject
to a temporal variation of 10 percent (expressed as the coefficient
of variation) during the period of concern and a 95 percent
confidence interval of +15 percent were desired, then, under a
117
-------
normality assumption, 4 samples would be required at each well
(see Table 25). Thus, a total of 40 samples would be required
at an expected analysis cost of J>44,000. Assuming (probably
conservatively) that the sampling costs for these samples was $200
per well, a cost of $2,000 would also be incurred. Thus, assuming
the 10 wells provided spacially representative data, the total cost
to develop representative concentration data under this scenario
(for ground water only) would exceed itllO,000, exclusive of handling,
QA/QC, interpretation, and administrative costs.
Average total costs to develop representative data over all
media can only be inferred, at best, from this limited information.
118
-------
6.0 SUMMARY OF PRECEDING DISCUSSIONS AND IMPLICATION FOR THE HRS
Representative data is necessary to have a high degree of
confidence in decisions made based on the results of analyses on
that data. The lower the degree of representativeness (i.e., the
lower the quality), the higher the uncertainty in the results and
the lower the confidence in the decisions. The point at which the
degree of representativeness of the data becomes too low to support
the decisions to be made is determined by the decision-maker. Thus,
it is EPA's responsibility to determine if the concentration data
developed as part of PA/SIs is of sufficient quality to support
particular uses of that data in the HRS to aid in deciding whether a
site is placed on the NPL. However, as indicated previously, truly
representative data may be very expensive to develop at some sites
and may be beyond the capabilities of the program to develop
routinely.
Many factors act to limit the quality of PA/SI concentration
data, thereby increasing its uncertainty. Some factors are inherent
in the nature of concentration data; such data alone cannot indicate
future conditions, the data rarely reflect actual exposure
concentrations, and there is always a chance that actual
concentrations are significantly higher or lower. Other factors are
programmatic in nature. They arise from constraints on time and
resources, variations and limitations in the level of experience of
those evaluating sites, and from any gaps in a comprehensive QA/QC
119
-------
program. Still others are technological in that there are gaps in
sampling equipment and standard analytical techniques for some field
situations and contaminants. A fourth set of factors arise from
difficulties in the acquisition, handling, analysis, and
interpretation of samples and sample data. There are numerous
opportunities for introducing error into the development of
concentration data. Undetected temporal and spatial variation in
environmental processes induce potentially undetected variation in
environmental concentrations. The lack of information on these
processes at each site increases the uncertainty associated with
concentration data developed for that site. Further, without
careful management, it is likely that samples will not be
representative of in-situ conditions and that they will become
contaminated. These problems of sampling management arise in all
areas except the interpretation component of the data development
process. Uncertainty in results is further increased by the
necessary assumptions that must be made in order to analyze sampling
results. The degree to which these assumptions diverge from reality,
increases the uncertainty still further.
Overall, there are numerous reasons to question the
representativeness of even the best concentration data set, i.e.,
those data sets that have been carefully developed and analyzed
subject to an extensive QA/QC program. Data developed without such
QA/QC can, at best, be considered suspect.
120
-------
This discussion indicates that the role that concentration data
play in the HRS and in NPL decision-making should not be too
extensive, even if a good QA/QC program were to be implemented as
part of the PA/SI program. Concentration data should be only one of
many factors considered in evaluating the risk posed by a site at
this stage of the assessment process. Further, the limitations in
the data that could reasonably be developed from a site inspection
program on a routine basis indicate that any concentration data
factors should not be weighted heavily in the HRS.
121
-------
7.0 OPTIONS FOR EMPLOYING CONCENTRATION DATA
This chapter presents several options for using concentration-
based factors in the HRS. These factors are designed to improve the
accuracy of the HRS in evaluating risk without exceeding the
limitations in concentration data discussed in the previous chapters.
The discussion also identifies changes in the site inspection program
that should be implemented if the options are incorporated into the
HRS. Without these programmatic changes, the uncertainty in the
concentration data would be very high and the use of these
concentration data in the factors presented would be questionable.
7.1 Basic Philosophy
A combination of factors determined the basic philosophy towards
employing concentration data that is reflected in the options
discussed below. These factors include the possible objectives of
the HRS, the wide variation in the characteristics of uncontrolled
waste sites and the limitations in concentration data discussed
previously. The objective of the HRS is assumed to be to
expeditiously identify sites for possible response action based on
their relative risks (U.S. House of Representatives, 1986). The
objective requires that the HRS be applicable to all potential
CERCLA sites. This, in turn, requires that any concentration
factors be applicable to all sites as well. In the interest of
maintaining simplicity in the HRS, no special options were developed
to assist in evaluating the occasional sites with extensive
123
-------
concentration and other environmental data. Rather, the options and
factors discussed below employ data that could be collected at any
site within the constraints of a reasonable site inspection.
The basic approach embodied in all of the options is to assign
additional points to sites with current problems, as indicated by
concentration data, without putting at a disadvantage those sites
that appear to lack current problems (as indicated by available
concentration data). There are several reasons for adopting this
type of approach. First, as discussed previously, sampling by its
nature is retrospective. It indicates past problems only. The data
can be used to infer present or future problems only by the use of
either implicit or explicit environmental models. Alone,
concentration data tell nothing about future problems. Second,
environmental concentration data can only be employed when they also
demonstrate an observed release. Concentrations that do not exceed
background levels cannot be reliably associated with the site in
question. Thus, the risk posed by these concentrations cannot be
associated with the site and should not be included in the evaluation
of the risk posed by the site being evaluated. Third, it is a
characteristic of any statistical analysis that many low values do
not completely rule out the existence of high values. This is
embodied by the statement that "lack of evidence is not evidence of
lack". Fourth, the data collected during site investigations can
rarely be considered as representative in all media. The cost and
124
-------
time requirements associated with developing concentration data that
is representative of environmental concentrations in both space and
time across the five media of concern are likely to be excessive.
Finally, the uncertainty in the environmental concentration data
developed during site inspections is such that the measured data
indicate at best the lower limit of the maximum possible exposure
concentrations, even considering possible dilution and mass removal
effects. If a particular contaminant concentration is measured in
the environment then the potential clearly exists for someone to
become exposed to that concentration either at the time the
measurement was taken or at some time in the future. Thus, for a
given contaminant, the maximum detected concentration determines a
"worst-known-case" of potential exposure. Moreover, the uncertainty
in the data generally indicate that the possibility exists for true
environmental concentrations to exceed detected concentrations
either at the same location (due to sampling, handling and analysis
uncertainty), at some other location (due to spatial variation), or
at some other time (due to temporal variation). It is unlikely that
sampling will find the true maximum concentration. For similar
reasons, the maximum detected waste concentration indicates a lower
bound on the maximum possible waste concentration.
These considerations taken together indicate that little
reliance should be placed on low concentration values determined
during site inspections in determining that little risk exists from
125
-------
the site. However, the existence of high concentration values does
indicate that some risk may exist. Thus, the approach taken in
these options is that sites with "high" observed contaminant
concentrations should be assigned extra points while those with
"low" contaminant concentrations should not be overly penalized.
The definitions of "high" and "low" concentrations require the
determination of benchmarks against which to assess the measured
concentration values.
7.2 Benchmarks
Benchmarks determine whether an observed concentration of a
contaminant can be considered to be "high" or "low" (or "medium,"
etc.). Benchmarks establish a metric to gauge the relative magnitude
of observed concentrations. For example, a concentration of
3
4,000 mg/m of sulfur hexafluoride in a workplace might be
3
considered low while a concentration of 500 mg/m of trichloro-
ethylene in the same workplace would be considered high. In the
former case the concentration is below the threshold limit value
(TLV) of 6,000 mg/m3 of sulfur hexafluoride while the latter
3
concentration exceeds the trichloroethylene TLV of 270 mg/m
(American Conference of Governmental Industrial Hygienists, 1985).
Thus, within the context of employing concentration data within
the HRS, risk-related benchmarks have to be determined. These
benchmarks need to be determined for each chemical of interest in a
consistent fashion. Further, since the HRS employs separate
126
-------
migration pathways (and to a lesser degree exposure pathways in
evaluating toxicity) and since the risk posed by an exposure to a
given concentration of a contaminant frequently differs depending on
whether the contaminant is ingested or inhaled, separate pathway
specific benchmarks have to be developed. Additionally, since the
HRS addresses both human health and environmental risks, separate
benchmarks reflecting these two types of risk may have to be
developed for each contaminant and pathway. The feasibility of
developing a collection of suitable benchmarks for use in the HRS is
problematic and remains an important outstanding issue in the use of
concentration data in the HRS. The options discussed below assume
that an acceptable set of benchmarks can be developed. The
development of such benchmarks is beyond the scope of this paper.
7.3 Options
Options for employing concentration data in the HRS were
developed in each of the three evaluation categories: release,
waste characteristics, and targets. The options were developed so
as to be employable in any of the HRS migration pathways. No special
ground water, surface water, or air options were developed. The
options are discussed, by HRS factor category, in the following
sections.
7.3.1 Options in the HRS Release Category
The HRS currently employs concentration data in determining
whether an observed release has occurred. The benchmark in this
127
-------
case is the background contaminant concentration. A release category
value of 45 is assigned whenever the observed site concentration
significantly exceeds the background concentration in the medium
being examined, else, the site is evaluated based on its potential
to release. The maximum potential to release value is also 45.
An alternate approach of similar structure can be readily
defined employing a risk based benchmark as well as the background
benchmark, as follows. If the observed concentration of any
contaminant in the environment* significantly exceeds the background
concentration for that contaminant (i.e., constitutes an observed
release) and also exceeds a benchmark concentration for that
contaminant and pathway, then a maximum observed release value of,
for example, 45 would be assigned. If the observed concentration
significantly exceeds the background concentration but is lower than
the benchmark, then a lower value of, for example, 40 would be
assigned. If the observed concentration does not significantly
exceed the background concentration, then the site would be
evaluated based on its potential to release since no observed
release can be attributed to the site. The maximum potential to
release value might be set equal to the value for an observed
release that does not exceed the benchmark or it might be set equal
to the maximum observed release value. Either variation is
reasonable.
*Waste concentrations cannot be used to evaluate the release category.
128
-------
Table 27 illustrates this scoring approach. At both Sites 1
and 2, the observed site concentrations significantly exceed background
and exceed the benchmark as well. Thus, a release category value of
45 is assigned to both the sites. At Sites 3 and 4, the observed
concentrations exceed background, but do not exceed the benchmark.
Thus, these sites are assigned values of 40. At Sites 5 and 6, the
observed concentrations do not significantly exceed background and
hence the sites are assigned values of 0 regardless of whether the
observed concentrations exceed the benchmark. (It is noted that a
background level above the benchmark would presumably trigger an
investigation to identify an alternate source.)
A shortcoming of this approach is evident in comparing the last
two hypothetical sites (7 and 8). At both sites, the observed site
concentration of the contaminant in question significantly exceeds the
background concentration. Thus, observed releases have occurred at
both sites. Also, the observed concentration at Site 7 significantly
exceeds the benchmark while at Site 8 it only slightly exceeds the
benchmark. Therefore, Site 7 poses a much greater risk than Site 8,
in principle, due to the presumably higher concentrations to which the
people around the site will be exposed, all other factors being the
same. Thus, Site 7 should receive a higher value, in principle, than
Site 8. However, due to the structure of this option both sites would
receive the same release category value (45) and site pathway scores,
all other factors being the same. This consideration argues for a
129
-------
TABLE 27
HYPOTHETICAL EXAMPLES OF RELEASE CATEGORY OPTION
Site
1
2
3
4
5
6
7
8
Benchmark
Concentration
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
Background
Concentration
0.1
1.0
0.01
0.01
5.0
0.1
1.1
0.1
Site
Concentration
2.0
10.0
0.5
0.1
5.0
0.1
11.0
1.1
Assigned
HRS Value
45
45
40
40
0
0
45
45
130
-------
more refined evaluation approach employing varying levels relative to
the benchmark with associated levels of valuation (e.g., one-tenth of
the benchmark could be assigned one-tenth of the value, i.e., 4.5).
However, such an option is not presented here since it would place
too great a reliance on the concentration data in determining scores.
Using data of debatable validity in such an important fashion would
be questionable.
One should note that the adoption of this option would change
somewhat the role that the release category plays in the HRS.
Currently, the release category reflects the likelihood that the
site has, is, or will release a significant quantity of contaminants
into the environment. By establishing two levels of release values,
an implicit evaluation of the magnitude of the release is made
current, detected releases above benchmarks are evaluated higher
than potential future releases or undetected, current releases that
may or may not exceed benchmarks. Thus, the release category would
address the probabilistic aspects of risk as well as a component of
the hazard aspects.
7.3.2 Options in the HRS Waste Characteristics Category
Two options for reflecting concentration data in the waste
characteristics category were developed. The first envisions a
separate concentration factor, the value of which could be added to
the toxicity and waste quantity factor values to determine the
overall waste characteristics value. As in the previous options,
131
-------
this factor would be evaluated based on environmental concentration
data, not waste constituent concentration data. The second envisions
modifying the waste quantity factor, changing it to reflect the
quantity of hazardous constituents found at the site rather than the
quantity of wastes containing hazardous substances. This latter
option uses waste concentration data. A third option for
incorporating concentration data in the toxicity factor is also
discussed. This option is presented as an example of a potential
approach that requires strict confidence in the representativeness
of the concentration data.
In the first option, the maximum detected environmental
concentration of each contaminant in an observed release would be
compared with both the background concentration for that contaminant
and the benchmark concentration. Those concentrations that do not
significantly exceed background (and thus are ilot "observed
releases") would be rejected from further consideration in
evaluating factors related to environmental concentrations. If all
concentration data are rejected or if no concentration data are
available, then the concentration factor is assigned a value of 0.
For each remaining concentration datum, the ratio of the datum to
the benchmark is calculated and the contaminant with the highest
ratio chosen. The concentration factor would then be evaluated
based on the calculated ratio for that contaminant using a factor
table such as is illustrated in Table 28. This option is designed
132
-------
TABLE 28
ILLUSTRATIVE CONCENTRATION FACTOR TABLE*
Ratio (0/B) Factor Value
Greater than 1 5
0.1 - 1.0 4
Less than 0.1 0
*Values are provided for illustrative purposes only.
B: Benchmark concentration.
0: Observed site concentration.
133
-------
to be used as an addition to the toxicity factor. An example of the
approach is given in Table 29.
The concentration of three contaminants in Table 29 (contaminants
1, 2, and 3) significantly exceed background and can be used in
evaluating the concentration factor under this option. Since the
maximum site concentrations of contaminants 4 and 5 do not
significantly exceed background, these concentrations cannot be
attributed to the site and hence would be rejected from further
consideration in evaluating the concentration factor. This rejection
is indicated despite the high site to benchmark concentration
associated with contaminant 4. There is probably a risk at the site
due to high environmental concentrations of contaminant 4, yet it
cannot be unquestionably associated with the site. Given the high
background concentrations, it is equally probable that the risk
arises from other contaminant sources.
The ratios of the concentrations of the three contaminants are
calculated and the contaminant with the largest ratio is identified
(contaminant 2, ratio 20). Using the illustrative factor table
presented on Table 28, the concentration factor score for this site
would be 5.
Although the values listed in Table 28 are illustrative and
subject to change, two relationships embodied in the example table
are important and should be maintained if the option is adopted.
First, the maximum score (5) is smaller than the current maximum
134
-------
TABLE 29
HYPOTHETICAL EXAMPLES OF CONCENTRATION FACTOR OPTION*
Contaminant
1
2
3
4***
5***
Benchmark
Concentration
1.0
0.5
0.1
1.0
1.0
Background
Concentration
0.1
1.0
0.01
45.0
0.1
Maximum Site
Concentration
2.0
10.0
0.1
40.0
0.1
Ratio**
2
20
1
40
1
*Environmental sample.
**Rejected since detected concentrations do not demonstrate an
observed release.
***Ratio of maximum site concentration to benchmark concentration.
135
-------
scores for the toxicity factor (18) and the waste quantity factor (8).
This embodies the conclusion that due to the uncertainty associated
with concentration data and the associated inferences about current
and future risk, this factor is less important than the others in
determining the site pathway score. Also, it minimizes the
disadvantage associated with sites that lack observed releases.
Second, a relatively small ratio (e.g., at least one-tenth of the
benchmark) is needed to receive nearly the maximum value for the
factor. This reflects a belief that the uncertainty in concentration
data is high and that concentration measurements near the benchmark
(i.e., within one-tenth of the benchmark) indicates the probable
presence of an exposure situation of concern. These relationships
are consistent with the basic philosophy discussed previously.
In the second option, the waste quantity factor would be
evaluated based on either the total quantity of wastes containing
hazardous substances at the site or the total quantity of hazardous
substances found in the wastes at the site. The former approach is
that currently used in the HRS. It would be employed whenever
detailed, complete, representative data on waste constituent
concentrations are lacking at the site. The latter approach could
be taken only if such data were available. Distributions of the
concentrations of the waste constituents would be developed from the
waste analyses. The mean (or some other measure of tendency such as
the median, the mode, or the 95th percentile) of the distribution of
130
-------
each substance would then be calculated and multiplied by the total
waste quantity to yield the quantity of that substance present at
the site. The sum of the quantities of each substance would then
be calculated and used to evaluate the waste quantity factor.*
The principal difficulty with this approach is the data
requirement. Such analyses are rarely available as they are
difficult and hazardous to perform, as discussed in Section 4. An
alternate approach would be to use the highest concentration of each
hazardous constituent detected in the waste instead of some parameter
of the distribution. These concentrations would then be multiplied
by the total quantity of wastes present to yield the quantity of each
substance present. These quantities would then be summed to form an
estimate of the total quantity of hazardous substances present. This
approach has the advantage of relying on the highest concentration
and would, in principle, tend to be conservative (i.e., over estimate
the quantity and thus the risk). However, unless a sufficient number
of samples have been taken at a sufficient number of locations, there
is no guarantee that the highest detected concentration is greater
than the actual average concentration. Further, both approaches
rely on the completeness of the analyses, i.e., chemical analyses
*It should be noted that this approach is potentially inconsistent
with the current definition of the HRS waste quantity factor. The
HRS currently evaluates waste quantity on an "as received" basis.
This may differ significantly due to temporal factors from the
waste quantity indicate as being present at the time of sampling.
137
-------
must be performed for each CERCLA hazardous substance. This
would prove difficult in those cases where acceptable methods of
analysis do not exist.
A final difficulty in these approaches lies in the definition of
the factor evaluation table. One approach would be to transform the
existing hazardous waste quantity table (47 FR 31180, 12 July 1982)
into a hazardous substance quantity table. The availability of
information required to do this translation is examined in
Wusterbarth, 1986. The possibility of employing other approaches
needs to be examined.
As stated previously, a third option for incorporating
concentration data into a toxicity factor was examined. In this
option, the toxicity factor would be evaluated in terms of a
concentration factor and a severity factor, as follows. Each
contaminant would be assigned a pathway-specific toxicity benchmark
and a severity index. The benchmark would be defined as discussed
previously while the severity index would reflect the severity of
the effects associated with exceedances of the toxicity benchmark
for any single contaminant (evaluated, for example, on a scale of
0 to 3). For each contaminant at the site, the ratio of the maximum
detected contaminant concentration to the benchmark concentration
would be determined as in the previously discussed concentration
factor. The concentration factor would then be evaluated using this
ratio as indicated in Table 30 (for example). The overall toxicity
138
-------
TABLE 30
ILLUSTRATIVE CONCENTRATION FACTOR*
Ratio (0/B) Factor Value
Greater than 10 5
1.0 - 9.99 4
0.1 - 0.99 3
0.01 - 0.099 2
0.001 - 0.0099 1
Less than 0.001 0
*Values are provided for illustrative purposes only.
B: Benchmark concentration.
0: Observed site concentration.
139
-------
factor value for each contaminant would then be determined for each
contaminant using Table 31 (for example). The highest calculated
toxlcity factor value for any single contaminant present in an
observed release would be used as the site toxicity factor value (as
in the current HRS toxicity factor). For sites lacking observed
releases, a presumptive concentration factor value of, for example,
three could be used as a default and the toxicity factor evaluated
as indicated.
There are several potential problems with this approach that
makes its implementation unreasonable. First, it may not be possible
to develop acceptable toxicity benchmarks for all CERCLA substances.
This problem affects the previously discussed options, as well,
although it is even more important in this option as the toxicity
factor is the major determinant of the waste characteristics
value. Problems in defining these benchmarks include differing
benchmarks for different effects from exposures to the same
contaminants, treatment of threshold versus nonthreshold contaminant/
effect pairings, establishment of de facto acceptable risk levels
for carcinogens, and treatment of contaminants/effects pairings with
nonlinear dose response relationships. Second, the feasibility of
defining a system for evaluating the severity of effects given
benchmark exceedances is also questionable. The development of such
a system would encounter the difficult problems of effectively
establishing the relative worth of different types of effects, for
140
-------
TABLE 31
ILLUSTRATIVE TOXICITY FACTOR MATRIX
Concentration Factor Value
Severity
Index
0
1
2
3
0
0
0
0
0
1
0
1
2
3
2
0
2
4
6
3
0
3
6
9
4
0
4
8
12
5
0
5
10
15
141
-------
example, losing ones eyesight as compared with contracting cancer.
Finally, this option would place too great an emphasis on the
concentration data. This approach relies heavily on the
representativeness of the concentration data; the weakest aspect of
the data. No provision would be Bade for reflecting the risks
associated with possibly every toxic contaminants found at the site
that were not detected in an observed release. Further, as with the
waste quantity factor, there would be no guarantee that the highest
detected concentration even exceeds the current, average exposure
concentration unless the data were truly representative. In such
circumstances, the toxicity factor value would probably understate the
risks to both maximum and "average" exposed individuals. Additionally,
as discussed previously, the present detected concentrations aay not
reflect future conditions and thus could understate (or overstate) the
risk from the site. In total, these considerations make the adoption
of this, otherwise attractive option, scientifically questionable.
7.3.3 Options in the Targets Category
Two options for reflecting concentration data in the targets
category were developed. The first is fairly simple. All of the
current target factor tables reflect the distance from the site to the
exposed population. One option for incorporating concentration data
would be to give added emphasis to those people who are known to be
exposed to contaminant concentrations that exceed background or
benchmarks by adjusting the values in the current MRS target distance
142
-------
factor value matrix (as illustrated in Table 32 for the ground water
pathway).
An alternate, more complex approach is described below. A two-
tiered approach could be adopted in evaluating the value for exposed
population. First, the number of people within a specified distance of
each sampling location with a concentration exceeding background and
benchmarks would be evaluated using a "currently exposed" population
factor table. The remaining people within another, greater specified
distance of the site would be evaluated using a second, "potentially
exposed" population factor table. The effective, per-person value
would be lower in the second table than in the first table. The
resulting "currently exposed" and "potentially exposed" values would
be added to form the population factor value. If the sum exceeds the
maximum for the factor (e.g., 30), the maximum factor value would be
used. An illustrative table for this option, using the air pathway,
is presented in Table 33. Using this example, if 350 people were
exposed to above benchmark contaminant concentrations arising from
the site, then a "currently exposed" population value of 24 would be
assigned. If an additional 15,000 people were potentially exposed
within a radius of 4 miles, then a "potentially exposed" population
value of 21 would be assigned. The sum of the exposed population
values (45) exceeds the maximum value (30) and therefore, a population
factor value of 30 would be assigned to the site.
143
-------
TABLE 32
ILLUSTRATIVE TABLE FOR EVALUATING POPULATION
(GROUND WATER PATHWAY)
Population Served
0
1-100
101-1,000
1,001-3,000
3,001-10,000
10,000+
0
0
30
35
40
40
40
2,000'
0
10
20
30
35
40
Distance
1 mile
0
8
16
24
32
35
to Nearest Well
1 mile
0
6
12
18
24
30
2-3 mile
0
4
8
12
16
20
3 mile
0
0
0
0
0
0
Adapted from 47 FR 31231 (July 16, 1982).
144
-------
TABLE 33
ILLUSTRATIVE TABLES FOR EVALUATING POPULATION (AIR PATHWAY)
Currently Exposed Population
Population within 1/2 mile* Value
0 0
1-10 15
11-100 18
101-300 21
301-500 24
501-1,000 27
1,000+ 30
Potentially Exposed Population
Distance to Population
from Hazardous Substance
0-4
miles
0
9
12
15
18
21
0-1
mile
0
12
15
18
21
24
0-1/2
mile
0
15
18
21
24
27
0-1/4
mile
0
18
21
24
27
30
Population
0
1-100
101-1,000
1,001-3,000
3,001-10,000
10,000+
*Total number of people residiag within 1/2 mile of any critical
location as well as transients such as workers in factories, offices,
restaurants, motels, or students. Critical locations are those
locations at which measured contaminant concentrations exceed
benchmarks and significantly exceed background.
145
-------
8.0 OVERALL CONCLUSIONS AND RECOMMENDATIONS
Several conclusions can be drawn from the previous discussion.
Concerning the quality of concentration data, the data currently
developed as part of site inspections is not of uniform quality and
is not generally of sufficient quality to support its expanded use
in the HRS. Further, within the constraints of a preliminary
assessment (PA) and site inspection (SI) program, it is unlikely
that concentration data that is representative in space and time
could be developed for most sites. The difficulties in developing
representative concentration data combine with resource and time
constraints to make development of such data infeasible at many
sites. Thus, in some cases, the concentration data developed during
site inspections would be of questionable scientific validity
overall. While it is possible to ensure that each individual datum
is scientifically valid, generally an insufficient number of samples
will be taken to ensure overall representativeness.
Despite these considerations, it is possible to employ
concentration data in an expanded fashion in the HRS. However, the
limitations of the data must be recognized and, thus, the uses of
the data must be limited accordingly. The options discussed
previously illustrate several appropriately limited methods for
using concentration data to more accurately reflect site risks
within the framework of the HRS.
147
-------
The adoption of any of these options would require changes in
the PA/SI program to ensure that the concentration data at all sites
across all media meet minimum quality standards. These changes would
include the development and distribution of uniform guidance for
preliminary assessments and site inspections. Such guidance should
include not only guidance on site sampling, sample handling, sample
analysis, and data interpretation but should also include guidance on
the development of background information necessary to develop a valid
sampling plan. The development of such plans should be required for
all, nonemergency, site inspections. Selected references that could
be employed in developing such guidance are provided in Table 34.
In order to ensure that the guidance is followed and quality
standards are maintained, a vigorous quality assurance (QA) and
quality control (QC) program should be implemented, particularly in
regard to site inspections not performed by EPA. Such a program
would encompass not only laboratory QA and the ongoing EPA
Headquarters QA of site documentation packages, but would also
include QA/QC of:
• Preliminary assessment and background site information.
• Site sampling plans and their execution.
* Sampling and handling procedures.
Without such a program, the concentration data developed will be of
varying quality and will too often be of insufficient quality to
warrant its use.
148
-------
TABLE 34
GUIDANCE ON CONCENTRATION DATA DEVELOPMENT:
SELECTIONS FROM THE LITERATURE
Multi-Media
Berg, Edward L., Handbook for Sampling and Sample Preservation of
Water and Wastewaters, (EPA-600/4-82-029), U.S. Environmental
Protection Agency, Cincinnati, OH, September 1982.
Ford, Patrick J., Paul J. Turina, and Douglas E. Seeley,
Characterization of Hazardous Waste Sites - A Methods Manual - Volume
II; Available Sampling Methods, Second Edition, (EPA-600/4-84-076),
U.S. Environmental Protection Agency, Las Vegas, NV, December 1984.
Ford, Patrick and Paul Turina, Characterization of Hazardous
Waste Sites - A Methods Manual - Volume I; Site Investigations,
(EPA-600/4-84-075), U.S. Environmental Protection Agency,
Las Vegas, NV, April 1985.
Gosse, Michelle et al., Continuing Releases at RCRA Facilities
Preliminary Draft, Remedial Investigation Guidance, (WR4818), GCA
Corporation, Bedford, MA, April 1986.
U.S. Environmental Protection Agency, Test Methods for Evaluating
Solid Waste; Physical/Chemical Methods, (SW-846), U.S. Environmental
Protection Agency, Washington, DC, May 1980.
U.S. Environmental Protection Agency, Guidance on Remedial
Investigations Under CERCLA, (EPA-540/G-85-002), U.S. Environmental
Protection Agency, Washington, DC, June 1985.
Groundwater
Barcelona, M. J. et al., Practical Guide for Ground-Water Sampling,
(EPA-600/2-85-104), U.S. Environmental Protection Agency, Ada, OK,
September 1985.
Claassen, Hans C., Guidelines and Techniques for Obtaining Water
Samples that Accurately Represent the Water Chemistry of an Aquifer,
(Open File Report 82-1024), U.S. Geological Survey, Lakewood, CO, 1982.
Gibb, James P., Rudolph M. Schuller, and Robert A. Griffin, Procedures
for the Collection of Representative Water Quality Data from Monitoring
Wells, (ISWS/COOP-7/81), Illinois State Water Survey. Champaign, IL,
1981.
-------
TABLE 34 (Continued)
Groundwater (Concluded)
Nacht, S. J., "Monitoring Sampling Protocols," Ground Water Monitoring
Review, Vol. 3, No. 3, Summer, 1983, pp. 23-29.
Popkin, Barney P., "Guidelines for Ground-Water Quality Assessments for
Hazardous Waste Facilities," Ground Water Monitoring Review, Vol. 3,
No. 2, Spring 1983, pp. 65-70.
Scalf, Marion R. et al., Manual of Ground-Water Quality Sampling
Procedures, (EPA-600/2-81-160), U.S. Environmental Protection Agency,
Ada, OK, September 1981.
Sisk, Steven W., NEIC Manual for Groundwater/Subsurface Investigations
at Hazardous Waste Sites, (EPA-330/9-81-Q02), U.S. Environmental
Protection Agency, Denver, CO, July 1981.
U.S. Environmental Protection Agency, Procedure Manual for Ground
Water Monitoring at Solid Waste Disposal Facilities, (SW-611), U.S.
Environmental Protection Agency, Washington, DC, 1977.
U.S. Environmental Protection Agency, RCRA Ground-Water Monitoring
Technical Enforcement Guidance Document, (Draft), (EPA-600/4-84-076),
U.S. Environmental Protection Agency, Washington, DC, August 1985.
Air
Riggin, R. M., Technical Assistance Document for Sampling and Analysis
of Toxic OrganicTompounds in Ambient Air, (EPA-600/4-83-027), U.S.
Environmental Protection Agency, Research Triangle Park, NC, June 1983.
Turpin, Rodney D., "ERT's Air Monitoring Guides for Uncontrolled
Hazardous Waste Sites," Proceedings of the Fourth National Conference
on Management of Uncontrolled Hazardous Waste Sites, Held on October 31-
November 2, 1983 in Washington, DC, Hazardous Materials Control
Research Institute, Silver Spring, MD, 1983, pp. 82-84.
U.S. Environmental Protection Agency, Field Standard Operating
Procedure for Air Surveillance, (F.S.O.P. 8), U.S. Environmental
Protection Agency, Washington, DC, undated.
Soil
Mason, Benjamin J., Preparation of Soil Sampling Protocol; Techniques
and Strategies, (EPA-600/4-83-020), U.S. Environmental Protection
Agency, Las Vegas, NV, May 1983.
150
-------
TABLE 34 (Concluded)
Wastes
deVera, Emil R. et al., Samplers and Sampling Procedures for Hazardous
Waste Streams, (EPA-600/2-80-018), U.S. Environmental Protection
Agency, Cincinnati, OH, January 1980.
Hanisch, R. C. and M. A. McDevitt, Protocols for Sampling and Analysis
of Surface Impoundments and Land Treatment/Disposal Sites for VOCs,
(DCN 84-222-078-11-12), Radian Corporation, Austin, TX, September 28,
1984.
Laboratory
Plumb, Russel H., Jr., Characterization of Hazardous Waste Sites - A
Methods Manual - Volume III; Available Laboratory Analytical Methods,
(EPA-600/4-84-038), U.S. Environmental Protection Agency, Las Vegas,
NV, May 1984.
B ioassessment
Porcella, D. B., Protocol for Bioassessment of Hazardous Waste Sites,
(EPA-600/2-83-054), U.S. Environmental Protection Agency, Corvallis,
OR, July 1983.
151
-------
9.0 REFERENCES AND BIBLIOGRAPHY
Absalon, John R. and Robert C. Starr, "Practical Aspects of Ground
Water Monitoring at Existing Disposal Sites," Proceedings of the
First National Conference on Management of Uncontrolled Hazardous
Waste Sites, Held on October 15-17, 1980 in Washington, DC,
Hazardous Materials Control Research Institute, Silver Spring, MD,
1980, pp. 53-58.
Adams, William M., Stephen W. Wheatcraft, and John W. Hess,
"Downhole Sensing Equipment for Hazardous Waste Site Investigations,'
Proceedings of the Fourth National Conference on Management of
Uncontrolled Hazardous Waste Sites, Held on October 31-November 2,
1983 in Washington, DC, Hazardous Materials Control Research
Institute, Silver Spring, MD, 1983, pp. 108-113.
Aller, Linda et al., DRASTIC; A Standardized System for Evaluating
Ground Water Pollution Potential Using Hydrogeologic Settings,
(EPA-600/2-85-018), U.S. Environmental Protection Agency, Ada, OK,
May 1985.
American Conference of Governmental Industrial Hygienists, Threshold
Limit Values and Biological Indices for 1985-1986, American
Conference of Governmental Industrial Hygienists, Cincinnati, OH,
1985.
Amster, Michael B., Nasrat Hijazi, and Rosalind Chan, "Real Time
Monitoring of Low Level Air Contaminants for Hazardous Waste Sites,"
Proceedings of the Fourth National Conference on Management of
Uncontrolled Hazardous Waste Sites, Held on October 31-November 2,
1983 in Washington, DC, Hazardous Materials Control Research
Institute, Silver Spring, MD, 1983, pp. 98-99.
Astle, Alice D., Richard A. Duffee, and Alexander R. Stankunas,
"Estimating Vapor and Odor Emission Rates from Hazardous Waste
Sites," Proceedings of the Third National Conference on Management
of Uncontrolled Hazardous Waste Sites, Held on November 29-
December 1, 1982 in Washington, DC, Hazardous Materials Control
Research Institute, Silver Spring, MD, 1982, pp. 326-330.
Baker, Lynton W., An Evaluation of Screening Models for Assessing
Toxic Air Pollution Downwind of Hazardous Waste Landfills, Masters
Thesis, Office of Graduate Studies and Research, San Jose State
University, San Jose, CA, May 1985.
153
-------
Balfour, W. David, Bart M. Elkund, and Shelly J. Williamson,
"Measurement of Volatile Organic Emissions from Subsurface
Contaminants," Proceedings of the Fifth National Conference on
Management of Uncontrolled Hazardous Waste Sites, Held on
November 7-9, 1984 in Washington, DC, Hazardous Materials Control
Research Institute, Silver Spring, MD, 1984, pp. 77-80.
Barcelona, M.J. et al., Practical Guide for Ground-Water Sampling,
(EPA-60U/2-85-104), U.S. Environmental Protection Agency, Ada, OK,
September 1985.
Barcelona, Michael J., "Chemical Problems in Ground-Water Monitoring
Programs," Proceedings of the 3rd National Symposium on Aquifer
Restoration and Ground Water Monitoring, Held on May 25-27, 1983 in
Columbus, OH, D. M. Nielsen, ed., Water Well Journal Publishing
Company, Worthington, OH, 1983, pp. 263-271.
Ben-Hur, David, James S. Smith, and Michael J. Urban, "Application
of Mobile MS/MS to Hazardous Waste Site Investigation," Proceedings
of the Fifth National Conference on Management of Uncontrolled
Hazardous Waste Sites, Held on November 7-9, 1984 in Washington, DC,
Hazardous Materials Control Research Institute, Silver Spring, MD,
1984, pp. 53-58.
Benson, Richard C. and Robert A. Glaccum, "Site Assessment:
Improving Confidence Levels with Surface Remote Sensing,"
Proceedings of the First National Conference on Management of
Uncontrolled Hazardous Waste Sites, Held on October 15-17, 1980 in
Washington, DC, Hazardous Materials Control Research Institute,
Silver Spring, MD, 1980, pp. 59-65.
Benson, Richard C., Charles M. Thomas, and Lynn Yuhr, "An Approach
to Long-Term Groundwater Monitoring," Proceedings of the Sixth
National Conference on Management of Uncontrolled Hazardous Waste
Sites, Held on November 4-6, 1985 in Washington, DC, Hazardous
Materials Control Research Institute, Silver Spring, MD, 1985,
pp. 112-115.
Benson, Richard, Robert Glaccum, and Paul Beam, "Minimizing Cost and
Risk in Hazardous Waste Site Investigations Using Geophysics,"
Proceedings of the Second National Conference on Management of
Uncontrolled Hazardous Waste Sites, Held on October 28-30,1981 in
Washington, DC, Hazardous Materials Control Research Institute,
Silver Spring, MD, 1981, pp. 84-88.
Berg, Edward L., Handbook for Sampling and Sample Preservation of
Water and Wastewa'ters, (EPA-600/4-82-029), U.S. Environmental
Protection Agency, Cincinnati, OH, September 1982.
154
-------
Berhardt, David E., Richard 0. Gilbert, and Paul B. Hahn,
"Comparison of Soil-Sampling Techniques for Plutonium at Rocky
Flats," TRAN-STAT Statistics for Environmental Studies, Number 22,
(PNL-SA-11034, DE83005561), Richard 0. Gilbert, coordinator, Batelle
Memorial Institute Pacific Northwest Laboratory, Richland, WA,
January 1983.
Blackman, William C. et al., "Chemical Composition of Drum Samples
from Hazardous Waste Sites," Proceedings of the Fifth National
Conference on Management of Uncontrolled Hazardous Waste Sites, Held
on November 7-9, 1984 in Washington, DC, Hazardous Materials Control
Research Institute, Silver Spring, MD, 1984, pp. 39-44.
Bruck, John M., Eugene W. Koesters, and William R. Parker, "Air
Monitoring at a Major Hazardous Waste Cleanup Site: Objectives/
Strategy/Results," Proceedings of the Fifth National Conference on
Management of Uncontrolled Hazardous Waste Sites, Held on
November 7-9, 1984 in Washington, DC, Hazardous Materials Control
Research Institute, Silver Spring, MD, 1984, pp. 72-76.
Bruehl, Donald H., Neville K. Chung, and Warren F. Diesl, "Geologic
Studies of Industrially-Related Contamination: Soil and Ground
Water Investigations," Proceedings of the First National Conference
on Management of Uncontrolled Hazardous Waste Sites, Held on
October 15-17, 1980 in Washington, DC, Hazardous Materials Control
Research Institute, Silver Spring, MD, 1980, pp. 78-84.
Bryan, Robert J., "Ambient Air Quality Surveillance," Air Pollution,
Third Edition Volume III Measuring, Monitoring, and Surveillance of
Air Pollution, Arthur C. Stern, ed., Academic Press, New York, NY,
1976, pp. 343-392.
Burger, Carol J. and Lawrence M. Kushner, Hazard Ranking System
Issue Analysis; An Assessment of the HRS Target Distance Limit for
Surface Water, (MTR-86W84), The MITRE Corporation, McLean, VA,
October 1986.
Caldwell, Steven, U.S. Environmental Protection Agency, Washington,
DC, personal communications to Thomas Wolfinger, The MITRE
Corporation, 1986.
California Department of Health Services, California Air Resources
Board, and South Coast Air Quality Management District, Ambient Air
Monitoring and Health Risk Assessment for Suspect Human Carcinogens
Around the BKK Landfill in West Covina, California Department of
Health Services, March 1983.
155
-------
Chang, Ruth and Robert D. Stephens, "The Application of Mass
Selective Detector for the Screening of Environmental Pollutants in
Toxic Waste," Proceedings of the Sixth National Conference on
Management of Uncontrolled Hazardous Waste Sites, Held on
November 4-6, 1985 in Washington, DC, Hazardous Materials Control
Research Institute, Silver Spring, MD, 1985, pp. 102-106.
Claassen, Hans C., Guidelines and Techniques for Obtaining Water
Samples that Accurately Represent the Water Chemistry of an Aquifer,
(Open File Report 82-1024), U.S. Geological Survey, Lakewood, CO,
1982.
Clay, Paul F. and Thomas M. Spittler, "The Use of Portable
Instruments in Hazardous Waste Site Characterizations," Proceedings
of the Third National Conference on Management of Uncontrolled
Hazardous Waste Sites, Held on November 29-December 1, 1982 in
Washington, DC, Hazardous Materials Control Research Institute,
Silver Spring, MD, 1982, pp. 40-44.
Cook, David K., "Selection of Monitoring Well Locations in East and
North Woburn, Massachusetts," Proceedings of the Second National
Conference on Management of Uncontrolled Hazardous Waste Sites, Held
on October 28-30, 1981 in Washington, DC, Hazardous Materials
Control Research Institute, Silver Spring, MD, 1981, pp. 63-69.
Core, John E., Receptor Modeling Technical Series, Volume 1;
Overview of Receptor Model Application to Particulate Source
Apportionment, (EPA-450/4-81-016a), U.S. Environmental Protection
Agency, Research Triangle Park, NC, July 1981.
Cox, Robert D., Kenneth J. Baughman, and Ronald F. Earp, "A
Generalized Screening and Analysis Procedure for Organic Emissions
from Hazardous Waste Disposal Sites," Proceedings of the Third
National Conference on Management of Uncontrolled Hazardous Waste
Sites, Held on November 29-December 1, 1982 in Washington, DC,
Hazardous Materials Control Research Institute, Silver Spring, MD,
1982, pp. 58-62.
Crawley, W. W., R. L. Shiver, and D. C. Anderson, "In Situ Sampling
of Hazardous Waste Surface Impoundments," Proceedings of the Sixth
National Conference on Management of Uncontrolled Hazardous Waste
Sites, Held on November 4-5, 1985 in Washington, DC, Hazardous
Materials Control Research Institute, Silver Spring, MD, 1985,
pp. 80-83.
Cullis, C. F. and J. G. Firth, Detection and Measurement of
Hazardous Gases, Heinemann, London, 1981.
156
-------
Dawson, Gaynor W., Jill M. Meuser, and Mary C. Lilga, Dioxin
Transport form Contaminated Sites to Exposure Locations; A
Methodology for Calculating Conversion Factors, (EPA-600/8-85-012),
U.S. Environmental Protection Agency, Washington, DC, June 1985.
deVera, Emil R. et al., Samplers and Sampling Procedures for
Hazardous Waste Steams, (EPA-600/2-80-018), U.S. Environmental
Protection Agency, Cincinnati, OH, January 1980.
Devary, Joseph L. and Ronald Schalla, "Improved Methods of Flow
System Characterization," Proceedings of the Fourth National
Conference on Management of Uncontrolled Hazardous Waste Sites, Held
on October 31-November 2, 1983 in Washington, DC, Hazardous
Materials Control Research Institute, Silver Spring, MD, 1983,
pp. 117-122.
Di Nitto, Richard G., William R. Norman, and M. Margret Hanley, "An
Approach to Investigating Groundwater Contaminant Movement in
Bedrock Aquifers: Case Histories," Proceedings of the Third
National Conference on Management of Uncontrolled Hazardous Waste
Sites, Held on November 29-December 1, 1982 in Washington, DC,
Hazardous Materials Control Research Institute, Silver Spring, MD,
1982, pp. 111-117.
Driscoll, Fletcher G., Groundwater and Wells Second Edition, Johnson
Division, St. Paul, MN, 1986.
Driscoll, J. N., A. G. Wilshire, and I. W. Bodenrader, "New
Continuous Monitoring Systems for Measurement of Hazardous
Pollutants," Proceedings; National Symposium on Recent Advances in
Pollutant Monitoring of Ambient Air and Stationary Sources, (EPA-600/
9-84-019), Held on May 8-10, 1984 in Raleigh, NC, U.S. Environmental
Protection Agency, Research Triangle Park, NC, November 1984,
pp. 91-95.
Duvel, William A., "Practical Interpretation of Groundwater
Monitoring Results," Proceedings of the Third National Conference on
Management of Uncontrolled Hazardous Waste Sites, Held on
November 29-December 1, 1982 in Washington, DC, Hazardous Materials
Control Research Institute, Silver Spring, MD, 1982, pp. 86-90.
Ehrenfeld, John and Jeffery Bass, Handbook for Evaluating Remedial
Action Technology Plans, (EPA-600/2-83-076), U.S. Environmental
Protection Agency, Washington, DC, August 1983.
157
-------
Engels, J. L. et al., "Survey of Mobile Laboratory Capabilities and
Configurations," Proceedings of the Fifth National Conference on
Management of Uncontrolled Hazardous Waste Sites, Held on
November 7-9, 1984 in Washington, DC, Hazardous Materials Control
Research Institute, Silver Spring, MD, 1984, pp. 45-48.
Evans, R. B., R. C. Benson, and J. Rizzo, "Systematic Hazardous
Waste Site Assessments," Proceedings of the Third National
Conference on Management of Uncontrolled Hazardous Waste Sites, Held
on November 29-December 1, 1982 in Washington, DC, Hazardous
Materials Control Research Institute, Silver Spring, MD, 1982,
pp. 17-22.
Everett, L. G. et al., Monitoring Groundwater Quality; Methods and
Costs, (EPA-600/4-76-023), U.S. Environmental Protection Agency, Las
Vegas, NV, May 1976.
Everett, L. G. et al., "Vadose Zone Monitoring Concepts at
Landfills, Impoundments, and Land Treatment Disposal Areas,"
Proceedings of the Third National Conference on Management of
Uncontrolled Hazardous Waste Sites, Held on November 29-December 1,
1982 in Washington, DC, Hazardous Materials Control Research
Institute, Silver Spring, MD, 1982, pp. 100-106.
Everett, L. G., L. G. Wilson, and E. W. Hoylman, Vadose Zone
Monitoring for Hazardous Waste Sites, (EPA-600/X-83-064), U.S.
Environmental Protection Agency, Las Vegas, NV, October, 1983.
Everett, Lome, Groundwater Monitoring, General Electric Company,
Schenectady, NY, 1980.
Everett, Lorne G., "Monitoring in the Zone of Saturation," Ground
Water Monitoring Review, Vol. 1, No. 1, Spring 1981, pp. 38-41.
Farthing, William E., "Particle Sampling and Measurement,"
Environmental Science and Technology, Vol. 16, No. 4, April 1982,
pp. 237A-244A.
Fellows, Charles R. and James H. Sullivan, "Refined Strategies for
Abandoned Site Discovery and Assessment," Proceedings of the Fourth
National Conference on Management of Uncontrolled Hazardous Waste
Sites, Held on October 31-November 2, 1983 in Washington, DC,
Hazardous Materials Control Research Institute, Silver Spring, MD,
1983, pp. 37-42.
Fetter, C. W., "Potential Sources of Contamination in Ground-Water
Monitoring," Ground Water Monitoring Review, Vol. 3, No. 2, Spring
1983, pp. 60-64.
158
-------
Fitzgerald, Jerry M., Hazard Ranking System Issue Analysis:
Laboratory Limits of Detection, (MTR-86W77), The MITRE Corporation,
McLean, VA, July 1986.
Flatman, George T., "Using Geostatistics in Assessing Lead
Contamination Near Smelters," Environmental Sampling for Hazardous
Wastes, (ACS Symposium Series 267), Glenn E. Schweitzer and John A.
Santolucito, eds., American Chemical Society, Washington, DC, 1984,
pp. 43-52.
Ford, Karl and Paul Gurba, "Health Risk Assessments for Contaminated
Soils," Proceedings of the Fifth National Conference on Management
of Uncontrolled Hazardous Waste Sites, Held on November 7-9, 1984 in
Washington, DC, Hazardous Materials Control Research Institute,
Silver Spring, MD, 1984, pp. 230-231.
Ford, Patrick J., Paul J. Turina, and Douglas E. Seeley,
Characterization of Hazardous Waste Sites - A Methods Manual -
Volume II; Available Sampling Methods, Second Edition, (EPA-600/
4-84-076), U.S. Environmental Protection Agency, Las Vegas, NV,
December 1984.
Ford, Patrick and Paul Turina, Characterization of Hazardous Waste
Sites - A Methods Manual - Volume I: Site Investigations, (EPA-600/
4-84-075), U.S. Environmental Protection Agency, Las Vegas, NV,
April 1985.
Friedman, Paul H. and Duane Geuder, "Quality Control Attributes of
Process Analytical Data," Proceedings of the Fifth National
Conference on Management of Uncontrolled Hazardous Waste Sites, Held
on November 7-9, 1984 in Washington, DC, Hazardous Materials Control
Research Institute, Silver Spring, MD, 1984, pp. 29-34.
Furst, George A., Valerie Tillinghast, and Thomas Spittler,
"Screening for Metals at Hazardous Waste Sites: A Rapid Cost-
Effective Technique Using X-Ray Fluorescence," Proceedings of the
Sixth National Conference on Management of Uncontrolled Hazardous
Waste Sites, Held on November 4-6, 1985 in Washington, DC, Hazardous
Materials Control Research Institute, Silver Spring, MD, 1985,
pp. 93-96.
GCA Corporation, Guidelines for Air Quality Maintenance and
Analysis. Volume 11; Air Quality Monitoring and Data Analysis,
(EPA-450/4-74-012), U.S. Environmental Protection Agency, Research
Triangle Park, NC, September 1974.
159
-------
Geraghty and Miller, Inc. and American Ecology Services, Inc., The
Fundamentals of Ground-Water Contamination, Geraghty and Miller,
Inc., Tampa, FL, 1985.
Gerstein, Robert E., Hazard Ranking System Issue Analysis;
Geohydrology, (MTR-86W62), The MITRE Corporation, McLean, VA,
October 1986.
Gibb, James P., Rudolph M. Schuller, and Robert A. Griffin,
Procedures for the Collection of Representative Water Quality Data
from Monitoring Wells, (ISWS/COOP-7/81), Illinois State Water
Survey, Champaign, IL, 1981.
Gilbert, Richard 0., "Field Sampling Designs, Simple Random and
Stratified Random Sampling," IRAN-STAT Statistics for Environmental
Studies, Number 24, (PNL-SA-11551, DE83016826), Richard 0. Gilbert,
coordinator, Batelle Memorial Institute Pacific Northwest
Laboratory, Richland, WA, August 1983.
Gilbert, Richard 0., "Field Sampling Designs: Systematic Sampling,"
TRAN-STAT Statistics for Environmental Studies, Number 26,
(PNL-SA-12180, DE84-008651), Richard 0. Gilbert, coordinator,
Batelle Memorial Institute Pacific Northwest Laboratory, Richland,
WA, April 1984.
Gillham, R. W. et al., Groundwater Monitoring and Sample Bias, (API
Publication 4367), American Petroleum Institute, Washington, DC,
June 1983.
Gilkeson, Robert H. and Keros Cartwright, "The Application of
Surface Electrical and Shallow Geothermic Methods in Monitoring
Network Design," Ground Water Monitoring Review, Vol. 3, No. 3,
Summer 1983, pp. 30-42.
Glaccum, Robert A., Richard C. Benson, and Michael R. Noel,
"Improving Accuracy and Cost-Effectiveness of Hazardous Waste Site
Investigations," Ground Water Monitoring Review, Vol. 2, No. 3,
Summer 1982, pp. 36-40.
Gleit, Alan, "Estimation for Small Normal Data Sets with Detection
Limits," Environmental Science and Technology, Vol. 19, No. 12,
December 1985, pp. 1201-1206.
Gosse, Michelle et al., Continuing Releases at RCRA Facilities
Preliminary Draft, Remedial Investigation Guidance, (WR4818), GCA
Corporation, Bedford, MA, April 1986.
160
-------
Gruenfeld, Michael et al., "Management of Analytical Laboratory
Support at Uncontrolled Hazardous Waste Sites," Proceedings of the
Second National Conference on Management of Uncontrolled Hazardous
Waste Sites, Held on October 28-30, 1981 in Washington, DC, Hazardous
Materials Control Research Institute, Silver Spring, MD, 1981,
pp. 96-102.
Gurka, D. F. et al., "Analytical and Quality Control Procedures for
the Uncontrolled Hazardous Waste Sites Contract Program," Proceedings
of the Third National Conference on Management of Uncontrolled
Hazardous Waste Sites, Held on November 29-December 1, 1982 in
Washington, DC, Hazardous Materials Control Research Institute,
Silver Spring, MD, 1982, pp. 45-51.
Hagger, Christopher and Paul F. Clay, "Hydrogeological Investigation
of an Uncontrolled Hazardous Waste Site," Proceedings of the Second
National Conference on Management of Uncontrolled Hazardous Waste
Sites, Held on October 28-30, 1981 in Washington, DC, Hazardous
Materials Control Research Institute, Silver Spring, MD, 1981,
pp. 45-51.
Hanisch, R. C. and M. A. McDevitt, Protocols for Sampling and
Analysis of Surface Impoundments and Land Treatment/Disposal Sites
for VOCs, (DCN 84-222-078-11-12), Radian Corporation, Austin, TX,
September 28, 1984.
Harman, H. Dan, Jr., "Detailed Stratigraphic and Structural Control:
The Keys to Complete and Successful Geophysical Surveys of Hazardous
Waste Sites," Proceedings of the National Conference on Hazardous
Wastes and Hazardous Materials, Held on March 4-6, 1986 in Atlanta,
GA, Hazardous Materials Control Research Institute, Silver Spring,
MD, 1986, pp. 19-21.
Harman, H. Dan, Jr. and Shane Hitchcock, "Cost Effective Preliminary
Leachate Monitoring at an Uncontrolled Hazardous Waste Site,"
Proceedings of the Third National Conference on Management of
Uncontrolled Hazardous Waste Sites, Held on November 29-December 1,
1982 in Washington, DC, Hazardous Materials Control Research
Institute, Silver Spring, MD, 1982, pp. 97-99.
Harrison, Roy M., Recent Advances in Air Pollution Analysis,
University of Lancaster, Lancaster, England, undated.
Hatayama, Howard K., "Special Sampling Techniques Used for
Investigating Uncontrolled Waste Sites in California," Proceedings of
the Second National Conference on Management of Uncontrolled Hazardous
Waste Sites, Held on October 28-30, 1981 in Washington. DC. Hazardous
Materials Control Research Institute, Silver Spring, MD, 1981,
pp. 149-153.
161
-------
Ho, J. S-Y, "Effect of Sampling Variables on Recovery of Volatile
Organics in Water," Journal of the American Water Works Association,
December 1983, pp. 583-586.
Holdren, M. W., D. L. Smith, and R. N. Smith, Comparison of Ambient
Air Sampling Techniques for Volatile Organic Compounds,(EPA-600/
4-85-067), U.S. Environmental Protection Agency, Research Triangle
Park, NC, October 1985.
Hwang, Jack C. et al., "Feasibility Studies of Groundwater Pollution
Source Identification from Actual Field Monitoring Well Data,"
Proceedings of the Fifth National Conference on Management of
Uncontrolled Hazardous Waste Sites, Held on November 7-9, 1984 in
Washington, DC, Hazardous Materials Control Research Institute,
Silver Spring, MD, 1984, pp. 1-5.
Hwang, Seong T., "Statistical Considerations in Groundwater
Monitoring," Proceedings of the Fifth National Conference on
Management of Uncontrolled Hazardous Waste Sites, Held on
November 7-9, 1984 in Washington, DC, Hazardous Materials Control
Research Institute, Silver Spring, MD, 1984, pp. 346-349.
laccarino, T. et al., "A Superfund Site Atmospheric Study:
Application to Remedial Response Decision-Making," Proceedings of
the Fifth National Conference on Management of Uncontrolled
Hazardous Waste Sites, Held on November 7-9, 1984 in Washington, DC,
Hazardous Materials Control Research Institute, Silver Spring, MD,
1984, pp. 66-67.
Isaacson, Peter J., William P. Eckel, and John F. Fisk, "Low
Occurrence Compounds: Analytical Problem or Environmental
Process?," Proceedings of the Sixth National Conference on
Management of Uncontrolled Hazardous Waste Sites, Held on
November 4-6, 1985 in Washington, DC, Hazardous Materials Control
Research Institute, Silver Spring, MD, 1985, pp. 130-135.
Jacot, Brian J., "OVA Screening at a Hazardous Waste Site,"
Proceedings of the Fourth National Conference on Management of
Uncontrolled Hazardous Waste Sites, Held on October 31-November 2,
1983 in Washington, DC, Hazardous Materials Control Research
Institute, Silver Spring, MD, 1983, pp. 76-78.
Janisz, Amelia A. and W. Scott Butterfield, "Biological Sampling at
Abandoned Hazardous Waste Sites," Proceedings of the Third National
Conference on Management of Uncontrolled Hazardous Waste Sites, Held
on November 29-December 1, 1982 in Washington, DC, Hazardous
Materials Control Research Institute, Silver Spring, MD, 1982,
pp. 52-56.
162
-------
Johnson, Michael G., "The Stratified Sample Thief - A Device for
Sampling Unknown Fluids," Proceedings of the Second National
Conference on Management of Uncontrolled Hazardous Waste Sites, Held
on October 28-30, 1981 in Washington, DC, Hazardous Materials
Control Research Institute, Silver Spring, MD, 1981, pp. 154-157.
Journel, Andre G., "New Ways of Assessing Spatial Distributions of
Pollutants," Environmental Sampling for Hazardous Wastes, (ACS
Symposium Series 267), Schweitzer, Glenn E. and John A. Santolucito,
eds., American Chemical Society, Washington, DC, 1984, pp. 109-118.
JRB Associates, Remedial Action Costing Procedures Manual, (Draft),
U.S. Environmental Protection Agency, Cincinnati, OH, September 1985.
K. B. Brown and Associates, Inc., Hazardous Waste Land Treatment,
(SW-874), U.S. Environmental Protection Agency, Washington, DC,
September 1980.
Kaczmar, Swiatoslav, Edwin C. Tifft, Jr., Cornelius B. Murphy, Jr.,
"Site Assessment Under CERCLA: 'The Importance of Distinguishing
Hazard from Risk'," Proceedings of the Fifth National Conference on
Management of Uncontrolled Hazardous Waste Sites, Held on
November 7-9, 1984 in Washington, DC, Hazardous Materials Control
Research Institute, Silver Spring, MD, 1984, pp. 221-225.
Katz, Morris, "Advances in the Analysis of Air Contaminants,"
Journal of the Air Pollution Control Association, Vol. 30, No. 5,
May 1980, pp. 528-557.
Kazmann, R. G., "An Introduction to Ground-Water Monitoring," Ground
Water Monitoring Review, Vol. 1, No. 1, Spring 1981, pp. 28-29.
Keith, Lawrence H. et al., "Principles of Environmental Analysis,
Analytical Chemistry, Vol. 55, 1983, pp. 2210-2218.
Keith, Susan J. et al., "Dealing with the Problem of Obtaining
Accurate Ground-Water Quality Analytical Results," Proceedings of
the 3rd National Symposium on Aquifer Restoration and Ground Water
Monitoring, Held on May 25-27, 1983, in Columbus, OH, D. M. Nielsen,
ed., Water Well Journal Publishing Company, Worthington, OH, 1983a,
pp. 272-283.
Keith, Susan J. et al., "Sources of Spatial-Temporal Variability in
Ground Water Quality Data and Methods of Control," Ground Water
Monitoring Review, Vol. 3, No. 2, Spring 1983b, pp. 21-32.
163
-------
Kerfoot, H. B., J. A. Kohout, and E. N. Amick, "Detection and
Measurement of Groundwater Contamination by Soil-Gas Analysis,"
Proceedings of the National Conference on Hazardous Wastes and
Hazardous Materials, Held on March 4-6, 1986 in Atlanta, GA,
Hazardous Materials Control Research Institute, Silver Spring, MD,
1986, pp. 22-26.
Kirchmer, Cliff J., "Quality Control in Water Analyses,"
Environmental Science and Technology, Vol. 17, No. 4, 1983,
pp. 174A-181A.
Kirchmer, Cliff J., Margaret C. Winter, and Barbara A. Kelly,
"Factors Affecting the Accuracy of Quantitative Analyses of Priority
Pollutants Using GC/MS," Environmental Science and Technology,
Vol. 17, No. 7, 1983, pp. 396-4U1.
Kolb, Sean, U.S. Environmental Protection Agency, Washington, DC,
personal communication to Thomas Wolfinger, The MITRE Corporation,
September 1986.
Kolybe, Albert C., Jr., "The Need for Accuracy in a Regulatory
Agency," Accuracy in Trace Analysis: Sampling, Sample Handling,
Analysis - Volumes I and II, (NBS Special Publication 422),
D. Philip LaFleur, ed., National Bureau of Standards, Gaithersburg,
MD, 1974, pp. 3-8.
LaFleur, Philip D., ed., Accuracy in Trace Analysis; Sampling,
Sample Handling, Analysis - Volumes I and II, (NBS Special
Publication 422), Proceedings of the 7th Material Research
Symposium, Held on October 7-11, 1987 in Gaithersburg, MD, National
Bureau of Standards, Gaithersburg, MD, 1974.
Lappala, Eric G. and Glenn M. Thompson, "Detection of Groundwater
Contamination by Shallow Soil Gas Sampling in the Vadose Zone Theory
and Applications, " Proceedings of the Fifth National Conference on
Management of Uncontrolled Hazardous Waste Sites, Held on
November 7-9, 1984 in Washington, DC, Hazardous Materials Control
Research Institute, Silver Spring, MD, 1984, pp. 20-28.
Liggett, Walter, "Detecting Elevated Contamination by Comparisons
with Background," Environmental Sampling for Hazardous Wastes, (ACS
Symposium Series 267), Glenn E. Schweitzer and John A. Santolucito,
eds., American Chemical Society, Washington, DC, 1984, pp. 119-128.
Lipsky, David and Brian Jacot, "Hazardous Emissions from Sanitary
Landfills," Proceedings for the 78th Annual Meeting of the Air
Polluation Control Association, Held on June 16-21, 1985 in Detroit,
MI, Air Pollution Control Association, Pittsburgh, PA, 1985.
164
-------
Liu, M. K. and J. Arvin, Methodology for the Design of an Optimal Air
Quality Monitoring Network, (EPA-600/4-81-002), U.S. Environmental
alogy
V-600/
Protection Agency, Las Vegas, NV, February 1981.
Liu, M. K. et al., "Methodology for Designing Air Quality Networks:
I Theoretical Aspects," Environmental Monitoring and Assessment,
Vol. 6, 1986, pp. 1-11.
Lodge, James P., Jr., "Accuracy in Air Sampling," Accuracy in Trace
Analysis; Sampling, Sample Handling, Analysis - Volumes I and II,
(NBS Special Publication 422), Philip D. LaFleur, ed., National
Bureau of Standards, Gaithersburg, MD, 1974, pp. 311-320.
Mack, James P. and Thomas J. Morahan, "Equipment for Data Collection
at Hazardous Waste Sites - An Overview for Environmental Management
Professionals," Proceedings of the National Conference on Hazardous
Wastes and Hazardous Materials, Held on March 4-6, 1986 in Atlanta,
GA, Hazardous Materials Control Research Institute, Silver Spring,
MD, 1986, pp. 1-7.
Mason, Benjamin J., Preparation of Soil Sampling Protocol;
Techniques and Strategies, (EPA-600/4-83-020), U.S. Environmental
Protection Agency, Las Vegas, NV, May 1983.
McKee, C. R. and A. C. Bumb, "The Importance of Unsaturated Flow
Parameters in Designing a Monitoring System for a Hazardous Waste
Site," Proceedings of the National Conference on Hazardous Wastes
and Environmental Emergencies, Held on March 12-14, 1984 in Houston,
TX, Hazardous Materials Control Research Institute, Silver Spring,
MD, 1984, pp. 50-58.
McKown, Gary L., Ronald Schalla, and C. Joseph English, "Effects of
Uncertainties of Data Collection on Risk Assessment," Proceedings of
the Fifth National Conference on Management of Uncontrolled Hazardous
Waste Sites, Held on November 7-9, 1984 in Washington, DC, Hazardous
Materials Control Research Institute, Silver Spring, MD, 1984,
pp. 283-286.
Mernitz, Scott, Roger Olsen, and Ken Staible, "Use of a Portable
X-Ray Analyzer and Geostatistical Methods to Detect and Evaluate
Hazardous Metals in Mine/Mill Tailings," Proceedings of the Sixth
National Conference on Management of Uncontrolled Hazardous Waste
Sites, Held on November 4-6, 1985 in Washington, DC, Hazardous
Materials Control Research Institute, Silver Spring, MD, 1985,
pp. 107-111.
165
-------
Mooney, Gregory A. and Russell W. Bartley, "Characterization of
Organic Wastes for Evaluation of Remedial Action Alternatives,"
Proceedings of the Fifth National Conference on Management of
Uncontrolled Hazardous Waste Sites, Held on November 7-9, 1984 in
Washington, DC, Hazardous Materials Control Research Institute,
Silver Spring, MD, 1984, pp. 35-38.
Moore, Stephen F. and Dennis B. Mclaughlin, "Mapping Contaminated
Soil Plumes by Kriging," Proceedings of the First National Conference
on Management of Uncontrolled Hazardous Waste Sites, Held on
October 15-17, 1980 in Washington, DC, Hazardous Materials Control
Research Institute, Silver Spring, MD, 1980, pp. 66-70.
Moss, M. E. and G. D. Tasker, "Progress in the Design of Hydrologic-
Data Networks," Reviews of Geophysics and Space Physics, Vol. 17,
No. 6, September 1979, pp. 1298-1307.
Morin, Joanne O'Neill, "Development and Application of an Analytical
Screening Program to Superfund Activities," Proceedings of the Sixth
National Conference on Management of Uncontrolled Hazardous Waste
Sites, Held on November 4-6, 1985 in Washington, DC, Hazardous
Materials Control Research Institute, Silver Spring, MD, 1985,
pp. 97-101.
Moylan, Craig A., "Hazardous Materials Training for First Responders:
Concepts and Future Needs," Proceedings of the Sixth National
Conference on Management of Uncontrolled Hazardous Waste Sites, Held
on November 4-6, 1985 in Washington, DC, Hazardous Materials Control
Research Institute, Silver Spring, MD, 1985, pp. 71-73.
Myers, Jeffrey C., "Objective Quantification of Sampling Adequacy and
Soil Contamination Levels Around Point Sources Using Geostatistics,"
Proceedings of the National Conference on Hazardous Wastes and
Hazardous Materials, Held on March 4-6, 1986 in Atlanta, GA,
Hazardous Materials Control Research Institute, Silver Spring, MD,
1986, pp. 92-97.
Nacht, S. J., "Monitoring Sampling Protocols," Ground Water
Monitoring Review, Vol. 3, No. 3, Summer 1983, pp. 23-29.
Nadeau, Royal J. et al., "Measuring Soil Vapors for Defining
Subsurface Contaminant Plumes," Proceedings of the Sixth National
Conference on Management of Uncontrolled Hazardous Waste Sites, Held
on November 4-6, 1985 in Washington, DC, Hazardous Materials Contro1
Research Institute, Silver Spring, MD, 1985, pp. 128-129.
166
-------
Nazar, Andrzej, James Prieur, and Daniel Threlfall, "An Integrated
Ground Water Monitoring Program Using Multilevel Gas-Driven Sampler
and Conventional Monitoring Wells," Ground Water Monitoring Review,
Vol. 4, No. 4, Fall 1984, pp. 43-47.
Nelson, James D. and Robert C. Ward, "Statistical Considerations and
Sampling Techniques for Ground-Water Quality Monitoring, Ground
Water, Vol. 19, No. 6, November-December 1981, pp. 617-625.
Oppenheimer, J. A., A. D. Eaton, and L. Y. C. Leong, Multielement
Analytical Techniques for Hazardous Waste Analysis; The State-of-
the Art. (EPA-600/4-64-028), U.S. Environmental Protection Agency.
Las Vegas, NV, April 1984.
Pellizzari, Edo D., Measurement of Carcinogenic Vapors in Ambient
Atmospheres, (EPA-600/7-78-062), U.S. Environmental Protection
Agency, Research Triangle Park, NC, April 1978.
Perazzo, James A., Richard C. Dorrler, and James P. Mack, "Long-Term
Confidence in Ground Water Monitoring Systems," Ground Water
Monitoring Review, Vol. 4, No. 4, Fall 1984, pp. 119-123.
Peters, James A., Keith M. Tackett, and Edward C. Eimutis,
"Measurement of Fugitive Hydrocarbon Emissions from a Chemical Waste
Disposal Site," Proceedings of the Second National Conference on
Management of Uncontrolled Hazardous Waste Sites, Held on
October 28-30, 1981 in Washington, DC, Hazardous Materials Control
Research Institute, Silver Spring, MD, 1981, pp. 123-128.
Plumb, Russel H., Jr., Characterization of Hazardous Waste Sites -
A Methods Manual - Volume III: Available Laboratory Analytical
Methods, (EPA-60Q/4-84-U38J, U.S. Environmental Protection Agency,
Las Vegas, NV, May 1984.
Popkin, Barney P., "Guidelines for Ground-Water Quality Assessments
for Hazardous Waste Facilities," Ground Water Monitoring Review,
Vol. 3, No. 2, Spring 1983, pp. 65-70.
Porcella, D. B., Protocol for Bioassessment of Hazardous Waste
Sites, (EPA-600/2-83-054), U.S. Environmental Protection Agency,
Corvallis, OR, July 1983.
Possin, Boyd N., "Groundwater Systems and Hazardous Wastes Sites - A
Basic Conceptual Framework," Proceedings of the Fourth National
Conference on Management of Uncontrolled Hazardous Waste Sites, Held
on October 31-November 2, 1983 in Washington, DC, Hazardous
Materials Control Research Institute, Silver Spring, MD, 1983,
pp. 114-116.
167
-------
Powell, David H. et al., "Use of Detector Ratios for Contaminant
Screening by High-Pressure Liquid Chromatography," Proceedings of
the Fourth National Conference on Management of Uncontrolled
Hazardous Waste Sites, Held on October 31-November 2, 1983 in
Washington, DC, Hazardous Materials Control Research Institute,
Silver Spring, MD, 1983, pp. 86-93.
Priznar, Francis J. and Lucy P. Sibold, "Site Inspection Sampling
Strategy to Support the Hazard Ranking System Scoring," Proceedings
of the Sixth National Conference on Management of Uncontrolled
Hazardous Waste Sites, Held on November 4-6, 1985 in Washington, DC,
Hazardous Materials Control Research Institute, Silver Spring, MD,
1985, pp. 74-79.
Provost, Lloyd P., "Statistical Methods in Environmental Sampling,"
Environmental Sampling for Hazardous Wastes, (ACS Symposium
Series 267), Gleen E. Schweitzer and John A. Santolucito, eds.,
American Chemical Society, Washington, DC, 1984, pp. 79-96.
Quimby, Jay M., Robert W. Cibulskis, and Michael Gruenfeld,
"Evaluation and Use of a Portable Gas Chromatograph for Monitoring
Hazardous Waste Sites," Proceedings of the Third National Conference
on Management of Uncontrolled Hazardous Waste Sites, Held on
November 29-December 1, 1982 in Washington, DC, Hazardous Materials
Control Research Institute, Silver Spring, MD, 1982, pp. 36-39.
Quinn, Kenneth J., Steven G. Wittmann, and R. Damon Lee, "Use of
Soil Gas Sampling Techniques for Assessment of Groundwater
Contamination," Proceedings of the Sixth National Conference on
Management of Uncontrolled Hazardous Waste Sites, Held on
November 4-6, 1985 in Washington, DC, Hazardous Materials Control
Research Institute, Silver Spring, MD, 1985, pp. 157-163.
Rhodes, Raymond C. and E. Gardner Evans, Precision and Accuracy
Assessments for State and Local Air Monitoring Networks, (EPA-600/
4-86-012), U.S. Environmental Protection Agency, Research Triangle
Park, NC, February 1986.
Riggin, R. M., Technical Assistance Document for Sampling and
Analysis of Toxic Organic Compounds in Ambient Air,(EPA-600/
4-83-027), U.S. Environmental Protection Agency, Research Triangle
Park, NC, June 1983.
Ritzert, C. John, "Sampling Plan Design for Hazardous Wastes,"
Proceedings of the National Conference on Hazardous Wastes and
Environmental Emergencies, Held on March 12-14, 1984 in Houston, TX,
Hazardous Materials Control Research Institute, Silver Spring, MD,
1984, pp. 141-145.
168
-------
Rogoff, Marc J., Summary of EPA Symposium on Land Disposal of
Hazardous Waste, (WP-81W406), The MITRE Corporation, McLean, VA,
July 31, 1981.
Russell, Sue, National Priorities List Quality Assurance Team, The
MITRE Corporation, McLean, VA, personal communications to Thomas
Wolfinger, The MITRE Corporation, 1986.
Ryan, Robert M., "Texas Experience in Ambient Air Sampling for Toxic
Wastes," Proceedings of the Sixth National Conference on Management
of Uncontrolled Hazardous Waste Sites, Held on November 4-6, 1985 in
Washington, DC, Hazardous Materials Control Research Institute,
Silver Spring, MD, 1985, pp. 125-127.
Sachs, Lothar, Applied Statistics; A Handbook of Techniques,
Springer-Verlag, New York, NY, 1982.
Sayers, William T. and Wayne R. Ott, "Use and Interpretation of
Water Quality Data," Accuracy in Trace Analysis; Sampling, Sample
Handling, Analysis - Volumes I and II, (NBS Special Publication
422), Philip D. LaFleur, ed., National Bureau of Standards,
Gaithersburg, MD, 1974, pp. 91-107-
Scalf, Marion R. et al., Manual of Ground-Water Quality Sampling
Procedures, (EPA-600/2-81-160), U.S. Environmental Protection
Agency, Ada, OK, September 1981.
Schmidt, Charles E., W. David Balfour, and Robert D. Cox, "Sampling
Techniques for Emissions Measurements at Hazardous Waste Sites,"
Proceedings of the Third National Conference on Management of
Uncontrolled Hazardous Waste Sites, Held on November 29-December 1,
1982 in Washington, DC, Hazardous Materials Control Research
Institute, Silver Spring, MD, 1982, pp. 334-339.
Schmidt, Charles E. and M. W. Eltgroth, "Off-Site Assessment of
Air Emissions from Hazardous Waste Disposal Facilities," Proceedings
of the Fourth National Conference on Management of Uncontrolled
Hazardous Waste Sites, Held on October 31-November 2, 1983 in
Washington, DC, Hazardous Materials Control Research Institute,
Silver Spring, MD, 1983, pp. 293-295.
Schuller, Rudolph M., James P- Gibb, and Robert A. Griffin,
"Recommended Sampling Procedures for Monitoring Wells," Ground Water
Monitoring Review, Vol. 1, No. 1, Spring 1981, pp. 42-46.
169
-------
Schweitzer, Glenn E., "Risk Assessment Near Uncontrolled Hazardous
Waste Sites: Role of Monitoring Data," Proceedings of the Second
National Conference on Management of Uncontrolled Hazardous Waste
Sites, Held on October 28-30, 1981 in Washington, DC, Hazardous
Materials Control Research Institute, Silver Spring, MD, 1981,
pp. 238-247.
Schweitzer, Glenn E. and Stuart C. Black, "Monitoring Statistics"
Environmental Science and Technology, Vol. 19, No. 11, November 1985,
pp. 1026-1030.
Schweitzer, Glenn E. and John A. Santolucito, eds., Environmental
Sampling for Hazardous Wastes, (ACS Symposium Series 267), American
Chemical Society, Washington, DC, 1984.
Seanor, Arthur M., Anna L. Pisano, and Larry K. Brannaka, "Quality
Assurance to Assure Quality of Field Sampling," Proceedings of
the National Conference on Hazardous Wastes and Environmental
Emergencies, Held on March 12-14, 1984 in Houston, TX, Hazardous
Materials Control Research Institute, Silver Spring, MD, 1984,
pp. 123-128.
Seanor, Arthur M. and Larry K. Brannaka, "Influence of Sampling
Techniques on Organic Water Quality Analysis," Proceedings of the
Second National Conference on Management of Uncontrolled Hazardous
Waste Sites, Held on October 28-30, 1981 in Washington, DC,
Hazardous Materials Control Research Institute, Silver Spring, MD,
1981, pp. 143-148.
Sgambat, Jeffrey and Jery R. Stedinger, "Confidence in Ground-Water
Monitoring, Ground Water Monitoring Review, Vol. 1, No. 1, Spring
1981, pp. 62-69.
Sherwani, Jabbar and David H. Moreau, Strategies for Water Quality
Monitoring, (UNC-WRRI-75-107), University of North Carolina Water
Resources Research Institute, June 1975.
Singh, Udai P. et al., "Sampling the Biscayne Aquifer for Toxic
Pollutants," Proceedings of the Second Conference on Management of
Municipal, Hazardous, and Coal Wastes, (DOE/METC-84-34), Held on
December 5, 1983, in Miami, FL, U.S. Department of Energy,
Washington, DC, September 1984, pp. 422-431.
Sisk, Steven W., NEIC Manual for Groundwater/Subsurface
Investigations at Hazardous Waste Sites, (EPA-330/9-81-002),
U.S. Environmental Protection Agency, Denver, CO, July 1981.
170
-------
Smart, Glenn R. and David K. Cook, "The Design of Monitoring Well
Systems to Meet RCRA Requirements," Proceedings of the National
Conference on Hazardous Wastes and Environmental Emergencies, Held
on March 12-14, 1984 in Houston, TX, Hazardous Materials Control
Research Institute, Silver Spring, MD, 1984, pp. 68-71.
Spear, Richard and Peter Franconeri, "Installing Groundwater
Monitoring Wells at a Hazardous Waste Site," Proceedings of the
Second National Conference on Management of Uncontrolled Hazardous
Waste Sites, Held on October 28-30, 1981 in Washington, DC,
Hazardous Materials Control Research Institute, Silver Spring, MD,
1981, pp. 89-95.
Spittler, Thomas M., "Field Measurements of PCB's in Soil and
Sediment Using a Portable Gas Chromatograph," Proceedings of the
Fourth National Conference on Management of Uncontrolled Hazardous
Waste Sites, Held on October 31-November 2, 1983 in Washington, DC,
Hazardous Materials Control Research Institute, Silver Spring, MD,
1983, pp. 105-107.
Spittler, Thomas M., Richard Siscanaw, and Moira Lataille,
"Correlation Between Field GC Measurements of Volatile Organics and
Laboratory Confirmation of Collected Field Samples Using the GC/MS
Extended Abstract," Proceedings of the Third National Conference on
Management of Uncontrolled Hazardous Waste Sites, Held on
November 29-December 1, 1982 in Washington, DC, Hazardous Materials
Control Research Institute, Silver Spring, MD, 1982, p. 57.
Stephens, Robert D., "The Stringfellow Industrial Waste Disposal
Site: A Technical Assessment of Environmental Impact," Proceedings
of the First National Conference on Management of Uncontrolled
Hazardous Waste Sites, Held on October 15-17, 1980 in Washington,
DC, Hazardous Materials Control Research Institute, Silver Spring,
MD, 1980, pp. 15-20.
Stern, Arthur C., ed., Air Pollution, Third Edition Volume III
Measuring, Monitoring, and Surveillance of Air Pollution, Academic
Press, New York, NY, 1976.
Stetter, J. R. et al., "Portable Instrument for the Detection and
Identification of Air Pollutants," Proceedings; National Symposium
on Recent Advances in Pollutant Monitoring of Ambient Air and
Stationary Sources, (EPA-600/9-84-019), Held on May 8-10, 1984 in
Raleigh, NC, U.S. Environmental Protection Agency, Research Triangle
Park, NC, November 1984, pp. 73-81.
171
-------
Stevens, Robert K., Sampling and Analysis of Atmospheric Aerosols,
Presented at the Advanced Research Workshop on Environmental Monitoring
of Architectural Conservation, Rome, Italy, June 11-14, 1984.
Sullivan, David A. and Jerome B. Strauss, "Air Monitoring of a
Hazardous Waste Site," Proceedings of the Second National Conference
on Management of Uncontrolled Hazardous Waste Sites, Held on
October 28-30, 1981 in Washington, DC, Hazardous Materials Control
Research Institute, Silver Spring, MD, 1981, pp. 136-142.
Swaroop, R. and 0. S. Ghuman, "Subsampling of Hazardous Waste,"
Proceedings of the Fifth National Conference on Management of
Uncontrolled Hazardous Waste Sites, Held on November 7-9, 1984 in
Washington, DC, Hazardous Materials Control Research Institute, Silver
Spring, MD, 1984, pp. 90-93.
Thibodeaux, Louis J. et al., "Air Emission Monitoring of Hazardous
Waste Sites," Proceedings of the Third National Conference on
Management of Uncontrolled Hazardous Waste Sites, Held on November 29-
December 1, 1982 in Washington, DC, Hazardous Materials Control
Research Institute, Silver Spring, MD, 1982, pp. 70-75.
Thomas, John M., "Field Sampling for Monitoring, Migration and Defining
the Areal Extent of Chemical Contamination," Proceedings of the Fifth
National Conference on Management of Uncontrolled Hazardous Waste
Sites, Held on November 7-9, 1984 in Washington, DC, Hazardous
Materials Control Research Institute, Silver Spring, MD, 1984,
pp. 85-89.
Tirsch, F. S. and J. W. Male, River Basin Water Quality Monitoring
Design, (Water Resources Research Center Publication Number 141),
Massachusetts University, Amherst, MA, March 1983.
Todd, David K. et al., Monitoring Groundwater Quality; Monitoring
Methodology, (EPA-600/4-7b-026), U.S. Environmental Protection Agency,
Las Vegas, NV, June 1976.
Townsend, Richard W., "Air Monitoring of Hazardous Waste Sites,"
Proceedings of the Third National Conference on Management of
Uncontrolled Hazardous Waste Sites, Held on November 29-December 1,
1982 in Washington, DC, Hazardous Materials Control Research
Institute, Silver Spring, MD, 1982, pp. 67-69.
Tucker, William A. and Carolyn Poppell, "Method for Determining
Acceptable Levels of Residual Soil Contamination," Proceedings of the
National Conference on Hazardous Wastes and Hazardous Materials, Held
on March 4-6, 1986 in Atlanta, GA, Hazardous Materials Control
Research Institute, Silver Spring, MD, 1986, pp. 87-91.
172
-------
Turpin, Rodney D., "ERT's Air Monitoring Guides for Uncontrolled
Hazardous Waste Sites," Proceedings of the Fourth National
Conference on Management of Uncontrolled Hazardous Waste Sites, Held
on October 31-November 2, 1983 in Washington, DC, Hazardous
Materials Control Research Institute, Silver Spring, MD, 1983,
pp. 82-84.
Turpin, Rodney D., "On-Site Air Monitoring Classification by Use of
the ERT Two-Stage Collection Tube," Proceedings of the Fourth
National Conference on Management of Uncontrolled Hazardous Waste
Sites, Held on October 31-November 2, 1983 in Washington, DC,
Hazardous Materials Control Research Institute, Silver Spring, MD,
1983, pp. 85.
U.S. Environmental Protection Agency, Ambient Air Monitoring
Guidelines for Prevention of Significant Deterioration (PSD),
(EPA-450/4-80-012), U.S. Environmental Protection Agency, Researeh
Triangle Park, NC, November 1980a.
U.S. Environmental Protection Agency, Enforcement Considerations for
Evaluations of Uncontrolled Hazardous Waste Disposal Sites, U.S.
Environmental Protection Agency, Denver, CO, April 1980b.
U.S. Environmental Protection Agency, Field Standard Operating
Procedure for Air Surveillance, (F.S.O.P. 8), U.S. Environmental
Protection Agency, Washington, DC, undated.
U.S. Environmental Protection Agency, Guidelines; Air Quality
Surveillance Networks, (AP-98), U.S. Environmental Protection
Agency, Research Triangle Park, NC, May 1971.
U.S. Environmental Protection Agency, Guidance for Air Quality
Monitoring Network Design and Instrument Siting (Revised), (Draft,
OAQPS Number 1.2-012), U.S. Environmental Protection Agency,
Research Triangle Park, NC, September 1975.
U.S. Environmental Protection Agency, Guidance on Remedial
Investigations Under CERCLA, (EPA-540/G-85-002), U.S. Environmental
Protection Agency, Washington, DC, June 1985a.
U.S. Environmental Protection Agency, Procedure Manual for Ground
Water Monitoring at Solid Waste Disposal Facilities, (SW-611), U.S.
Environmental Protection Agency, Washington, DC, 1977.
173
-------
U.S. Environmental Protection Agency, Proceedings; National
Symposium on Recent Advances in Pollutant Monitoring of Ambient Air
and Stationary Sources, (EPA-600/9-84-019), Held on May 8-10, 1984
in Raleigh, NC, U.S. Environmental Protection Agency, Research
Triangle Park, NC, November 1984.
U.S. Environmental Protection Agency, RCRA Ground-Water Monitoring
Technical Enforcement Guidance Document, (Draft), (EPA-600/4-84-076),
U.S. Environmental Protection Agency, Washington, DC, August 1985b.
U.S. Environmental Protection Agency, Resource Document for the
Ground-Water Monitoring Strategy Workshop, U.S. Environmental
Protection Agency, Washington, DC, March 1985c.
U.S. Environmental Protection Agency, Site Inspection Training
Course, U.S. Environmental Protection Agency, Washington, DC,
undated.
U.S. Environmental Protection Agency, Test Methods for Evaluating
Solid Waste; Physical/Chemical Methods, (SW-846), U.S.
Environmental Protection Agency, Washington, DC, May 1980c.
U.S. Environmental Protection Agency, Third Annual National Symposium
on Recent Advances in Pollutant Monitoring of Ambient Air and
Stationary Sources Agenda, Abstracts, Attendee List, Held on
May 3-6, 1983 in Raleigh, NC, U.S. Environmental Protection Agency,
Research Triangle Park, NC, May 1983.
U.S. House of Representatives, Superfund Amendments and
Reauthorization Act of 1986 Conference Report, (Report 99-962),
U.S. House of Representatives, Washington, DC, 1986.
Vukovich, Fred M., Walter D. Bach, Jr., and C. Andrew Clayton,
Optimum Meteorological and Air Pollution Sampling Network Selection
in Cities Volume 1; Theory and Design for St. Louis, (EPA-600/
4-78-030), U.S. Environmental Protection Agency, Research Triangle
Park, NC, June 1978.
Vukovich, Fred M., Walter D. Bach, Jr., and C. Andrew Clayton,
Optimum Meteorological and Air Pollution Sampling Network Selection
in Urban Areas, (NSF/RA-77Q454), National Science Foundation,
Washington, DC, December 1977.
Wallace, Lance A. et al., "Personal Exposures, Indoor-Outdoor
Relationships, and Breath Levels of Toxic Air Pollutants Measured
for 355 Persons in New Jersey/1 Atmospheric Environment, Vol. 19,
No. 10, 1985, pp. 1651-1661.
174
-------
Walther, Eric G., "Study of Subsurface Contamination with Geophysical
Monitoring Methods at Henderson, Nevada," Proceedings of the Fourth
National Conference on Management of Uncontrolled Hazardous Waste
Sites, Held on October 31-November 2, 1983 in Washington, DC,
Hazardous Materials Control Research Institute, Silver Spring, MD,
1983, pp. 28-36.
Weber, Dennis D. et al., "Spatial Mapping of Conductive Ground Water
Contamination With Electromagnetic Induction," Ground Water
Monitoring Review, Vol. 4, No. 4, Fall 1984, pp. 70-77.
Weber, Robert C. and Kenneth Mims, Evaluation of the Walkthrough
Survey Method for Detection of Volatile Organic Compound Leaks,
(EPA 600/2-81-073), U.S. Environmental Protection Agency,
Cincinnati, OH, April 1981.
Wetzel, Roger et al., "Drum Handling Practices at Abandoned Sites,"
Land Disposal of Hazardous Waste - Proceedings of the Ninth Annual
Research Symposium, (EPA-600/9-83-018), Held on May 2-4, 1983 in
Ft. Mitchell, KY, U.S. Environmental Protection Agency, Cincinnati,
OH, September, 1983, pp. 414-426.
Wetzel, Roger, Kathleen Wagner, and Anthony N. Tafuri, "Drum
Handling Practices at Abandoned Sites," Proceedings of the Third
National Conference on Management of Uncontrolled Hazardous Waste
Sites, Held on November 29-December 1, 1982 in Washington, DC,
Hazardous Materials Control Research Institute, Silver Spring, MD,
1982, pp. 169-174.
Wilson, L. G., Monitoring in the Vadose Zone; A Review of Technical
Elements and Methods, (EPA-6QO/7-89-134), U.S. Environmental
Protection Agency, Las Vegas, NV, June 1980.
Wolfinger, Thomas F., Hazard Ranking System Issue Analysis;
Carcinogenic Risk Analysis of the Air Pathway Target Distance Limit,
(MTR-86W140), The MITRE Corporation, McLean, VA, October 1986.
Wusterbarth, Arlene, Hazard Ranking System Issue Analysis;
Relationship Between Waste Quantity and Hazardous Constituent
Quantity, (MTR-86W141), The MITRE Corporation, McLean, VA,
September 1986.
Wyeth, Robert K., "The Use of Laboratory Screening Procedures in the
Chemical Evaluation of Uncontrolled Hazardous Waste Sites,"
Proceedings of the Second National Conference on Management of
Uncontrolled Hazardous Waste Sites, Held on October 28-30, 1981 in
Washington, DC, Hazardous Materials Control Research Institute,
Silver Spring, MD, 1981, pp. 107-109.
175
-------
Zachowskl, Michael S. and Stephen A. Borgianini, "Problems and
Pitfalls of Trace Ambient Organic Vapor Sampling at Uncontrolled
Hazardous Waste Sites," Proceedings; National Symposium on Recent
Advances in Pollutant Monitoring of Ambient Air and Stationary
Sources, (EPA-60U/9-84-019J, Held on May 8-10, 1984 in Raleigh, NC,
U.S. Environmental Protection Agency, Research Triangle Park, NC,
November 1984, pp. 82-90.
Zar, Jerrold H., "Power and Statistical Significance in Impact
Evaluation," Ground Water Monitoring Review, Vol. 2, No. 3, Summer
1982, pp. 33-3T:
176
-------
APPENDIX A
COMMENTS ON THE HRS RELATED TO HAZARDOUS WASTE CONCENTRATIONS
Comments on the HRS related to hazardous waste concentrations
have been related to the appropriateness and the manner of
considering concentration data in the scoring of Superfund sites
(Federal Register, 1982). A number of these comments have called
for the inclusion of concentration data in at least three factors
used in the HRS; however, the comments have not specified how the
data should be obtained, analyzed, or incorporated in the HRS. This
section briefly describes the current rating factors considered in
the HRS, outlines comments addressing the modification of three of
these factors to account for concentration data, and summarizes
commentors1 suggestions for implementing the modifications.
The HRS assigns three scores to a hazardous site, or facility.
One of these scores is designed to reflect the potential for harm
from substances that can explode or cause fires, and a second is
designed to reflect the potential harm from direct contact with
hazardous substances at a facility. Since these two scores are
used to identify facilities requiring emergency attention, and since
the commentors did not indicate that concentration data should be
obtained to determine these scores, they will not be considered
here. The third score, however, is designed to rank facilities
across the nation of possible inclusion on the National Priorities
List, which includes those sites slated for further government
177
-------
investigation and possible remedial action (Casler and Ramsey,
1985). This "site migration score" reflects the potential for harm
to humans or the environment from migration of a hazardous substance
away from the facility by three routes: ground water, surface
water, or air. To determine the site migration score, a separate
score is first determined for each of the three separate migration
routes using a two-step process. First, numerical values are
assigned to the factors for each migration route. Second, the
values for each site are combined to yield the individual migration
route scores, none of which can exceed 100. The site migration
score is then determined by combining the three route scores into a
single score that also cannot exceed 100 Federal Register, 1982).
Comments regarding the incorporation of concentration data in
the MRS have focused on the surface water and ground water routes.
Both of these routes are scored by considering three major
categories of factors: "release" (incorporating the subordinate
categories "observed release," route characteristics," and
"containment"), "waste characteristics," and "targets". The
comments received by the EPA nave concerned observed releases and
waste characteristics. (For a detailed explanation of the HRS and
the scoring of rating factors, refer to Federal Register, 1982).
The HRS considers that observed releases of hazardous
substances have occurred if hazardous substances attributable to the
site are found in the environment at a "significantly higher level
178
-------
than the background level" (Federal Register, 1982), regardless of
regulatory limits (e.g., drinking water standards). If there is
direct evidence of an observed release, a value of 45 is assigned to
the "observed release" category. If no such evidence exists, a
score of 0 is assigned to the factor. (However, in this case,
alternative values for "route characteristics" and "containment" are
assigned, with a combined maximum value of 45). Commentors have
objected to the lack of consideration given to concentration levels
in the assignment of the score for observed release, arguing that
low concentrations of hazardous substances can be detected in many
locales (i.e., not just Superfund sites) and that these low
concentrations are not necessarily harmful. One of the commentors,
the Chemical Manufacturers' Association, suggested that scores from
0 to 45 be assigned to the "observed release" category according to
6 concentration ranges (Chemical Manufacturers' Association, 1982).
No rationale for this suggestion, or the specified ranges, was
provided.
The HRS also considers waste characteristics, which include the
two rating factors, "toxicity/persistence" and "quantity". For both
the surface water and ground water routes, the maximum values for
these factors are 18 and 8, respectively. The score for the
category is derived by summing the two factor values; therefore, the
maximum category score is 26. The single toxicity/persistence value
is determined by first assigning separate numerical values to the
179
-------
toxicity and the persistence of the hazardous substance using scales
of 0 to 3, and then extracting an appropriate value (0 to 18) from a
matrix having toxicity and persistence values along its axes. (The
numericl values assigned to toxicity and persistence are determined
by consulting guidance materials provided by the EPA). The hazardous
waste quantity value is determined by assigning a numerical value to
the estimated quantity of hazardous waste at the site, using an
integer scale from 0 to 8 (Federal Register, 1982).
Commentors have objected to the lack of consideration given
to concentration data in the assignment of values for toxicity/
persistence and quantity, arguing the toxicity and the actual
quantity of hazardous substances in the waste are both functions of
concentration. More specifically, these commentors have stated that
concentrations of hazardous substances could be within regulatory
limits, and thus not pose a substantial hazard to human health
(Homestake Mining Company, 1982). Commentors also have stated that
estimating waste quantities without consideration of waste
concentrations result in incorrect estimates and possible inequities.
They indicate that such results are possible since sites are evaluated
on the total quantity of waste (e.g., quantity of hazardous compounds
and the nonhazardous material with which it was mixed before the
mixture was dumped on the site) rather than the "actual" quantity of
the hazardous substance (i.e., excluding the nonhazardous material)
(American Mining Congress, 1983; Chemical Manufacturers' Association,
1982).
180
-------
No guidance was given by any of the commentors regarding the
methods of obtaining appropriate concentration data or incorporating
the data in the HRS. Additionally, the commentors did not suggest
methods of considering the many substances for which there are no
regulatory limits.
In summary, the issues raised by commentors regarding the
incorporation of concentration data in the HRS reflect concern about
observed releases and the "actual" quantity and toxicity of hazardous
wastes at Superfund sites. Since concentration data are not used in
the HRS, except to identify observed releases, commentors contend
that sites with low concentrations of wastes could receive site
scores that are inappropriately high. None of the commentors,
however, have addressed the issues of how concentration data should
be obtained. This is important, as will be "discussed in the
following section, since there is great variability in the size,
nature, and characteristics of Superfund sites, and since the extent
of concentration data required to characterize a site also may vary
along a broad continuum.
181
-------
APPENDIX B
DATA QUALITY
An assessment of the quality of waste site data generally
requires the comparison of certain characteristics of the data with
specified standards. These standards are designed to ensure that
the data are representative of the conditions at the site or in the
vicinity of the site, as applicable. A set of concentration data
for a particular contaminant from a site would be considered
representative if the distribution of the data in time and space,
within the data set, matched the temporal and spatial distribution
of true concentrations from the site. As has been noted by others
(e.g., Barcelona et al., 1985), representativeness must be assessed
subjectively, as the true distribution of conditions from the site
is not known.
There are six aspects of representativeness of a data set that
are amenable, to a varying degree, to objective assessment (adapted
from Sachs, 1982 and Ford and Turina, 1985):
• Specificity
• Homogeneity
• Accuracy
• Precision
• Sensitivity
• Completeness
183
-------
An individual contaminant concentration measurement would be
considered specific if the measurement reflected the level of that
particular contaminant in the environment and was free from
interferences caused by the presence of other contaminants. In part,
a measurement is specific if it reflects what it claims to reflect.
Thus, for example, a set of benzene data would be considered specific
if the data reflected the levels of benzene, and only benzene, in
the environment. Alternately, it would be considered not specific
if the data that were labeled as "benzene" actually reflected levels
of total aromatic hydrocarbons rather than benzene. It is not
always possible to completely rule out interferences and thus ensure
that data are specific. However, this should be done to the extent
possible.
An additional aspect of specificity is related to the concept
of homogeneity. A set of data is considered homogeneous if the
individual data points are derived from samples that were drawn from
the same distribution (i.e., to the extent possible, they have the
same characteristics; all controllable factors are the same). Data
sets that combine samples from multiple distributions are considered
to be nonhomogeneous. As related to specificity, the distribution
reflected in a homogeneous data set must be the intended
distribution. For example, if two wastes are present on an
uncontrolled waste site and a differentiation between the
concentrations of contaminants in these wastes is needed, then two
184
-------
distinct, homogeneous data sets should be developed, one reflecting
samples of the first waste, the second reflecting samples from the
second waste. In this case, a single combined nonhomogeneous data
set would be useless in determining concentrations in the individual
wastes.
Accuracy is related to the statistical concept of bias. An
individual measurement is considered accurate if it can be expected
that the measured value is the true value. Operationally, a data
set is considered sufficiently accurate if the average difference
(or ratio) between the measured values and the true value is
acceptable based on the intended use of the data. An important
characteristic of accurate data is that it tends to be equally
distributed above and below the true value. It is free from
systematic error; error that causes a generally consistent bias
upwards or downwards. Also, data can be accurate even though the
individual measurements differ significantly from each other (on
average the differences must cancel each other out).
An immediately apparent problem with the concept of accuracy is
that its assessment requires that the true value be known at the
time of measurement. If the true value were available, the average
difference between the measured values and the true value could be
used as a measure of accuracy. This estimate of the average
difference could then be used to determine whether a data set were
185
-------
sufficiently accurate. However, the true value is rarely known,
generally only when standard solutions are being used.
As an illustration of the concept of accuracy, consider the
benzene data set exhibited in Table B-l. Assume these values
represent the results of 10 replicate benzene samples taken in
surface water at the same place and time. Assume also that the true
value of 25 mg/1 was also known. These data would be considered
accurate since their average is equal to the true value.
Alternately, if their average were actually 24, the data might be
considered sufficiently accurate depending on the purpose for which
they are intended. Accuracy is typically a relative measure rather
than an absolute measure.
The concept of precision is frequently confused with that of
accuracy as both address the extent to which the measured values
deviate from the true values. As noted above, accurate data may
exhibit wide variations between individual data points. Precision
is a measure of the degree of this variation. It is a measure of
the degree of agreement between separate measurements of the same
quantity. The most common measure of precision is the square root
of the average of the squared deviations of eacn data point from the
sample average (i.e., the sample standard deviation of the data).
If the estimation method is accurate, then precision can be estimated
as the average distance between the individual measurements and the
average of the measurements. Precise data is characterized by a
186
-------
TABLE B-l
ILLUSTRATIVE EXAMPLE OF BENZENE CONCENTRATION DATA
ACCURACY AND PRECISION
Sample Number Value (mg/1)
1 17
2 43
3 26
4 9
5 7
6 19
7 31
8 49
9 21
10 28
Average 25
Standard Deviation 13.5
Coefficient of Variation 54%
187
-------
clustering of most of the measurements around the average value.
An excellent discussion of the differences between accuracy and
precision, using a "bulls-eye" analogy is presented in Kirchmer, 1983.
Again referring to the example data set in Table B-l, the
estimated precision of this data is 13.5, or as it is often stated,
54 percent of the mean value of 25.
Estimation of precision requires multiple measurements at the
same point in space and time or an assumption of spatial or temporal
stability of the factor being measured. The acquisition of multiple
samples may be very difficult or the assumption of stability
unacceptable.
Sensitivity, as described by Sachs (1982), is a measure of the
minimum quantity of a substance that can be differentiated from
zero. If the data set contains measurements listed as "not detected"
or "not quantified" then the data set cannot be considered to be
completely representative of the actual environmental concentrations.
The "not detected" or "not quantified" reading may be as low as zero
or as high as the detection or quantification limits, respectively.
Thus, some error must be introduced in either assuming "not detected"
concentrations equal zero or equal the detection limit (the two most
common approaches to addressing "not detected" data). Special
interpretative techniques that may require additional sampling,
analysis, or interpretative assumptions (e.g., knowledge of the
188
-------
conditional distribution of the concentrations below the detection
limit) are needed to analyze data containing readings below levels
of detection or quantification (Gleit, 1985).
The final important aspect of representativeness of a data set
is completeness. There are no objective measures of completeness.
A complete data set would reflect all significant variations in
environmental or waste concentrations that occur in both space and
time. Further, the data set would include measurements of all
contaminants of concern, across all media of concern. Without
complete knowledge about the site, estimates of the degree of
completeness of a data set are not determinable.
Overall, the principal consideration in determining the quality
of concentration data in terms of its intended usefulness in a
particular application is the degree to which the data are
representative of the actual concentration levels from the site.
In assessing the degree of representativeness, several questions
must be asked reflecting the five most important aspects of
representativeness:
• Specificity: Do the data represent the parameters that they
should, and, if not, is the difference
acceptable?
• Homogeneity: Are the data homogeneous, do they come from
the same distributions? Were sampling'
conditions the same for all samples within
each data set?
189
-------
• Accuracy:
• Precision:
• Sensitivity:
• Completeness:
Are the data sufficiently accurate? Do we
expect the estimated values to be the true
values on average? Have systematic errors
been eliminated or controlled to an
acceptable degree?
Are the data reproducible? Do replicate
values cluster around the average or are they
widely variable? Have all source of
unintended variation been eliminated or
controlled to an acceptable degree?
Are the levels of detection and quantification
low enough to show important variations in
concentrations?
Are all significant variations in
concentrations reflected in the data? Have
all contaminants of concern been measured?
Have all media of concern been sampled?
190
------- |